Abstracts of the 50th EASD Annual Meeting

Diabetologia (2014) 57:[Suppl1]S1–S564 S 1 1 C MINUTES OF THE 49th GENERAL ASSEMBLY OF THE EUROPE- AN ASSOCIATION FOR THE STUDY OF DIABETES held in the Pi i Sunyer Hall, Fira de Barcelona, Barcelona, Spain on Thursday 26 September, 2013 at 18:30 Present: Dr. Andrew J.M. Boulton (President) Dr. Stefano Del Prato (Vice President) Dr. Bernard Thorens (Vice President) Dr. Michael Roden (Honorary Treasurer) Dr. Mark Walker (Honorary Secretary) Dr. Cees J. Tack (Chair, PGEC) Dr.Viktor Jörgens (Executive Director) Dr. Monika Grüsser (Vice Director) and 51 members thanked all members of the EASD Office and the Execu- tive Committee for their commitment and hard work. b) Honorary Treasurer i) Result of tax audit 2008-2010/actions to be taken In autumn 2012 for the second time, a control of the EASD by the Inland Revenue took place. These controls will al- ways occur every three years due to the large turnover of the Association. In August 2013, EASD received the draft conclusion of this tax control. The non-profit status of the EASD is beyond all question; the main issue is the question of the taxation of the income from industry ex- hibition and symposia. EASD itself does not handle the industry exhibition and the symposia organized by third parties; these activities are taken care of by a professional congress organizer which is actually Interplan in Munich. In common with other medical associations in Germany, EASD has a contract, in our case with Interplan, giving them permission to organize these activities on their own responsibility without interfering with EASD concerning the organization and the fund raising. The basis of this col- laboration is a loan contract and the income from such a loan contract is considered to be tax free. This legal and taxation construction is the basis of a healthy financial situation of German academic medical societies. The Düs- seldorf tax authorities questioned this regulation and even asked their superiors for comments. In August 2013, they came up with the opinion that all incomes of EASD from these loan contracts, starting with 2008, are considered by the Inland Revenue as a taxable income. Taking into ac- count corporation and local business tax, the result is that a total of Euro 5.4 million will likely have to be paid at the beginning of 2014. Following the advice of legal and The President, Dr. Boulton, welcomed everyone to the 49th General Assembly. He asked those present to stand in memory of the following members, who had passed away: Drs. Georg Eisenbarth, John Hutton, Harry Keen, Carol Lurie, Irina G. Obrosova, Samuel Rahbar, Jean-Louis Richard, Richard Rubin and Patrick Vexiau. 1. MINUTES 48th GENERAL ASSEMBLY 2012 Since there were no comments, the minutes were unani- mously approved and officially signed as a correct re- cord. 2. REPORTS a) President The President’s report to the members on the activities of EASD was given in the President’s Address before the Minkowski Lecture. It is available under: http://www.eas- dvirtualmeeting.org/resources/6926 The President announced the Claude Bernard Lecturer 2014: Dr. Domenico Accili. The President announced the Medical Devices in Diabe- tology meeting on 26/27 February 2014. He said better post-marketing surveillance was required and the Swedish project of regulating pumps would be looked at. The President expressed his thanks to all partners. Dr. Boulton reported that as expected the EASD Annual Meet- ing in Barcelona was doing very well and the number of delegates attending had slightly increased. Dr. Boulton DOI 10.1007/s00125-014-3355-0 S 2 1 C taxation advisors, EASD will oppose this decision and will eventually go to court. Nevertheless, this amount will be paid in advance to avoid the potential addition of interest on this sum should we elect to await the outcome of our appeal and should this be negative. As we allowed for this potential outcome last year, no transfers have been carried forward from EASD to EFSD resulting in a reserve of Euro 5.2 million being kept on the accounts which is more than usual. Now EASD, following the income from the last Annual Meetings, will be able to cover the taxation imposed, but the Association will nev- ertheless have a hard time to finance its activities in the first six months of 2014 until a substantial income from the Meeting in Vienna has occurred. The final decision regarding these matters will be made by the Düsseldorf Inland Revenue and their report could reach EASD at the beginning of 2014. Following the advice from a specialist lawyer in Bonn, it was unanimously decided that when requested, the cur- rent tax bill of Euro 5.4 million should be paid; after that, we will challenge the decision in court. It was also unan- imously decided to seek further advice from the lawyer concerning the way forward. Dr. Roden explained that with the possibility of this tax invoice in mind, EASD had made no financial transfer to EFSD in 2012, resulting in EASD having enough funds to pay the tax bill. However, this will necessitate making savings in the first half of 2014 until an income has been received from the Annual Meeting in Vienna. Such sav- ings could include reducing the costs of the Annual Meet- ing by no longer printing a Provisional Programme and no longer funding a Welcome and Farewell party. The deci- sion had already been made to increase registration fees moderately. Dr. Boulton thanked the Honorary Treasurer for his report and asked if there were any questions. Dr. Lenzen asked if there were any plans to relocate EASD. Dr. Boulton said that there were no relocation plans at that time. c) Honorary Auditors The President asked the Honorary Auditor, Dr. Peter Diem, for his report. Dr. Diem confirmed that the accounts had been checked carefully by Dr. Luis Gardete-Correia and were in perfect order. Dr. Boulton asked for the vote to accept the accounts. The Honorary Treasurer was unanimously discharged (35 votes in favour and 6 abstentions). d) Honorary Secretary The Honorary Secretary reported that 2321 abstracts had been received for the Annual Meeting in Barcelona. Of these, 1360 were accepted: 264 oral presentations and 1096 poster presentations. The main countries submitting abstracts are the United States, the United Kingdom, Japan and Germany. The palm questionnaire rating the 48th An- nual meeting in Berlin gave excellent feedback on the ven- ue facilities, the scientific programme, the Prize Lectures, the EASD symposia, the oral presentations and abstracts, the poster presentations, the Industry symposia and the As- sociations’ Village. The Rising Star Symposium continues to identify promising and innovative young researchers in Europe. The Honorary Secretary concluded by reporting that the Programme Committee for 2014 had been appointed and had had its first meeting during the Annual Meeting in Barcelona. It would then meet in May 2014 to select the abstracts on an anonymous basis for the Annual Meeting in 2014. Dr. Walker closed his report by thanking all those who had forwarded ideas for the scientific programme and all mem- bers of the EASD staff, in particular Mrs. H. Goliberzuch and Mrs. M. Toledo, for their outstanding help and support with the organisation of the EASD Annual Meetings. Dr. Boulton thanked Dr. Walker for his diligence and asked if there were any questions. There were no further comments. e) Editor-in-Chief, Diabetologia In the absence of Dr. Juleen Zierath, Editor-in-Chief, Dia- betologia, the President reported the following informa- tion: The Committee on Publication Ethics had put together ethical guidelines for peer-reviewers. From July 2013, all papers returned for revision underwent plagiarism detec- tion. Springer is as yet unable to check images for possible manipulation, but are aware such a service is of high prior- ity for the journal. The Scientific Integrity Panel (SIP) set up a policy outlining what is expected of authors and how to deal with cases of misconduct which was published on the Diabetologia website in April 2013. Mr. Tony Kirby, a former press officer with the Lancet, has helped to publi- cise papers by issuing press releases. In total, 1,990 articles were submitted to Diabetologia in 2012 and 17% were published. The proportion of papers triaged increased to 63% in the first half of 2013. On aver- age, papers are triaged in 3 days and peer-reviewed in 30 days. Diabetologia (2014) 57:[Suppl1]S1–S564 S 3 1 C Dr. Boulton expressed his thanks to Dr. Zierath in her ab- sence for her co-operation and hard work. Thanks were also expressed to the team in Bristol, especially to Dr. Judy Naylor, and the outgoing associate editors. f) Chair, Postgraduate Education Committee Dr. Tack reported on the courses that had taken place in 2012 and 2013: Almaty/Kazakhstan, Lviv/Ukraine and Kazan/Russia. Dr. Tack said all courses in 2012 and 2013 had been very successful and had attracted delegates from many of the countries neighbouring the one where the course was held. Dr. Tack said a course was being planned in Budapest/Hungary in November 2014 and additional courses will be agreed upon with the incoming PGEC Chair. Dr. Boulton thanked Dr. Tack for his co-operation and commitment to postgraduate education. g) Chair, Extra-European Postgraduate Activities Dr. Czupryniak reported that a course had successfully been organised in Dubai/United Arab Emirates in 2012. In August 2013, a course was held in India when 4 cities (Chennai, Ahmedabad, Lucknow and Chandigarh) were visited. The main Extra-European course in collaboration with IDF and ADA in 2013 will take place in Sri Lanka in November. Also in November, another course will be held in the Gulf Region (Oman). Dr. Czupryniak brought his report to an end by thanking Dr. Boulton for his support and the EASD team for their help in organising the courses. Dr. Boulton thanked Dr. Czupryniak for his co-operation and hard work. 3. ELECTIONS Honorary Secretary (2013-2016) The election of Dr. Per-Henrik Groop was unanimously approved with 40 votes. Chair, PGEC (2013-2016) The election of Dr. Leszek Czupryniak was unanimously approved with 41 votes and 1 abstention. Council Members (2014 - 2017) The election of Drs. Bo Ahrén, Sehnaz Karadeniz, Michele Solimena, Nicholas Tentolouris was unanimously ap- proved with 37 votes each and 4 abstentions. Editor-in-Chief, Diabetologia (2013-2015) extension of office The election of extension of office of Dr. Juleen Zierath was unanimously approved with 41 votes. 4. STUDY GROUPS i) Non-Alcoholic Fatty Liver Disease (NAFLD) Study Group The General Assembly was informed of the decision by the Executive Committee and Council to endorse the initiation of this EASD Study Group. ii) EASD Diabetes and Cancer Study Group The General Assembly was informed of the decision by the Executive Committee and Council to endorse the initiation of this EASD Study Group. iii) Gluco-Incretin Biology and Clinical Applications Study Group The General Assembly was informed of the decision by the Executive Committee and Council to endorse the initiation of this EASD Study Group. iv) Metabolic Imaging Study Group The General Assembly was informed of the decision by the Executive Committee and Council to endorse the initiation of this EASD Study Group. 5. HONORARY MEMBERSHIP The nomination of Drs. Anthony Cerami, Willy Malaisse, John D. Ward was unanimously approved. Dr. Boulton thanked them for their outstanding contribution to diabe- tes research. 6. ANY OTHER BUSINESS Dr. Boulton reported that from January 2014 only an elec- tronic version of the journal would be available for mem- bers. He added that this was the correct way forward both economically and ecologically. He said this would not affect re-prints which would continue to be available. An email will be sent to all members informing them of this situation and a monthly email will be sent with the table of contents for each month’s journal. The President thanked Dr. Walker, retiring Honorary Sec- retary, and Dr. Cees Tack, retiring Chair of the PGEC, for their friendly collaboration and diligence during their term of office. He also expressed his sincere gratitude to the Local Organising Committee for their outstanding contri- bution to the organisation of the 49th EASD Annual Meet- ing. He warmly thanked Dr. Jörgens, Dr. Grüsser and the EASD team in Düsseldorf for their dedicated work. Dr. Boulton brought the General Assembly to a close at 19:25. Diabetologia (2014) 57:[Suppl1]S1–S564 S 4 1 C Agenda for the 50th General Assembly of the European Association for the Study of Diabetes to be held in the Randle Hall at Messe Wien Exhibition & Congress Center, Vienna, Austria on Tuesday, 16 September 2014 at 18:30 1. Minutes of the 49th General Assembly, Barcelona, Spain 2013 2. Reports a) President Andrew J.M. Boulton i) Membership fees 2016 ii) Retirement Executive Director b) Honorary Treasurer Michael Roden c) Honorary Auditors Peter Diem Luis M. Gardete-Correia d) Honorary Secretary Per-Henrik Groop e) Editor-in-Chief, Diabetologia Juleen Zierath f) Chair, Postgraduate Education Committee Leszek Czupryniak 3. POSITIONS To be elected for retiring member(s): 3.1 Elections a) President Extension of office (2014-2015) Andrew JM Boulton b) President (2015-2018) Andrew JM Boulton c) Vice President (2014-2017) Stefano Del Prato d) Honorary Treasurer Extension of office (2014-2016) Michael Roden d) Honorary Auditors (2014-2017) Peter Diem and Luis M. Gardete-Correia e) Council Members (2015-2018) Henning Beck-Nielsen Svitlana Bolgarskaya Anna Novials Raimund Weitgasser 4. Study Groups a) Reactive Metabolites in Diabetes Study Group 5. Honorary Membership 6. Any other business Diabetologia (2014) 57:[Suppl1]S1–S564 S 5 1 C 50th EASD Annual Meeting of the European Association for the Study of Diabetes Vienna, Austria, 15 – 19 September 2014 OP 46 New approaches and clinical outcomes of islet or pancreas transplantation OP 47 Pharmacogenetics and disease progression OP 48 Novel targets for anti-inflammatory therapies Index of Poster Sessions PS 001 Pragmatic prediction and prevention of type 2 diabetes PS 002 Aetiological epidemiological studies of type 2 diabetes PS 003 Type 1 diabetes: epidemiology PS 004 Epidemiology of obesity and ectopic fat PS 005 Descriptive epidemiology of diabetes PS 006 Epidemiology of diabetes: comorbities and complications PS 007 Pharmaco-epidemiology PS 008 Type 1 diabetes: genes and biomarkers PS 009 Old genes, new phenotypes PS 010 Genomics and epigenetics of type 2 diabetes PS 011 New and old genes for monogenic diabetes PS 012 Monogenic diabetes across the world PS 013 Islet gene expression in type 2 diabetes PS 014 Beta cell proliferation and differentiation PS 015 Metabolic coupling in insulin secretion PS 016 Modulation of islet function through cell surface receptors PS 017 Transgenic animal models of type 1 and type 2 diabetes PS 018 Control of insulin exocytosis PS 019 Cytokine-induced beta cell death PS 020 Clinical immunology of type 1 diabetes PS 021 Measuring and preserving insulin secretion in type 1 diabetes PS 022 Beta cell ER stress and apoptosis PS 023 Mechanisms of lipotoxicity PS 024 Islet and pancreas transplantation PS 025 Hypoglycaemia rates and their economic burden PS 026 Insulin secretion in animal models PS 027 Determinants of insulin secretion in humans PS 028 Insulin secretion in diabetes PS 029 Gut hormones PS 030 Gut endocrinology in vivo PS 031 Gut hormone action PS 032 Novel approaches for identifying and targeting determinants of glucose tolerance PS 033 Determinants of insulin sensitivity and glycaemic control PS 034 Glucose uptake and glucose action Abstracts Index of Oral Presentations OP 01 SGLT2 inhibitors: new outcome studies OP 02 Nephropathy: biomarkers and infections OP 03 Evolving tools in education OP 04 Brown adipose tissue OP 05 Factors driving islet cell development OP 06 Novel mechanism of glucose tolerance OP 07 GLP-1 analogues: clinical efficacy OP 08 Matters of the heart OP 09 Diabetic retinopathy OP 10 Entero-pancreatic endocrinology OP 11 Lifestyle factors and prediction of type 2 diabetes OP 12 The many faces of advanced glycation OP 13 Clinical studies with GLP-1 analogues OP 14 Weight regulation and obesity OP 15 Diabetic nephropathy: epidemiology and genetics OP 16 Mechanisms of cardiovascular disease OP 17 Intra- and inter-islet cell signalling OP 18 Novel genes for type 2 diabetes and its complications OP 19 Insulin: clinical decision making OP 20 Novel compounds on the horizon OP 21 Physiological adaptation to bariatric surgery OP 22 Neuropathy: mechanisms and outcomes OP 23 Novel regulators of insulin and glucagon secretion OP 24 Genes and biomarkers for type 2 diabetes OP 25 Novel insulin formulations and combinations OP 26 Pregnancy and diabetes OP 27 Protecting the periphery OP 28 Ectopic lipids and type 2 diabetes OP 29 Autoimmune diabetes OP 30 Integrative physiology OP 31 GLP-1 analogues: non-glycaemic endpoints OP 32 Complications: expect the unexpected OP 33 Device utilisation and outcomes OP 34 Novel approaches for beta cell protection OP 35 Hypertension in diabetes OP 36 Molecular mechanisms of insulin action in vitro OP 37 Metformin: new insights into an old drug OP 38 Novelties in risk prediction OP 39 Inflammation in obesity and type 2 diabetes OP 40 Translational immunology of type 1 diabetes OP 41 The human methylome in diabetes OP 42 Liver metabolism OP 43 GLP-1 analogues: novel formulations OP 44 Diabetes and cancer OP 45 Diabetic foot: mechanisms of wound healing Diabetologia (2014) 57:[Suppl1]S1–S564 S 6 1 C PS 035 Mechanisms of insulin action in non-human models PS 036 Mechanisms of insulin resistance in vivo PS 037 Mechanisms of insulin action and inter-organ cross-talk PS 038 Insulin resistance and adipose tissue PS 039 Glucagon PS 040 Exercise physiology I PS 041 Exercise physiology II PS 042 Consequences of hypoglycaemia PS 043 Nocturnal and drug-induced hypoglycaemia PS 044 Endocrine control of glucose tolerance PS 045 Lipid metabolism in humans PS 046 Brain and glucose metabolism PS 047 Adipose tissue function in humans PS 048 Brown, beige or brite? PS 049 Animal models of obesity and type 2 diabetes PS 050 Adipose tissue function: animal models PS 051 The role of the liver in glucose and lipid metabolism PS 052 Mitochondrial function, oxidative stress and glucose metabolism PS 053 Metabolic effects of dietary composition, caloric restriction and bariatric surgery PS 054 Inflammation and beta cell in type 2 diabetes PS 055 Inflammation in adipose tissue and muscle PS 056 Novel adipokines PS 057 Nutrition and diet I PS 058 Nutrition and diet II PS 059 Metformin PS 060 SGLT2 inhibitors: safety PS 061 Insulin secretagogues PS 062 SGLT2 inhibitors: non-glycaemic endpoints PS 063 GLP-1 based therapies: efficacy I PS 064 GLP-1 based therapies: efficacy II PS 065 SGLT2 inhibitors: efficacy PS 066 Novel therapies PS 067 Non-glycaemic effects of DPP-4 inhibitors PS 068 Incretin based therapies in special populations PS 069 Safety of DPP-4 inhibitors PS 070 Clinical studies with DPP-4 inhibitors PS 071 Non-glycaemic endpoints PS 072 Incretin based therapies basic: basic science PS 073 Incretin based therapies: mechanistic / miscellaneous PS 074 The potential future of insulin therapy PS 075 Insulin treatment in the “real world” PS 076 Glucose variability in insulin treatment PS 077 Old and new insulin formulations PS 078 Insulin immunogenicity and kinetics PS 079 Insulin treatment and hypoglycaemic reactions PS 080 Gastrointestinal liners and e-health opportunities PS 081 Improvements in continuous glucose monitoring PS 082 Improving insulin pump treatment and continuous glucose monitoring use PS 083 On the way to optimise pump treatment PS 084 Supporting tools for insulin treatment PS 085 Psychological distress PS 086 Psychology and quality of life PS 087 Individual perception of diabetes PS 088 Lifestyle and delivery of care PS 089 Met and unmet needs in diabetes care PS 090 Individualised care PS 091 Tailored diabetes care PS 092 Screening and risk factors for gestational diabetes mellitus PS 093 Gestational diabetes mellitus management PS 094 Pregnancy outcomes in gestational diabetes mellitus PS 095 Postpartum outcomes PS 096 Pregnancy in type 1 diabetes PS 097 Neuropathy: biomarkers and mechanisms PS 098 Mechanisms and outcomes of neuropathy PS 099 Diabetic foot: prevention and treatment PS 100 Management of foot ulcers PS 101 Diabetic eye disease PS 102 Mechanisms of microangiopathy: experimental PS 103 Mechanisms of nephropathy: experimental PS 104 Nephropathy: genes and mechanisms PS 105 Diabetic nephropathy: clinical PS 106 Clinical: hypertension in diabetes PS 107 Predicting complications I PS 108 Predicting complications II PS 109 Predicting complications III PS 110 Biomarkers and complications I PS 111 Biomarkers and complications II PS 112 Clinical observations in children PS 113 Clinical observations in children and women PS 114 Clinical observations in type 2 diabetes PS 115 Liver PS 116 Vascular calcification PS 117 Vascular dysfunction PS 118 Advanced glycation PS 119 Metabolism and complications PS 120 Animal models of complications I PS 121 Animal models of complications II Diabetologia (2014) 57:[Suppl1]S1–S564 S 7 1 C OP 01 SGLT2 inhibitors: new outcome studies 1 Fixed dose combinations of empagliflozin/linagliptin for 52 weeks as add-on to metformin in subjects with type 2 diabetes S. Patel1, R. DeFronzo2, A. Lewin3, D. Liu4, R. Kaste4, H.J. Woerle5, U.C. Broedl5; 1Boehringer Ingelheim Ltd., Bracknell, Berkshire, UK, 2University of Texas Health Sciences, San Antonio, USA, 3National Research Institute, Los Angeles, USA, 4Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, USA, 5Boehringer Ingelheim, Ingelheim, Germany. Background and aims: A randomized, double-blind, parallel group Phase III study evaluated the efficacy and safety of fixed dose combinations (FDCs) of empagliflozin/linagliptin (EMPA/LINA) as add-on to metformin in subjects with type 2 diabetes (T2DM). Materials and methods: Subjects were randomized to EMPA 25 mg/LINA 5 mg (n=137), EMPA 10 mg/LINA 5 mg (n=136), EMPA 25 mg (n=141), EMPA 10 mg (n=140), or LINA 5 mg (n=132) as add-on to stable-dose met- formin for 52 weeks. Primary analysis was at week 24. Exploratory endpoints at week 52 were changes from baseline in HbA1c, body weight, systolic and diastolic blood pressure (SBP and DBP), and percentage of subjects with baseline HbA1c ≥7% who reached HbA1c S 8 1 C reduction, weight loss and BP reduction with a low risk of hypoglycaemia compared with GLIM. Clinical Trial Registration Number: NCT01167881 Supported by: Boehringer Ingelheim and Eli Lilly 3 Energy balance following sodium-glucose co-transporter-2 (SGLT2) inhibition G. Ferrannini1, T. Hach2, S. Crowe2, E. Ferrannini1; 1University of Pisa, Italy, 2Boehringer Ingelheim, Ingelheim, Germany. Background and aims: SGLT2 inhibitors lower glycaemia by inducing uri- nary glucose excretion (UGE), with the attendant calorie loss. Evidence sug- gests that the resulting weight loss (WL) is less than expected from UGE. Materials and methods: To quantify this phenomenon we analysed data from 86 type 2 diabetic (T2D) patients (39 women, age = 58 ± 9 years, BMI = 29.8 ± 4.5 kg/m2, HbA1c = 7.8 ± 0.8%, FPG = 169 ± 41 mg/dL, eGFR = 89 ± 19 ml.min-1.1.73m2, µ ± SD), the per-protocol completers cohort of a clinical tri- al who received empagliflozin (25 mg/day) for 90 weeks with frequent (n=11) assessments of body weight, eGFR, and FPG. Time-dependent glucose filtra- tion was calculated as the product of eGFR and FPG, time-dependent UGE was estimated by assuming - from previous direct measurements - a quasi- linear relationship between fractional UGE and glycaemia. The relation of calorie-to-weight changes was estimated using a mathematical model (http:// bwsimulator.niddk.nih.gov) that simulates the time-course of WL for a given change in calorie balance. Results: At week 90, WL averaged -3.2 ± 4.2 kg (range -17.0 to +5.5); over 90 weeks, UGE averaged 54 ± 15 g/day (fractional UGE = 45 ± 4%). The observed WL corresponded to a calorie deficit of -78 ± 103 kcal/day. On the other hand, the observed calorie loss (-217 ± 59 kcal/day) predicted a WL of -8.7 ± 2.4 kg (range -4.0 to -15.3 kg) over 90 weeks. Thus, patients lost only 38 ± 53% of the WL predicted by their glycosuria. As previous studies showed that empagliflozin does not affect either resting or meal-induced energy ex- penditure, patients likely increased their energy intake (by an estimated +138 ± 116 kcal/day). This excess calorie intake was inversely related to baseline BMI (partial r=-0.33, p S 9 1 C discontinuation rates were 4.6%, 10.2%, and 6.8%, and serious AE rates were similar across groups. CANA 100 and 300 mg were associated with a higher incidence than PBO of urinary tract infections (14.5%, 16.5%, 10.1%) and genital mycotic infections in females (23.9%, 18.7%, 4.3%) and males (5.6%, 10.9%, 1.4%). Incidences of osmotic diuresis-related AEs (eg, pollakiuria [in- creased urine frequency] and thirst; 9.1%, 12.3%, 5.5%) and AEs related to re- duced intravascular volume (eg, postural dizziness and orthostatic hypoten- sion; 5.4%, 5.9%, 1.7%) were higher with CANA 100 and 300 mg than PBO, but these AEs led to few discontinuations. The incidence of documented hy- poglycaemia (≤3.9 mmol/L) was modestly higher with CANA 100 and 300 mg than PBO in patients on an AHA associated with hypoglycaemia (54.1%, 61.0%, 48.9%) as well as those on an AHA not associated with hypoglycaemia (18.3%, 10.9%, 6.6%); severe hypoglycaemia rates were low across groups. Conclusion: CANA 100 and 300 mg improved glycaemic control, reduced body weight and BP, and were generally well tolerated in older patients with T2DM over 104 weeks, consistent with 26-week results; the AE profile was consistent with that seen in a broad range of CANA-treated patients. Clinical Trial Registration Number: NCT01106651 Supported by: Janssen Research & Development, LLC 6 Safety of dapagliflozin in patients with type 2 diabetes mellitus and hypertension inadequately controlled by a renin-angiotensin system blocker with/without a second agent T.A. Mansfield1, M.A. Weber2, J.F. List1, A. Ptaszynska1; 1Bristol-Myers Squibb, Princeton, 2SUNY Downstate College of Medicine, New York, USA. Background and aims: Hypertension is common in patients with type 2 dia- betes mellitus (T2DM) and is often treated with an ACE inhibitor (ACEi) or an angiotensin receptor blocker (ARB) plus other antihypertensive (AHT) agents if needed. Dapagliflozin inhibits renal glucose reabsorption causing glucosuria, weight loss, diuresis and decreased BP. Dapagliflozin significantly reduced HbA1c and seated and ambulatory systolic BP versus placebo in clinical studies of patients with T2DM and hypertension. Here we describe key safety data from these studies. Materials and methods: In two, randomized, double-blind, 12-week studies, patients with T2DM and inadequate glycaemic control (HbA1c 7.0-10.5%) and BP (seated systolic BP / diastolic BP: 140-164 / 85-104 mmHg) receiving an ACEi or an ARB (Study 1) plus a 2nd AHT drug (Study 2) were rand- omized to dapagliflozin 10 mg or placebo over 12 weeks. Preliminary efficacy results have been previously presented. Results: In Study 1, 613 patients on an ACEi or ARB were randomized to treatment. In Study 2, 449 patients on an ACEi or ARB plus a 2nd AHT drug were randomized to treatment. In each study, similar proportions of dapagli- flozin and placebo experienced adverse events (AEs), with few serious AEs reported. In Study 1 and 2, fewer dapagliflozin (1.0 and 0.4%, respectively) versus placebo (1.3 and 1.8%) patients withdrew due to an AE. Few hypo- glycaemic events occurred (3.3 and 5.8% with dapagliflozin vs 1.3 and 2.7% with placebo), none of which led to discontinuation. No serious BP related safety issues were noted; one orthostatic hypotension AE occurred with da- pagliflozin in Study 1. Serum uric acid decreased with dapagliflozin but there was no effect on serum sodium, potassium, or calcium, despite its diuretic effect (Table). Conclusion: Dapagliflozin had a good safety profile over 12 weeks when used in combination with an ACEi or ARB ± 1AHT in patients with T2DM and inadequately controlled hypertension. Clinical Trial Registration Number: NCT01137474, NCT01195662 Supported by: BMS/AstraZeneca Diabetologia (2014) 57:[Suppl1]S1–S564 S 10 1 C OP 02 Nephropathy: biomarkers and infections 7 Uric acid serum level affects serum level of cystatin C, independently of GFR, in patients with type 2 diabetes F. Iliadis1, T. Didangelos1, A. Ntemka1, C. Margaritidis1, A. Zantidis1, E. Moralidis2, A. Makedou3, A. Gotzamani-Psarakou2, D. Grekas1, A. Hatzitolios1; 1First Propedeutic Department of Internal Medicine, 2Laboratory of Nuclear Medicine, Medical School, 3Laboratory of Lipids, 2nd Paediatric Department, AXEPA Hospital, Aristotle University of Thessaloniki, Greece. Background and aims: Cystatin C has been reported to be a reliable marker of GFR in both type1 and type 2 diabetic patients with mild-to-moderate CKD (stages 2-3). Recently, elevated serum level of cystatin C has also been identified as a significant prognostic indicator for the development of cardio- vascular disease in people with diabetes. However, there are limited data on factors, other than GFR, that influence its serum concentration, especially in adult patients with type 2 diabetes. The aim of our study was to identify such factors. Materials and methods: In this cross-sectional study, 560 consecutive type 2 diabetic patients (252 men, 308 women), aged 65.0±10.0 years (mean±SD) were recruited from our outpatient diabetic clinic. All participants were Eu- ropids. GFR was measured using 51Cr-EDTA (mGFR). Serum cystatin C was related to several clinical and biochemical parameters. Multivariate analysis was performed in order to identify factors independently associated with cys- tatin C serum level beyond mGFR. SPSS 18.0 was used for statistical analysis. A value of p S 11 1 C Conclusion: High levels of MR-proADM and sTNFR1 were independently associated with an increased risk of renal event in patients with type 2 diabe- tes. With this approach, usCT-proAVP did not carry additional information regarding of renal event when adjusting on these 2 biomarkers. Supported by: PHRC2007 10 Lipoprotein(a) predicts new onset of chronic kidney disease in patients with type 2 diabetes mellitus: a ten-year follow-up study J.-S. Yun1, K.-H. Song1, Y.-M. Park2, S.-H. Ko1, Y.-B. Ahn1; 1Internal Medicine, 2Preventive Medicine, The Catholic University of Korea, Seoul, Republic of Korea Background and aims: We investigated factors that might influence the de- velopment of chronic kidney disease (CKD) in patients with type 2 diabetes. Materials and methods: From January 2000 to December 2002, a total of 1,367 patients with type 2 diabetes without CKD (estimated glomerular filtra- tion rate [eGFR] ≥ 60 ml/min/1.73m2) were consecutively enrolled. Patients were divided into two groups according to their baseline serum Lp(a) levels (Lp(a) >30mg/dL vs Lp(a) ≤30mg/dL). The estimated GFR was measured an- nually, and new onset CKD was defined as eGFR < 60 ml/min/1.73m2 using a Modification of the Diet in Renal Disease formula. We used a Cox propor- tional hazard regression analysis to test associations between new onset CKD and potential explanatory variables. Results: Of the 1,367 patients who were enrolled in this study, 904 (66.1%) completed the follow-up evaluation. The median follow-up time was 9.8 years. The mean age was 56.0 ± 10.3 years, and the duration of diabetes was 8.5 ± 6.9 years. The baseline eGFR was 98.3 ± 25.6 ml/min/1.73m2. During the follow-up period, 234 patients (25.9%) progressed to chronic kidney disease. The patients in the CKD group were older (P < 0.001), had hypertension (P < 0.001), a longer duration of diabetes (P < 0.001), higher baseline A1C levels (Lp(a) >30mg/dL vs Lp(a) ≤30mg/dL: hazard ratio 3.4; 95% CI 2.50 - 4.54; P < 0.001) after adjusting for sex, age, diabetic duration, mean A1C, albuminuria, treatment of insulin, ACE inhibitor/ARB and lipid lowering agents. Conclusion: In conclusion, the development of CKD from normal renal function was independently associated with serum Lp(a) level in patients with type 2 diabetes. 11 Lipidomic biomarkers associated with and predictive of rapid progression of renal disease in type 2 diabetes M. Colombo1, H.C. Looker2, B. Farran2, F. Agakov3, P.E. Sanders4, M.S. Kuo4, M.K. Thomas4, P. McKeigue1, H.M. Colhoun5, on behalf of the SUMMIT Investigators; 1Centre for Population Health Sciences, University of Edinburgh, UK, 2Population Health Sciences, University of Dundee, UK, 3Pharmatics Ltd, Edinburgh, UK, 4Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, USA, 5Population Health Sciences, University of Dundee, UK. Background and aims: To identify lipidomic biomarkers associated with and/or predictive of rapid progression of nephropathy in subjects with type-2 diabetes mellitus. Materials and methods: 288 serum samples (134 cases and 154 controls) from the Genetics of Diabetes Audit and Research Tayside Study (Go-Darts) study were selected based on changes in eGFR following baseline sampling. All cases and controls had a baseline eGFR of 30-60ml/min with cases los- ing >40% of baseline eGFR over a maximum follow-up of 3.5 years, whilst controls lost a maximum of 5% baseline eGFR after >3.5 years follow-up. 221 lipid species, belonging to 9 lipid classes, were measured using a mass spec- trometry platform. We assessed the association with the rapid eGFR loss with individual lipid species, the sum of the lipid class and ratios of individual li- pids to the sum of lipid class, by univariate tests and using predictive selection models. We applied two complementary approaches to biomarker selection: forward selection using logistic regression, and sparse logistic regression with the L1 (LASSO) regularization penalty, with models adjusted for standard clinical covariates including age, sex, diabetes duration, baseline eGFR, al- buminuria, HbA1c and use of ACE Inhibitors or ARBs. Model performance was assessed using Area Under the Receiver Operating Characteristic curve (AUROC) from cross-validated models. Results: A model limited to clinical covariates only had an AUROC of 0.695. A class-by-class analysis of lipids reveals that the most predictive associations are with the Phosphatidylcholines lipids (PC) (AUROC = 0.739 with forward selection, 0.764 with LASSO-regularized selection). Other lipid classes which help improving predictions are Sphingomyelins (SPM) (AUROC = 0.724 and 0.726). Among the Phosphatidylcholines PC26.6 was inversely associated with progression in univariate analysis (Odds Ratio 0.66, 95% Confidence Interval 0.49, 0.86, p 83 years vs. ≤62 years (1.38/1.51/1.65), female gender (1.32), previous UTI in 2010/11 (5.13), other non-UTI infections in 2010/11 (1.63), and CCI>8 (1.67). In the subgroup with laboratory data available (217,022 pts), a higher UTI risk in 2012 was additionally associated with a mean HbA1C of 9.5-10.0/>10.0% vs. 7.0-7.5 in 2011 (1.31/1.33; p S 12 1 C Conclusion: Among a real-life T2DM cohort from 2010-12, UTI event risk was very high compared to previous studies which reported event rates 47-60 cases/1,000 patient years. This may be due to the older and more comorbid sample which reflects a real life cohort of T2DM patients in Germany. Age, female gender, health status and previous infections are the greatest risk fac- tors associated with the UTI event risk. OP 03 Evolving tools in education 13 Internet intervention designed to identify and reduce risk of diabetic driving mishaps, a randomised clinical trial D.J. Cox1, L. Gonder-Frederick1, L.M. Ritterband1, B. Kovatchev1, K. Ingersoll1, H. Singh1, D. Ford1, K. Schmidt2, T. Banton1; 1Psychiatry & Neurobehavioral Sciences, 2Psychology, University of Virginia, Charlottesville, USA. Background and aims: While driving mishaps are more common among those who have Type 1 Diabetes (T1D), in part due to extreme blood glucose (BG), this increased risk is believed to be restricted to a subgroup of drivers who are more vulnerable to moderate hypoglycaemia. We tested the hypothe- ses that an interactive internet intervention, DiabetesDriving.com (DD.com), could: 1) identify high-risk drivers, and 2) aid high-risk drivers to diminish their risk by better anticipating, preventing, detecting, and treating extreme BG while driving. Materials and methods: 1739 drivers with T1D from all 50 US states who registered for DD.com were screened for inclusion criteria, and for High or Low risk for future driving mishaps based on the Risk Assessment for Diabet- ic Drivers (RADD). 379 High-Risk drivers were randomized to Routine Care (RC), DD.com, or DD.com + Motivational Interviews administered once before and once after DD.com. 122 Low Risk drivers were assigned to RC. DD.com was administered during months 1-2. During months 3-14, all par- ticipants completed monthly on-line driving diaries in which the frequency of the following driving mishaps was reported: Severe hypoglycaemia while driving, Loss of vehicle control without hitting anything, Collisions, Being stopped by the police while driving, Automatic driving due to extreme BG, Unintentional stops while driving, and Requiring assistance from someone else while driving due to extreme BG. Results: Analyses of partial data (4,036 diaries) indicate High-Risk RC driv- ers reported more than twice as many driving mishaps/month than Low-Risk RC drivers (0.54 vs 0.21 mishaps/month, p S 13 1 C duration of three years or more and between 50 and 70 years of age, recruited via their general practitioner, were randomised to receive the peer support intervention (n=101) or one educational meeting (n=132). Outcomes were measured one month before, and 6, and 12 months after baseline by means of self-reported questionnaires (PAID; WHO-5; diabetes specific CIDS). Base- line characteristics between intervention and control group were compared and multilevel linear regression models were performed on complete cases to investigate the effect of the intervention on diabetes-related distress (n=118), psychological well-being (n=116) and diabetes-specific self-management (n=107) (Mlwin2.22). Unadjusted analysis as well as analysis adjusted for dif- ferences in baseline characteristics are presented. Results: Mean age of the study participants was 64 years (SD: 5) in both the control and intervention group. More men were randomised to the inter- vention group (69.3%) and more participants were lower educated (60.4%). Other characteristics did not differ between the two groups at baseline. No effects were seen for diabetes-related distress or well-being. However, the in- tervention group showed an increase in diabetes-specific self-management after 6 months of follow-up (β= 4.6; 95% CI: -5.4 to 14.6) and between base- line and 12 months of follow-up (β= 5.1; 95% CI: 1.2 to 9.0). Adjustment for differences in sex and educational level between intervention and control group at baseline slightly increased the differences in effects after 6 months (β = 6.5; 95% CI: -2.8 to 15.7) and after 12 months (β=7.0; 95% CI:3.01 to 11.02). Conclusion: Results of this study indicate that a group-based, peer support program had no effect on diabetes distress nor on well-being but a modest beneficial effect on diabetes-specific self-management in type 2 diabetes pa- tients was seen, which sustained for at least six months after the last peer group session. Clinical Trial Registration Number: NTR3474 Supported by: Dutch Diabetes Research Foundation 15 The efficacy of a new education and treatment programme (PRIMAS) for people with type 1 diabetes in daily routine and RCT: a health care research trial N. Hermanns, N. Bergis-Jurgan, D. Ehrmann, T. Haak; Research Institute of Diabetes Academy Mergentheim (FIDAM), Bad Mergentheim, Germany. Background and aims: It´s a common result of health care research that ef- ficacy of treatments assessed in randomized trials (RCT) is superior to ef- ficacy of these treatments in routine care. We evaluated if a new education and teaching programme for type 1 diabetes patients (PRIMAS) had a similar efficacy in daily routine than in the RCT. Materials and methods: In this health care research trial (HCRT) 255 people with type 1 diabetes from 42 diabetologists practices participated (age 43 ± 14 yrs.; 54% male gender; diabetes duration 13 ± 12 years; HbA1c 8.1 ± 1.4%). Participants took part in the education and treatment programme (PRIMAS), which consisted of 12 lessons. The outcomes were assessed 6 months after the education programme was finished. In both trials HbA1c was assessed in a central laboratory. Patients also completed a Hypoglycaemia Aware- ness Questionnaire, the Center of Epidemiological Studies-Depression Scale (CESD), the Diabetes Distress Scale (DDS), the Diabetes Self-efficacy Scale and the Empowerment Scale. Central outcomes were the difference in HbA1c reduction between the RCT and this health care research trial. Secondary outcomes were improvements of the hypoglycaemia awareness score, depres- sive symptoms, diabetes related distress, self-efficacy and empowerment. Results: HbA1c showed an equivalent improvement in RCT and HCRT (-0.36 ± 1.1 vs. -0.39 ±1.2 percentage points, Δ 0.03 95% CI -0.27 to 0.33 per- centage points, p=.827). The confidence interval of HbA1c-differences was beneath the 0.4 % non-inferiority threshold as defined by the ADA. Similar results were obtained regarding the hypoglycaemia unawareness score (RCT -0.52 ± 1.42 vs. HCRT -0.36 ±1.36, p=.375), for depressive symptoms score (RCT -1.18 ± 7.93 vs. HCRT -1.85 ±8.49, p=.542), for diabetes distress score (RCT -0.32 ± 0.76 vs. HCRT -0.18 ±0.66, p=.180) for self-efficacy score (RCT +1.39 ± 3.56 vs. HCRT -1.11 ±5.88, p=.631) and for empowerment score (RCT +2.61 ± 5.93 vs. HCRT -2.34 ±7.01, p=.752). Conclusion: The PRIMAS diabetes education and treatment programme had similar effects under routine care conditions than observed in a RCT. Thus, PRIMAS can contribute to an improvement of routine health care in people with type 1 diabetes. Clinical Trial Registration Number: NCT01220557 Supported by: Berlin Chemie AG 16 Effects of a patient-oriented decision aid for prioritising treatment goals in diabetes, a randomised controlled trial P. Denig1,2, J. Schuling3,2, F.M. Haaijer-Ruskamp1,2, J. Voorham1,2; 1Clinical Pharmacy and Pharmacology, University Medical Center Groningen, 2University of Groningen, 3General Practice, University Medical Center Groningen, Netherlands. Background and aims: Decision aids (DA) can encourage patient-provider discussions about disease management by presenting available treatment op- tions and expected outcomes for the individual patient. The newest genera- tion DA for diabetes patients and providers prioritise between clinical do- mains to guide treatment decisions. Our aim was to assess the effects of such a patient-oriented DA in comparison to usual care on shared goal-setting and treatment decisions. Materials and methods: We conducted a randomised controlled trial includ- ing 18 general practices in the north of the Netherlands. Four presentation formats of the DA were compared to usual care. Practices were randomly allocated to a computer-screen or printed version of the aid. Within each practice, patients were randomised to receiving either the short or extended version of the aid or usual care. Effects on patient empowerment for making shared decisions about treatment goals (primary outcome), smoking status, prescribing of glucose-regulating, blood pressure-regulating, lipid-regulating and albuminuria-regulating drugs (secondary outcomes) were tested in intention-to-treat and per-protocol analyses. Data were collected through structured questionnaires and automated data extraction from electronic health records in 6 months before and after the intervention. Results: Of 665 eligible patients with type 2 diabetes being ≤65 years at time of diagnosis, 344 consented to participate; 225 were allocated to the interven- tion groups and 119 to the usual care group. The mean empowerment score increased 0.1 on a 5-point scale in the overall intervention group. This effect was not significantly different from the control group in the intention-to- treat analysis but was significant in the per-protocol analysis (n=103 patients, who received the DA as planned vs 119 control patients, p=0.031). Lipid- regulating drug treatment was intensified in 24% of intervention and 11% of control patients with elevated cholesterol levels, which was a significant dif- ference in the intention-to-treat (p=0.041) as well as the per-protocol analysis (p=0.038). No significant changes were seen for the other drug treatments or in smoking status. Conclusion: Our study showed that a patient-oriented treatment DA can lead to improved drug treatment and has potential for increasing patient em- powerment, provided it is used as intended. The effects of the DA may have been limited due to its single use within the setting of this randomised trial. Clinical Trial Registration Number: NTR1942 Supported by: ZonMW 17 Health economic implications of education-based flexible insulin therapy versus conventional or technology-based approaches in type 1 diabetes H. Zulewski1, M. Brändle2, A.-E. Minder1, B. Hunt3, W.J. Valentine3; 1University Hospital Basel, 2Kantonsspital St. Gallen, 3Ossian Health Economics and Communications, Basel, Switzerland. Background and aims: To estimate the long-term clinical and cost outcomes associated with flexible intensive insulin therapy (FIT) in relation to stand- ard management and a continuous subcutaneous insulin infusion with con- tinuous glucose monitoring (CGM-CSII) for patients with type 1 diabetes in Switzerland. Materials and methods: Evaluation of a long-term cost-effectiveness model with parameters estimated from three landmark type 1 diabetes trials: The Basel FIT study, STAR 3 and EDIC, with clinical outcomes extrapolated using the published and validated CORE Diabetes Model. Main outcome measures included life expectancy, quality-adjusted life expectancy, incidence of dia- betes-related complications, direct costs, and incremental cost-effectiveness ratios (ICERs). Results: FIT and CGM-CSII were associated with clinical benefits over standard management, improving life expectancy (14.43 years and 14.20 years versus 14.07 years) and quality-adjusted life expectancy (10.14 quality- adjusted life years [QALYs], and 9.97 QALYs versus 9.75 QALYs). Improve- ments were driven by reduced incidences of diabetes-related complications and severe hypoglycaemic events, as a result of improved glycaemic control Diabetologia (2014) 57:[Suppl1]S1–S564 S 14 1 C over standard management. FIT was associated with reduced direct medical costs over standard management (CHF 186,373 versus CHF 192,588), and therefore was considered to dominate standard management. Costs were higher in the CGM-CSII (CHF 245,983) arm as a result of increased costs of CSII pumps and CGM consumables, with an ICER of CHF 248,602 per QALY gained versus standard management. Conclusion: The education-based approach of FIT and the technology-based approach of CGM-CSII are both likely to produce improvements in clinical outcomes, but the high cost of CGM-CSII represents a barrier to its use, par- ticularly compared to the cost savings associated with FIT. Investment of phy- sician time in patient education may be more cost-effective than investment in CGM-CSII technology, in terms of optimizing management of patients with type 1 diabetes in Switzerland. Supported by: Novo Nordisk 18 Effects of multidisciplinary and structured education programme on glycaemic control during transition of young adults with type 1 diabetes A. Da Porto, E. Tommasi, S. Cum, E. Del Forno, E. Caroli, E. Manca, A. Petrucco, R. Candido; Diabetes Unit, ASS 1 Triestina, Trieste, Italy. Background and aims: The transition from pediatric to adult diabetes care represents a high risk period for a person with diabetes. Prior studies high- light the risk of gaps in medical follow-up and adverse diabetes related out- comes in emerging adults, including poor glycaemic control, appearance of long term diabetes complications and early mortality. Very little is known about the specific aspects of transitions process, in particular about the role of structured education during transition. Aim of our study was to evaluate the relationship between multidisciplinary structured education during transi- tion and post-transition glycaemic control. Materials and methods: We evaluated baseline clinical characteristics, gly- caemic control, glycaemic variability, hypoglycaemia frequency and their variation one year after multidisciplinary structured education program in 55 young adults with type 1 diabetes who have done transition process. Data for Glycaemic variability (SD of average glucose) and hypoglycaemia frequency were obtained from the analysis of glycaemic samples from per- sonal blood glucose meters during a period of 90 days. Multidisciplinary structured education was done at the first visit (and carried on during the following moths) by Diabetologist, Dietician and Nurse with particular atten- tion to CHO counting, hypoglycaemia correction and modulation of insulin therapy during exercise. Every 2 months correct use of CHO count or adher- ence to constant CHO diet were verified by dietician. Descriptive statistics were presented as means and SD or proportions. Chi square test was used to compare proportions. Multivariate logistic regression model was used to verify the association between correct management of CHO and glycaemic control at 1 year (variable included: age, sex, ∆ Hb1c ∆ glycaemic variability, ∆ hypoglycaemic frequency). Results: Average population age was 27,8±;10,1 years with mean diabetes du- ration of 17,3±9,9 years. Average transition GAP from the last pediatric dia- betes visit and first adult visit was 6, 24±9, 12 months. At the first visit to adult center mean Hba1c was 7, 9±1, 2% , on average worsen of 0,32 ±1,8% during transition GAP. Glycaemic variability was 83,1±23 mg/dl and hypoglycae- mic frequency was 12,6 ±8,2%. Only 23,6% of patients correctly apply CHO counting and 11% employ constant CHO diet at baseline. Education was ac- cepted by 85% of patients. One year later 34,6% of patients correctly apply CHO counting and 25,2% employ constant CHO diet (p 0.014 and 0,037 vs baseline). Hb1c and glycaemic variability decreased significantly from base- line (-0,72±;0,65% p 0.009 and - 12% p 0,041 respectively) . Hypoglycaemia frequency decreased not significantly (-2% p 0,124). Correct CHO manage- ment was associated to reduction of Hb1c (r 0,293 p 0,012) less glycaemic variability (r - 0,366 p =0,021) and lower hypoglycaemia frequency (r - 0,534 p 0,032) after 12 months in univariate model but loose statistical relevance in multivariate one. Baseline Hb1ac was the only predictor of glycaemic control after 12 months of follow-up (R =0,0659 p 0,0001). Conclusion: In our study sample age of transition and transition GAP was high, during transition GAP glycaemic control worsen. A multidisciplinary structured education during transition, if accepted by patient , lead to a bet- ter CHO management that is associated to a better postransition glycaemic control. OP 04 Brown adipose tissue 19 Impaired brown adipose tissue fatty acid metabolism in obese subjects T. Saari1, J. Raiko1, T. Niemi2, M. Taittonen3, J. Laine4, N. Savisto1, M. Haaparanta-Solin1, P. Nuutila1, K. Virtanen1; 1Turku PET Centre, 2Department of Plastic and General Surgery, 3Department of Anesthesiology, 4Department of Pathology, Turku University Hospital, Finland. Background and aims: Brown adipose tissue (BAT) activation is commonly found during cold exposure: BAT glucose uptake and blood flow are in- creased in cold conditions. Obese subjects have a blunted glucose metabolism in BAT. The role of fatty acid metabolism in BAT function is uncertain. The aim of this study was to quantify NEFA uptake using PET-CT in lean and obese subjects during cold exposure and in room temperature (RT), and to measure the differences in NEFA uptakes between lean and obese subjects. Materials and methods: NEFA uptake was quantified in lean (n=12, 5F/7M, ages: 34.4±11.8 years, BMI: 24.1±1.4 kg/m2) and obese subjects (n=9, 4F/5M, ages: 45.7±5.8 years, BMI 32.6±4.4 kg/m2) using 18F-FTHA, a palmitate analog. Each subject was imaged twice, once at RT and once during cold ex- posure. Cold exposure was started 2 hours before the imaging session. Tissue specific NEFA uptakes were calculated for supraclavicular adipose deposits, deltoid muscles and subcutaneous white adipose tissue (WAT). A biopsy was performed to confirm the existence of BAT in the supraclavicular area. Sub- jects were grouped to BAT positive (BAT+) and BAT negative (BAT-) groups based on their biopsy result. Biopsy site was determined from CT images. Results: Three lean subjects (2F/1M) had biopsy proven BAT. They had 7-fold higher NEFA uptake during cold compared to obese¬ subjects (5.56±1.15 vs. 0.74±0.26 µmol/100g/min, P=0.01) and 4-fold compared to lean BAT- sub- jects (1.43±1.00 µmol/100g/min, P=0.02 vs. BAT+). Cold exposure increased BAT NEFA uptake in BAT+ subjects 3-fold compared to RT (P=0.02). BAT+ subjects had higher BAT NEFA uptake in RT compared to obese subjects (1.69±0.57 vs. 0.57±0.55 µmol/100g/min, P=0.01). BAT FA uptake in cold was higher in lean BAT- subjects compared to all obese subjects (P=0.01). Lean and obese groups had similar muscle and WAT NEFA uptakes in cold and RT. Conclusion: Cold stimulation increases NEFA uptake of BAT, but not skel- etal muscles or WAT. All subjects who had functionally active BAT were lean, and all lean subjects had higher tissue activity during cold exposure in the supraclavicular area compared to the obese subjects. BAT+ subjects had in- creased BAT NEFA uptake in RT. This would suggest that BAT takes in NEFA even without cold exposure. These results suggest that obese subjects have an impaired NEFA uptake in the supraclavicular area, typical BAT region. Supported by: ACADFI, EU FP7 DIABAT, TDRF, FCF 20 Cold exposure increases the oxygen uptake rate of brown adipose tissue in humans M. U Din1, J. Raiko1, T. Saari1, T. Tolvanen1, V. Oikonen1, J. Teuho1, H. Sipilä1, N. Savisto1, P. Nuutila1,2, K.A. Virtanen1; 1Turku PET Centre, Turku University Hospital and University of Turku, 2Turku University Hospital, Department of Endocrinology, Finland. Background and aims: Brown adipose tissue (BAT) activity has been sug- gested to have direct relation with insulin sensitivity and inverse association with obesity. BAT is more active during cold environment as it produces heat by the oxidation of glucose and fatty acids, a phenomenon known as non- shivering thermogenesis. Oxygen uptake rate can be a marker of BAT ther- mogenic potential as an indicator of oxidative metabolism. In our present study we aimed to determine whether cold exposure increases the oxygen uptake rate in BAT and how much BAT oxygen uptake rate is associated with blood flow and NEFA uptake rate during cold. Materials and methods: Healthy lean and obese study subjects (n = 8, age: 34.5 ± 11.0 years, BMI range: 23.2 - 31.1 kg/m2) of both genders (3F/5M) were studied at two different scanning sessions, 1) at room temperature (RT) and 2) with acute cold exposure, using PET-CT. Radioactive oxygen [15O]O2, [15O]H2O and [ 18F]FTHA were given for the measurements of oxygen uptake rate, blood flow and NEFA uptake rate of BAT, respectively. Indirect calorim- etry was performed to assess the differences in whole body energy expendi- Diabetologia (2014) 57:[Suppl1]S1–S564 S 15 1 C ture between RT and cold environment. Data was compared with two tailed T-test and correlation analysis were performed with Pearson’s correlations. Results: Cold exposure increased oxygen uptake rate by 22.4 % (from 2.21 ± 0.68 to 2.85 ± 0.59 ml/100g/min, P = 0.001) in 16 BAT regions within 8 subjects, while oxygen uptake tended to be higher during RT in deltoid mus- cle (2.55 ± 1.52 vs. 3.13 ± 1.85 ml/100g/min, cold vs. RT, P = 0.06); however, no significant difference was found in white adipose tissue (1.76 ± 0.21 vs. 1.78 ± 0.19 ml/100g/min, cold vs. RT, P = 0.76). Cold exposure also tended to increase whole body energy expenditure (from: 7.07 ± 1.10 to 8.24 ± 2.43 MJ/24h, P = 0.09). During cold exposure, oxygen uptake rate correlated with blood flow (r = 0.51, P = 0.05) and NEFA uptake rate in BAT (r = 0.76, P = 0.01, n = 5). Conclusion: High oxygen demand during cold in brown adipose tissue is most likely due to increased oxidation of available substrates (either glucose or fatty acids) for thermogenesis. Positive correlation between NEFA and oxygen uptakes suggests that NEFAs that enter brown adipocytes during cold also undergo β-oxidation within mitochondria. Supported by: Academy of Finland, EU FP7 DIABAT, The Diabetes Research Foundation, SKR 21 Loss of sympathetic drive may explain loss of brown adipose tissue activity in elderly but not in obese males L. Bähler, H.J. Verberne, W.M. Admiraal, M.R. Soeters, J.B.L. Hoekstra, F. Holleman; Academic Medical Center, Amsterdam, Netherlands. Background and aims: Metabolically active brown adipose tissue (BAT) could facilitate weight loss by increasing energy expenditure. Cold is a potent stimulator of BAT, activating BAT primarily through the sympathetic nerv- ous system (SNS). Old or obese individuals have less metabolic BAT activity than lean and young, but the role of the SNS in this decrease is unknown. We determined whether the lower metabolic BAT activity can be explained by a lower SNS response to cold. Materials and methods: We studied 10 young obese (26 [21-31] years, BMI 32 [31-39] kg/m2), 10 old lean (55 [51-60] years, BMI 23 [22-25] kg/m2) and 14 young lean (26 [21-28] years, BMI 22 [21-23] kg/m2) males. Metabolic BAT activity was measured as maximal standardised uptake value and vol- ume (SUVmax, SUVvol) on 18F-Fluorodeoxyglucose positron emission to- mography CT (FDG). SNS activity was measured as semiquantative uptake values of 123I-metaiodobenzylguanidine (MIBG) on single photon emission computed tomography scans, with the mediastinum as reference region. Scans were made after an overnight fast and 2 hours of cold exposure. Results: Metabolic BAT activity and volume were different between young vs old (median [IQR] SUVmax (g/L) 7.9[4.2-17.3] vs 3.0[0.0-4.2] p S 16 1 C 23 Brown adipose tissue activity associates with insulin sensitivity and liver adiposity independently of visceral fat mass in healthy adults J.R.H. Raiko1, T. Saari1, V. Saunavaara1, R. Parkkola2, P. Nuutila1,3, K. Virtanen1; 1Turku PET Centre, University of Turku, 2Department of Radiology, Tampere University Hospital, 3Department of Endocrinology, Turku University Hospital, Finland. Background and aims: Brown adipose tissue (BAT) activity associates with insulin sensitivity and inversely with obesity. However, it remains unclear if the association between BAT and insulin sensitivity is mediated by low intra- abdominal adipose tissue mass and low liver adiposity in subjects with active BAT. In this study, we examined associations between BAT activity during cold exposure and glycaemic control, whole body insulin sensitivity, intra- abdominal adipose tissue mass and liver adiposity. Materials and methods: 22 healthy adult subjects (M/F: 8/14, 41.0 ± 8.3 years, BMI 27.3 ± 5.3 kg/m2) underwent positron emission tomography with [15O]-H2O to measure BAT perfusion and [ 18F]-FTHA to measure NEFA uptake in the supraclavicular BAT depot during cold exposure. Cold expo- sure had started two hours prior to and continued during the PET imag- ing. Whole body insulin sensitivity (hyperinsulinemic euglycemic clamp) and HbA1c were measured. MRI imaging was used for the measurement of intra-abdominal adipose tissue volume and liver adiposity was adjusted with MR spectroscopy. Results: Cold-induced NEFA uptake in BAT was 2.25±1.75 umol/100g/min and perfusion was 22.0±27.0 ml/100g/min. HbA1cwas 5.4±0.4%, M value was 6.8±3.2 mg/kg/min, visceral adipose mass was 3.3±2.0 kg and liver fat con- tent was 5.7±6.9%. NEFA uptake in BAT correlated with whole body insulin sensitivity (r = 0.50, P = 0.02), intra-abdominal fat mass (r = -0.43, P = 0.04) and liver adiposity (r = -0.47, P=0.04). Cold-induced BAT perfusion correlat- ed with HbA1c (r = -0.57, P = 0.01). These associations (perfusion vs. HbA1c, and NEFA uptake vs. whole body insulin sensitivity) were independent of intra-abdominal fat volume or liver adiposity (P always < 0.05). Associations between NEFA uptake and liver adiposity were also independent of intra- abdominal adipose tissue volume (P = 0.03). Conclusion: These results show that cold-induced BAT activity is associated with glycaemic control and insulin sensitivity independently of visceral adi- pose tissue mass and liver adiposity in health. Subjects with high BAT activ- ity might have lower visceral fat mass due to increased energy expenditure. Additionally, BAT NEFA uptake might prevent fat accumulation in liver in visceral obesity. Although, liver fat content and visceral obesity correlate in- versely with BAT activity, BAT may associate with insulin sensitivity via a separate mechanism. These may include yet unrecognized BAT-derived cy- tokines affecting muscle metabolism. Supported by: DIABAT, Acad. Finland 24 Inorganic nitrate promotes the browning of white adipose tissue through the nitrate-nitrite-nitric oxide pathway L.D. Roberts1, T. Ashmore2, A.O. Kotwica2, S.A. Murfitt3, B.O. Fernandez4, M. Feelisch4, A.J. Murray2, J.L. Griffin1; 1Medical Research Council - Human Nutrition Research, University of Cambridge, 2Physiology, Development and Neuroscience, University of Cambridge, 3Biochemistry, University of Cambridge, 4Faculty of Medicine, Clinical & Experimental Sciences, University of Southampton, UK. Background and aims: Inorganic nitrate is considered an oxidation end- product of nitric oxide with little biological activity. However, recent studies demonstrate that dietary nitrate, largely derived from green leafy vegetables, modulates mitochondrial function in man and is effective in reversing fea- tures of the metabolic syndrome in mice. The role of nitrate in white adipose tissue (WAT) function remains poorly characterized. The development of a brown-like phenotype in white adipocytes (“beige” or “brite” cells), a process known as “browning”, includes the induction of thermogenesis, the dissipa- tion of chemical energy as heat. Activation of the browning response in WAT may represent a process underlying the altered systemic energy balance ob- served with nitrate treatment. Materials and methods: Wistar rats were treated with 0.35 mM, 0.7 mM or 1.4 mM sodium nitrate (NaNO3) in drinking water for 18 days (n = 6/ group). Nitrate induced browning of WAT was assessed using both RT qPCR and mass spectrometry based metabolomics. In vitro browning of primary adipocytes treated with 25 μM, 50 μM and 500 μM NaNO3 was analysed us- ing respirometry, metabolomics, stable-isotope labelling techniques and RT qPCR. The mechanism of nitrate induced browning was defined in primary white adipocytes using pharmacological inhibitors of nitric oxide (NO), gua- nylate cyclase and protein kinase G (PKG). The effect of hypoxia on nitrate induced browning of WAT was determined in vivo in Wistar rats and in vitro using primary adipocytes. Results: Nitrate dose-dependently increased the expression of brown-adipo- cyte specific genes in WAT of nitrate treated rats (UCP-1, CIDEA, PGC-1α, CYCS, CPT1, ACADvl, Two-way ANOVA, Control vs. 0.35 mM NaNO3 P ≤ 0.01, Control vs. 0.7 mM NaNO3 P ≤ 0.0001, Control vs. 1.4 mM NaNO3 P ≤ 0.001). Metabolomic analysis demonstrated that treatment with nitrate also decreased the total triacylglycerol content within WAT (0.95-fold, 0.35 mM NaNO3; 0.94-fold, 0.7 mM NaNO3; 0.98-fold, 1.4 mM NaNO3; ANOVA, Control vs. 0.35 mM P ≤ 0.01, Control vs. 0.7 mM P ≤ 0.01). These find- ings were reproduced in vitro using primary white adipocytes treated with nitrate and analysed using RT qPCR, stable isotope labelling and metabo- lomics, and were accompanied by an increase in oxygen consumption, as- sessed by respirometry (Control 4.2 nmoles O2/min/106 cells, 50 μM NaNO3 6.6 nmoles O2/min/106 cells, 500 μM NaNO3 7.7 nmoles O2/min/106 cells, ANOVA, P = 0.02)). Using pharmacological inhibitor assays in primary white adipocytes, nitrate was found to induce these phenotypic changes in WAT independently of NO synthase, through the recently identified xanthine oxi- doreductase catalysed nitrate-nitrite-NO pathway and downstream cGMP/ PKG signalling. Furthermore, the nitrate induced browning of WAT, both in vitro and in vivo, was enhanced in hypoxia. Conclusion: Since beige/brite cells exhibit anti-diabetic and anti-obesity ef- fects and WAT from obese individuals is characterized by hypoxia, nitrate may be an effective means of targeting the induction of browning to white adipocytes located in obese and hypoxic adipose depots to treat the metabolic syndrome. Supported by: MRC-HNR Elsie Widdowson Fellowship Diabetologia (2014) 57:[Suppl1]S1–S564 S 17 1 C OP 05 Factors driving islet cell development 25 G1 lengthening promotes pancreatic progenitor cell differentiation in mouse embryonic development N.A.J. Krentz, M. Tang, A. Watanabe, F.C. Lynn; University of British Columbia, Vancouver, Canada. Background and aims: Cellular-based therapies for diabetes mellitus, such as the differentiation of human embryonic stem cells (hESCs), require an in-depth understanding of pancreatic development. During early pancreatic development, the Pdx1+Cpa1+ tip multipotent progenitor cells give rise to all three cell types of the pancreas, exocrine, endocrine and ductal cells, while the trunk Pdx1+Cpa1- cells give rise to endocrine or ductal cells only. The process that regulates the proliferation and differentiation of these progeni- tors populations is not fully known. There is evidence during neurogenesis that the length of the G1 phase of the cell cycle can directly influence the differentiation of neural precursors. Thus, we hypothesize that the cell cycle length regulates pancreas organogenesis. Materials and methods: Pregnant CD-1 mice (P12.5) were injected with the thymidine analog 5’Ethynyl-2’-deoxyuridine (EdU) every 1.5 hours starting at 9 am. Embryos were collected at evenly spaced intervals from 9:30 am to 8 pm. The number of Pdx1+Cpa1+ and Pdx1+Cpa1- cells labelled with EdU was determined using immunofluorescence and confocal microscopy (n>4). The lengths of the G1, S and G2/M phases were determined mathematically from the length of time required to label all dividing cells, the number of cells labeled at time 0, and the proportion of dividing cells. Results: We determined that the G1 length of multipotent progenitor cells at E12.5 was 3.6 hours while the bi-potent trunk cells had a G1 length of 5.8 hours. We did not find any significant change in the lengths of either the S or G2/M phases of the cell cycle. Consistent with findings in neural develop- ment, the difference in total cell cycle length of these two progenitor popu- lations at E12.5 was largely due to a lengthening of the G1 phase. Intrigu- ingly, the maximal proportion of EdU labeled cells or the growth fraction was significantly different (p S 18 1 C Conclusion: We have investigated the expression and potential function of MEN1 in the early to mid-gestation human fetal pancreas. Functional knock- down and overexpression studies both suggest that Menin promotes human fetal pancreatic cell survival and proliferation, maintains the islet progenitor pool and regulates islet differentiation, very different from its anti-tumori- genic role in the adult pancreas. Supported by: CIHR 28 SOX9 and WNT signalling during human foetal pancreatic development R. Wang1, J.A. Belo1, J. Li1, G.F. Fellows2, C.G. Goodyer3; 1Physiology & Pharmacology, Western University, 2Obstetrics and Gynaecology, Western University, London, 3Pediatrics, McGill University, Montreal, Canada. Background and aims: Regulation of pancreatic progenitor proliferation and differentiation is crucial when generating an appropriate beta-cell mass. SOX9 is an important factor in pancreatic progenitor maintenance as well as endocrine differentiation; however, the signaling pathways that regulate expression of SOX9 remain unclear. A study of murine duodenal epithelium demonstrated that Wnt signaling regulated progenitor expansion and differ- entiation in addition to Sox9 expression. In the present study, the co-locali- zation and inter-relationship between SOX9 and Wnt/β-catenin factors and their targets in the developing human pancreas were examined. Materials and methods: Human fetal pancreata (8-21 weeks fetal age) were examined for SOX9 and Wnt signaling molecules using immunofluores- cence, western blot and qRT-PCR approaches. Isolated human fetal (18-21 week) islet-epithelial cell clusters were also treated with or without recombi- nant WNT3A in a dose- and time-dependent fashion. In addition, cells were treated with either a GSK3β inhibitor (1-Akp) or a Wnt signaling inhibitor (FZD8-CRD). Results: Half of the SOX9+ cells expressed WNT3A at 8-12 weeks but the numbers decreased with age (p S 19 1 C captured alpha cells (RPKM GCG >100, INS, PPY and SST 100, PPY and SST S 20 1 C genital insulin deficiency could extend lifespan and reverse hyperglycaemia. Ins1-/-Ins2-/- (InsKO) mice were maintained on twice daily insulin injections until 14 days of age then transplanted with ~100-150 islets into the anterior chamber of the eye. At ~10 weeks of age InsKO mice were treated daily with an i.p. injection of 10 µg/day PEG-ylated leptin or vehicle and vehicle treated Ins1-/-Ins2+/- (Het) littermates were used as controls. On day 4 of treatment the transplanted eye was enucleated to render the mice entirely insulin defi- cient. Survival, body weight, and fed and fasting blood glucose were assessed. Results: In STZ-treated mice both leptin and insulin treatment reduced fast- ing blood glucose by day 2 compared to vehicle treated controls (22.6±0.7 mM STZ-vehicle, 10.8±2.0 mM STZ-leptin, 12.5±2.0 mM STZ-insulin). In- jection of S961 increased blood glucose in the insulin treated mice after 2 hours (17.9±4.5 mM STZ-insulin + S961, 4.8±0.5 mM STZ-insulin + vehi- cle, P S 21 1 C that of LDL particles increased (0.12 SD units, P=0.008). Cholesterol content in very large HDL particles decreased (-0.36 SD units, P=3.12e-6) but rose in medium-sized particles (0.19 SD units, P=0.016). Healthy obese MZ co- twins (without IR) had lower HDL cholesterol levels throughout the OGTT, especially in large HDL and HDL2 (P S 22 1 C Conclusion: The addition of LIRA to basal insulin analogues ± metformin significantly improved glycaemic control, which can be attributed to the ef- fect of LIRA on both FPG and post-prandial glucose levels. Additionally, LIRA induced greater weight loss and a reduction in systolic blood pressure and selected lipids compared to PLAC. Typical gastrointestinal symptoms and minor hypoglycaemia were more frequent with LIRA than PLAC. No severe hypoglycaemic events were reported during this trial. Clinical Trial Registration Number: NCT01617434 Supported by: Novo Nordisk 38 Efficacy and safety of once weekly dulaglutide vs insulin glargine in combination with metformin and glimepiride in type 2 diabetes patients (AWARD-2) F. Giorgino1, M. Benroubi2, J.-H. Sun3, A.G. Zimmermann4, V. Pechtner5; 1Endocrinology & Metabolic Diseases, University of Bari, Italy, 2Evangelismos-Polyclinic, Athens, Greece, 3Chang Gung Memorial Hospital, Taoyuan Hsien, Taiwan, 4Eli Lilly and Company, Indianapolis, USA, 5Eli Lilly and Company, Neuilly-Sur-Seine Cedex, France. Background and aims: This Phase 3, 78-week, parallel-arm, open-label (double-blinded to dulaglutide [DU] doses) study compared two doses of the once weekly GLP-1 receptor agonist DU with insulin glargine titrated to fast- ing glucose target, in patients with type 2 diabetes inadequately controlled with maximally tolerated doses of metformin and glimepiride. Metformin and glimepiride were to be continued throughout the study. Materials and methods: Patients (N = 807; mean baseline characteristics: age 57 years; duration of diabetes 9.1 years; HbA1c 8.1%; body weight 86.3 kg; BMI 31.6 kg/m2) were randomised (1:1:1) to once weekly DU 1.5 mg, DU 0.75 mg, or once daily insulin glargine. The primary objective was to demonstrate DU 1.5 mg was noninferior (margin 0.4%) to insulin glargine for HbA1c change from baseline at 52 weeks. Additional analyses were carried out at 52 and 78 weeks. Results: At 52 weeks, DU 1.5 mg was superior and DU 0.75 mg was nonin- ferior to insulin glargine on HbA1c change from baseline. The mean insulin glargine dose was 29.4 U. Body weight decreased with both DU doses and increased with insulin glargine. Over the 52-week period, the mean rate of documented symptomatic hypoglycaemia (≤3.9 mmol/L) was 2.0, 2.0, and 3.3 events/patient/year for DU 1.5 mg, DU 0.75 mg, and insulin glargine, re- spectively. At 78 weeks, HbA1c, body weight, and hypoglycaemia results were similar to 52 weeks, and the mean insulin glargine dose was 31.4 U. Through 78 weeks, four events of severe hypoglycaemia occurred: 2 in patients treated with DU 1.5 mg and 2 in patients treated with insulin glargine. Nausea and diarrhoea were more common with DU 1.5 mg (15.4% and 10.6%) and DU 0.75 mg (7.7% and 9.2%) versus insulin glargine (1.5% and 5.7%) through 78 weeks. Conclusion: DU 1.5 mg demonstrated superior and DU 0.75 mg noninferior glycaemic control compared with insulin glargine, and this was associated with weight loss, reduced incidence of hypoglycaemia, and an acceptable safety profile. Clinical Trial Registration Number: NCT01075282 Supported by: Eli Lilly and Company 39 Liraglutide 3.0 mg for weight management in obese/overweight adults with type 2 diabetes: SCALE diabetes 56-week randomised, double-blind, placebo-controlled trial M.J. Davies1, R. Bergenstal2, B. Bode3, R. Kushner4, A. Lewin5, T.V. Skjøth6, C.B. Jensen6, R. DeFronzo7; 1University Hospitals of Leicester, UK, 2Park Nicollet Institute, St Louis Park, USA, 3Atlanta Diabetes Associates, USA, 4Northwestern University, Chicago, USA, 5National Research Institute, Los Angeles, USA, 6Novo Nordisk A/S, Søborg, Denmark, 7Texas Diabetes Institute, San Antonio, USA. Background and aims: Liraglutide at doses up to 1.8 mg is approved for the treatment of T2D. This study investigated the efficacy and safety of liraglutide 3.0 mg and 1.8 mg, as adjunct to diet and exercise, for weight management in obese/overweight adults with T2D. Materials and methods: In this 56-week, randomised, double-blind, place- bo-controlled trial, adults with T2D (on diet and exercise alone or with 1-3 oral antidiabetic drugs [OADs], HbA1c 7-10%, BMI ≥27.0 kg/m2) were ran- domised 2:1:1 to receive liraglutide 3.0 mg, 1.8 mg or placebo. All subjects received diet (500 kcal/day deficit) and exercise instruction. Results: 846 individuals were randomised: age 54.9 (18.0-82.0) years, 50% male, BMI 37.1 (27.0-67.6) kg/m2, HbA1c 7.9% (6.4-10.3%), fasting plasma glucose (FPG) 8.8 (4.2-17.3) mmol/l, T2D for 7.3 (0.2-36.5) years, 11.5% on diet and exercise, 57.3% on metformin only, 31.2% on combination OADs. Liraglutide 3.0 mg and 1.8 mg were superior to placebo, and 3.0 mg was superior to 1.8 mg on mean and categorical weight loss at week 56 (Table). Liraglutide 3.0 mg also achieved superior glycaemic control vs. placebo and liraglutide 1.8 mg (change in HbA1c and FPG, proportion reaching HbA1c ≤6.5%, and [vs. placebo only] postprandial plasma glucose [PG] increment; Table). The safety profiles with liraglutide 3.0 mg and 1.8 mg were similar, although gastrointestinal disorders were more frequent with 3.0 mg (65% of individuals) than 1.8 mg (56%) and placebo (39%). No cases of pancreatitis were reported during the trial. An increase in mean serum lipase activity was seen with liraglutide 1.8 mg and 3.0 mg; the increase was not dose-dependent and few individuals (7.7% and 9.8% on liraglutide 1.8 mg and 3.0 mg, respec- tively, vs. 6.3% on placebo) had levels ≥3 times the upper normal range at any time during treatment. Rates of documented symptomatic hypoglycaemia (PG S 23 1 C Conclusion: Liraglutide 3.0 mg, as adjunct to diet and exercise, was effica- cious and well-tolerated for weight management over 56 weeks in obese/ overweight individuals with T2D. Clinical Trial Registration Number: NCT01272232 Supported by: Novo Nordisk 40 Efficacy and safety of once weekly dulaglutide versus once daily liraglutide in type 2 diabetes (AWARD6) S. Tofe Povedano1, K.M. Dungan2, T. Forst3, J.G. González González4, C. Atisso5, W. Sealls5, J.L. Fahrbach5; 1Endocrinologia, Clinica Juaneda, Palma de Mallorca, Spain, 2The Ohio State University, Columbus, USA, 3Profil Mainz GmbH & Co. KG, Mainz, Germany, 4Universidad Autόnoma de Nuevo Leόn, Monterrey, Mexico, 5Eli Lilly and Company, Indianapolis, USA. Background and aims: This Phase 3, randomised, open-label, parallel-arm 26-week (wk) study compared the efficacy and safety of once weekly dula- glutide (DU) 1.5 mg, a long-acting GLP-1 receptor agonist, with once daily liraglutide (LIRA) 1.8 mg in metformin-treated (≥1500 mg) patients with type 2 diabetes. Materials and methods: Patients (N=599; mean baseline age, 57 years; HbA1c, 8.1 %; weight 94.1 kg) were randomised to DU 1.5 mg or LIRA 1.8 mg in a 1:1 ratio. The primary objective was HbA1c change from baseline at 26 wks tested for noninferiority (margin 0.4%); DU 1.5 mg vs LIRA 1.8 mg. Results: DU 1.5 mg was noninferior to LIRA 1.8 mg at 26 wks as measured by HbA1c change from baseline (between-group HbA1c change -0.06; 95% CI [-0.19, 0.07]) (Table 1). While both groups experienced significant weight reduction, LIRA-treated patients demonstrated a 0.71 kg greater weight re- duction than DU-treated patients (p=0.01). The most common treatment- emergent gastrointestinal adverse events for DU 1.5 mg and LIRA 1.8 mg, respectively, were nausea (20.4%, 18.0%), diarrhoea (12.0%, 12.0%), dyspep- sia (8.0%, 6.0%), and vomiting (7.0%, 8.3%). Patients who discontinued study and/or study drug due to gastrointestinal adverse events were similar (DU 1.5 mg [3.0%], LIRA 1.8 mg [4.3%]). Rates of hypoglycaemia (≤3.9 mmol/L ± symptoms) were 0.34 events/pt/yr (DU 1.5 mg) and 0.52 (LIRA 1.8 mg) events/pt/yr. No severe hypoglycaemia was reported. Conclusion: Once weekly DU 1.5 mg demonstrated noninferior glycaemic control compared to once daily LIRA 1.8 mg with a similar safety and toler- ability profile. Clinical Trial Registration Number: NCT01624259 Supported by: Eli Lilly and Company 41 Harmony 1 year 3 Results: albiglutide vs placebo in patients with type 2 diabetes mellitus not controlled on pioglitazone (pio) ± metformin (met) C.M. Perkins1, B.W. Bode2, M.W. Stewart3, D.T. Cirkel4, F. Yang3, C.R. Perry3, P.D. Ambery5; 1PPD, Morrisville, 2Atlanta Diabetes Associates, 3GlaxoSmithKline, King of Prussia, USA, 4GlaxoSmithKline, Stevenage, 5MedImmune, Cambridge, UK. Background and aims: This 3 year (y), randomized, double-blind, placebo (Pbo) controlled study evaluated efficacy & safety of once weekly GLP-1 re- ceptor agonist albiglutide 30 mg vs Pbo in patients (pts) inadequately con- trolled (A1c 53-85.8 mmol/mol [7-10%]) on Pio ± Met. Materials and methods: Pts could continue if hyperglycaemic rescue was needed. Primary endpoint (PE) was A1c change from baseline at week 52. Results: Baseline demographics were similar between groups; mean A1c 65 mmol/mol (8.1%); age 55 y; 80% on Pio + Met. PE showed Albi superior to Pbo (treatment difference: -8.2 mmol/mol [−0.75%], P S 24 1 C Materials and methods: Patients (N = 884; mean baseline characteristics: age 59.4 years; duration of diabetes 12.7 years; HbA1c 8.5%; body weight 91.1 kg; BMI 32.5 kg/m2; total daily insulin dose 56 U) were randomised (1:1:1) to once weekly DU 1.5 mg, DU 0.75 mg, or bedtime insulin glargine titrated- to-target. The primary objective was to compare the change in HbA1c from baseline of DU 1.5 mg with insulin glargine at 26 weeks for noninferiority (margin 0.4%) and if met, then superiority was tested. Results: At 26 and 52 weeks, both DU doses were statistically superior to insulin glargine for HbA1c change from baseline. Insulin glargine was associ- ated with greater fasting serum glucose reduction compared with both DU doses. The mean prandial insulin doses at 26 weeks were 93 U for DU 1.5 mg, 97 U for DU 0.75 mg, and 68 U for insulin glargine. The insulin glargine dose was 65 U. Similar insulin doses were observed at 52 weeks. Body weight decreased with DU 1.5 mg and increased with DU 0.75 mg and insulin glar- gine at 52 weeks. The rate of documented symptomatic hypoglycaemia (≤3.9 mmol/L) at 52 weeks was 31.0, 35.0,and 39.9 events/patient/year for DU 1.5 mg, DU 0.75 mg, and insulin glargine, respectively. The number of severe hypoglycaemia events was 11 for DU 1.5 mg, 15 for DU 0.75 mg, and 22 for insulin glargine. Nausea, diarrhoea, and vomiting were more common with DU 1.5 mg (25.8%, 16.6%, and 12.2%, respectively) and DU 0.75 mg (17.7%, 15.7%, and 10.6%) versus insulin glargine (3.4%, 6.1%, and 1.7%). Conclusion: DU compared to insulin glargine, both combined with insu- lin lispro, resulted in better glycaemic control, less body weight gain, no in- creased risk of hypoglycaemia, and more common reporting of gastrointes- tinal adverse events. Clinical Trial Registration Number: NCT01191268 Supported by: Eli Lilly and Company OP 08 Matters of the heart 43 Does diabetes affect the efficacy of dual antiplatelet therapy in patients with acute coronary syndromes? M. Samos1, L. Duraj2, M. Fedor2, F. Kovar1, P. Galajda1, J. Fedorova3, J. Stasko1, T. Bolek1, M. Mokan1, P. Kubisz1, M. Mokan1; 1Jessenius Faculty of Medicine in Martin, Comenius University in Bratislava, 2Department of Hematology and Blood Transfusion, Jessenius Faculty of Medicine in Martin, Comenius University in Bratislava, 3HemoMedika, Center of Thrombosis and Hemostasis, Martin, Slovakia. Background and aims: Diabetes is a strong, independent risk factor of acute coronary syndromes (ACS). Diabetes aggravates the course of ACS and in- creases the risk of its complications. Recently, there is growing amount of data about failure in antiplatelet response, which is specifically associated with insulin resistance and diabetes. This incomplete antiplatelet response may contribute to a worse prognosis of ACS in diabetic patients. The aim of this study was to clarify the impact of presence of diabetes on the efficacy of dual antiplatelet therapy given in standard doses in patients with ACS. Materials and methods: 82 patients with ACS (53 men, 29 women, mean age 65 years) were enrolled in this preliminary prospective observational study. Patients were treated with aspirin loading dose (400 mg) and ADP recep- tor antagonist loading dose: in 62 patients clopidogrel (600 mg) and in 20 patients prasugrel (60 mg) was used. 21 patients had diabetes. Coronary an- giography and percutaneous coronary intervention of culprit coronary lesion was subsequently performed. Light transmission aggregometry (LTA) with specific inducers and VASP phosphorylation assessment was chosen for anti- platelet therapy efficacy testing. Samples were taken after first maintance dose administration (sample 1) and on 30th day from loading dose administration (sample2). Results: Mean LTA measured platelet reactivity was 32.7±18.9% in sample1 and 28.4±14.4% in sample2 respectively. No significant difference in antiplate- let response on ADP receptor antagonist (ADP-RA) between diabetic and non-diabetic patients was found (sample1: 30.7±19.8% versus 33.5±18.8%; sample2: 28.4±12.1% versus 28.4±15.7%). Totally 38 non-responders on ADP-RA were identified; ADP-RA unresponsiveness was not associated with the presence of diabetes. Prasugrel therapy generally showed better plate- let inhibition than clopidogrel administration (sample1: 22.8±13.7± versus 35.8±19.3%, p S 25 1 C 2011 (median 22.5 months). Differences in background characteristics were adjusted for in a logistic regression model. Results: Patients with diabetes were younger (75 vs. 77 years) and more often had preserved renal function (>60 ml/min; 44 vs. 38%), however hyperten- sion was more common in those with diabetes (59 vs. 45%). EF did not dif- fer, 17% in both groups had an EF ≥50%, however those with diabetes had more of severe heart failure symptoms (NYHA III-IV; 53 vs. 46%). Among those with diabetes, 88% received beta-blockade, 61% ACE inhibitors, 67% Statins and 71% Aspirin. Kaplan-Meyer curves of mortality are presented in Figure 1. The unadjusted and adjusted* ORs (95% CI) for mortality were 1.32 (1.24-1.41) and 1.71 (1.56-1.86). In those revascularised (50%), unadjusted and adjusted* ORs for mortality were 1.52 (1.39-1.67) and 1.63 (1.45-1.84). Conclusion: Diabetes is an independent predictor of long-term mortality in patients with ischemic heart failure. Revascularisa- tion did not abolish the impact of diabetes. The use of anti-ischem- ic secondary preventive treatment seemed somewhat low consid- ering all patients were classified to have ischemic heart disease * Adjusted for gender, age, duration of heart failure, weight, blood pressure, hypertension, atrial fibrillation, pulmonary disease, EF class, revascularisa- tion, eGFR class, Hb class and pharmacological treatment. Figure 1. Kaplan Meyer-curve showing long-term survival by diabetes (DM) in patients with ischemic heart failure. Odds ratio (OR), *=adjusted 45 Impact of diabetes mellitus on long-term prognosis in patients with preserved heart failure: a report from the Swedish Heart Failure Registry (S-HFR) I. Johansson1, M. Edner1, L. Rydén1, P. Näsman2, U. Dahlström3,4, A. Norhammar1; 1Cardiology Unit, Department of medicine, Karolinska Institutet, Karolinska University Hospital Solna, Stockholm, 2Centre for Safety Research, KTH Royal Institute of Technology, Stockholm, 3Division of Cardiovascular Medicine, Department of Medicine and Health Sciences, Linköping University, Linköping, 4Department of Cardiology UHL, County Council of Östergötland, Linköping, Sweden. Background and aims: Patients with diabetes are at increased risk for de- veloping heart failure. We investigated the impact of diabetes on long-term prognosis in patients with heart failure and preserved left ventricular func- tion from an everyday life perspective. Materials and methods: Patients with EF ≥50%, with (n=1658) and without (n=5047) type 2 diabetes included in the Swedish Heart Failure Registry (S- HFR) 2003-2011 were followed for mortality until 30 September 2011 (me- dian 22.5 months). Differences in background characteristics were adjusted for in a logistic regression model. Results: Patients with diabetes were younger (76 vs. 78 years), more often had known ischemic heart disease (47 vs. 36%), hypertension (68 vs. 52%), and more often had preserved renal function (eGFR > 60ml/min, 45 vs. 38%). NYHA classes III and IV were somewhat more common in those with diabe- tes (44 vs. 39%). Kaplan-Meyer curves for mortality are presented in Figure 1. The unadjusted and adjusted* ORs (95% CI) for mortality were 1.02 (0.92- 1.15) and 1.39 (1.20-1.61). Conclusion: Our data support that diabetes is an independent predictor of mortality in patients with heart failure and preserved left ventricular func- tion. As much as 50% of patients with preserved left ventricular function had reported ischemic heart disease, which questions the concept of a pure dia- betic cardiomyopathy. * Adjusted for gender, age, duration of heart failure, weight, blood pressure (systolic and diastolic), ischemic heart disease, hyper- tension, atrial fibrillation, pulmonary disease, revascularisation, eGFR class, Hb class, ACE inhibitors, ARBs, beta-blockers, mineralocorticoid receptor antagonists, diuretics, digitalis, nitrates, statins and antithrombotic agents. Figure: Kaplan-Meier curve showing survival by diabetes in patients with heart failure and preserved ventricular function. Odds ratio (OR), *=adjusted 46 Methylglyoxal-induced endothelial dysfunction has a role in the development of heart failure in a mouse model of type 1 diabetes B. Vulesevic1, B. McNeill1, F. Giacco2, M. Brownlee2, R. Milne3, E. Suuronen1; 1Cardiac Surgery, University of Ottawa Heart Institute, Ottawa, Canada, 2Diabetes Research Center, Albert Einstein College of Medicine, Bronx, USA, 3Diabetes and Atherosclerosis Laboratory, University of Ottawa Heart Institute, Canada. Background and aims: Diabetes mellitus increases the risk of heart failure independent of coronary artery disease or hypertension. Endothelial dys- function (ED) and inflammation, both common in diabetes, are partly caused by the accumulation of methylglyoxal (MG), the primary substrate of the en- zyme glyoxalase 1 (GLO1). We have generated transgenic mice that carry the human GLO1 transgene under the control of the preproendothelin promoter. Using this mouse model we examined if the over-expression of GLO1 could reduce MG-induced ED and prevent the development of heart failure caused by streptozotocin (STZ)-induced hyperglycaemia. Materials and methods: GLO1 transgenic mice and their wildtype (WT) lit- termates were treated with STZ (WT-diabetic and GLO1-diabetic mice) or vehicle (WT-control and GLO1-control mice). GLO1 activity was measured in endothelial cells (ECs) and cardiomyocytes isolated from the hearts. Eight weeks after STZ treatment, blood serum levels of the ED markers E-selectin, ICAM and VCAM were determined by ELISA. Heart function was assessed using echocardiography. Protein expression of the receptor for advanced gly- cation end products (RAGE) and neurogulin in the heart was determined by western blot. Apoptosis, detected by TUNEL assay, and capillary numbers, expressed by number of CD31+ cells were determined by immunohisto- chemistry. QPCR for the Bcl-2 pro-survival gene was done on ECs collected from the digested hearts. Results: Increased GLO1 activity was confirmed to be 5.6-fold greater in aor- tic endothelial cells (ECs) of GLO1-mice compared to their WT littermates. Also, the over-expression and increased activity of GLO1 in the heart was restricted to ECs, and not observed in cardiomyocytes. Eight weeks post-STZ elevated serum levels of all 3 ED markers was observed in WT-diabetic mice compared to all other groups (E-selectin ≥1.3-fold, VCAM ≥1.1-fold, and ICAM ≥6.7-fold; p≤0.04). RAGE levels were significantly higher in hearts of WT-diabetic mice compared to all other groups (≥1.8-fold, p≤0.03). The WT- diabetic group had lower left ventricular ejection fraction (EF; 43.6%) and fractional shortening (FS; 28.6%), compared to the other 3 groups (≥67.5% (EF) and ≥42.7% (FS); p≤0.01). The number of CD31+ ECs in WT-diabet- ic hearts was reduced by 46% compared to both non-diabetic groups and GLO1-diabetic mice (p=0.03). Apoptotic cells were more numerous in the hearts of WT-diabetic mice (0.44% of cells), compared to non-diabetic WT (0.03%) and GLO1-diabetic mice (0.07%; p≤0.04). The reduced number of Diabetologia (2014) 57:[Suppl1]S1–S564 S 26 1 C ECs seen in WT-diabetic mice may also be involved in the 3-fold reduction of neurogulin, an EC-produced protein that supports cardiomyocyte survival (p=0.04). The improved survival of ECs in GLO1-diabetic mice may be re- lated to the preserved mRNA levels of Bcl-2, which can be modified by MG. Bcl-2 was down-regulated 2-fold in WT-diabetic ECs compared to the other groups (p=0.04). Conclusion: Taken together, these results suggest that increased GLO1 activ- ity in ECs reduced ED, cell death and inflammation, and preserved cardiac function, despite high blood glucose levels. Supported by: Heart and Stroke 47 A relationship of asymptomatic coronary artery disease and type 2 diabetes in acute ischaemic stroke patients; cerebral angiography and coronary angiography study H. Kwon, J. Lim, J. Shin, J. Son, S. Lee, S. Kim, S. Yoo; Division of Endocrinology and Metabolism, Department of Internal Medicine, The Catholic University of Korea, Bucheon, Republic of Korea Background and aims: Stroke and coronary artery disease (CAD) share sim- ilar atherosclerosis risk factors such as smoking, hypertension, dyslipidaemia, and diabetes. CAD is considered the leading cause of mortality and morbidity in patients with ischemic stroke. However, data about the prevalence of CAD in the patients with acute ischemic stroke remains limited. Furthermore, characteristics of CAD in patients with diabetes are asymptomatic onset and multiple vessels involvement. But it remains unclear whether the impact of diabetes on CAD is different in ischemic stroke. The aim of this study was to determine the relation of type 2 diabetes and acute ischemic stroke to the presence and extent of asymptomatic CAD by coronary angiography. Materials and methods: Acute ischemic stroke patients without known CAD were enrolled between October 2008 and September 2013 in our hospital. Patients with high Framingham Risk Score (FRS; a 10 year-risk of coronary heart disease ≥20%) underwent cerebral angiography and coronary angiog- raphy at the same time. We analyzed the coronary angiography and other classic atherosclerosis risk factors in 187 patients retrospectively. The patients were divided into diabetes group and non-diabetes group, and the data was assessed according to the following classifications: (1) presence of coronary artery stenosis, (2) presence of significant stenosis (≥50%), (3) presence of multiple coronary artery stenosis (≥ 2 vessels), and (4) presence of significant multiple coronary artery stenosis (≥ 2 vessels). We calculated the prevalence of CAD with 95% confidence intervals. All data was analyzed after adjusting classic atherosclerosis risk factors such as age, sex, hypertension and history of past stroke. P value S 27 1 C OP 09 Diabetic retinopathy 49 Validation of model to estimate risk of progression of diabetic retinopathy using screening and clinical data in 3 cohorts I.M. Stratton1, S.J. Aldington1, R. Cragg2, T. Gazis3, G. Russell4, C. Connor5, S. Sivaprasad6, P.H. Scanlon1; 1Gloucestershire Diabetic Retinopathy Research Group, Gloucestershire Hospitals, Cheltenham, 2Nottingham Diabetic Retinopathy Service, Nottingham University Hospitals NHS Trust, 3Diabetes and Endocrinology, Nottingham University Hospitals NHS Trust, 4Health Intelligence Ltd, Sandbach, 5Diabetic Eye Screening Programme for Lambeth, Southwark and Lewisham, Guy’s and St Thomas NHS Foundation Trust, London, 6Ophthalmology, King’s College Hospital NHS Foundation Trust, London, UK. Background and aims: Diabetic retinopathy (DR) is a microvascular compli- cation of diabetes and can lead to vision loss and blindness. Annual screening with 2 field digital retinal imaging after mydriasis is recommended by the UK National Screening Committee. This is becoming difficult to achieve because of resource limitations and increasing numbers of people with diabetes. We have developed and validated a model to estimate risk of progression to sight threatening DR (STDR) using results from one screening episode, HbA1c in twelve months prior to screening and duration of diabetes in one Eng- lish screening programme. We know that grading protocols differ between screening programmes and rates of DR are affected by duration of diabe- tes, ethnicity, deprivation and diabetic control. Here we examine the perfor- mance of the model in 3 more English programmes. Materials and methods: Data were obtained from 3 English screening pro- grammes and data for a subset of patients was extracted from primary care. Patients free of STDR were categorised into those with No DR, those with mild non proliferative DR (NPDR) in one eye and those with Mild NPDR in both eyes. Using the risk estimation algorithm the risk score in those free of STDR was estimated and progression to STDR in quintiles of risk obtained. Results: The programmes had 8632, 1084 and 1377 people respectively. There were few non White Caucasian patients in the first 2 programmes but 30% of those in the third programme were of African or Afro-Caribbean ethnicity. Duration of diabetes was 6 (2 to 11) (median (25th to 75th centile)), 2.9 (0.6 to 6.2) and 3.6 ( 1.4 to 6.8) years, HbA1c 56 (48 to 66), 53 (46 to 53) and 52 (45 to 64), follow-up from date of index screening 2.9 (2.1 to 3.0), 4.2 (2.2 to 5.3) and 4.1 (2.1 to 6.9) years. In the first programme the rate of progression to STDR in the lowest risk quintile was 4.1 per 1000 patient years and in highest quintile 74.0 per 1000 patient years, in the second programme 2.4 and 79.2 respectively and in the third programme 1.7 and 49.0 respectively. Area un- der the ROC curve was 0.82, 0.87. and 0.77. Conclusion: Within each of the three programmes examined the risk model discriminates well between those with very low and with high risk of progres- sion to STDR. The algorithm would be suitable for calculation of personal- ised screening intervals. Further validation in other screening programmes and ethnic groups is required. Supported by: NIHR HTA 50 Importance of using HbA1c data of total diabetes duration in studying the relation between HbA1c value and retinopathy in type 1 diabetes A. Hirose1, S. Kitano1, Y. Uchigata2; 1Diabetic Ophthalmology, Diabetes Center, 2Diabetes Center, Tokyo Women’s Medical University, Japan. Background and aims: Glycaemic memory, in which past glycaemia could affect future retinopathy, is a potential issue in studying the relation between glycosylated haemoglobin A1c (HbA1c) value and retinopathy. We have demonstrated that mean HbA1c value (mA1C) covering total diabetes du- ration, which might reflect the full effect of glycaemic memory, could sub- stantially predict retinopathy, but the longer the period without HbA1c data following diabetes onset in simulation, the less accurate the prediction in our type 1 diabetes mellitus (T1DM) patients. To confirm the importance of us- ing HbA1c data of total diabetes duration, we examined the relation between mA1C and retinopathy in the public data of DCCT/Epidemiology of Diabe- tes Interventions and Complications (EDIC) studies. Materials and methods: From the primary prevention cohort, we included T1DM patients with the shortest diabetes duration (≤12 months) at DCCT baseline. We developed a window which we named ‘half-yearly visit (HYV)’ to calculate the mA1C and to determine the retinopathy steps of the final Ear- ly Treatment Diabetic Retinopathy Study (ETDRS) severity scale for persons. HbA1c values of an even number of quarterly visits (QVs) were applied to those of HYVs: for example, the value of QV00 (DCCT baseline) to HYV00, QV02 to HYV01, and so on during DCCT. During EDIC, HbA1c values were applied to the corresponding HYVs according to their data collection dates from DCCT baseline: for example, 11.75≤X S 28 1 C porters hTHTR1 and hTHTR2 (coded by SLC19A2 and SLC19A3, respective- ly), and/or their transcription factors Sp1 and Sp2 (coded by SP1 and SP2), are associated with proliferative diabetic retinopathy (PDR) and/or diabetic nephropathy (DN). Materials and methods: The patient population consisted of 1568 cases with PDR, and 217 controls with no/mild retinopathy from the Finnish Diabetic Nephropathy (FinnDiane) Study. PDR was defined as ETDRS-score of 53 or worse or scattered laser treatment. Controls were required to have mini- mal diabetes duration of 20 years, maximal ETDRS-score of 30 on altered ETDRS-scale (corresponding to mild retinopathy) and no laser treatment. We chose HapMapII imputed SNPs from all four candidate genes and 10kb up and downstream of them (n=134). Logistic regression adjusted for sex, age, diabetes duration, and first ten genetic principal components were used for statistical calculations. In addition, association was tested between patient groups with different states of DN, and intersection of classes with both se- vere PDR and end-stage renal disease (ESRD) (n=369) vs. normo/microalbu- minuric controls with no/mild retinopathy (n=190). Results: Two SNPs, rs12694743 and rs6713116, showed association with PDR (rs12694743: OR = 1.91, CI95% = 1.62-2.19, P = 1.02×10-5; rs6713116: OR = 2.34, CI95% = 1.97-2.72, P = 9.10×10-6). Both SNPs are located in in- tronic regions of SLC19A3. The signal was also noticeable in DN phenotypes, but was not significant after correcting for multiple testing. In analysis of in- tersection of extreme phenotypes the signal was amplified (rs12694743: OR = 3.18, CI95% = 2.76-3.60, P = 7.51×10-8; rs6713116: OR = 3.87, CI95% = 3.33-4.40, P = 7.49×10-7) providing proof of association not only with retin- opathy, but both retinopathy and nephropathy. Replications in DCCT/EDIC and WESRD cohorts are ongoing. Conclusion: The results of this study suggest that mutations in SLC19A3 may increase susceptibility to PDR and especially to the combined phenotype of PDR and ESRD. 52 Microvesicles derived from mesenchymal stem cells in diabetic-like conditions increase permeability in a retinal blood-barrier model A. Mazzeo, E. Beltramo, T. Lopatina, M. Porta; Dept of Medical Sciences, University of Turin, Italy. Background and aims: Loss of pericytes in the early phases of diabetic retin- opathy may disrupt their stable association with endothelial cells (EC), lead- ing to EC proliferation and, eventually, angiogenesis. Microvesicles (MV) are small membrane particles derived from different cells which contain biologi- cally active proteins and RNA and are known to promote phenotypic changes in target cells. MV derived from injured cells may induce dedifferentiation of pericytes, allowing their detachment from vessels. We have previously shown that MV derived from mesenchymal stem cells (MSC), but not from EC, induce pericyte detachment and that diabetic-like conditions (high glucose and hypoxia) play a synergistic role in this destabilizing influence. This study aimed at evaluating whether MV produced by MSC in hypoxia and/or high glucose are able to influence the retinal blood-barrier permeability and to explore the possible role of matrix metalloproteases (MMP) in MV-induced pericyte detachment. Materials and methods: We used commercially available human microvas- cular EC (HMEC) and MSC from bone marrow, while human retinal peri- cytes (HRP) had been previously immortalized in our laboratory. A blood- barrier model was established by seeding HMEC on the porous membrane of transwell inserts, letting them adhere for 24 hrs and then adding HRP into the same insert. MV were extracted from the supernatant of MSC cultured in 1) physiological conditions (NG) 2) hypoxia (hypo) 3) high glucose (HG) 4) HG + hypo. These MV were subsequently added to the confluent HMEC/ HRP co-cultures. After 2 hrs of MV exposure, FITC was added into the up- per chamber and fluorescence measured in the lower chamber of the inserts after another 30’, 1, 2, 3, 4 and 24h. MMP expression in both MV and super- natants of HRP exposed to MV was evaluated by Zimography. Results were confirmed by pre-treatment of MV with batimastat, a MMP inhibitor, and subsequent exposure of HRP to them. Results: Permeability of EC-HRP co-cultures was increased by exposure to MSC-derived MV obtained in all the above conditions, the highest percent- age increase occurring after 6h of total exposure to MV (2h pre-treatment + 4h FITC): NG-MV 134.42 ± 17.46% (p S 29 1 C 54 Diabetic macular oedema and diabetic retinopathy: treatment outcomes with aflibercept do not depend on systemic diabetes control C. Metzig, for the VIVID-DME/VISTA-DME Study Investigators; Bayer HealthCare Pharmaceuticals, Berlin, Germany. Background and aims: Diabetic macular edema (DME) is a major cause of vision loss in patients with diabetic retinopathy resulting from poorly con- trolled diabetes mellitus. Vascular endothelial growth factor plays a key role in the pathogenesis of DME. Materials and methods: Two similarly designed Phase 3 trials, VIVID-DME and VISTA-DME, evaluated the efficacy and safety of intravitreal aflibercept (IVT-AFL) injection for the treatment of DME. These trials randomised 872 patients with DME 1:1:1 to receive either IVT-AFL 2 mg every 4 weeks plus sham laser (2q4), IVT-AFL 2 mg every 8 weeks following 5 initial monthly doses plus sham laser (2q8), or macular laser photocoagulation plus sham IVT treatment. The primary efficacy endpoint was the mean change in best- corrected visual acuity (BCVA) from baseline at Week 52. An exploratory analysis examined BCVA and diabetic retinopathy severity score (DRSS) out- comes in subgroups of patients in these studies with baseline haemoglobin A1c (HbA1c) levels ≤8% and >8%. Results: Overall, 65% and 35% of patients had HbA1c ≤8% and >8% at base- line, respectively. The mean change in BCVA from baseline to Week 52 in the 2q4 and 2q8 groups versus the laser group was +12.3 and +10.9 versus +1.1 letters (P8%). Furthermore, the DRSS improvement indicates an effect of IVT-AFL not only on DME, but also on the underlying diabetic retinopathy. Clinical Trial Registration Number: NCT01331681, NCT01363440 OP 10 Entero-pancreatic endocrinology 55 Severely disrupted islet function in mice with global deletion of GLP-1 and GIP receptors B. Omar, B. Ahrén; Clinical Sciences, Lund University, Sweden. Background and aims: The incretin hormones glucagon like peptide 1 (GLP- 1) and glucose dependent insulinotropic polypeptide (GIP) potentiate glu- cose stimulated insulin secretion. Both incretin hormones promote prolifera- tion of pancreatic beta cells and prevent beta cell apoptosis. Incretins have also long been known to mediate the majority of insulin secretion during a meal or oral glucose challenge, however a role for the endogenous and basal incretin hormones in the long term regulation of beta cell function has not been characterized. We have therefore evaluated the role of basal incretin hormones on islet function by using mice with targeted disruption of both incretin hormone receptors. Materials and methods: Double incretin receptor knockout (DIRKO) mice with global deletion of both the GLP-1 receptor and GIP receptor genes were used to study the effect of chronic deficiency of incretin action on beta cell function ex vivo and in vivo. Ex vivo insulin secretion was determined in stat- ic batch incubations of isolated islets from wild type and DIRKO mice with stimulatory glucose concentrations (16 mmol/l). In vivo insulin secretion was assessed by hyperglycaemic clamp in which, after a 20 minute baseline, mice were given a bolus injection of glucose (0.35mg/kg) together with infusion of 30% glucose in saline and clamped at a target blood glucose concentration of 16.7 mmol/l. The steady state period was between 20 and 105 minutes. Plasma insulin was measured at multiple time points throughout the steady state period. Insulin sensitivity from the clamp was determined by the insulin sensitivity index, calculated as the mean glucose infusion rate during the last 40 minutes of the steady state period divided by the mean plasma insulin concentration for the same period. Results: Insulin secretion was markedly lower in islets from DIRKO mice compared to islets from wild type mice during static incubation of isolated islets with stimulatory glucose (16.7 mmol/l) (395 ± 80 vs 1171 ± 131 pmol/ islet/hour, p = 0.02). In order to specifically study beta cell function in vivo, hyperglycaemic clamp experiments were performed in which mice were clamped with a target blood glucose concentration of 16.7 mmol/l. This tar- get was closely achieved in both wild-type and DIRKO mice (17.0 ± 0.3 vs 17.1 ± 0.2 mmol/l). Insulin secretion during the hyperglycaemic clamp was severely impaired in DIRKO mice compared to control mice. The acute insu- lin response (AIR) was blunted in DIRKO mice compared to controls (199 ± 88 vs 540 ± 108 pmol/l, p = 0.005). The total area under the insulin secretion curve (AUC) during the steady state period was significantly lower in DIRKO mice compared to wild-type mice (25.5 ± 1.8 vs 37.0 ± 3.6 nmol/l*min, p = 0.008). Insulin sensitivity, derived from clamp data, was not significantly dif- ferent between wild-type and DIRKO mice (0.029±0.004 vs 0.036±0.005 mg glucose/kg/min/pmol insulin). Diabetologia (2014) 57:[Suppl1]S1–S564 S 30 1 C Conclusion: Normal islet function is heavily dependent on both incretin hormones, whereas insulin sensitivity is not altered by genetic deletion of incretin hormone receptors Supported by: LU, ALF/Region Skane 56 GIP contributes to the protection from hypoglycaemia of DPP-4 inhibition S. Malmgren, B. Ahrén; Department of Clinical Sciences in Lund, Lund University, Sweden. Background and aims: Intensive glucose lowering therapies in type 2 diabe- tes (T2D) increases the risk of hypoglycaemia, which is associated with acute unpleasant consequences and increased long-term risk for cardiovascular diseases. A major defence mechanism to combat hypoglycaemia is glucagon counter-regulation; however, this mechanism is commonly compromised in persons with T2D. Therefore, there is a need to develop glucose-lowering ther- apies which sustain the endogenous counter-regulations to hypoglycaemia. Inhibition of dipeptidyl peptidase-4 (DPP-4) prevents the regulatory inac- tivation of the incretin hormones and is a glucose-lowering therapy in T2D with low risk for hypoglycaemia. We have previously shown that DPP-4 in- hibition protects from insulin-induced hypoglycaemia in mice. The aim of the present study is to examine the mechanisms behind this protection and particularly the contribution of the incretin hormone glucose-dependent in- sulinotropic polypeptide (GIP). Materials and methods: Anesthetized C57BL/6 mice were infused with 15 mU/kg/min of insulin and variable rates of glucose to target a steady state blood glucose of 2.5 mmol/L between minute 60 and 90 of the hyperinsuline- mic hypoglycaemic clamp. To explore GIP in the endogenous defence against hypoglycaemia we performed hypoglycaemic clamps in GIPR -/- mice lack- ing the GIP receptor (n=5) or wild type (wt) counterparts (n=7). To further explore the effect of GIP, wt mice were infused with synthetic GIP (50 pmol/ kg/min) during hypoglycaemic clamp (n=10). Glucose, insulin and glucagon levels were measured and glucose infusion rate (GIR) during the clamp was estimated. To explore the relationship of GIP expression to glucagon secre- tion in humans, correlation analysis was performed in islet gene expression and secretion data from isolated human islets from 69 individuals. Values are presented as Mean±S.E and statistical significance assessed using Student’s t-test. Results: GIPR -/- mice required increased GIR compared to matched wt ani- mals to maintain blood glucose during the clamp (6.0±1.2 vs. 2.5±0.5 mg/kg/ min; p=0.03). Both wt and GIPR -/- mice had similar increases in endpoint glucagon after 90 minutes of clamp compared to baseline levels (2.3±0.5- fold and 1.7±0.2-fold respectably; p=0.50). During GIP infusion, GIR was reduced to 4.2±0.9 compared to 6.9±0.8 mg/kg/min (p=0.040) in animals in- fused with saline. This was accompanied by a tendency towards an increase in glucagon levels at 60 (11.5±2.0 vs 6.2±1.9 pg/mL; p=0.072) and 90 (17.5±4.0 vs 8.6±2.1 pg/mL; p=0.062) minutes of the clamp. Clamp insulin levels did not differ between groups. In human islets, there was a significant positive correlation between GIP expression and glucagon secretion at 1 mM of glu- cose (rho=0.345, p=0.004). Conclusion: GIPR -/- mice were more sensitive to insulin-induced hypogly- caemia while GIP infusion during clamp protects wt mice from hypogly- caemia. Results from glucagon measurements suggests that GIP infusion protects from hypoglycaemia via increases in glucagon levels but also raise the possibility of glucagon independent mechanisms further explaining the protective mechanism of DPP-4 inhibition. Furthermore, the positive GIP expression correlation with increased glucagon secretion in human islets sug- gests relevance in humans. Based on these findings we conclude that GIP is a protective factor from hypoglycaemia and suggest that it contributes to the low risk of hypoglycaemia that can be seen with diabetes treatment using DPP-4 inhibition, thus elevating GIP levels. Supported by: Kungl. Fysiografiska Sällskapet i Lund 57 Altered enteroendocrine expression of glucagon and somatostatin in the gut of patients with type 2 diabetes compared with healthy matched controls T. Jorsal1,2, N.A. Rhee1,2, J. Pedersen1,2, S.L. Jepsen2, C.D. Wahlgren2, B. Mortensen1,2, P. Vilman3,4, H. Hassan3,4, J.W. Hendel3,4, S.S. Poulsen2, J.J. Holst2, T. Vilsbøll1, F.K. Knop1,2; 1Diabetes Research Division, Department of Medicine, Gentofte Hospital, University of Copenhagen, Hellerup, 2Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, 3Unit of Enteroscopy, Department of Medicine, Gentofte Hospital, University of Copenhagen, Hellerup, 4Endoscopy Unit, Department of Gastroenterology, Herlev Hospital, University of Copenhagen, Herlev, Denmark. Background and aims: Hormones secreted from enteroendocrine cells in the gut play a crucial role in regulation of whole-body glucose metabolism. Lit- tle is known about the distribution of these enteroendocrine cells along the intestinal tract, and it is currently unknown whether the distribution and/ or phenotype of these cells is altered in type 2 diabetes (T2D). We aimed to evaluate the expression of products from enteroendocrine D, K and L cells, respectively, along the entire length of the small and large intestines in pa- tients with T2D and in healthy control subjects Materials and methods: We used double-balloon enteroscopy (DBE) to col- lect mucosal gut biopsies. The study involved 12 subjects diagnosed with T2D and 12 age and BMI-matched non-diabetic controls. All subjects underwent anterograde and retrograde DBE. Biopsies were collected from every 30 cm from pylorus to the ileocecal valve and from 6 specific sites in colon. RNA was isolated and analysed for expression of glucagon (GCG), peptide YY (PYY), glucose-dependent insulinotropic polypeptide (GIP) and somatostatin (SST). Results: Expression of markers for L cells (GCG, PYY) and K cells (GIP) showed significantly higher expression of L cell genes in the distal versus the proximal parts of the small intestine and significantly higher expression of the K cell marker (GIP) in the proximal part of the small intestine. Expression of the L cell marker (GCG) was significantly higher in the proximal small intes- tine in subjects with T2D compared to healthy controls (337% (113%;1002%) p S 31 1 C 58 Role of the somatostatin in the control of glucagon secretion by glucose and KATP channel blockers A. Gómez-Ruiz, K.B. Lai, H.-Y. Chae, N. Antoine, P. Gilon; Pôle d’endocrinologie, diabète et nutrition, Université catholique de Louvain, Brussels, Belgium. Background and aims: It is well established that glucose (G) inhibits gluca- gon release by pancreatic α-cells. However, it has sometimes been reported that high concentrations of the sugar (>20mM) stimulate glucagon secretion. The mechanisms behind these effects are unknown. In particular, it is still unclear whether G controls glucagon release by modulating KATP channels of α- and non-α-cells within the islet and whether somatostatin (SST) released from δ-cells is involved in the control of glucagon release. In the present study, we evaluated the role of SST in the control of glucagon secretion by various G concentrations and KATP channel blockers. Materials and methods: The effects of G and KATP channel modulators were tested on glucagon secretion of control islets and islets devoid of SST parac- rine influence (islets pre-treated for 24h with 200ng/ml pertussis toxin (PTX) or from Sst-/- mice). Unless indicated, all the experiments were performed in the presence of a 6 mM mixture of amino acids (Mix AA: 2 mM alanine, 2 mM glutamine and 2 mM arginine). Results: The role of SST in the control of glucagon secretion by various G concentrations (0, 1, 2, 5, 7, 15 and 30 mM) was first evaluated in incubation experiments with control islets or islets preincubated with PTX to remove the paracrine influence of SST. In control islets, G dose-dependently inhibited glucagon secretion with a maximal effect obtained already at around 7 mM. In PTX-pre-treated islets, glucagon secretion was larger at all G concentra- tions and the dose-response relationship for G-regulated glucagon secretion displayed a U-shape reflecting an inhibition of glucagon release at G concen- trations up to 7 mM followed by a progressive decrease of the glucagonostatic effect at G concentrations >7 mM. Perifusion experiments with islets of Sst+/+ or Sst-/- mice confirmed these results. Thus increasing the G concentration of the medium from 1 to 7 or 30 mM equally inhibited glucagon release of Sst+/+ islets, whereas G7 inhibited glucagon release more potently than G30 in Sst- /- islets. Moreover, increasing the G concentration of the perifusion medium from 7 to 30 mM did not affect glucagon release of Sst+/+ islets, whereas it stim- ulated that of Sst-/- islets. These results suggest that SST is partially involved in the glucagonostatic effect of high G concentrations. We next studied the role of SST in the control of glucagon secretion by tolbutamide (Tolb, a KATP chan- nel blocker). In Sst+/+ islet, Tolb (500 μM) did not affect glucagon release at G1 and slightly stimulated it at G7. By contrast, in Sst-/- islets, Tolb stimulated glucagon secretion both at G1 and G7. The glucagonotropic effect of Tolb in Sst-/- islets was much more evident in the absence of Mix AA. It is unrelated to Epac2 activation because a much lower concentration of Tolb (10 μM) and gli- clazide (10 μM) which do not activate Epac2 also strongly stimulated glucagon release of Sst-/- islets perifused with G7 and without MixAA. Conclusion: G dose-dependently inhibits glucagon release of control islets. This inhibition is independent of SST for intermediate G concentrations (0-7 mM) but starts to involve SST for G concentrations > 7 mM. Pharmacological closure of KATP channels controls glucagon secretion by two mechanisms: a direct stimulation of α-cells (independent of Epac2 activation and observed in the absence of SST influence) and an indirect inhibition via SST released from δ-cells. Supported by: FRSM, ARC and EFSD/Boehringer-Ingelheim 59 Effect of glucose and free fatty acid on PC1/3-PC2 expression in pancreatic alpha cell V. Sancho Bornez, R. Lupi, V. Garau, S. Del Guerra, D. Lucchesi, A. Dardano, L. Giusti, R. Miccoli, G. Penno, S. Del Prato; Endocrinology and Metabolism, University of Pisa, Italy. Background and aims: An intra-islet incretin system has been recently sug- gested to operate through modulation of the expression and activity of pro- convertase 1/3 and 2 (PC1/3, PC2) in pancreatic alpha-cell accounting for local release of GLP-1. Little is known, whether this alpha-cell activity can be affected by the metabolic alterations occurring in type 2 diabetes, such as hyperglycaemia, hyperlipidemia or hyperglucagonemia. Materials and methods: AlphaTC1/6 cells from a mice pancreatic cell line were incubated in the presence of two glucose (G) concentration (5.5 and 16.7 mM) for 16 h with or without free fatty acid (FA, 2:1 palmitate:oleate, 0.5 mM), IL6 (200 ng/ml) or glucagon (GLU: 250 pg/ml). GLP-1 secretion was measured by ELISA and expression of PC1/3 and PC2 by RT-PCR and western blot; cell viability was determined by MTT method, Reactive Oxygen Species generation (ROS) by H2DCFDA fluorescence and apoptosis by An- nexin staining and Propidium Iodine (PI) fluorescence. Results: Upon 16.7G incubation, GLP-1 secretion (total and active) was sig- nificantly increased (130±9% and 158±19%, respectively, p S 32 1 C OP 11 Lifestyle factors and prediction of type 2 diabetes 61 Trajectories of glycaemic traits in south Asians and whites before diabetes diagnosis: a longitudinal analysis from the Whitehall II study A.G. Tabák1,2, E.J. Brunner2, D.R. Witte3, K. Faerch4, D. Vistisen4, S. Ikehara5, M. Kivimaki2, A. Hulmán6; 11st Department of Medicine, Semmelweis University, Faculty of Medicine, Budapest, Hungary, 2Department of Epidemiology and Public Health, University College London, UK, 3Centre de Recherche Public de la Santé, Strassen, Luxembourg, 4Steno Diabetes Center, Gentofte, Denmark, 5Osaka University, Japan, 6University of Szeged, Hungary. Background and aims: An increased risk of type 2 diabetes among south Asians compared to whites is well established. Most of the literature suggests that this is related to lower insulin sensitivity among south Asians however the role of insulin secretion is rarely assessed. Therefore, we aimed to inves- tigate ethnic differences in trajectories of fasting (FPG) and 2-h post-load plasma glucose (2hPG), log-transformed homeostasis model assessment in- sulin sensitivity (HOMA-S) and secretion (HOMA-B) prior to the diagnosis of type 2 diabetes. Materials and methods: We analysed trajectories of glycaemic traits before the diagnosis of diabetes by fitting mixed-effects models to longitudinal data (with up to 4 repeat measures within individuals) from 101 south Asian and 764 white participants of the Whitehall II study, who developed type 2 diabe- tes during follow-up between 1992 and 2009. Results: South Asians had an almost 4 times increased risk of incident di- abetes compared to whites during follow-up (26.4% vs 10.2%, P < 0.001). According to age and sex-adjusted mixed-effects models, South Asians had a faster increasing FPG trajectory before diagnosis (slope difference: 0.34 mmol/l per decade, 95% CI: 0.05, 0.64; P = 0.022) and 0.36 mmol/l (95% CI: 0.10, 0.63; P = 0.007) higher FPG levels at diagnosis than whites. There were no differences in 2hPG trajectories between the two ethnic groups. South Asians had significantly lower HOMA-S already 15 years before diagnosis and the difference increased even further until diagnosis (difference at di- agnosis: 0.30, 95% CI: 0.14, 0.46; P < 0.001). HOMA-B trajectories had dif- ferent quadratic characteristics (P = 0.04 for ethnicity x time2 interaction). It increased in both ethnicities until 7 years before diagnosis however the increase was faster in whites, while it followed a similar decreasing trajectory thereafter until diagnosis. Differences in FPG and HOMA-B trajectories were robust to adjustment for measures of obesity and lifestyle measures. Conclusion: Insulin sensitivity falls more rapidly among south Asians than whites. In contrast to whites, who could increase insulin secretion up to 7 years before diagnosis, this compensatory mechanism is can be hardly seen in south Asians, hence a more rapidly increasing FPG trajectory develops prior to the diagnosis of diabetes. These findings extend our prior observation of inadequate β-cell compensation with aging in healthy south Asians. Supported by: MRC, NHLBI, BHF, NIA 62 Food sources of fat may clarify the earlier inconsistent role of dietary fat intake for incidence of type 2 diabetes U. Ericson1, S. Hellstrand1, L. Brunkwall1, E. Sonestedt1, P. Wallström2, B. Gullberg2, E. Wirfält2, M. Orho-Melander1; 1Department of Clinical sciences in Malmö, Diabetes and Cardiovascular disease, Genetic Epidemiology, Lund University Diabetes Center (LUDC), 2Department of Clinical sciences in Malmö, Nutrition Epidemiology, Lund University, Malmö, Sweden. Background and aims: Dietary fats could affect glucose metabolism and in- sulin sensitivity and may thereby have a crucial role in the development of type 2 diabetes (T2D). Studies have indicated that replacing saturated fat with monounsaturated and polyunsaturated fats might be favorable in the preven- tion of T2D. In line with this, plant sources of fat have been suggested to be a better choice compared with animal sources. Indeed, high intakes of red meat and meat products show positive associations with risk of T2D. Never- theless, several epidemiological studies have indicated that a high intake of dairy products may be protective. Subsequently, the importance of dietary fat content and food sources of fat remains to be clarified. Our aim was to exam- ine intakes of main dietary fat sources, classified according to fat content, in relation to incident T2D. Materials and methods: In total 26 930 individuals (60% women), 45 -74 years, from the population-based Malmö Diet and Cancer cohort, were included. Dietary data was collected with a modified diet history method. During 14 years follow-up, 2860 incident T2D cases were identified. Cox proportional hazards regression model was used to estimate hazard ratios (HR) of diabetes incidence in quintiles of energy adjusted dietary intakes. The multivariate model included adjustments for age, sex, season, diet assess- ment method version, total energy intake, BMI, leisure time physical activity, smoking, alcohol consumption and education. Results: High intake of high-fat dairy products was associated with lower in- cidence of T2D (HR for the highest (median=8 portions/day) compared with the lowest (median=1 portion/day) intake quintile: 0.77; 95% CI: 0.68-0.87; P for trend S 33 1 C Results: Linear regression analysis showed that compared with those who had the most minutes with movements during sedentary time, participants with fewer movement minutes (quartiles 3, 2, 1) had a higher BMI (B= 1.43; B= 1.97; B= 3.38; all p S 34 1 C 66 Effect of lifestyle modification on the prevention of type 2 diabetes and impaired glucose tolerance in a young healthy urban South Asian population J. Karalliedde1, M. Wijesuriya2, L. Vasantharajah2, M. Gulliford3, G. Viberti1, L. Gnudi3; 1Diabetes, King’s College London, UK, 2Diabetes, Diabetes Association of Sri Lanka, Colombo, Sri Lanka, 3King’s College London, UK. Background and aims: There is an increasing incidence of type 2 diabetes mellitus (T2DM) and related cardiovascular disease (CVD) in young South Asian subjects. Several studies have confirmed that lifestyle modification (LSM) is very effective in preventing T2DM in older subjects with impaired glucose tolerance (IGT). The effect of LSM on the prevention of cardio-meta- bolic disease in young healthy urban South Asian subjects is unknown. Materials and methods: A randomised controlled clinical trial to compare an intensive 3-monthly LSM (I-LSM) with a less-intensive 12 monthly LSM (LI-LSM) for a primary composite cardio-metabolic endpoint of new onset T2DM, IGT, impaired fasting glycaemia, hypertension, initiation of lipid low- ering therapy and cardio-renal disease in subjects aged 5 to 40 years with at least 2 of the following risk factors: raised body mass index, raised waist circumference (WC), first degree family history of T2DM and physical inac- tivity. A cluster sampling strategy was used to select a representative sample of healthy at risk subjects with two or more of the above risk factors. We ran- domised 4606 subjects of whom 3,685 (48% males) qualified for analyses. The study was performed at a single centre in Colombo, Sri Lanka between 2010 to 2013. Intervention: Each subject received LSM advice aimed at reducing weight, improving diet and increasing physical activity 3 monthly (I-LSM) or 12 monthly (LI-LSM) for 4 years. Results: There were no significant baseline differences in anthropometric, clinical and demographic measures between groups [I-LSM (n=1807, mean ± SD age, 22.4±10 yrs.) and LI-LSM (n=1878 age, 22.4±9.8 yrs]. The cumula- tive incidence of the primary end point after 4 years was n=270 in I-LSM vs. n=302 in LI-LSM, a 9% (95% CI 1.0% to 16%) relative risk reduction (RRR), which was independent of baseline age, gender, p=0.02. Similarly there was a significant 26% RRR (95% CI 2% to 44%) in new onset T2DM p=0.04, and 18% RRR (95% CI 7% to 28%) in new onset IGT p=0.002. I-LSM did not sig- nificantly reduce the other components of the primary composite endpoint. Conclusion: Our results demonstrate for the first time that in a young healthy but at risk South Asian population intensive LSM significantly reduces the development of T2DM, IGT and a composite endpoint of cardio-metabolic disease. These results highlight the importance of intervening early with life- style modification to prevent and reduce the burden of T2DM in young at risk subjects. Clinical Trial Registration Number: SLCTR/2008/World Health Organization (WHO) Supported by: BRIDGES grant from IDF and National Diabetes Centre Sri Lanka OP 12 The many faces of advanced glycation 67 Exogenous and endogenous hydrogen sulfide protects the cardiomyocyte cell line H9c2 from methylglyoxal-mediated damage T. Ghela1, J.G. Mabley2; 1Brighton and Sussex Medical School, 2School of Pharmacy & Biomolecular Sciences, University of Brighton, UK. Background and aims: Methylglyoxal (MGO), a glycolysis derived reactive dicarbonyl compound, has been implicated as a meditator of diabetic car- diovascular complications. MGO has already been shown to cause endothe- lial cell dysfunction as well as cardiomyocyte contractile dysfunction via in- creased cellular oxidative stress. Hydrogen sulfide is a gaseous transmitter found in the vasculature and synthesised from the amino acid L-cysteine. Hydrogen sulfide has been demonstrated to have a protective role in the vas- culature acting as a vasodilator and protecting cells from increased oxida- tive stress. Hydrogen sulfide has already been shown to protect endothelial cells from MGO-mediated damage. The aim of this study is to investigate whether hydrogen sulfide generated exogenously by a chemical donor or endogenously from L-cysteine is also able to protect cardiomyocytes from MGO-mediated dysfunction. Materials and methods: The cardiomyocyte cell line H9c2 was exposed to increasing concentrations of MGO (0.1-1 mM) for 24h prior to measuring cell viability and apoptosis. In a second series of experiments H9c2 cells were exposed to MGO 0.6 and 0.8 mM in combination with 0.25 or 0.5 mM so- dium hydrogen sulfide or L-cysteine. Cell viability was determined using the MTT assay and cell death by propidium iodide/Hoechst staining. Endoplas- mic Reticulum (ER) stress by Western blotting for CHOP. Results: Methylglyoxal dose-dependently reduced H9c2 cell viability. MGO 0.6 and 0.8 mM reduced cell viability to 55±3% and 47±2% respectively, sodi- um hydrogen sulfide protected against this loss of cell viability with 0.25 mM returning viability to 65±5% and 57±4% and 0.5 mM returning viability to 70±4% and 63±3% respectively (p S 35 1 C of vascular complications of diabetes. Glyoxalase1 (Glo1) and glyoxalase 2 (Glo2) of the cytoplasmic glyoxalase system metabolises MG and prevents AGE formation. The aim of this investigation was to produce a mathematical model of MG concentration and related AGE formation in human vascular endothelial cells in vitro from experimental measurements of glucose metab- olism and activities of Glo1 and Glo2 in normal glucose concentration and also high glucose concentration to model hyperglycaemia. The model is then validated by experimental measurement of cellular MG and AGEs Exposure to increased AGEs may thereby by predicted for a given level of hyperglycae- mia and the level of pharmacological induction of Glo1 induction required to prevent increased AGEs predicted. Materials and methods: Human microvascular endothelial HMEC-1 cells were incubated in MCDB131 media with 10% serum containing low glu- cose or high Glucose (5 mM, LG and 30 mM, HG, respectively) for 6 days at 37°C. Similar experiments were performed with human aortic endothelial cells (HAECs) in primary culture with 5 mM or 20 mM glucose. Rate of glu- cose consumption, activities of Glo1 and glyoxalase 2, cellular and medium concentrations of MG, flux of MG (measured as the surrogate accumulation of D-lactate), D-lactate metabolism and cell protein content and flux of the major MG-derived AGE, hydroimidazolone MG-H1 were determined. Single and two compartment models of MG metabolism by the glyoxalase system and AGE formation were constructed using the COPASI programme. Results: Incubation of HMEC-1 cells in HG produced a progressive decrease in Glo1 activity of 31% after 6 days. The rate of glucose metabolism was in- creased 70% and the flux of formation of MG was increased 66%. A math- ematical model taking into account flux of MG formation, reversible binding of MG to GSH and protein thiols, metabolism by Glo1 and rate of MG glyca- tion predicted an increase in cellular MG concentration and related AGEs in HG 0f 134%. Experimental validation produced similar results: increased in MG content of cells (pmol/106 cells) of 114% (LG, 2.08 ± 1.15; HG, 4.76 ± 0.38, P S 36 1 C individual dicarbonyls metabolism in tissue is not fully clarified. In our study we analyzed level of individual reactive dicarbonyls - methylglyoxal (MG), glyoxal (GL) and 3-deoxyglucosone (3-DG) in relation to lipid disorders and after metformin administration in non-obese rats with hereditary hypertri- glyceridemia (HHTg rats). Materials and methods: Adult Wistar rats (controls) and HHTg rats, which exhibit genetic fixed hypertriglyceridemia, tissue insulin resistance, impaired glucose tolerance, hyperinsulinemia and ectopic lipid accumulation, were fed a standard laboratory diet with or without metformin (300mg/kg b.wt.) for 4 weeks. The concentration of dicarbonyls were determined by derivatisation with 1,2-diamino-4,5-dimethoxy-benzene and HPLC-method with fluores- cence detection. Results: Compared with controls, HHTg rats exhibited markedly elevated se- rum levels of triacylglycerol (4.05±0.39 vs 1.88±0.23 mmol/l, p S 37 1 C OP 13 Clinical studies with GLP-1 analogues 73 Liraglutide 3.0 mg reduces the prevalence of prediabetes and delays onset of type 2 diabetes in overweight/obese adults: the SCALE obesity and prediabetes trial X. Pi-Sunyer1, K. Fujioka2, C. le Roux3, A. Astrup4, F. Greenway5, A. Halpern6, M. Krempf7, D.C.W. Lau8, R. Violante Ortiz9, C.B. Jensen10, S.K. Lillleøre10, J.P.H. Wilding11; 1Division of Endocrinology, Diabetes and Nutrition, St. Luke’s-Roosevelt Hospital Center, New York, USA, 2Scripps Clinic, La Jolla, USA, 3University College Dublin, Ireland, 4University of Copenhagen, Denmark, 5Louisiana State University, Baton Rouge, USA, 6University of São Paulo Medical School, Brazil, 7Université de Nantes, France, 8University of Calgary, Canada, 9Instituto Mexicano del Seguro Social, Ciudad Madero, Mexico, 10Novo Nordisk A/S, Søborg, Denmark, 11University of Liverpool, UK. Background and aims: This trial investigated the effects of liraglutide 3.0 mg, as adjunct to diet and exercise, on weight loss, prediabetes prevalence and onset of T2D (ADA 2010) over 56 weeks. The effects of liraglutide 3.0 mg cessation were investigated in a subsequent 12 week re-randomised period. Materials and methods: Adults (BMI ≥27 kg/m2 + ≥1 comorbidity or ≥30 kg/m2) were advised on a 500 kcal/day deficit diet and exercise programme and randomised 2:1 to once daily sc liraglutide 3.0 mg (n=2487) or placebo (n=1244). Randomisation was stratified by prediabetes status (ADA 2010). At week 56, individuals without prediabetes on liraglutide 3.0 mg were re- randomised 1:1 to liraglutide 3.0 mg or placebo (diet and exercise continued). Results: Baseline characteristics: age 45.1 yr, 78.5% female, body weight 106.2 kg, BMI 38.3 kg/m2, 61.2% with prediabetes. At week 56, individuals on liraglutide 3.0 had lost 8.0% weight compared to 2.6% with placebo (es- timated treatment difference [ETD] -5.4%, p S 38 1 C (p=0.003). Total glucose CMR was inversely correlated with RaOr (r=-0.63; p S 39 1 C Conclusion: 12 wks after liraglutide cessation, the beneficial treatment effects on weight were reduced and effects on FPG were reversed, emphasising the need for continued treatment. Clinical Trial Registration Number: NCT01272232 Supported by: Novo Nordisk 77 DURATION-1 extension: efficacy and tolerability of exenatide once weekly over 6 years in patients with type 2 diabetes mellitus E. Klein1, R.R. Henry2, J. Malloy3, P. Griffin3, M. Zhou4, N. Iqbal5; 1West Olympia Internal Medicine, Olympia, 2Center for Metabolic Research, VA San Diego Healthcare System and University of California, 3Bristol-Myers Squibb, San Diego, 4Bristol-Myers Squibb, Hopewell, 5Bristol-Myers Squibb, Princeton, USA. Background and aims: In a 30-week controlled, Phase III trial (DURA- TION-1), exenatide once weekly (QW) exhibited greater reduction in HbA1c than exenatide twice daily (−1.9% vs −1.5%; P=0.002), with similar weight loss, in 295 intention to treat (ITT) patients with type 2 diabetes mellitus not controlled with diet and exercise, or single or combination oral glucose-low- ering therapies. Here, we report data from patients who completed 6 years of treatment, the longest assessment of the efficacy and safety of a glucagon-like peptide-1 (GLP-1) receptor agonist to date. Materials and methods: In the open-ended extension of DURATION-1, all patients received exenatide QW. Results: In total, 127 patients (43%) completed 6 years of treatment; baseline characteristics of these patients were similar to the ITT population. In com- pleters, baseline [mean ± SD] values were: HbA1c 8.2 ± 0.9%; fasting plasma glucose (FPG) 9.22 ± 2.26 mmol/L; weight 101 ± 17 kg; diabetes duration 7 ± 6 years. Withdrawal from the extension was most often due to withdrawal of consent (23.4%), adverse events (AEs; 5.8%) or investigator decision (5.1%). Among 6 year completers, HbA1c improved significantly from baseline (least squared mean −1.6% [95% CIs −1.9, −1.4]), 45% achieved HbA1c S 40 1 C OP 14 Weight regulation and obesity 79 Genetic risk score of 31 BMI loci and substantial weight change over a period of 50 years in the Malmö Diet and Cancer Study G. Rukh1, U. Ericson1, G. Hindy1, O. Melander1, P. Nilsson2, M. Orho-Melander1; 1Department of Clinical Sciences in Malmö, Lund University Diabetes Center (LUDC), 2Department of Internal Medicine, Skåne University Hospital, Lund University, Malmö, Sweden. Background and aims: Body weight gain has increasingly become a public health threat worldwide as it is associated with many metabolic abnormalities conveying an increased morbidity and mortality. Our aim was to evaluate the contribution of genetic susceptibility defined as a genetic risk score (GRS) of 31 BMI associated single nucleotide polymorphisms (SNPs) to BMI at young, middle-age and older age, and to substantial weight gain (SWG) covering a mean time period of over 50 years starting from age of 20 years in a large Swedish cohort. Materials and methods: Totally, 22634 non-diabetic participants (62% fe- males) from the population based Malmö Diet and Cancer Study (MDCS) with baseline examinations 1990-1996 (age 57±8y, BMI 26±4kg/m2) and who in a questionnaire reported weight at 20 years of age (BMI 21±3kg/m2) as well as if their weight after 20 years of age had been stable, unstable, increased or decreased. A random sample of the MDCS (N=6,103), who were alive and had not emigrated from Sweden (N=4,924) were invited to a follow-up (FU) re-examination 2007-2012. Of these, 3734 subjects attended and 2676 with information of self-reported BMI from 20y were included in this study (58% females, age 73±6y, BMI 26±4kg/m2, FU time from baseline 16.5±1.5y and from 20y of age 52.8±5.6y). SWG was defined as gaining (i) ≥10% of self- reported weight at 20y until baseline, (ii) ≥10% of baseline weight until FU and (iii) ≥10% of self-reported weight at 20y until FU. A weighted GRS based on 31 GWAS identified BMI susceptibility loci was created. Linear regression was used to analyze the associated effect sizes (β) per increasing quintile of GRS on BMI at 20 years of age, at baseline and at the end of FU. Further, we used logistic regression to analyze the risk (OR) per GRS quintile for belong- ing to the self-reported unstable weight group (23.7% of all individuals), and for SWG from 20 years age to baseline and to FU, and from baseline to FU. All analyses were adjusted for age, sex and FU time when applicable. Results: The GRS associated with higher BMI at all ages and with highest ef- fect size at middle-age; 20y (β 0.03±0.01, p=2.0x10-37), baseline (β 0.23±0.02, p=1.8x10-34) and FU (β 0.20±0.06, p=0.003). The GRS associated with 10% increased risk per GRS quintile for having unstable weight from 20y to base- line (OR 1.10 [1.07-1.13], p=1.1x10-9) and with 4% increased risk of SWG (OR 1.04 [1.02-1.07], p=0.002) from 20y to baseline. The risk increase was 42% for unstable weight reporters (p=4.1x10-7) and 15% for SWG (p=0.016) for individuals in the highest GRS quintile compared to lowest. In contrast to this, the GRS did not associate with SWG from 20y to end of FU (OR 0.97 [0.90-1.05], p= 0.47) and it associated with 10% decreased risk of SWG from baseline to FU (OR 0.90 [0.84-0.97]; p=0.004) and 26% decreased risk comparing highest to lowest quintiles of GRS (OR 0.74 [0.55-1.01]; p=0.061). Conclusion: Our data suggest that BMI GRS associates with BMI at all adult ages and with unstable weight and substantial weight gain until the later mid- dle age. However, our results indicate inversed association with weight gain at older ages. Whether the latter can be explained by the BMI GRS accentuating the age related loss of muscle mass, other age related weight loss, age related effects on appetite regulation, or other reasons, needs to be investigated in future studies. Supported by: VR, HLF, NNF, SDF, PF, KAWF, LF 80 Effects of genetic and environmental influences on abdominal adipose tissue compartments and hepatic lipid accumulation: a classical twin study A.L. Jermendy1, Z.D. Drobni1, T. Horvath1, C. Celeng1, A. Panajotu1, F.I. Suhai1, A.D. Tarnoki2, D.L. Tarnoki2, B. Merkely1, P. Maurovich-Horvat1, G. Jermendy3; 1MTA-SE Lendulet Cardiovascular Imaging Research Group, 2Semmelweis University Department of Radiology and Oncotherapy, 3Bajcsy-Zsilinszky Hospital, Budapest, Hungary. Background and aims: In patients with diabetes and/or obesity, accumu- lation of abdominal adipose tissue and non-alcoholic fatty liver disease (NAFLD) are linked to increased cardiometabolic risk. Little is known about the genetic and environmental effects on the distribution of the abdominal adipose tissue compartments and hepatic lipid accumulation. The aim of the study was to assess the magnitude of genetic and environmental impact on the size of various abdominal adipose tissue compartments and the hepatic lipid accumulation within a cohort of healthy twin pairs. Materials and methods: In this classical twin study, 136 adult twin subjects (58.8% women; age: 56.8±9.3 years, weight: 77.1±17.2 kg, BMI 27.3±4.9 kg/ m2 [x±SD], 37 monozygotic [MZ] and 31 dizygotic [DZ] pairs) were in- volved. The twin pairs were investigated with a 256-slice CT-scanner. A 2 mm thick axial slice was acquired at the level of L3-L4. Subsequently a 50 mm wide axial image slab was acquired below the diaphragm. For each patient CT-based measurement of waist circumference, subcutaneous adipose tis- sue (SAT) and visceral adipose tissue (VAT) quantification were performed. Liver and spleen attenuation was determined by calculating the average of three 300 mm2 ROIs (regions of interest). Hepatic lipid accumulation was characterized by attenuation ratios (CTL/S) and ratio of ≤0.9 was assessed as sign of NAFLD. Concordance between MZ and DZ pairs was assessed by Pearson correlations. For assessing heritability of abdominal adipose tissue compartments and that of hepatic lipid accumulation, the structural equation (A-C-E) model was used. Results: Comparing MZ to DZ twin pairs, no significant differences were found in age (55.9±9.7 vs. 58.2±8.8 years), in BMI (27.2±3.9 vs. 26.5±4.0 kg/ m2), in waist circumference (94.0±12.9 vs. 95.4±13.1 cm), in SAT (206.0±79.9 vs. 200.9±83.1 cm2), in VAT (159.9±91.0 vs. 143.0±77.6 cm2), and in CTL/S ratio (1.1±0.2 vs 1.2±0.2); p>0.05 for all comparison. Strong correlations among BMI, SAT and VAT values were found in MZ twin pairs (r=0.63 [95% CI 0.34 - 0.84], r=0.74 [95% CI 0.54 - 0.90], r=0.60 [95% CI 0.34 - 0.79], re- spectively) whereas these correlations were weak or absent in DZ twin pairs (r=0.08 [95% CI -0.43 - 0.43], r=0.35 [95% CI 0.00 - 0.64], r=0.20 [95% CI -0.16 - 0.51], respectively). As for hepatic lipid accumulation, correlations among CTL/S values were absent in both MZ pairs (r=0.30, 95% CI -0.16 - 0.67) and DZ pairs (r=0.15, 95% CI -0.15 - 0.55). Using the structural equa- tion (A-C-E) model, relatively strong heritability index was found regarding BMI (58%, 95% CI 18-85%), SAT (74%, 95% CI 43-93%) and VAT (59%, 95% CI 22-82%) whereas environmental influences predominated in hepatic lipid accumulation (additive genetic effect 30% [95% CI 0-75%], shared environ- mental effect 1% [95% CI 0-41%], unique environmental effect 69% [95% CI 32-100%). Conclusion: Both BMI and abdominal adipose tissue compartments (SAT and VAT) have relatively strong heritability whereas hepatic lipid accumu- lation (presence of NAFLD) is predominantly influenced by environmental factors. Supported by: EFSD New Horizons 81 Endogenous GLP-1 alters brain activations in response to visual food-cues in reward and satiety circuits in humans J.S. ten Kulve1, R.G. Ijzerman1, L. van Bloemendaal1, F. Barkhof2, D.J. Veltman3, M. Diamant1; 1Diabetes centre / Internal medicine, VU University Medical Centre, 2Departement of Radiology, VU University Medical Centre, 3Departement of Psychiatry, VU University Medical Centre, Amsterdam, Netherlands. Background and aims: The central nervous system (CNS) plays a major role in the regulation of feeding and maintenance of body weight. Food inges- tion activates the secretion of gut-hormones, such as glucagon-like peptide-1 (GLP-1). GLP-1 has been proposed to be involved in the CNS regulation of feeding, by relaying information about the nutritional status to the CNS. We Diabetologia (2014) 57:[Suppl1]S1–S564 S 41 1 C hypothesised that endogenous GLP-1 has effects on central reward and satiety circuits in overweight individuals with diabetes and healthy, lean individuals. Materials and methods: We included overweight patients with type 2 diabe- tes (T2DM) (n=20, mean ± SD age 59.3 ± 4.1 yrs, BMI 32.0 ± 4.7 kg/m2, 11 males) and age matched healthy, lean controls (n=20, mean ± SD age 56.3 ± 6.2 yrs, BMI 22.5 ± 1.7 kg/m2, 10 males). Using functional MRI (fMRI), we determined the effects of blocking endogenous GLP-1 on CNS responses to visual food-cues before and after a standardized liquid meal. To block the endogenous GLP-1 effects, intravenous administration of the GLP-1 recep- tor antagonist exendin 9-39 (ex9-39) was used and compared to placebo in- fusion. During the fMRI session, subjects were presented pictures of high calorie, low calorie and non-food objects. Imaging data were analysed using SPM8 and activation contrasts were computed (food vs. non-food). Results: In the fasting state, obese T2DM patients versus lean individu- als showed increased brain activation in response to food pictures within left amygdala, right orbitofrontal cortex and bilateral insula. The stand- ardized meal reduced these hyperactivations in bilateral insula in patients with T2DM. Blocking endogenous GLP-1 with ex9-39 partly prevented this meal induced reduction in brain activations. In healthy lean individuals, the standardized meal also reduced activation in the insula in response to food pictures, however to a lesser extent and only in right insula. Blocking endog- enous GLP-1 with ex9-39 did not prevent this meal induced reduction in brain activations. Conclusion: Patients with T2DM showed increased brain activations in re- sponse to food pictures in areas involved in reward and satiety while fasted. Intake of a meal reduced these brain responses and this effect was blunted with blockade of endogenous GLP-1. These effects could not be detected in healthy, lean individuals. The lower brain activation in response to food pic- tures in healthy lean individuals at baseline, may have decreased the likeli- hood to detect changes in these brain activation after a meal with or without blocking endogenous GLP-1. Our findings in patients with T2DM support the hypothesis that endogenous GLP-1 is involved in the regulation of central reward and satiety. Clinical Trial Registration Number: NCT 01363609 82 A high fat diet during pregnancy and lactation affects the metabolic fate of Gipr-/- mice via hypothalamic insulin signalling and DNA-methylation of lipid metabolism genes F. Keyhani Nejad1,2, F. Isken1,2, M.A. Osterhoff1,2, B. Nitz3,4, T. Ludwig4, H. Grallert3,4, A.F.H. Pfeiffer1,2, M. Kruse1,2; 1Department of Clinical Nutrition, German Institute of Human Nutrition, Nuthetal, 2Department for Endocrinology, Diabetes and Nutrition, Charité – University of Medicine, Berlin, 3Research Unit of Molecular Epidemiology, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, 4Institute of Epidemiology II, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany. Background and aims: Maternal high-fat (HF) feeding during intra-uterine (IU) and lactation (L) periods predisposes the offspring for obesity and im- paired glucose homeostasis in mice. Inhibition of Glucose-dependent insu- linotropic polypeptide (GIP) signaling in GIP receptor knockout mice (Gipr- /-) leads to protection from HF induced obesity. We reported that Gipr-/- mice exposed to HF during IU/L were no longer protected from diet induced obesity and had a decreased glucose tolerance in adulthood. We hypothesized that the metabolic programming of IU/L HF consumption triggers hypotha- lamic insulin resistance and promoter DNA methylation of key genes in fat metabolism and their subsequently altered gene expression in Gipr-/- mice. Materials and methods: Female GIP receptor heterozygous (Gipr+/-) mice were fed either a HF (60% fat) or control (C, 10% fat) diet for 2 weeks prior to gestation and during IU/L. After weaning, male Gipr-/- and wild type (WT) offspring were kept on normal chow until the age of 25 weeks, after which all offspring were exposed to HF for the following 20 weeks. This resulted in Gipr-/- which were exposed to either a C (KO Ciu-HF) or a HF (KO HFiu- HF) diet during IU/L and a HF later in adulthood. WT mice fed a control diet during IU/L and a HF in adulthood served as controls (WT Ciu-HF). A glucose tolerance test (GTT) was performed by i.p. injection of 2 g/kg BW glucose. At 45 weeks of age, body fat content, adipocyte size and genes in- volved in hypothalamic insulin sensitivity and DNA-methylation of fatty acid oxidation genes in muscle were analyzed. Results: Body fat content significantly increased in KO HFiu-HF com- pared to KO Ciu-HF (22.6 ± 2.5 g vs. 19.5 ± 0.8 g, respectively; p S 42 1 C Conclusion: Endogenous opioid system plays an important role in the pathophysiology of obesity while the role of dopaminergic pathways remains questionable. Because bariatric surgery and concomitant weight loss recov- ers MOR availability, lowered MOR availability is a consequence of obesity and may mediate maintenance of excessive energy uptake. Understanding the opioidergic contribution to overeating is thus critical for developing new treatments for obesity. Clinical Trial Registration Number: NCT00793143 Supported by: Academy of Finland, The Finnish Diabetes Research Foundation, CLIGS 84 KBP-042 lowers body-weight and sustains weight-loss in high fat-diet rats S.T. Hjuler, S. Gydesen, K.V. Andreassen, M.A. Karsdal, K. Henriksen; Diabetes group, Nordic Bioscience, Herlev, Denmark. Background and aims: KBP-042 is a dual amylin- and calcitonin receptor agonist, with superior activity compared to salmon calcitonin. In this study we evaluated the long term potential of KBP-042 as a treatment against obe- sity. We evaluated body weight, adiposity, glucose tolerance and insulin ac- tion in a rat model of obesity. Materials and methods: Male sprague-Dawley rats were fed a high fat-diet for ten weeks resulting in an obese and glucose intolerant phenotype. Based on body weight the rats were randomized into the treatment groups: Vehicle (saline), 0.625 µg/kg, 1.25 µg/kg, 2.5 µg/kg, 5.0 µg/kg, 10.0 µg/kg KBP-042 s.c. once daily (n=10), as well as pair-fed controls for the two highest con- centration groups (pair-fed 10 µg/kg and pair-fed 5 µg/kg) which were food restricted to match the food intake of the 10 µg/kg and 5 µg/kg groups. Results: A dose-dependent and sustained weight-loss was obtained. For the KBP-042 10 µg/kg group the treatment resulted in a 4.7 % weight reduction (20 % vehicle-corrected), while only a transient change in food intake. Fur- thermore, while the pair-fed groups did not lower body weight, they did not match the extent of their corresponding treatment group, +5.6 % and +7.1 % respectively (-5.6 % and -8.0 % vehicle-corrected). Adiposity was vastly im- proved after treatment with KBP-042. The visceral fat depots were significant reduced (~40% perirenal AT: p S 43 1 C OP 15 Diabetic nephropathy: epidemiology and genetics 85 Prevalence of nonalbuminuric chronic kidney disease (CKD) is increasing in patients with type 1 diabetes mellitus E. Russo1, G. Penno1, M. Garofolo1, D. Lucchesi1, L. Pucci2, V. Sancho Bornez1, L. Giusti1, R. Bellante1, R. Russo2, R. Miccoli1, S. Del Prato1; 1Endocrinology and Metabolism, University of Pisa, 2Institute of Agricultural Biology and Biotechnology, CNR, Pisa, Italy. Background and aims: According to traditional paradigms of diabetic ne- phropathy, albuminuria precedes glomerular filtration rate (GFR) loss in the progression to CKD. However, even in type 1 DM (T1DM), recent findings (DCCT/EDIC) demonstrate that GFR loss may occur in normoalbuminuria. We compared prevalence of different CKD phenotypes in two cohorts of Ital- ian T1DM subjects. Materials and methods: The first cohort (C1) consisted of 777 T1DM re- cruited 2001-2009 at our Metabolic Unit; the second cohort (C2) consisted of 936 T1DM belonging from the EURODIAB IDDM Complications Study (they attended 9 Italian centres in 1989-1991). Eligibility criteria were the same for both cohorts. However, C2 was stratified at enrollment by sex, age (15-29, 30-44, 45-59 yrs) and duration (1-7, 8-14, ≥15 yrs). Methods em- ployed to assess diabetic complications were similar. Results: Gender distribution (M/F: 52/48% vs 51/49%) and current smok- ers (29.6% vs 30.4%) were similar. Due to inclusion criteria, C1 were older (40.2±11.7 vs 32.1±10.5) and had longer diabetes duration (19.4±12.2 vs 14.4±9.0 years; both p S 44 1 C 87 Folkhälsan Research Foundation, the Wilhelm and Else Stockmann Foundation. Kidney injury molecule - 1 is linked to the loss of kidney function and life span, in patients with type 1 diabetes N.M. Panduru1,2, N. Sandholm2, C. Forsblom2, V. Harjutsalo2,3, A. Bierhaus4, M. Saraheimo5, P. Humpert6, P.-H. Groop5,2; 12nd Clinical Department - Diabetes, Nutrition and Metabolic Diseases Chair, “Carol Davila” University of Medicine and Pharmacy, Bucharest, Romania, 2Folkhälsan Institute of Genetics, Folkhälsan Research Center, Helsinki, Finland, 3Diabetes Prevention Unit, National Institute for Health and Welfare, Helsinki, Finland, 4Department of Medicine I and Clinical Chemistry, University of Heidelberg, Germany, 5Division of Nephrology - Department of Medicine, Helsinki University Central Hospital, Finland, 6Stoffwechselzentrum Rhein Pfalz, Mannheim, Germany. Background and aims: We previously showed that kidney injury molecule 1 (KIM-1) predicts progression to ESRD, but its relation with kidney function and mortality in patients with type 1 diabetes is not very clear. The aim of this study is to investigate if KIM-1 predicts and has a causal role in the loss of kidney function or life span by a Mendelian randomisation approach. Materials and methods: We enrolled at baseline 1573 patients with type 1 diabetes divided in three groups: 953 patients with normal AER, 269 patients with microalbuminuria and 350 patients with macroalbuminuria. KIM-1 was measured at baseline, by ELISA and normalized with urinary creatinine. Kid- ney function was evaluated by estimated GFR acording to CKD-EPI formula. Life span was considered the vital time. The predictive value of KIM-1 for the loss of kidney function and life span was evaluated by linear regression. Mul- tiple linear regression models estimated the observed effect (association) of KIM-1 on eGFR or life span. We assessed the causal effect of KIM-1 on eGFR and life span, by Instrumental variable analysis, using two stage least squares method (2SLS), with the top SNP associated with KIM-1 from our GWAS. Results: In linear regression analysis KIM-1 predicted the loss of GFR in univariate analysis (β=-4.522; P 98%. Results: Seventy-five renal events (66 doublings of serum creatinine concen- tration, and 9 cases of end-stage renal failure) occurred in genotyped patients during the study. The 400K allele was significantly associated with a higher risk of incident renal event: sex and age adjusted OR 1.66, 95%CI 1.15-2.39, P=0.007. This association was still significant after multiple additional adjust- ments for values at baseline (BMI, blood lipids, estimated glomerular filtra- tion rate, urinary albumin excretion): OR 1.57, 95%CI 1.07-2.31, P= 0.02. There was a trend toward an interaction with ramipril treatment (P interac- tion= 0.06). The 400K allele was associated with a higher risk in the ramipril Diabetologia (2014) 57:[Suppl1]S1–S564 S 45 1 C treated group (OR 2.41, 95%CI 1.41-4.13, P 10 years. The study included four cohorts of T2D patients: GoDARTS (n=3240), SDR (n=1830), Steno (n=294) and MNI (n=353). We analyzed ~9.2 million single nucleotide polymorphisms (SNPs), imputed based on the 1000G (March 2012) refer- ence panel, using logistic regression, adjusting for sex, age at onset and dura- tion of diabetes, for five phenotypes: DKD, CKD, MiAU, MaAU+ESRD and ESRD. Meta-analyses were performed using a fixed effects model. We also performed joint analyses with four T1D studies analyzed using similar meth- ods and phenotype definitions: FinnDiane (n=3415), Eurodiab (n=789), SDR (n=556) and Cambridge (n=396). Results: In T2D, rs2206136 (OR=1.2, p=2.1x10-8) near PLCB4 was signifi- cantly associated with CKD (3094 cases, 2906 controls). Nominal associa- tions were also seen with DKD, MiAU and ESRD (p0.1) suggesting an effect specifically on late stages of disease. Conclusion: We have identified two new candidate loci for DKD situated near the PLCB4 (phospholipase C, beta 4) and STXBP5L (syntaxin binding protein 5-like) genes. PLCB4 has previously been shown to be differentially expressed in DKD, making it a strong candidate gene, possibly affecting dis- ease risk via the DAG/PKC pathway. Supported by: IMI, European Commission’s FP7 (the SUMMIT consortium, IMI-2008/115006) OP 16 Mechanisms of cardiovascular disease 91 Low levels of C-peptide production protect from complications and improve HbA1c control in longstanding type 1 diabetes D.L. Faustman1, P.E. Reinhold1, S.L. Washer1, E. Hsu1, M. Zhao1, D. Burger1, H. Zheng2; 1Immunobiology Laboratory, Massachusetts General Hospital/Harvard Medical School, Charlestown, 2Biostatistics, Massachusetts General Hospital/Harvard Medical School, Charlestown, USA. Background and aims: Low levels of C-peptide (2.5-50 pmol/L) are pro- duced for decades after the onset of type I diabetes, but the clinical signifi- cance is unknown. This study seeks to understand the clinical significance of extremely low levels (< 10 pmol/L) of C-peptide production. Materials and methods: We evaluated the relationship between low levels of C-peptide and age of onset (n=1273), diabetes complications (n=324), HbA1c control (n=807), risk of hypoglycemia (n=331), and response to a mixed meal tolerance test (MMTT, n=9) in patients with longstanding type I diabetes. C-peptide samples were tested using a regular or ultrasensitive C-peptide ELISA kit. Hypoglycemia risk was determined using a validated survey. For the MMTT, each subject had three MMTT tests performed over a three-month period to understand biological variation in stimulated C- peptide at low ranges of basal pancreas function. Results: After adjusting for disease duration, we found that extremely low levels of C-peptide were associated with risk for diabetes-related complica- tions (e.g., nephropathy, neuropathy, cardiovascular disease)(p=0.03) and poorer metabolic control captured by HbA1c (p=0.01). There was no asso- ciation between C-peptide < 10 pmol/L and risk of hypoglycemia. Even at extremely low levels of C-peptide production, beta-islet cells responded to a glucose challenge from a MMTT by secreting insulin, indicating that pan- creatic function is preserved, but patients with undetectable C-peptide ( 40 mg/24 hr) (P= 0.0038) which remained significant after adjustment for DCCT HbA1c. In EDIC, LW-1 correlated with retinopathy progression at EDIC Year 13-16, intima media thickness(IMT) at Yr 6 (n=147, P=0.014) and left ventricular mass (EDIC Yr 14-16) adjusted for EDIC A1c (P=0.004). LW-1 correlated highly (P less than 0.0001) with collagen modifications in the order glucosepane > pento- Diabetologia (2014) 57:[Suppl1]S1–S564 S 46 1 C sidine > pepsin insolubility > collagen fluorescence > MG-H1, and weaker with fructose-lysine (P = 0.0021), and explained almost 10% of the variability in DCCT HbA1c. Conclusion: LW-1 is a novel robust marker which predicts progression of retinopathy and nephropathy and future progression of IMT and LVM in- crease independently of the effects of HbA1c. Supported by: NIH/NIDDK, JDRF 93 Plasma levels of MMP-2, -3 and -10, and of TIMP-1 are associated with vascular complications in patients with type 1 diabetes: the EURODIAB prospective complications study S.A. Peeters1,2, L. Engelen1,3, J. Buijs2, N. Chaturvedi4, J.H. Fuller5, C.G. Schalkwijk1,3, C.D.A. Stehouwer1,3; 1Department of Internal Medicine, Maastricht University Medical Centre, Netherlands, 2Department of Internal Medicine, Atrium Medical Centre, Heerlen, Netherlands, 3CARIM School for Cardiovascular Diseases, Maastricht University Medical Centre, Netherlands, 4Institute of Cardiovascular Sciences, University College, London, UK, 5Department of Epidemiology and Public Health, University College, London, UK. Background and aims: Extracellular matrix remodeling by matrix metallo- proteinases (MMPs) and tissue inhibitor of metalloproteinase (TIMP) may lead to micro- and/or macrovascular complications in type 1 diabetes. Evi- dence so far with regard to the associations between plasma MMPs and vas- cular complications is contradictory, as both positive and null associations have been described in relatively small studies. Therefore we have investigat- ed the associations between plasma MMP-1, -2, -3, -9 and -10, and TIMP-1 on the one hand and cardiovascular disease (CVD) and microvascular com- plications on the other in a cohort of type 1 diabetic patients, and the extent to which such associations may be explained (i.e. mediated) by low-grade inflammation (LGI) and/or endothelial dysfunction (ED), as estimated by plasma markers. Materials and methods: The study included 493 type 1 diabetic patients (39.5 ± 9.9 years old, 51% men) from the EURODIAB Prospective Complica- tions Study. We used linear regression analyses to investigate the differences in plasma levels of MMP-1, -2, -3, -9 and -10, and of TIMP-1 between pa- tients with vs. without CVD, albuminuria (normo, micro and macro) and retinopathy (no, non-proliferative and proliferative). All analyses were ad- justed for age, sex, duration of diabetes and Hba1c and additionally for other cardiovascular risk factors and other vascular complications, as appropriate. Standardized concentrations of plasma CRP, IL-6 and TNF-α were averaged to compose an LGI score and standardized concentrations of plasma sV- CAM-1 and sE-selectin composed an ED score. These LGI and ED scores were then added to the fully adjusted model. Results: Patients with CVD (n=118) had significantly higher levels of TIMP- 1 [β = 0.32 SD (95%CI 0.12; 0.52)] than those without CVD (n=375). Higher plasma levels of MMP-2, MMP-3, MMP-10 and TIMP-1 were associated with increasing levels of albuminuria (p-trends were 0.034, 0.004, 0.004 and 0.001, respectively). The severity of retinopathy was significantly associated with higher levels of MMP-2 (p-trend=0.022). MMP-1, -3, -9 and -10, and TIMP- 1 were significantly and positively associated with the LGI score. MMP-2, MMP-10 and TIMP-1 were significantly and positively associated with the ED score. Nevertheless, the significant associations between plasma levels of MMPs and TIMP-1 on the one hand and CVD, albuminuria and retinopathy on the other were largely independent of LGI and ED. Conclusion: In patients with type 1 diabetes, prevalent CVD, albuminuria and retinopathy were significantly associated with higher plasma levels of MMP-2, MMP-3, MMP-10 and TIMP-1, and these associations were largely independent of LGI and ED. Thus, these data support the hypothesis that extracellular matrix remodeling, by the action of MMPs and TIMP-1, is in- volved in vascular complications in type 1 diabetes. 94 Acute glucose and insulin changes during OGTT relate to LV-myocardial deformation changes, untwisting and coronary-flow-reserve through increased arterial stiffness F. Kousathana1, I. Ikonomidis2, G. Pavlidis2, C. Koukoulis2, M. Varoudi2, G. Matsaggouras1, H. Triantafyllidi2, J. Lekakis2, G. Dimitriadis1, V.A. Lambadiari1; 12nd Dept of Internal Medicine, Research Institute and Diabetes Center, Attikon University Hospital, 22nd Cardiology Dept, University of Athens, Attikon Hospital, Athens, Greece. Background and aims: Insulin resistance is linked to endothelial dysfunc- tion and increased arterial stiffness. Increased Arterial stiffness may impair LV function. We investigated whether first-degree relatives of diabetic pa- tients have similarly impaired coronary microcirculation, LV-myocardial strain and twisting with diabetic patients, as assessed after an oral glucose tolerance test (OGTT). Materials and methods: In 76 subjects without known diabetes a standard 75-gr OGTT was performed. Glucose and insulin levels, pulse wave veloc- ity (PWVa) and augmentation index (AI) (Arteriograph, Tensiomed) were measured at 0, 30, 60, 90 and 120 min after glucose load. At 0 and 120 min, we measured: a) E’ and A’ mitral annular velocities and their ratio E’/A’ using tis- sue Doppler imaging, b) LV longitudinal (GLS-%), strain, systolic(LGSr) and diastolic strain rate (LGSrE), twisting (Tw -deg), peak twisting (Tw-deg/sec) velocity, and peak untwisting (unTw) velocity using speckle tracking echo- cardiography and c) coronary flow reserve (CFR) of the LAD after adenosine infusion using Doppler echocardiography. We assessed insulin resistance us- ing insulin sensitivity index (ISI) which includes both insulin and glucose levels at baseline and 120 min after OGTT. Results: Of the 76 subjects, 36 who were first degree relatives of diabetics had normal OGTT (relatives), 20 had normal OGTT and no family history of dia- betes (normals), and 20 had abnormal OGTT (diabetics). Age, sex and BMI were similar between subgroups (p=ns). Compared to normals, diabetics and relatives had both higher baseline PWVa (9.3±2 vs. 8.1±2 vs. 7.2±1.6 m/ sec,), AI (23±9 vs. 24±14, 18±15%,), insulin (median 14 vs. 15 vs. 10 μU/ml, p S 47 1 C 95 Vascular and cellular ageing in patients with type 2 diabetes mellitus N.V. Brailova1,2, E. Dudinskaya1, O. Tkacheva1, S. Boytsov1, I. Strazhesko1, D. Akasheva1, M. Shestakova2,3, Y. Brailov3, M. Pokrovskaya1; 1National Research Centre for Preventive Medicine, 22National Research Center for Endocrinology, 3Align Technology Inc., Moscow, Russian Federation. Background and aims: It is known that glucose disturbances contribute to micro- and macrovascular complications and vascular aging. The length of telomere (TL) is considered as a biomarker for vascular and cellular aging. But the interrelation of vascular aging and cellular aging in type 2 diabetes mellitus (T2DM) and pathogenic mechanisms of this interrelation remains a challenge. The aim of our study was to determine mechanisms of TL shorten- ing and vascular aging in patients with T2DM. Materials and methods: The study group included 50 patients with T2DM in mean age 58.4±7.83 years and 156 healthy patients in mean age 57.04±7.7 years. TL and telomerase activity (TA) was assessed by quantitative polymer- ase chain reaction (PCR). Intima-media thickness (IMT) and plaque pres- ence (PP) were determined by ultrasonography in both left and right carotid arteries. Arterial stiffness (AS) was appreciated by aortic pulse wave velocity (PWV) measuring by SphygmoCor (AtCor Medical). Endothelial dysfunc- tion as assessed by flow-mediated endothelium-dependent dilation (FMV) in response to reactive hyperemia and endothelium-independent vasodilation in response to nitroglycerine (NDV). Oxidative stress was assessed by malon- dialdehyde measuring, inflammation was estimated by interleukin-6 (IL-6), C-reactive protein (CRP), fibrinogen measuring. Results: We found in group with T2DM compared with the control group of healthy patients a greater telomere shortening (9,57 vs 9,75, р=0,0051) and TA reduction (0,33 vs 0,49, р=0,0023). Vascular aging are more pronounced in patients with T2DM, than without diabetes: PWV СРПВ 12,3 m/s vs 11,3 м/с (р=0,0032), IMT 0,93 mm vs 0,77 mm (р S 48 1 C OP 17 Intra- and inter-islet cell signalling 97 LKB1 and AMPK regulate Nptx2 expression and glutamate signalling in pancreatic beta cells M.-S. Nguyen-Tu1, C. Kantor1, S. Sayers1, T.J. Pullen1, G. Sun1, M. Kone1, M. Ibberson2, J. Ferrer3, B. Thorens4, A. Swisa5, Y. Dor5, T. Hildebrandt6, I. Uphues6, G.A. Rutter1; 1Section of Cell Biology, Department of Medicine, Imperial College London, UK, 2Swiss institute of Bioinformatics, University of Lausanne, Switzerland, 3Endocrinology and Metabolic Medicine, Department of Medicine, Imperial College London, UK, 4Center for integrative Genomics, University of Lausanne, Switzerland, 5Hebrew University of Jerusalem, School of Medicine-IMRC, Israel, 6Boerhringer Ingelheim International GmbH, Ingelheim, Germany. Background and aims: Inactivation in beta cells of the tumour suppressor Liver kinase B1 (LKB1/STK11) or the downstream enzyme AMP-activated protein kinase (AMPK) exerts dramatic effects on beta cell growth and in- sulin secretion. Here, we used massive parallel sequencing (RNASeq), and subsequent functional analyses, to identify gene clusters which may mediate these effects. Materials and methods: Mice null for LKB1 or both AMPK catalytic (α1, α2) subunits were generated by Ins1Cre-mediated (beta cell-selective) deletion of flox’d alleles. Islets were isolated from four 12 week old mice per geno- type. After 24 h culture at 11 mM glucose RNA was extracted (RNAEasy) before deep sequencing (RNASeq) on a Hiseq 2000. Cytosolic free Ca2+ was measured using fura-2 using an Olympus IX81 microscope with microman- ager-controlled data capture via an Andor Zyla cMOS camera. Immunohisto- chemial analysis of pancreatic slices was performed using a rabbit anti-Nptx2 antibody on a Zeiss-200M microscope (Zen software), and analysed using ImageJ. Insulin secretion was measured during static incubations (30 min) at 3 or 16 mM glucose using radioimmunoassay. Results: Amongst the mRNAs most strongly up-regulated by LKB1 deletion was a cluster involved with glutamate signalling, including Nptx2, encoding neuronal pentraxin 2 (11.2-fold, E-value S 49 1 C depleted: 90±10%, n=7-8, P S 50 1 C eNpHR3.0-EYFP following excision of a loxP-flanked STOP cassette. To al- low user-directed single cell silencing within the field of view, a diffraction- limited 585nm laser was coupled via a fibre optic to a custom-manufactured dichroic array. Results: Beta cells form a scale-free network which supports the synchro- nous propagation of glucose-stimulated Ca2+ waves by efficiently connect- ing distant regions of the intact islet (n = 12 recordings from 5 animals) (power law fit; R² = 0.7247). A typical feature of such a topology was the non-random appearance of superconnected hub cells whose firing activity repetitively preceded that of the remainder of the population. Online map- ping of islet functional architecture using MATLAB routines coupled directly to the imaging setup revealed the presence of a statistically-stable hub cell population. Silencing of individual identified hubs using a pinpointing laser had catastrophic consequences for coordinated islet responses to glucose, and this could be reversed simply by ceasing illumination (11.5 ± 1.8 versus 3.1 ± 0.8 % correlated cell pairs, laser OFF versus laser ON, respectively; P0.05) (n = 7 recordings from 4 animals). Further supporting a role for distinct wiring patterns in glucose-regulated islet connectivity, low-grade cytokine (IL-1β and IL-6) treatment resulted in a dramatic and rapid collapse in correlated cell-cell activity due to impaired hub function (9.6 ± 0.9 versus 5.1 ± 0.6 % correlated cell pairs, 0 versus 4 hrs cytokine exposure, respec- tively; P< 0.01) (n = 7 recordings from 3 animals). Conclusion: The intra-islet circuitry is dominated by superconnected hub cells which dictate population responses to glucose. Notably, these hubs are vulnerable to pro-inflammatory T2D insults and may contribute to the re- duced functional beta cell mass that accompanies glucose intolerance. Supported by: Diabetes UK, Wellcome Trust, MRC and the Royal Society. OP 18 Novel genes for type 2 diabetes and its complications 103 A common Greenlandic TBC1D4 variant confers muscle insulin resistance and type 2 diabetes N. Grarup1, I. Moltke2, M.E. Jørgensen3, P. Bjerregaard4, J.T. Treebak1, T.S. Korneliussen5, M.A. Andersen1, T.S. Nielsen1, N.T. Krarup1, A. Linneberg6, J. Wang7, O. Pedersen1, R. Nielsen8, A. Albrechtsen9, T. Hansen1; 1The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Denmark, 2Department of Human Genetics, University of Chicago, USA, 3Steno Diabetes Center, Gentofte, Denmark, 4National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark, 5Centre for GeoGenetics, Natural History Museum of Denmark, University of Copenhagen, Denmark, 6Research Centre for Prevention and Health, Glostrup University Hospital, Glostrup, Denmark, 7BGI-Shenzhen, China, 8Department of Integrative Biology, University of California, Berkeley, USA, 9The Bioinformatics Centre, Department of Biology, University of Copenhagen, Denmark. Background and aims: Previous studies have shown that founder popula- tions can be valuable for genetic association studies, because they have in- creased levels of linkage disequilibrium and may harbor deleterious muta- tions of higher frequency. The same can be argued for historically small and isolated populations. Motivated by this, and by the dramatically increased prevalence of type 2 diabetes in the small Greenlandic population, we per- formed association mapping of four type 2 diabetes-related quantitative traits, plasma glucose and serum insulin levels at fasting and at two hours during an OGTT, in Greenlandic individuals without known diabetes. Materials and methods: Discovery analyses were performed in up to 2,757 participants of the Inuit Health in Transition cohort and replication in up to 1,064 participants from the B99 cohort. Participants underwent an OGTT. Samples were genotyped by the Illumina CardioMetabochip. Association testing was done using a linear mixed model in GEMMA to control for ad- mixture and relatedness. Exome sequencing of selected individuals was done by Agilent SureSelect capture and Illumina sequencing. Results: Applying array-based genotyping and exome sequencing, we dis- covered a nonsense variant in TBC1D4 with an allele frequency of 17% in the Greenlandic population. Under a recessive model, this variant strongly increases the levels of plasma glucose (β=3.8 mmol/L, P=2.5×10-35) and se- rum insulin (β=165 pmol/L, P=1.5×10-20) two hours after an oral glucose load, while reducing fasting plasma glucose (β=-0.18 mmol/L, P=1.1×10-6) and fasting serum insulin (-8.3 pmol/L, P=0.0014). Even the heterozygous carriers have an increased 2 hour plasma glucose (β=0.43 mmol/L, P=5.3×10- 5). The same polymorphism was associated with an increased risk of type 2 diabetes risk (OR 10.3, P=1.6×10-24). All findings were replicated in the B99 cohort. Analyses of muscle biopsies from mutation carriers showed de- creased mRNA and protein expression of the long muscle-specific isoform of TBC1D4 and decreased muscle protein abundance of GLUT4 projecting a decreased insulin-stimulated glucose uptake in skeletal muscle. Conclusion: We establish a single variant which imposes effect sizes several times larger than any findings of large-scale genome-wide association studies and accounts for more than 10% of all cases of type 2 diabetes in a popula- tion. This finding constitutes further proof that it is valuable to conduct ge- netic association studies outside the traditional setting of large homogeneous populations. 104 Fine-mapping type 2 diabetes susceptibility loci with high-density imputation A.P. Morris1,2, T.M. Teslovich3, T. Ferreira2, A. Mahajan2, Y. Lee3, N.W. Rayner2, R. Magi4, on behalf of the DIAGRAM Consortium; 1Department of Biostatistics, University of Liverpool, 2Wellcome Trust Centre for Human Genetics, University of Oxford, UK, 3Department of Biostatistics, University of Michigan, Ann Arbor, USA, 4Estonian Genome Centre, University of Tartu, Estonia. Background and aims: Genome-wide association studies (GWAS) have been successful in identifying loci for type 2 diabetes (T2D). These loci are typi- Diabetologia (2014) 57:[Suppl1]S1–S564 S 51 1 C cally characterised by common lead SNPs with association signals that extend over large genomic intervals. Consequently, there has been limited progress in localising causal variants for T2D and establishing the functional impact of these loci in the pathogenesis of the disease. We aimed to: (i) identify multiple signals of association with T2D in established loci; (ii) assess the evidence for association with low-frequency (LF) variants, minor allele frequency (MAF) 6.3), a signal driven exclusively by East Asians (odds ratio=1.79[1.59-1.99], MAF~10%). Reciprocal conditional analyses at these loci revealed the novel coding vari- ants to be either independent of, or indistinguishable from previously report- ed lead GWAS SNPs, and implicate several novel genes - including RREB1, THADA, and TM6SF2 - in T2D pathogenesis. The remaining six variants, mapping to three loci, were located entirely outside known GWAS regions. These include two variants in PAM (D563G, log10BF=9.96) and PPIP5K2 (S1207G, log10BF=7.47), for which a T2D-association was recently described in the Icelandic population. Three further highly-correlated coding variants at MTMR3/ASCC2 were associated with T2D (Europeans specific signal). The MTMR3 variant (log10BF=6.46) has the strongest residual association signal after accounting for the other variants and represents the most plausible can- didate at this novel locus. The last novel locus was FAM63A (log10BF=5.24) and signal was driven by a LF variant. The T2D risk-allele has a MAF of 1% in Europeans, and was otherwise only detected in South Asian sequence data (11 copies, p=0.02 for association). Conclusion: Here we show that rare and LF coding variants of moderate to large effect are not major contributors to T2D risk. We identify 26 coding variants with exome-wide significant associations with T2D and highlight the promise of this approach to provide insight into the pathophysiology of the disease. 106 Meta-analysis of birth weight genome-wide association studies identifies two novel loci extending links between early growth and adult metabolic diseases M. Horikoshi1,2, M.N. Kooijman3, J.P. Bradfield4, D. Strachan5, N.V. Tejedor6, E. Kreiner-Møller7, P. Joshi8, V. Lindi9, N. Grarup10, S.F.A. Grant4, M.I. McCarthy1,2, I. Prokopenko11, R.M. Freathy12, A.P. Morris1,13, EGG Consortium; 1Wellcome Trust Centre for Human Genetics, University of Oxford, 2Oxford Centre for Diabetes, Endocrinology & Metabolism, UK, 3Department of Epidemiology, Erasmus Medical Center, Rotterdam, Netherlands, 4Center for Applied Genomics, The Children’s Hospital of Philadelphia, USA, 5Population Health Research Institute, St George’s, University of London, UK, 6Center for Research in Environmental Epidemiology (CREAL), Barcelona, Spain, 7COPSAC, the Copenhagen Prospective Studies on Asthma in Childhood; Health Sciences, University of Copenhagen, Denmark, 8Centre for Population Health Sciences, University of Edinburgh, UK, 9Department of Physiology, Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland, 10The Novo Nordisk Foundation, Center for Basic Metabolic Research, Faculty Of Medical and Health Sciences, University of Copenhagen, Denmark, 11Department of Genomics of Common Disease, Imperial College London, 12University of Exeter Medical School, 13Department of Biostatistics, University of Liverpool, UK. Background and aims: Lower birth weight (BW) is associated with increased risk of future type 2 diabetes (T2D) and cardiovascular diseases. Based on HapMap 2 imputation, we previously reported 7 loci associated with BW, of which two (ADCY5 and CDKAL1) have been implicated in T2D and one (ADRB1) in hypertension. We have expanded this effort by increasing sample size and undertaking imputation up to 20.8M SNPs from the more dense 1000 Genomes Project reference panels. With these data we aimed to 1) dis- cover novel loci; 2) assess evidence of association with low-frequency variants (minor allele frequency S 52 1 C 107 Functional SNPs as focused gene probes: results from SUMMIT on diabetic nephropathy, coronary and lower extremity artery disease D. Ziemek1, E. Fauman2, N.R. Van Zuydam3, N.W. Rayner3, C.L. Hyde2, N. Sandholm4, C. Forsblom4, E. Ahlquist5, C. Ladenvall5, SUMMIT Consortium; 1Pfizer, Berlin, Germany, 2Pfizer, Cambridge, USA, 3OCDEM and WTCHG, University of Oxford, UK, 4FINNDIANE, Folkhalsan Institute of Genetics, University of Helsinki, Finland, 5Diabetes Center, Lund University, Sweden. Background and aims: Diabetic nephropathy (DN), coronary (CAD) and lower extremity artery disease (LEAD) are devastating, but, at least partially, heritable complications of diabetes with a poorly understood mechanistic basis. To move towards more effective preventative and therapeutic strate- gies, the SUMMIT consortium has conducted large-scale genome-wide as- sociation studies. Despite large sample sizes for CAD (4009 cases and 7918 controls), LEAD (2345 cases and 8706 controls), and specific sub-phenotype interrogation in DN (5908 cases and 4965 controls), only 3 genome wide sig- nificant signals (p S 53 1 C OP 19 Insulin: clinical decision making 109 Clinical factors influencing the basal rate profile in subjects with type 1 diabetes treated with continuous subcutaneous insulin infusion (insulin pumps) M. Kahle, U. Buss, A. Varnhorn, B. Nawrodt, M.A. Nauck; Diabeteszentrum Bad Lauterberg, Germany. Background and aims: Defining an individually appropriate basal rate pro- file in patients with type 1 diabetes requires knowledge about the general shape of such a profile, and needs to take into account patient factors with a significant influence on (a) the total dose of insulin administered as a basal rate and (b) the circadian variation in insulin needs. We assessed these influ- ences based on a retrospective analysis of 339 patients, in whom the adequacy of basal rates was checked by means of a 24 h fasting period. Materials and methods: Laboratory-based blood glucose profiles determined during regular day and night (eating meals) and during a 24 h fasting period (6 p.m. to 6 p.m.) and basal rate profiles corrected after the fasting period were entered into our database for 339 patients with type 1 diabetes during inpa- tient treatment in a specialized diabetes centre (183 women, 156 men; age 41 ± 14 years, HbA1c 8.3 ± 1.6 %; BMI 26.1 ± 4.9 kg/m 2; diabetes duration 20 ± 12 years; insulin dose 0.55 ± 0.20 IU/kg body weight per day). Patients were divided according to gender, or into tertiles of age, body-mass-index, and dia- betes duration. A significant influence of the characteristic in question was as- sessed by repeated-measures ANOVA on hourly basal rates and also expressed as a percentage of the overall 24 h basal rate for each individual hour. Results: A 24 h fasting period resulted in relatively stable blood glucose val- ues of 6.6 ± 2.8 mmol/L with 11.5 % of the values > 9.9 mmol/L and 11.3 % of the values < 3.7 mmol/L. Basal rates after correction for low or high blood glucose values during the fasting period showed typical circadian rhythms with “dawn-” and “dusk-” phenomena and a total (mean ± SD) basal rate of 21.5 ± 9.2 IU/day. There were significant differences due to gender (more prominent dawn phenomenon in males; p < 0.0001)), age (more prominent dawn phenomenon in younger subjects; p = 0.036), body mass index (higher overall basal rates in more obese subjects; p < 0.0001; with no influence on the circadian distribution p = 1.00), and diabetes duration (more prominent dawn phenomenon in those with shorter diabetes duration; p < 0.0001). Conclusion: Our large database helps define basal rate profiles for subjects with type 1 diabetes treated with continuous subcutaneous insulin infusions. It also allows to take into account clinical characteristics of the patients with a significant influence on the basal rate profile (both total amount of insulin and circadian profile) to individually predict insulin requirements during in- sulin pump treatment. Based on multivariate regression analysis, it may be possible to predict individual hourly basal rates with reasonable confidence intervals. 110 Relationship of HbA1c with body weight change over 4 years from starting insulin therapy in people with type 2 diabetes: the CREDIT study B. Balkau1, N. Freemantle2, F. Calvi-Gries3, V. Pilorget4, M. Vincent4, P. Home5; 1INSERM, U1018, University Paris Sud 11, UMRS 1018, Villejuif, France, 2University College London, UK, 3AtlanStat, Rezé, France, 4Sanofi, Paris, France, 5Newcastle University, Newcastle upon Tyne, UK. Background and aims: Insulin treatment is often delayed for people with type 2 diabetes; fear of weight gain and any consequences of glycaemic con- trol can factor in that delay. Clinical trials provide some insight into this issue but have limited generalisability due to restrictions on participant selection and treatments used. Here, we examine the relationship between change in body weight and HbA1c over 4 years of insulin therapy in routine clinical practice. Materials and methods: CREDIT was a noninterventional study of people with type 2 diabetes beginning insulin therapy in Europe, North America, and Asia. Physicians were free to choose and adapt treatments in accordance with usual practice; data were collected from physician records. Univariate and multivariable analyses were used to determine the relationship between change in body weight and HbA1c over 4 years. Results: Determinants of HbA1c were studied in 1973 people with 4-year data from starting insulin; 47% were men, baseline mean age 61 ± 10 (SD) years, BMI 29.2 ± 6.1 kg/m², 10 ± 8 years of diabetes, HbA1c 9.5 ± 1.9 % (80 ± 21 mmol/mol), 69% continuing oral glucose-lowering therapies. Half were started on basal insulin alone, 23% on premix insulin alone, and 15% on basal + meal-time insulin; over all insulin regimens, starting insulin dose was 0.26 ± 0.18 U/kg. In the first year, HbA1c declined, then remained stable (Figure), with a decrease at year 4 of -2.0 ± 2.1 % (-22 ± 23 mmol/mol). Body weight increased an average of 1.9 ± 4.7 kg in the first year, but < 1.0 kg over the sub- sequent 3 years; over the 4 years, weight increased by 2.6 ± 7.4 kg. Daily insu- lin dose increased 0.15 ± 0.23 U/kg over the first year and by 0.27 ± 0.31 U/kg over the 4 years. This was accompanied by a change in insulin regimen from starting insulin to 4 years for 44% of population (30% on basal alone, 33% on basal + meal-time, and 25% on premix alone at year 4). After adjusting for region and centre, 4-year change in weight was predictive of HbA1c at year 4 (P < 0.0001), but with a small effect size: a weight gain of 2.6 kg was associated with a 0.005 %-unit (0.05 mmol/mol) higher HbA1c at year 4. Other variables were associated with HbA1c at year 4: baseline HbA1c, BMI, age, and at 4 years, insulin dose, number of oral glucose-lowering drugs and symptomatic hypoglycaemia. Taking these variables into account, weight change remained predictive of HbA1c at year 4 (P < 0.04). Symmetrically, the odds ratio (95% CI) for an HbA1c below 7.0 % vs. HbA1c ≥ 7.0 % at year 4 following a 1.0 kg increase in weight over the 4-year period, was 0.97 (0.96, 0.99; P = 0.0015 ), after adjusting for other predictive variables. Conclusion: After 4 years of insulin treatment in people with type 2 diabe- tes, weight gain was associated with minimally higher HbA1c values at year 4. These results should allay fears of weight gain having a strong impact on glycaemic control. Supported by: Sanofi 111 The INITIATOR study: real-world treatment patterns and outcomes in patients with type 2 diabetes initiating insulin glargine or liraglutide P.A. Levin1, W. Wei2, E. Buysman3, S. Thayer3, L. Brekke3, W.H. Crown3, M. Grabner4, X. Ke4, R.A. Quimbo4, M.J. Cziraky4, W. Hu2, R. Cuddihy2, J.W. Chu5; 1MODEL Clinical Research, Baltimore, 2Sanofi US, Bridgewater, 3Optum, Eden Prairie, 4HealthCore, Wilmington, 5Monterey Endocrine & Diabetes Institute, Monterey, USA. Background and aims: Patients with type 2 diabetes mellitus (T2DM) not achieving glycaemic control on oral antidiabetic drugs (OADs) can initiate injectable therapy with insulin or a glucagon-like peptide-1 receptor agonist (GLP-1) analogue. As the first large real-world observational longitudinal study of T2DM patients Initiating New Injectable Treatment Introduced after Anti-diabetic Therapy with Oral-only Regimens (INITIATOR), this analysis aimed to provide a comprehensive understanding of patient characteristics, treatment patterns, and associated outcomes. Materials and methods: T2DM patients from 2 of the largest US commer- cial health insurers who were aged ≥ 18 years, previously on OADs only, and initiated either insulin glargine pen (GLA) or the GLP1 analogue liraglutide (LIRA) between April 1, 2010-March 31, 2012 were included. Patients had continuous healthcare coverage during the 6 months before (baseline [BL]), Diabetologia (2014) 57:[Suppl1]S1–S564 S 54 1 C and 12 months after initiation (follow-up). Health claims and medical chart data from eligible patients were extracted by OP and HC from their respec- tive health plans. We compared differences in BL characteristics between treatment groups and 1-year follow-up measures to BL within each group. Treatment patterns, and clinical and economic outcomes at follow-up were assessed descriptively within each health plan. Results: 4,490 patients were included (45.5% women; mean age 52.7 years; mean number of OADs 2). Significant BL differences existed: GLA patients had higher comorbidity burden, higher HbA1c, lower weight, and were less often women (Table). A total of 23.8% of LIRA patients had a HbA1c level of < 7.0% and 21.4% reported ‘improve weight control’ as a reason for initiation. At 1 year follow-up, overall treatment persistence was 64.0% for GLA and 49.4% for LIRA patients; 3.5% of GLA patients were on twice-daily GLA. Both groups had significant HbA1c reductions. GLA patients had a slight weight gain whereas LIRA patients lost weight. Overall follow-up hypo- glycaemia rates were 16% in GLA and 9% in LIRA patients; rates of severe hypoglycaemia were low. Compared with BL, LIRA patients had significant increases in total healthcare costs (OP and HC), in contrast to GLA patients who had fewer diabetes-related hospitalizations (OP only) and no increase in total healthcare costs. LIRA (OP and HC) and GLA patients (OP only) incurred higher diabetes drug costs. Conclusion: These findings showed clinically relevant differences in BL characteristics and follow-up outcomes in T2DM patients initiating GLA or LIRA. This study highlights challenges in translating clinical trial findings into real-world settings, and in conducting comparative effectiveness studies where important BL group differences may exist. Supported by: Sanofi US, Inc. 112 Characteristics of patients with type 2 diabetes mellitus (T2DM) on basal insulin who do not achieve glycaemic goals S. Brunton1, L. Blonde2, P. Chava2, R. Zhou3, J. Meyers4, K. Davis4, M. Dalal5, A. DiGenio6; 1University of North Carolina, Chapel Hill, 2Ochsner Medical Center, New Orleans, 3Medpace Inc., Cincinnati, 4RTI Health Solutions, Research Triangle Park, 5Sanofi US, Inc., Bridgewater, 6Isis Pharmaceuticals, Inc., Carlsbad, USA. Background and aims: Despite the efficacy of basal insulin therapy in indi- viduals with T2DM, a significant number of patients may not achieve gly- caemic goals. Patients taking basal insulin who reach their fasting plasma glucose (FPG) goal but not their glycated haemoglobin A1c (HbA1c) target represent an important unmet T2DM management need. This study used both randomized clinical trial (RCT) and “real-world” data to compare the baseline characteristics of T2DM patients with HbA1c ≥ 7.0% achieving FPG levels < and ≥130 mg/dL. Materials and methods: Characteristics, including age, gender, BMI, and weight, of patients on basal insulin who do or do not achieve HbA1c < 7.0% were analysed using 11 pooled RCTs assessing insulin glargine (6-months follow-up) and real-world data from the GE Centricity electronic medical records (EMR) database (6- and 12-months follow-up). Patients with HbA1c ≥ 7.0% were stratified by FPG level (< 130 or ≥ 130 mg/dL). Results: Patients with high HbA1c at follow-up also had high baseline HbA1c (HbA1c ≥ 7.0% vs < 7.0%: RCT, 9.1% vs 8.5%; EMR: 9.0% vs 8.1% [6-months follow-up] and 9.0% vs 8.0% [12-months follow-up]). About 50% of the RCT patients reached HbA1c goal vs < 30% of EMR patients. Of patients with HbA1c ≥ 7%, about 50% (RCT) and < 30% (EMR) achieved FPG < 130 mg/ dL. RCT patients tended to be younger, less obese, and weigh less vs EMR patients (mean: age 2-4 years younger; BMI ~3 kg/m2 less; and weight ~6 kg less). Patients with HbA1c < 7.0% were somewhat older than those with HbA1c ≥ 7.0% (mean age in years for HbA1c < 7.0% vs ≥ 7.0%: RCT, 58.4 vs 57.7; EMR, 62.3 vs 60.2 [6-months follow-up] and 62.7 vs 59.8 [12-months follow-up]). Little difference existed in gender distribution or mean BMI for patients with HbA1c < 7.0% versus HbA1c ≥ 7.0%. For both RCT and EMR patients with HbA1c ≥ 7.0%, those with FPG < 130 mg/dL were slightly older than those with FPG ≥ 130 mg/dL (mean age for FPG < 130 vs ≥ 130 mg/ dL: RCT, 58.7 vs 56.4 years; EMR, 62.3 vs 59.6 [6-months follow-up] and 62.4 vs 59.0 [12-months follow-up]). Differences were also seen in baseline insulin usage in the RCT patients, with more patients with FPG ≥ 130 mg/ dL using a fast-acting insulin analogue (71.2% vs 28.9%). For both RCT and EMR patients, other demographic characteristics did not differ between the FPG groups. Conclusion: We present differences in characteristics between patients on basal insulin who do or do not achieve HbA1c < 7.0%. Patients with a high- er baseline HbA1c were less likely to achieve target HbA1c < 7.0% on basal insulin alone; more intensive treatment regimens may be required. A large proportion of the HbA1c > 7.0% patients were also above target for FPG at follow-up and would likely benefit from addition basal insulin titration. In patients achieving FPG < 130 mg/dL, despite above target HbA1c values, postprandial glucose (PPG) control could be used to help bring their HbA1c- within recommended limits. Appropriate therapeutic choices for those not reaching HbA1c < 7.0% should be based on assessment of FPG and PPG, as well as of HbA1c. A better understanding of patient characteristics and factors involved in goal achievement might increase the numbers of patients reach- ing targets and reduce the time they are exposed to hyperglycaemia and its adverse consequences. Supported by: Sanofi US, Inc. 113 People with type 2 diabetes with lower HbA1c using insulin experience fewer cardiovascular events and death: results from the CREDIT study N. Freemantle1, N. Danchin2, F. Calvi-Gries3, M.-P. Dain4, M. Vincent4, P. Home5; 1Department of Primary Care and Population Health, University College London, UK, 2Universite Paris 5, 3AtlanStat, Rezé, 4Sanofi, Paris, France, 5Institute for Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK. Background and aims: CREDIT (Cardiovascular Risk Evaluation in peo- ple with Type 2 Diabetes on Insulin Therapy) was a noninterventional study designed to examine relationships between HbA1c and cardiovascular (CV) events in 2999 patients beginning insulin in real-world practice in Europe, North America, and Asia, with up to 54 months’ follow-up. The study did not impact normal practice, having no intervention or formal visits-instead, data were collected from physician reports. Materials and methods: Primary outcome was the composite of nonfatal stroke or myocardial infarction (MI) or CV death. Events were blindly adjudicated by an endpoints committee. Relative hazard of CV events was described with Cox proportional hazards models, including patient risk factors, and updated mean HbA1c as a time-dependent covariate. Primary outcome components were de- scribed separately. The relationship of severe and symptomatic hypoglycaemia (collected observationally for prior 6 months) to CV and all-cause mortality was examined by adding patient-level covariates to the Cox models. Results: In total, 147 primary events were accrued during study follow-up. There were 60 CV deaths, 44 nonfatal MIs, and 57 nonfatal strokes, with 148 all-cause deaths. There was a significant positive relationship between up- dated mean HbA1c and primary outcome; hazard ratio (HR) 1.25 (95% CI: 1.12-1.40; P S 55 1 C 114 Intensive versus conservative glucose control in patients undergoing coronary artery bypass graft surgery (GLUCO-CABG Trial) D. Smiley1, S. Cardona1, F. Pasquel1, F. Farrokhi1, S. Jacobs1, L. Peng2, M. Halkos1, J.D. Puskas1, R.A. Guyton1, V. Thourani1, G.E. Umpierrez1; 1Medicine-Division of Endocrinology, Emory University, 2Biostatistics, Rollins School of Public Health, Atlanta, USA. Background and aims: This randomized controlled trial aimed to determine whether intensive glucose control (intensive, BG target: 5.6-7.8 mmol/L) re- duces perioperative complications compared to conservative glucose control (conservative, BG target: 7.8-10 mmol/L) in hyperglycaemic patients under- going CABG. Materials and methods: After ICU care, subjects were transitioned to the same treatment regimen targeting BG S 56 1 C prove markers of insulin sensitivity in T2D patients. Here, we investigated acute metabolic and vascular effects of tadalafil in T2D patients after a mixed meal. Materials and methods: We performed a randomised, double blind, pla- cebo controlled trial, in parallel groups. Twenty-six T2D patients (Age: 40- 70 (male) and 50-70 years (female); BMI: 27-40 kg/m2; HbA1c < 60 mmol/ mol) were enrolled and they received either 20 mg tadalafil (n=14) or pla- cebo (n=12) 30 min prior to a mixed meal. To assess forearm glucose uptake (FGU) and muscle capillary recruitment (PSglu), intramuscular microdialy- sis was combined with arterial and venous blood sampling, as well as plethys- mography monitoring of forearm blood flow (FBF). Blood samples and tissue dialysates were repeatedly collected for 5 hours. We used an ANOVA and Mann-Whitney U-test for all statistical analyses. Furthermore, we performed an ad hoc analysis excluding patients on ACE-inhibitors because this family of drugs have known positive effects on glucose metabolism. Accordingly, 10 patients in the tadalafil group and 9 patients in the placebo group were evalu- ated in a subgroup analysis. Results: The groups were comparable regarding gender, age, BMI and HbA1c. ITT analyses showed that T2D patients in the tadalafil group increased FBF (p < 0.05), whereas this was not observed in the placebo group. Moreover, incremental area under the curve (IAUC) for FGU and IAUC for PSglu dur- ing 5 hrs after the meal were similar in the two groups. However, the ad hoc analysis showed a significant increase (p S 57 1 C releasing properties of xenin-25(Gln) (10-12 to 10-6 mol/l) were determined in BRIN-BD11 cells (n=8; 20 min incubation) at 5.6 and 16.7 mmol/l glucose concentrations in the absence and presence of (d-Ala2)GIP. Acute effects of xenin-25(Gln) (25 nmol/kg bw; ip) on plasma glucose and Insulin concen- trations were examined in NIH Swiss mice fed a high fat diet (45% fat, 35% carbohydrate and 20% protein) for 140 days prior to experiment. Persistent effects of xenin-25(Gln) on blood glucose concentrations were examined in lean C57Bl/6J mice. Results: Xenin-25(Gln) was resistant to plasma-mediated degradation for up to 24 hours whereas native xenin-25 was sequentially degraded by murine plasma. In BRIN-BD11 cells, xenin-25(Gln) significantly stimulated (1.5 to 2.9-fold; P < 0.05 to P < 0.001) Insulin secretion in a concentration-depend- ent manner at both 5.6 and 16.7 mmol/l glucose. Moreover, xenin-25(Gln) significantly augmented (1.4 to 1.7-fold; P< 0.05 to P < 0.01) the insulinotrop- ic response of (d-Ala2)GIP in vitro. Acute administration of xenin-25(Gln) to high fat fed mice significantly reduced plasma glucose concentrations com- pared to glucose alone (56% reduction; P < 0.001) and xenin-25 (38% reduc- tion; P < 0.05) treated mice. Correspondingly, the overall Insulin secretory response was significantly enhanced (3.4-fold increase; P < 0.05) in xenin- 25(Gln) treated mice compared to high fat fed mice administered glucose alone. Furthermore, xenin-25(Gln) elicited a protracted glucose-lowering ef- fect (40% reduction; P 30 pmol/l (median difference [95% confidence interval (CI)]: -6.6 minutes [-8.0; -5.0]) and greater exposure during the first 2 hours with the largest difference in the first 15 minutes, whereas the overall phar- macokinetic exposure was similar between the two insulins (mean ratio [95% CI] AUC0-15 minutes: 3.14 [2.59; 3.80]; AUC0-30 minutes: 1.93 [1.64; 2.26]); AUC0-1 hour: 1.30 [1.15; 1.46]; AUC0-2 hours: 1.13 [1.03; 1.24]; AUC0-10 hours: 0.99 [0.93; 1.06]). The faster absorption led to a greater reduction in postprandial blood glucose (BG) with FIA(B) versus IAsp, indicated by lower post-meal AUCBG over 2 and 6 hours (reduction by 26% and 33%, respectively; Figure), and by lower BG values 1 and 2 hours postprandially (mean difference [95% CI] BG1 hour: -1.24 mmol/l [-2.01; -0.46]; BG2 hours: -1.45 mmol/l [-2.49; -0.42]). No safety or tolerability issues were identified; in particular, no injection site reactions were observed. Conclusion: FIA(B) was absorbed faster and had an increased early exposure than insulin aspart, leading to improved postprandial glycaemic control ver- sus insulin aspart. Clinical Trial Registration Number: NCT01121276 Supported by: Novo Nordisk Diabetologia (2014) 57:[Suppl1]S1–S564 S 58 1 C OP 21 Physiological adaptation to bariatric surgery 121 Bariatric surgery improves whole body and femoral subcutaneous adipose tissue insulin sensitivity independently of whole body weight loss L. Landini1,2, M. Bucci1, T. Pham1, J.C. Hannukainen1, P. Salminen3, P. Iozzo2, P. Nuutila1,4; 1Turku PET Centre, University of Turku, Finland, 2National Research Council, Pisa, Italy, 3Department of Surgery, 4Department of Endocrinology, Turku University Hospital, Finland. Background and aims: The effect of weight reduction by dietary treatment does not reverse adipose tissue insulin resistance, as opposed to skeletal mus- cle in which it makes a big improvement. To the best of our knowledge, there is no evidence of how bariatric surgery modulates femoral adipose tissue glucose uptake. We examine how bariatric surgery modifies femoral adipose tissue glucose uptake both in fasting condition and during euglycemic hyper- insulinemic clamp in severely obese patients and healthy lean subjects and we study if the changes in adipose tissue glucose uptake were related with the remission of diabetes. Materials and methods: Femoral subcutaneous adipose tissue and muscle glucose uptake were studied with positron emission tomography using [18F] fluorodeoxyglucose (Figure 1) during fasting and euglycemic hyperinsuline- mic clamp in 25 morbidly obese patients (BMI of 43.2±3.6 kg/m2) before and six months after the bariatric surgery. Ten age-matched lean subjects (BMI: 23.7±1.8 kg/m2) served as controls. All data are presented as mean ± SE. Results: At baseline, nine patients had type 2 diabetes mellitus (T2DM) and seven had impaired fasting glucose (IFG) and/or impaired glucose tolerance (IGT). After surgery, remission of T2DM or IFG/IGT was observed in ten patients. Two patients dropped out from the study. At baseline, obese patients had impaired whole body (M value: 12.2±1.14 vs 40.3±3.00 µmol/(min*kg), p S 59 1 C 123 Beta cell function improvements in subjects with type 2 diabetes 1 year after biliopancreatic diversion A.C.J. Vasques1, J.C. Pareja2, M.S. Oliveira2, F.S. Novaes2, M.M.O. Lima2, É.A. Chaim3, F. Piccinini4, C. Dalla Man4, C. Cobelli4, B. Geloneze2; 1Laboratory of Investigation on Metabolism and Diabetes (LIMED), School of Applied Sciences, Limeira, 2Laboratory of Investigation on Metabolism and Diabetes (LIMED), Faculty of Internal Medicine, Campinas, 3Department of Surgery, Faculty of Internal Medicine, Campinas, Brazil, 4Department of Information Engineering, University of Padua, Italy. Background and aims: Bariatric surgery is an alternative therapeutic ap- proach for obese patients with poorly controlled type 2 diabetes (T2DM). To provide evidence for the underlying mechanisms associated with T2DM remission after biliopancreatic diversion (BPD), this study aimed to evaluate the long term effect of BPD on β-cell function parameters and insulin sensi- tivity in grade I and II obese patients with T2DM. Materials and methods: Sixty-eight women were divided into three groups: 19 LeanNGT (BMI: 23±2 kg/m²), 18 ObeseNGT (BMI: 35±5 kg/m²) both with normal glucose tolerance (NGT), and 31 ObeseT2DM (BMI: 36±4 kg/m²). Twenty ObeseT2DM women underwent BPD and were reassessed 1 year after surgery. OGTTs and hyperglycaemic clamps were performed. Mathematical modeling, based on insulin and C-peptide serum levels, was used to analyze β-cell function, insulin sensitivity (ISoral and ISclamp) and delay time of β-cell response to a specific plasma glucose concentration. The basal (DIb), dynamic (DId), static (DIs) and total disposition indexes (DIoral and DIclamp) represents β-cell function adjusted to IS and were calculated by multiplying β-cell responsivity indices by IS. Results: After BPD, BMI (pre: 36 ± 4 vs post: 28 ± 3 kg/m²), fasting glycae- mia (pre: 133±38 vs post: 88 ± 13 mg/dl) and HbA1c levels (pre: 7.2 ± 1.3 vs post: 5.1 ± 0.9% were reduced (p < 0.001). At baseline, ISoral and ISclamp were reduced in the obese groups compared with the LeanNGT group (p < 0.001). After BPD, IS indexes increased approximately 6-fold. ISoral became comparable to LeanNGT, and ISclamp became increased in comparison with the LeanNGT [29 (9-89) vs 15 (8-24) dl/kg/min per µU/ml, p < 0.01], respec- tively. At baseline, the basal disposition index (DIbasal) was reduced in the ObeseT2DM [67 (44-145) dl/kg/min² per pmol/l] compared with the NGT groups [Lean: 258 (170-455) and Obese: 169 (119-326) dl/kg/min² per pmol/l; p < 0.001]. After surgery, the DIbasal increased 3-fold reaching similar levels to that of both NGT groups (p = 0.167). At baseline, the DId, DIs and DIoral derived from the OGTT were reduced in the ObeseTDM2 compared with both NGT groups (p < 0.001). After surgery, DId and DIoral were similar to ObeseNGT and remained reduced compared with the LeanNGT, while the DIs was restored to the levels of both NGT groups. In the clamp test, at base- line, DId, DIs and DIclamp of the ObeseT2DM were similar to ObeseNGT and reduced compared with the LeanNGT (p < 0.05). After surgery, all of the DI were similar to LeanNGT levels and were increased compared with the ObeseNGT. The delay time, at baseline, was markedly increased in the Obe- seT2DM compared with the NGT subjects in both dynamic tests (p < 0.001). After BPD, the delay time was decreased (Clamp: 100 (80-108) vs 29 (8-100) min.; OGTT: 15 (9-74) vs 9 (7-10) min.; p < 0.05) and reached similar levels to both NGT groups (p = 0.8). Conclusion: In long term, the BPD resulted in positive physiological adapta- tions in patients with grade I and II obesity and T2DM, with improvements in a wide range aspects of β-Cell function and IS which probably contributed to the improved glycaemic control. Clinical Trial Registration Number: U1111-1137-0489 Supported by: FAPESP - Grants 2008/09451-7 and 2008/07312-0 124 Restoration of beta cell function in severely obese type 2 diabetic patients after gastric bypass surgery is accompanied by improved insulin processing E. Svehlikova1, A. Tuca1, V. Höller1, B. Obermayer-Pietsch1, O. Freisinger2, F. Tadler3, B. Ernst4, B. Wilms5, M. Thurnheer4, B. Schultes4, T.R. Pieber1; 1Department of Internal Medicine, Division of Endocrinology and Metabolism, Medical University of Graz, 2Department of Surgery, Medical University of Graz, 3Department of Surgery, Krankenhaus der Elisabethinen, Graz, Austria, 4Interdisciplinary Obesity Center, eSwiss Medical & Surgical Center, St. Gallen, 5Exercise Physiology Lab, Institute of Human Movement Sciences, ETH Zurich, Switzerland. Background and aims: Gastric bypass surgery improves glycaemic control, but the underlying mechanisms are incompletely understood. The aim of the study was to quantify changes in beta cell function and insulin sensitivity before, early and 1 year after gastric bypass in type 2 diabetic (DM) and non- diabetic (ND) morbidly obese patients. Materials and methods: Before, 8 to 21 days and 1 year after the surgery, 34 T2DM (17 with diabetes duration ≥8 years) and 21 ND patients underwent an oral glucose tolerance test (OGTT) and a botnia clamp combining an in- travenous glucose tolerance test (IVGTT) with a subsequent hyperinsulinae- mic-euglycaemic clamp. Results: In DM patients, glucose tolerance gradually improved after the sur- gery, fasting insulin was decreased at 1 year (p S 60 1 C 125 Mechanisms of post-prandial hypoglycaemia after Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (LSG) M. Nannipieri1, A. Belligoli2, A. Mari3, D. Moriconi1, S. Baldi4, D. Colligiani4, M. Anselmino5, M. Foletto6, E. Zabeo7, R. Vettor7, E. Ferrannini8; 1Clinical and Experimental Medicine, University of Pisa, 2Medicine, University of Padua, 3Biomedical Engineering, CNR of Padua, 4Dpt Clinical and Experimental Medicine, University of Pisa, 5Azienda Ospedaliera Universitaria Pisana, 6Oncological, Surgical and Gastroenterological Sciences, University of Padua, 7Medicine, University of Padua, 8Dpt Clinical and Experimental Medicine, University of Pisa and CNR of Pisa, Italy. Background and aims: Symptomatic hypoglycaemia (Hypo) is a well recog- nised complication of RYGB surgery. Data on the development of postpran- dial Hypo after LSG are scanty. We investigated the mechanisms of postpran- dial Hypo in patients undergoing RYGB or LSG. Materials and methods: 32 obese non-diabetic subjects treated with RYGB and 39 with LSG received a 3-h OGTT before and 12-18 months after sur- gery. Hypo was defined as plasma glucose ≤2.7 mmol/L. Insulin sensitivity was assessed by OGIS index and ß-cell function by modelling analysis of the C-peptide response to glucose load. Results: Postprandial Hypo occurred in 20 of 32 RYGB patients and in 13 of 39 LSG patients. Age did not discriminate Hypo from non-Hypo (NH) sub- jects. Presurgery BMI was lower in RYGB-Hypo than RYGB-NH (43.8±5.2 vs 49.7±6.1 kg/m2, p=0.004), but not in in LSG-Hypo vs LSG-NH patients. Similarly, baseline insulin sensitivity was higher in RYGB-Hypo than RYGB- NH (386±53 vs 325±44 ml.min-1.m-2, p=0.004), but was similar in LSG- Hypo and LSG-NH subjects. After either operation, insulin sensitivity im- proved (p S 61 1 C OP 22 Neuropathy: mechanisms and outcomes 127 Diabetic neuropathy and urologic complications in men with type 1 diabetes in the DCCT/EDIC study R. Pop-Busui1, B. Braffett2, J. Hotaling3, P. Cleary2, C. Martin1, R. Dunn4, H. Wessels5, A. Sarma4, DCCT/EDIC; 1Department of Internal Medicine, University of Michigan Medical School, Ann Arbor, 2Department of Biostatistics, George Washington University, Rockville, 3Department of Urology, University of Utah, Salt Lake City, 4Department of Urology, University of Michigan Medical School, Ann Arbor, 5Department of Urology, University of Washington, Seattle, USA. Background and aims: Impaired genital sensory or motor function may promote sexual or urinary dysfunction, but data in persons with type 1 dia- betes (T1D) are limited. We evaluated associations between diabetic periph- eral neuropathy (DPN), erectile dysfunction (ED) and lower urinary tract symptoms (LUTS) in men with T1D participating in Diabetes Control and Complications Trial/Epidemiology of Diabetes Intervention and Complica- tions (DCCT/EDIC) Materials and methods: DPN was evaluated with a composite of symptoms, neurological examination and nerve conduction studies (NCS) performed at baseline and biennially during DCCT, and at years 13/14 during EDIC. Con- firmed DPN was defined as NCS abnormalities in ≥ 2 nerves and ≥ 2 posi- tive responses among symptoms, sensory signs, or reflex changes. ED was assessed using the question from the International Index of Erectile Function “Over the past 4 weeks, how would you rate your confidence that you can get and keep erection?” at EDIC year 17. Answers “very low-low” were classified as ED. LUTS was assessed by the American Urological Association Symptom Index (AUASI). AUASI scores 0-7 were defined as no LUTS, and AUASI of 8-35 as LUTS. Results: ED and LUTS data were available from 635 men (mean age 52 years, mean duration 30 years, mean DCCT/EDIC A1c 7.9%). ED only was report- ed by 193 (30%), LUTS only by 61 (10%) and both ED and LUTS by 97 (15%) men. Men with confirmed DPN at year 13/14 were more likely to report ED (41%), LUTS (31%) or both ED and LUTS (62%) at EDIC year 17 compared to those without ED or LUTS (22%) (p S 62 1 C [3] cm; p S 63 1 C 132 Abnormal neural plasticity and cortical reorganisation in underlies the painful/painless leg D. Selvarajah1, I.D. Wilkinson2, E. Bolan1, S. Tesfaye3; 1Department of Human Metabolism, University of Sheffield, 2Academic Department of Radiology, University of Sheffield, 3Academic Department of Radiology, Sheffield Teaching Hospital NHS Foundation Trust, UK. Background and aims: The clinical paradox of chronic painful neuropathic symptoms in patients with insensate feet (the so called ‘painless/painful leg’) is well recognised. Using recent advances in functional magnetic resonance imaging (fMRI) we sought to phenotype the pattern of brain activation in these patients with painful diabetic neuropathy (DN) to provide new insights into this poorly understood condition. In the present study, painful stimuli of the foot and thigh were examined in painful DN subjects with and without retained foot sensation using fMRI. Materials and methods: 26 subjects (Painless DN, n=9; Painful DN sensate, n=9; Painful DN insensate, n=8) with Type 1 diabetes and 17 healthy volun- teers underwent detailed clinical and neurophysiological assessments (vibra- tion detection thresholds, sural, common peroneal and tibial nerve conduc- tion studies using standard procedures). Painful neuropathic symptoms were assessed using the NTSS-6 questionnaire. All Painful DN subjects had severe neuropathic pain below the knees. Before fMRI heat pain was applied to the right anterior thigh (control region) and dorsum of the foot to establish a threshold of noxious thermal stimulation capable of eliciting a response of at least 7 on an 11 point Likert scale. Painful DN insensate patients were defined as those unable to perceive a noxious thermal stimulus applied to the foot at the maximum temperature setting (39.9oC). This was repeated inside the MR scanner (Acheiva 3T, Philips Healthcare) at the predetermined threshold alternating with a pain-free baseline condition in a pseudo-randomised box- car design. Images were analysed using FSL (FEAT, FMRIB, Oxford). Results: Thermal noxious stimulus delivered to the foot and thigh was associ- ated with significantly increased neuronal response in both the sensory dis- criminatory (left somatosensory cortex [stereotactic coordinates: -12,-44,78], left thalamus [-12,-12,4] and right midbrain [8,-18,-14]; corrected p S 64 1 C tic targets. Recently-identified common genetic variants in the ARAP1 locus affect fasting proinsulin levels (rs11603334) and glucose-induced insulin se- cretion (rs1552224) in man. Here we investigate the roles of two genes impli- cated at this locus: ARAP1 and STARD10. Materials and methods: Arap1 and StarD10 expression was silenced by siRNA and proteins overexpressed via a coding plasmid transfected by Lipo- fectamine 2000 (cell lines) or adenoviral infection (mouse and human islets). Total and secreted insulin were measured by radioimmunoassay. Results: Expression quantitative trait loci (eQTL) analysis by qRT-PCR in human islets from 49 donors revealed increased expression of the short form of Arap1 (variant 1: Arap1v1) in the carriers of the risk variant at SNP rs1552224 (p S 65 1 C indicate that elevated expression of CaV3.1 channels causes impaired glucose homeostasis. Conclusion: Elevated expression of CaV3.1 channels impairs glucose-stimu- lated insulin secretion through downregulation of the exocytotic machinery and plays an important role in the pathogenesis of diabetes. Supported by: KI Stiftelser, SVLS 137 NGF secretion from beta cells contributes to fine tuning of insulin secretion A. Pingitore1, M.C. Caroleo2, E. Cione2, S. Persaud1; 1Diabetes Research Group, Diabetes and Nutritional Sciences, King’s College London, UK, 2Pharmacy and NHS, University of Calabria, Rende, Italy. Background and aims: Islet β-cells express a range of peptides such as nerve growth factor (NGF), which was originally identified as a signaling protein regulating neuronal survival. It is now recognised that NGF is a more com- plex messenger involved in controlling a broad range of homeostatic respons- es through binding to two trans-membrane receptors, trkA and p75NTR, both expressed in islets. The role of NGF in islet function has not been estab- lished so we investigated whether β-cells possess an autocrine loop through which NGF release modulates insulin secretion. Materials and methods: Studies were conducted with isolated mouse islets (MI) and human islets (HI), and with MIN6 pseudoislets (PI) that were gen- erated by growing MIN6 β-cells on bacterial dishes for 10 days. NGF release from islets was determined by western blotting and quantified by ELISA, and insulin secretion was quantified by radioimmunoassay. The biological activity of β-cell-derived NGF was blocked via a NGF neutralising antibody (AbNGF) at 5µg/mL and trkA blockade was achieved with 5nM K252a. Im- munoprecipitation studies were performed using INS-832/13 β-cells with an anti-RhoGDI antibody followed by electrophoresis and western blotting probing with anti- p75NTR. Results: NGF expression and release from islets was identified by western blotting, and a sensitive ELISA indicated that NGF was secreted from MI in a glucose-dependent manner (2mM glucose: 17.8±6.1 pg/mL; 20mM glucose 56.7±2.8 pg/mL). Blockade of secreted NGF action in islets and PIs with AbNGF increased insulin output at 2mM glucose (MI 381.0±114.0%, p S 66 1 C OP 24 Genes and biomarkers for type 2 diabetes 139 A beta cell specific protein subnetwork significantly enriched for association with GLP-1 stimulated insulin secretion: a DIRECT study V. Gudmundsdottir1, H.K. Pedersen1, L.M. ‘t Hart2, K. Banasik3,4, D. Boomsma5, E. de Geus5, M. Eekhoff6, M. Diamant6, M. McCarthy3,7, E. Pearson8, C. Workman1, R. Gupta1, S. Brunak1; 1Technical University of Denmark, Kgs. Lyngby, Denmark, 2Leiden University Medical Center, Netherlands, 3University of Oxford, UK, 4University of Copenhagen, Denmark, 5VU University Amsterdam, 6VU University Medical Center, Amsterdam, Netherlands, 7Oxford Centre for Diabetes, Endocrinology and Metabolism, 8University of Dundee, UK. Background and aims: The heritability of insulin secretion response (ISR) during a modified hyperglycaemic clamp with glucose, glucagon-like pep- tide-1 (GLP-1) and arginine stimulation has been shown to be considerable (h2=0.52 to 0.80) and it is of great interest to identify genetic variants in- fluencing these responses. Evaluating ISR with a modified hyperglycaemic clamp is not feasible in larger cohorts of sizes preferred for genome-wide association studies (GWAS). When sample sizes are small, GWAS may be complemented with systems biology approaches to aid the prioritisation of genetic variants. Within the DIRECT consortium a GWAS was performed on GLP-1 stimulated ISR and here we aimed to use data integration to add biological context to the results and facilitate variant prioritisation. Materials and methods: GLP-1 stimulated ISR was measured with a modi- fied hyperglycaemic clamp in 130 twins and sibs from the Netherlands twin register. The cohort was genotyped using the Illumina HumanCore Exome BeadChip and association analysis was performed using the QTassoc soft- ware and adjusted for age, sex, familial relationships and insulin sensitivity index. Gene-based P-values were mapped onto a β-cell specific protein-pro- tein interaction (PPI) network, which was created by pruning high confi- dence PPIs from InWeb 3.0 using published β-cell RNAseq data. Connected components in the network enriched for high scoring genes were identified with the Cytoscape plugin jActiveModules. The significance of the top-scor- ing network was evaluated by converting the gene P-values to z-scores and estimating the probability of observing a combined z-score of same size or higher with 10,000 degree-preserved randomly sampled subnetworks from the β-cell PPI. Gene set enrichment analysis was performed on the top sub- network using DAVID. Results: None of the variants tested in the GWAS reached a genome-wide significance of P ≤ 5.0E-8. However, the top scoring subnetwork had a sig- nificantly higher combined z-score than expected by random (P ≤ 1.0E-4). It contained 25 genes and was most strongly enriched for the Gene Ontology terms “plasma membrane part” (Padj = 2.2E-4), “cell junction” (Padj = 3.6E- 4) and “cell projection” (Padj = 4.9E-3). The subnetwork contained a number of genes known to affect β-cell mass and function (FOXO1), insulin secretion (WFS1, RYR2) or to be implicated in type 2 diabetes (MAGI2, CTNNA2 and PTPRD). Conclusion: We have identified a β-cell PPI network enriched for genes with nominal associations with GLP-1 stimulated ISR, demonstrating how data integration can highlight biological mechanisms underlying a phenotype where GWAS results on their own may be insufficient. Furthermore, the net- work can be used to prioritise genetic variants to take forward for replication in independent cohorts. Supported by: EU IMI-DIRECT project 140 FOXA2 bound sites are enriched for type 2 diabetes risk variants K.J. Gaulton1, A.P. Morris2, M.I. McCarthy1, the DIAGRAM consortium; 1Wellcome Trust Centre for Human Genetics, Oxford, 2Department of Biostatistics, University of Liverpool, UK. Background and aims: Risk variants for type 2 diabetes (T2D) are primarily non-coding and are enriched at regulatory features; however, specific causal variants and regulatory factors underlying risk variant activity are largely unknown. We aimed to integrate systematic regional genotyping data with genomic annotation to identify (i) regulatory factor binding sites broadly enriched for T2D risk and (ii) causal regulatory variants driving enriched signals. Materials and methods: We used “credible sets” of SNPs that account for 99% of the probability of including the causal variant for 39 T2D association signals identified in a study of 27,116 T2D cases and 57,574 controls (Me- tabochip imputed up to 1000G). We then identified chromatin immunopre- cipitation (ChIP) binding sites of 141 proteins obtained from ENCODE and several independent studies. Credible sets were tested for enrichment in the sum of posterior probabilities for variants overlapping binding sites for each factor pooled across all assayed cell types compared to variants in shuffled site locations (within 100kb). We performed enrichment tests both across all loci together and, for globally-enriched factors, at each locus individually. Results: We identified significant enrichment for FOXA2 sites (assayed in pancreatic islets and HepG2 cells) across all risk loci (P=3x10-4). No other factor was significantly enriched. Among individual loci, there was nominally significant enrichment (P S 67 1 C tential direct involvement of terminal complement activation in the aetiology of human fatty liver disease, which deserves further investigation aiming to unravel the potential mechanisms involved. Supported by: NWO, DFN, NHS, CTMM-PREDICCt 142 Novel metabolite models that distinguish impaired glucose tolerance (IGT) from normal glucose tolerance (NGT) J.E. Cobb1, A. Eckhart1, K.-P. Adam1, R. Perichon1, J. Wulff1, M. Mitchell1, R. Elverson1, R. Wolfert1, K. Lawton1, E. Button1, Z. Li1, B. Carr2, M. Sinnott2, E. Ferrannini3; 1Metabolon, Durham, USA, 2Vhi Healthcare, Dublin, Ireland, 3University of Pisa, Italy. Background and aims: Metabolomic profiling studies have identified a num- ber of metabolites where fasting levels are associated with dysglycemia and type 2 diabetes. The aim of this study was to identify metabolites and metab- olite-based models that distinguish IGT from NGT in non-diabetic subjects. Materials and methods: Using the stable isotope dilution technique, quanti- tative assays were developed for a set of 23 candidate biomarker metabolites previously linked to dysglycemia. This set included: α-hydroxybutyric acid (AHB), linoleoylglycerophosphocholine (LGPC), oleic acid, α-ketoglutaric acid, 2-aminoadipic acid, glycine, aromatic amino acids, and the 3 branched- chain amino acids and several of their catabolites. These metabolites were measured in fasting plasma samples taken just prior to an OGTT from 1,679 non-diabetic subjects: 979 from the RISC Study 3 year follow up (11.5% have IGT) and 679 subjects from the DMVhi cohort in the DEXLIFE project (11.8% have IGT). Results: Random forest decision tree analysis using subject metabolite, an- thropometric, and metabolic characteristic data were generated to rank vari- ables for their ability to distinguish IGT from NGT. The top 4 were found to be AHB, fasting glucose, LGPC, and oleic acid in RISC and fasting glucose, age, LGPC, and AHB in DMVhi. Multivariate models for estimating risk of IGT were evaluated in both cohorts using AUCs calculated from the cor- responding ROC curves. AUCs generally did not increase in models having more than 7 variables. A number of 7 variable, metabolite only models were developed that had AUCs >0.80. For example, in RISC, a model consisting of AHB, glucose, LGPC, oleic acid, glycine, creatine, and 2-oxoisoleucine had an AUC of 0.825. In contrast, a model consisting of age, sex, BMI, glucose, insulin and HDL had a significantly lower (p=0.0001) AUC of 0.731. Conclusion: AHB and LGPC have been identified as metabolite markers of IGT and metabolite-based models may be useful for identifying IGT in non- diabetic subjects Supported by: DEXLIFE project (EU FP7 programme, Grant agreement no: 279228) 143 Causal effect of decreased LDL cholesterol and increased blood pressure on higher incidence of type 2 diabetes by Mendelian randomisation in the Malmö Diet and Cancer Study G. Hindy, G. Rukh, P. Almgren, U. Ericson, O. Melander, M. Orho-Melander; Department of Clinical Sciences in Malmö, Lund University Diabetes Center, Skåne University Hospital, Lund University, Malmö, Sweden. Background and aims: Obesity, dyslipidemia and hypertension associate with type 2 diabetes (T2D). However, the relationship between these traits and T2D may not be causal due to confounding and reverse causation. In the Mendelian randomization approach genetic variants are assumed to be randomly distributed, and can be used as instrumental variables to estimate causal relationships. The aim of this study was to investigate if BMI, systolic blood pressure (SBP) and lipid traits are causally associated with T2D in- cidence by using genetic risk scores (GRSs) of single nucleotide polymor- phisms (SNPs) identified in genome wide association studies (GWAS) for these traits as instrumental variables. Materials and methods: In total, 27254 non-diabetic participants (61% fe- males) from the population based Malmö Diet and Cancer Study (MDCS), collected at baseline during 1990-1996 (age 58±8 years, BMI 26±4 kg/m2), were genotyped for SNPs associating with BMI (N=31), high density lipopro- tein cholesterol (HDLC, N=41), low density lipoprotein cholesterol (LDLC, N=32), triglycerides (TG, N=26) and SBP (N=29). Trait-specific weighted GRSs were created using PLINK software. During a mean follow-up time of 15±4 years, 3248 incident T2D and 3977 cardiovascular disease (CVD) cases were identified. At baseline, BMI and SBP were measured for all individuals and fasting blood lipid levels were analyzed in a random sub-cohort (MDC- CC) of 5284 individuals of which 781 developed T2D and 754 CVD. COX- regression was used to analyze the association between BMI, SBP, HDLC, LDLC and TG, and incidence of T2D per standard deviation (SD). For Men- delian randomization, a two-stage least square regression method was used with the GRSs as instrumental variables. The predicted values for the first stage of regression of the lipid traits on their respective GRSs were estimated from MDC-CC for all MDCS. Adjustments were made for age, sex, lipid- lowering and antihypertensive medication. Results: As expected, baseline levels of higher BMI, SBP, LDLC and TG, and lower HDLC associated with higher incidence of T2D (HR 1.83 [1.78-1.88], P S 68 1 C stronger in females than in males, no SNP gender interaction was observed. Furthermore, rs7806429 was associated with mRNA expression of RARRES2 in visceral adipose tissue in women (p S 69 1 C 146 Similar efficacy and safety with LY2963016 insulin glargine compared with insulin glargine in patients with type 1 diabetes mellitus: the ELEMENT 1 study R.K. Pollom1, T. Blevins2, D. Dahl3, J. Rosenstock4, W.J. Huster1, L.L. Ilag1, M.J. Prince1; 1Eli Lilly and Company, Indianapolis, 2Texas Diabetes & Endocrinology, Austin, USA, 3Gemeinschaftspraxis für Innere Medizin und Diabetologie, Hamburg, Germany, 4Dallas Diabetes and Endocrine Center, USA. Background and aims: LY2963016 (LY IGlar) and insulin glargine (Sanofi- Aventis; IGlar) are both insulin glargine products, with identical amino acid sequences. Even with identical primary structure, protein-based therapeutics manufactured by distinct processes must be shown to be clinically similar. Materials and methods: A 52-week, Phase 3, randomized, open-label, par- allel study, with a 24-week treatment period and a 28-week extension, was undertaken to compare the efficacy and safety of LY IGlar once daily (QD) vs IGlar QD in combination with pre-meal insulin lispro thrice daily in pa- tients with T1DM (HbA1c ≤11.0 %). LY IGlar and IGlar were administered with prefilled pen injectors. The primary aim was to test the non-inferiority (0.3% margin) of LY IGlar to IGlar as measured by change in HbA1c, from baseline to 24 weeks. Testing for non-inferiority of IGlar to LY IGlar was also performed and pre-specified as a complementary hypothesis, which if met along with the primary aim, would demonstrate equivalent efficacy between LY IGlar and IGlar. Insulin doses were adjusted during the study to achieve glycemic targets (HbA1c S 70 1 C years, HbA1c 8.1 % [65 mmol/mol]) were randomised 1:1:1:1 to once-daily Gla-300 or Gla-100, morning or evening, while continuing mealtime insulin. Results: Overall, Gla-300 was non-inferior to Gla-100 for HbA1c change from baseline (primary endpoint) (LS mean change [SE] −0.40 [0.05] % (−4.4 [0.6] mmol/mol) and −0.44 [0.05] % (−4.8 [0.6] mmol/mol); LS mean difference 0.04 [95% CI: −0.10 to 0.19] % (0.4 [−1.1 to 2.1] mmol/mol)). Event rates of confirmed (≤3.9 mmol/l [≤70 mg/dl]) or severe hypoglycaemia at any time of day (24 h) were similar for the two groups, while nocturnal hypoglycaemia was lower in the Gla-300 group compared with the Gla-100 group during the first 8 weeks of the study (Table). Severe hypoglycaemia was observed in 6.6% (Gla-300) and 9.5% (Gla-100) of participants. Neither glycaemic con- trol nor hypoglycaemia differed between insulins or times for morning and evening injection groups. Total insulin dose increased to a somewhat greater extent for Gla-300 compared with Gla-100 (change from baseline +0.19 ver- sus +0.10 U/kg). Weight gain was statistically significantly lower with Gla- 300 versus Gla-100 (LS mean difference −0.56 [95% CI: −1.09 to −0.03] kg, p=0.037). There was no difference in adverse events between the two groups. Conclusion: In conclusion, new insulin glargine 300 U/ml provided compa- rable glycaemic control versus glargine 100 U/ml, while nocturnal hypogly- caemia was less frequent during the first 8 weeks of treatment. Clinical Trial Registration Number: NCT01683266 Supported by: Sanofi 149 Recombinant human hyaluronidase pretreatment of CSII cannula sites provides comparable glycaemic control with reduced hypoglycaemia in T1DM compared to usual CSII J.S. Skyler1, B. Bode2, I.B. Hirsch3, J.B. Buse4, S.K. Garg5, D.E. Vaughn6, X.W. Wu6, D.B. Muchmore6; 1University of Miami, 2Atlanta Diabetes Associates, 3University of Washington, Seattle, 4University of North Carolina, Chapel Hill, 5University of Colorado Denver, Aurora, 6Halozyme Therapeutics, Inc., San Diego, USA. Background and aims: Recombinant human hyaluronidase (rHuPH20) is FDA-approved to increase dispersion and absorption of injected or infused drugs. In CSII, a single pretreatment of the cannula site with rHuPH20 ac- celerates exposure to and action of bolus doses of rapid analogs for up to 3 days of catheter use. This study was performed to characterize the clinical attributes of this treatment in an outpatient cohort of T1DM. Materials and methods: 456 subjects with T1DM (age 48±13 years, BMI 28.5±5.1, screening A1C 7.8±0.7) were randomized 3:1 to CSII with rHuPH20 pretreatment or usual CSII for 6 months. Results: A1C fell 0.14% from 7.69% with rHuPH20 and 0.18% from base- line 7.70% for CSII alone. The primary endpoint of A1C noninferiority (0.4% margin) was achieved with a treatment difference of 0.05% (95% CI -0.08 to 0.18) with similar % of subjects reaching A1C S 71 1 C OP 26 Pregnancy and diabetes 151 Pregestational diabetes (type 1 and 2), gestational diabetes: data from the French population in 2011 C. Billionnet1, A. Weill1, P. Ricordeau1, F. Alla1, D. Mitanchez2,3, A. Hartemann4,3, S. Jacqueminet4; 1Caisse Nationale de l’Assurance Maladie, 2Néonatalogie, Hôpital Armand Trousseau, 3Sorbonne Universités UPMC, 4Diabétologie, Institute of Cardiometabolism and Nutrition, Hôpital Pitié-Salpétrière, Paris, France. Background and aims: Pregestational diabetes increases the risks of congen- ital malformations and perinatal complications. The risks incurred by new- born infants in the case of gestational diabetes (GD) are debated. There are no national data available in France except for data from multicenter cohorts. We evaluated the risks of complications according to the type of maternal diabetes from the French birth cohort in 2011. Materials and methods: Data were obtained from the PMSI (French hos- pital discharge database) and the SNIIRAM (French national health insur- ance information system). All childbirths and terminations of pregnancy (TOP) after 22 weeks of gestation, due to medical reasons were selected. The mother’s diabetes was identified by an algorithm based on the consumption of antidiabetics and hospitalization diagnoses before and during pregnancy. An identifier in the PMSI links mothers and infants, thus enabling analyses of associations between the mother’s diabetes and outcome. Results: 806 579 childbirths / TOP > 22 weeks were identified in the PMSI. The Mother - infant chaining was obtained for 474 614 births in public in- stitutions. 1257 (0.16%) type 1 diabetes (T1D), 1896 (0.24%) type 2 diabetes (T2D) and 51 701 (6.4%) GD were identified, with a mean age respectively of 30.3, 33.3 and 32 years. In the case of T1D and T2D, the risks were respec- tively increased for the following complications (OR adjusted on mother’s age [95%CI]): preterm birth (gestational age < 38 weeks) (6.6[5.9-7.4] and 3.7[3.3-3.9]), caesarean section (4.3[3.8-4.8] and 2.9[2.7-3.2]), preeclampsia and eclampsia (6.7 [5.6-8.2] and 3.9 [3.2-4.7]), macrosomia (birth weight (BW) > 90th percentile) (7.0 [6.1-8.0] and 3.9 [3.4-4.4]) perinatal death (2.2 [1.4-3.4] and 3.0 [2.2-4.1]), perinatal asphyxia (3.3 [2.2-5.1] and 2.5 [1.6- 3.7]), respiratory distress syndrome (OR adjusted on mother’s age and ges- tational age): 2.6 [2.0-3.4] and 1.9 [1.5-2.5]), brachial plexus trauma and/or collarbone fractures in cases of vaginal delivery (8.5 [4.9-14.8] and 2.9 [1.5- 5.9]), cardiac malformations (4.4 [3.0-6.5] and 3.2 [2.2-4.7]). Data on mac- rosomia are respectively for DT1 and DT2: BW > 4 kg 16.7% and 13.4%; BW > 90th percentile 42.5 % (n=354) and 30.4% (n=348). In the case of GD, the risk of certain outcomes was lower: prematurity (1.35 [1.32-1.38]), caesar- ean section (1.46 [1.44-1.49]), preeclampsia and eclampsia (1.55 [1.46-1.65]), macrosomia (BW > 90th percentile) (1.7 [1.6-1.8]) respiratory distress syn- drome (1.2 [1.1-1.3]). The risks were not increased for the others complica- tions compared to the population without diabetes. Concerning macrosomia, 9.0% of the newborn had a BW > 4 kg in GD, and 6.6% in the absence of diabetes. 15.7% of the newborn had a BW > 90th percentile in GD and 9.4% in the absence of diabetes. Conclusion: The rate of perinatal complications remains high for both types of pregestational diabetes, especially T1D. The link between gestational dia- betes and macrosomia is confirmed. Some complications associated with gestational diabetes are slightly increased (caesarean section, prematurity, eclampsia, respiratory distress syndrome), others are not (brachial plexus trauma and/or collarbone fractures, death, perinatal asphyxia). 152 Maternity care at a superspecialised unit for women with type 1 diabetes results in pregnancy outcome comparable to background population M. Landin-Olsson1, H. Strevens2, D. Ursing1; 1Dept of Endocrinology, Skane University Hospital, 2Dept of Obstetrics and Gynaecology Lund, Skane University Hospital, Lund University, Sweden. Background and aims: Pregnancies in women with Type 1 diabetes are more often complicated by preeclampsia, prematurity, caesarean section, malfor- mation, macrosomia, stillbirth and postnatal hypoglycaemia in the infant. The rate of complications is mainly depending on the blood glucose regula- tion. Our model for pregnant women with Type 1 is based on a specialised maternity unit including obstetrician, diabetologist, midwife and dietician. Women are required to take seven self monitored glucose measurements daily, all values are sent electronically to the hospital and followed up weekly by a phone call from a diabetologist. The insulin doses are adjusted frequently to optimize glucose levels. An extra ultrasound is done late in pregnancy and if growth is rapid there is a liberality to induce labour. This study is aimed to compare outcome of pregnancy in women with Type 1 diabetes with the Swedish background women and also to evaluate if higher insulin dose could further improve outcome. Materials and methods: All consecutive pregnant women with Type 1 dia- betes who delivered infants during 2000-2013 (n=266) at our hospital were compared to all women in Sweden who delivered infants during the same period. Data from background population was obtained from the Swedish Medical Birth Register, which has a very high rate of ascertainment. In the next step we compared outcome in pregnancies in Type 1 diabetic women who delivered in 2000-2006 (n=1418) with those who delivered in 2007-2013 (n=148). For women in the later time period we tried to further increase in- sulin doses to improve metabolic regulation. Results: For women with Type 1 diabetes birth weight was only 59 g higher (3566 compared to 3507), despite a shorter gestational length (37.7 vs. 39.3 weeks; p=0.001) and higher frequency of sectio (44% vs. 17%; p S 72 1 C ceiving appropriate screening [odds ratio 4.01; CI 2.38-6.75 (p S 73 1 C to assess the association between GDM exposure in-utero and development of ASD in offspring. Materials and methods: This retrospective longitudinal cohort included children who were born as a singleton at 28-44 weeks gestation in hospitals from a large integrated health plan between January 1, 1995 and December 31, 2009. Children were required to have the health plan membership by age 1-2 years, when screening for developmental delays and ASD are initiated per the health plan guidelines. Children were prospectively followed from birth using electronic medical record until any one of following criteria was met: (1) a clinical diagnosis of ASD identified by ICD-9 codes 299.x, (2) >4-months of inactive health plan membership, (3) death from any cause, or (4) December 31, 2012. Children who were born to women with pre-existing diabetes were excluded. Survival analysis was used to estimate the incidence rate of ASD and Cox proportional hazard regression was used to estimate the relative risk (RR) of developing ASD associated with GDM. Results: A total of 336,164 children (26,897 [8%] GDM exposed) met the cohort inclusion criteria with a median 5.5 years (range 1.0-18.0 years) of follow-up during which 4,526 children developed ASD. The incidence rate of ASD was 2.36/1000 per year (2.81/1000 for GDM and 2.32/1000 for non- GDM group, respectively). In the unadjusted analysis, the incidence of ASD in the GDM exposed group was 20% higher than that in the GDM unexposed group (RR= 1.20, 95% CI: 1.08-1.32, p=0.0004). However, the RR was reduced to 1.06 (95% CI: 0.96-1.18, p=0.22) after adjustment for maternal age, parity, education, household income, race/ethnicity and child gender. The greatest reduction in the RR was due to the adjustment for maternal age (from 1.20 to 1.09) as mothers who had GDM were on average 3.2 years older at delivery than mothers without GDM. Additional adjustment for child birth weight, gestational weeks at delivery and congenital abnormalities at birth, factors known to be associated with maternal hyperglycaemia, slightly reduced the RR association between GDM and ASD (to 1.04, p=0.46). Excluding children with congenital abnormalities at birth had no impact to the results. Analysis by ASD sub-type reached similar conclusions. Conclusion: Data from a large multi-ethnic and population-based clinical care system did not support significant association between GDM exposure in-utero and risk of development of ASD in offspring. The slightly elevated risk for ASD in offspring exposed to GDM was largely due to older age of women with GDM. OP 27 Protecting the periphery 157 mRNA expression profiling in perivascular and subcutaneous adipose tissue of patients with carotid artery (ACI) stenosis D. Schleinitz1, P. Büttner2, A. Körner3, B. Gutsmann1, M. Fasold4, M. Eszlinger5, S. Rohm6,7, O. Richter6,7, G. Aust6, P. Kovacs1; 1IFB Adiposity Diseases, University of Leipzig, 2Cardiocenter Leipzig, Clinic for Internal Medicine and Cardiology, Department of Rhythmology R&L, University of Leipzig, 3Hospital for Children and Adolescents, University Hospital Leipzig, Center for Pediatric Research, University of Leipzig, 4ecSeq Bioinformatics, 5Division of Endocrinology and Nephrology, University of Leipzig, 6Research Laboratories, Clinic for Visceral, Transplantation, Thoracic and Vascular Surgery, University of Leipzig, 7Ev. Diakonissenkrankenhaus Leipzig, Clinic for Vascular Surgery, Center for Surgery, Leipzig, Germany. Background and aims: There is growing evidence for a vasoregulatory and atherosclerosis-inducing role of local fat deposits around vessels. Secreted adipokines may contribute directly or indirectly to the regulation of vessel wall homeostasis. Investigating molecular mechanisms specific for perivas- cular fat could therefore contribute to better understanding of its role in the pathophysiology of atherosclerosis and its sequelae such as stroke or myocar- dial infarction. Aim of the study was to investigate gene expression profiles in paired human samples of subcutaneous (sc) and perivascular (pv) adipose tissue (AT) and to link it to clinical and anthropometric characteristics of carotid stenosis patients. Materials and methods: A RNA/cDNA bank was established from paired sc (cervical) and pv (ACI) AT samples of patients who underwent carotid en- darterectomy. Tissue specific marker genes (e.g. MYH2 for muscle, COMP for fibroblasts or GPM6A for nerve) have been measured to control for impuri- ties of “non-adipocyte” cells. The expression median with appropriate ranges for every marker gene was set specifically for each fat depot. Sixty paired sam- ples passed these criteria and were assayed on Illumina HT12 microarrays. P- values for differential expression between AT depots and phenotypic groups (e.g. symptomatic vs. asymptomatic) were calculated using background-cor- rected, quantile-normalized expression values and paired/standard t-test. To correct for multiple testing an experiment based genome wide significance level and FDR methodology were applied. Gene expression was correlated to anthropometric, metabolic and clinical parameters. Best hits underwent validation using TaqMan qPCR technology. Results: We found 1100 genes with significant differential expression be- tween sc and pc AT clearly distinguishing both AT types from each other. The top hits with >2 fold changes are represented by developmental genes like HOX genes, TBX15 or WNT5A but also genes found to be involved in athero- sclerosis (CDKN2B, TRIB1) and coronary artery disease (PTGS2, CARD8). Intra-depot comparison of e.g. symptomatic vs. asymptomatic, lean vs. obese or diabetic vs. non-diabetic patients revealed genes with nominal differences in mRNA levels, which however, correlated with anthropometric and meta- bolic parameters (BMI, triglycerides, cholesterol, percent stenosis). Conclusion: Our data revealed fat depot-specific mRNA expression of devel- opmental genes and genes associated with atherosclerosis or coronary artery disease, supporting the relevance of perivascular adipose tissue in the patho- genesis of atherosclerosis. Supported by: Junior Scientist Program, University Leipzig, Med. Faculty: formel.1 158 rs196462, near SMOC2, is associated with lower extremity arterial disease in patients with diabetes N.R. van Zuydam1, M. de Andrade2, E. Vlachopoulou3, V. Salomaa3, A. Paterson4, H. Colhoun5, I.J. Kullo2, GoLEAD, SUMMIT; 1University of Oxford, UK, 2Division of Cardiovascular Diseases and the Gonda Vascular Center and Department of Health Sciences, The Mayo Clinic, Rochester, USA, 3National Institute for Health and Welfare, Helsinki, Finland, 4Hospital for Sick Kids and DCCT/EDIC, Toronto, Canada, 5Medical Research Institute, University of Dundee, UK. Background and aims: Lower extremity arterial disease (LEAD) is a com- mon macrovascular complication of diabetes and the most common cause of amputations in patients with diabetes. LEAD in patients with diabetes is Diabetologia (2014) 57:[Suppl1]S1–S564 S 74 1 C often accompanied by dyslipidaemia, renal disease and neuropathy. Despite large sample sizes, previous GWA studies have identified 1 genome wide sig- nificant (p S 75 1 C arteries is associated with chronic renal insufficiency and residual stenosis. Timely rеintervention in diabetic patients with recurrent CLI promotes op- timal WH, LS and CS. 161 Optimal blood pressure targets for prevention of peripheral artery disease in patients with type 2 diabetes and hypertension A. Hishida1, M. Tochiya2, R. Koezuka2, Y. Oohata2, T. Tamanaha2, H. Makino2, I. Kishimoto2; National Cerebral and Cardiovascular Center, Suita, Japan Background and aims: Peripheral artery disease (PAD) is one of the mac- rovascular complications of diabetes, and often associated with considerable functional limitation. Although the benefits of reducing high blood pressure (BP) on the risks of cardiovascular disease and cerebral artery disease are well known, the optimal BP goal is not established. The guideline recommended BP target for patients with diabetes had recently been changed from S 76 1 C OP 28 Ectopic lipids and type 2 diabetes 163 Effect of bariatric surgery on hepatic fatty acid uptake and blood flow H.M. Immonen1, J.C. Hannukainen1, N. Kudomi2, R. Parkkola1, V. Saunavaara1, P. Iozzo3, P. Nuutila1; 1Turku PET Centre, University of Turku, Finland, 2Faculty of Medicine, University of Kawaga, Japan, 3Institute of Biomedical Engineering, National Research Council, Padua, Italy. Background and aims: Increased delivery of free fatty acids (FFAs) to the liv- er plays a role in the pathogenesis of type 2 diabetes. Bariatric surgery is asso- ciated with high rates of diabetes remission, but the underlying mechanisms are not fully elucidated. Liver fatty acid uptake is crucial in understanding the mechanism of postoperative metabolic changes. This study’s objective was to assess the surgery-induced changes in hepatic FFA uptake, together with quantification of hepatic perfusion and liver fat content (LFC). Materials and methods: We measured hepatic FFA uptake, hepatic portal, and arterial perfusion with positron emission tomography using [18F]- FTHA and [15-O]-H2O in 22 morbidly obese subjects before and 6 months after bariatric surgery and in 14 healthy lean volunteers. LFC was quantified by MRS and liver volume by MRI. Results: Baseline: Plasma FFA concentration was elevated in the obese group compared to the lean control group (0.67 ± 0.22 vs. 0.48 ± 0.21 mmol/l, p = 0.02). Liver FFA uptake, when measured per depot or per unit mass, was higher in obese subjects compared to controls (depot: 254 ± 67 vs. 107 ± 39 μmol/min, p < 0.001 and per unit mass: 14.4 ± 3.6 vs. 10.0 ± 3.7 μmol/ min/100 ml, p < 0.01, respectively). LFC in the obese group was significantly higher than in controls (5.7 ± 4.7 vs. 1.5 ± 1.5%, p < 0.01), and 9 of the obese patients had liver steatosis (LFC ≥ 5%). When obese subjects were divided into two groups, with LFC below (LFC-low) or above (LFC-high) the median of 4.4%, the LFC-high group had an increased FFA depot uptake compared to LFC-low group (p = 0.09). When patients were grouped according to diabetic status, no difference in FFA uptake was observed. In pooled data, FFA uptake correlated with intraperitoneal fat mass (r = 0.40, p = 0.02). Compared to controls, both portal and arterial blood flow were higher in the obese group (5.2 ± 2.1 vs. 3.6 ± 1.0 l/min, p = 0.01 and 0.34 ± 0.16 vs. 0.19 ± 0.10 l/min, p < 0.001, respectively). At 6 months after surgery: BMI decreased from 41 ± 4kg/m2 to 32 ± 4kg/m2 (p < 0.001) and the remission of diabetes was observed in 8 out of 10 patients. Plasma FFA concentration was unchanged (p = NS vs. baseline). The liver depot uptake decreased by 15% (p = 0.02 vs. baseline), but remained twice as high compared to the control group (p < 0.001 vs. con- trols). LFC was decreased by 69% (p = 0.001 vs. baseline) and was now similar to that of lean controls (p = NS vs. controls). None of the subjects had liver steatosis postoperatively. Portal blood flow decreased by 55% (p < 0.0001) and arterial flow by 76% (p < 0.0001), and both were reduced compared to levels measured in controls (p < 0.05 for both). Reduction in intraperitoneal fat mass was related to changes in LFC (r = 0.59, p = 0.02) and in portal blood flow (r = 0.69, p = 0.001). An association between the change in portal blood flow and improvement in insulin sensitivity index was observed (r = 0.52, p = 0.02). Conclusion: Hepatic fatty acid uptake was decreased, but did not normalize to controls, when measured six months after bariatric surgery. After surgery, fatty acid uptake changes may contribute to the resolution of liver steatosis. Morbidly obese subjects showed increased portal vein flood flow, which re- sulted in a higher supply of fatty acids for hepatic FFA uptake. Postoperatively blood flow decreased along with the reduction in visceral fat mass and im- provement in insulin sensitivity. Clinical Trial Registration Number: NCT01373892 Supported by: Academy of Finland, Sigrid Juselius Foundation, EUFP7 Hepadip 164 High-fat diet increases autophagic flux in pancreatic beta cells in vivo K.-Y. Chu, T.J. Biden; Garvan Institute, Darlinghurst, Australia. Background and aims: Autophagy is an important cellular survival mecha- nism that responds to changes in cellular nutrients, and potentially coun- ter-regulates endoplasmic reticulum (ER) stress, previously implicated in lipotoxic beta-cell death. Because it is extremely difficult to measure true autophagic flux in vivo, it is still unclear how high-fat feeding or fatty acids regulate autophagy in pancreatic beta-cells. Here, we used both in vivo and in vitro models to study autophagic flux; and also investigated the upstream signaling pathways triggered by fatty acids. Materials and methods: For unequivocal estimation of autophagic flux, GFP-LC3 mice were fed with chow or high-fat diet for 10 weeks and then, for 5 consecutive days before sacrifice, injected with 100mg/kg chloroquine to block clearance of autophagic markers. The pancreata and livers were col- lected and fixed for cryosectioning. The beta-cells of the pancreatic sections were labeled with the insulin antibody, and the GFP-LC3 signaling analysed by fluorescence microscopy. Autophagic and ER stress markers were detected by western blot following acute (2h) treatment with chloroquine using both islets ex vivo from the high-fat fed mice, and mouse clonal MIN6 beta-cells treated with oleate and palmitate for 0-48h. Results: Autophagic flux, assessed from the amount of GFP puncta, was in- creased in pancreatic acinar tissue and beta-cells after high-fat feeding and chloroquine injection. In contrast, GFP signal in the liver was markedly at- tenuated under identical conditions. LC3 levels in the isolated mouse islets were also increased by high-fat diet and further enhanced by chloroquine incubation ex vivo. ER stress, indicated by the markers CHOP and phospho- eIf2α levels, in isolated islets was not augmented by high-fat feeding alone, but was induced by autophagic inhibition with chloroquine and further en- hanced by high-fat feeding. This reveals a reciprocal relationship between ER stress and autophagy in beta-cells and suggests that autophagy is not solely secondary to ER stress in the context of lipotoxicity in vivo. In MIN6 cells, oleate increased LC3-II levels in both the presence and absence of chloro- quine (indicative of enhanced flux and steady state autophagy respectively) whereas palmitate augmented the autophagic flux alone. ER stress was in- creased by palmitate but not oleate, which concurs with the in vivo model that lipids/fatty acids could induce autophagy without the activation of ER stress. Nevertheless, neither fatty acid modulated classical mTOR signaling illustrated by the downstream phospho-4EBP levels. Conclusion: Using GFP-LC3 mice with chloroquine injection we show for the first time that beta-cells have a very low basal autophagic flux but, in con- trast to liver, this is increased by high-fat feeding. This appears to act as a protective mechanism prior to the activation of ER stress, and is triggered independently of the classical mTOR pathway. Supported by: NHMRC 165 Sphingosine kinase 1 promotes hepatic steatosis via up-regulation of PPARgamma P. Xia1,2, J. Chen2, W. Wang1, G. McCaughan2, X. Gao1, M. Vadas2; 1Endocrinology and Metabolism, Fudan University Zhongshan Hospital, Shanghai, China, 2Centenary Institute, the University of Sydney, Australia. Background and aims: Sphingolipid metabolites have emerged playing key roles in the pathogenesis of metabolic diseases, including obesity, diabetes and nonalcoholic fatty liver disease (NAFLD). The current study aimed to ex- plore the underlying mechanisms and discover a potentially druggable inter- mediate in the steatotic pathways associated with sphingolipid metabolites. Materials and methods: We utilized pharmaceutical inhibitors or genetic means to manipulate sphingosine kinase 1 (SphK1), one of the key enzymes that control sphingolipids metabolism. Both cellular and animal models were applied in the study investigating the role of SphK1 in hepatic lipid metabo- lism and NAFLD development. Results: A nearly 2-fold increase in the expression level of SphK1 was ob- served in either cellular or animal models of NAFLD, without alterations in SphK2 (another isoenzyme of SphK), suggesting a specific link of SphK1 to NAFLD. Enforced overexpression of SphK1 significantly promoted lipid ac- cumulation in mouse primary hepatocytes, whereas loss of SphK1 gene ex- pression inhibited the process compared with the control hepatocytes (both p S 77 1 C bition by using its siRNA or antagonists markedly reduced SphK1-depentant hepatic lipid accumulation (all p S 78 1 C Materials and methods: Ten healthy lean sedentary volunteers (age 25±1 years, body mass index 24.8±0.3 kg/m2) were randomized in a cross-over manner to an oral lipid emulsion (50% fat) enriched in saturated fatty acids (SAFA: 92 g palm oil), omega-6 polyunsaturated fatty acids (PUFA: 92 g saf- flower oil) and water as control (CON). Skeletal muscle biopsies were per- formed before and 2.5 hours after interventions. Euglycemic-hyperinsuline- mic clamps with infusion of deuterated glucose were initiated 6 hours after interventions to assess whole-body (M-value) and hepatic insulin sensitivity (insulin-mediated suppression of endogenous glucose production). High- resolution respirometry was applied to assess ex-vivo muscle mitochondrial respiratory capacity. Results: After both SAFA and PUFA, whole-body insulin sensitivity was lower compared to CON (CON: 7.3±0.8; PUFA: 5.8±0.8; SAFA: 4.3±0.5 mg.kg-1.min-1, p S 79 1 C healthy gut microbiota, and thus prevent the development to type 1 diabetes associated islet autoimmunity (IA) in young children. The aim of this study was to examine the association between early probiotic use and IA. Materials and methods: The Environmental Determinants of Diabetes in the Young (TEDDY) study prospectively follows 8502 children with type 1 diabe- tes (T1D) -associated HLA-DR-DQ alleles in Finland, Germany, Sweden and the US. Blood samples were collected every 3 months from birth to evaluate the primary outcome of IA, defined as appearance of one or more of the islet autoantibodies GADA, IAA, or IA-2A confirmed at two consecutive visits. The introduction of either probiotic supplement or infant formula containing probiotics was classified as: early introduction (at the age of S 80 1 C 173 Family history of type 1 and type 2 diabetes and the risk of LADA-results from a population-based study of incident cases S. Carlsson1, T. Andersson1, P.-O. Carlsson2, M. Dorkhan3, L. Groop3, J. Edwall Löfvenborg1, R. Hjort1, M. Martinell4, B. Rasouli1, P. Storm3, T. Tuomi5,6; 1Division of epidemiology, Institute of Environmental Medicine, Stockholm, Sweden, 2Department of Medical Sciences, Uppsala, Sweden, 3Department of Clinical Sciences, Clinical Research Centre, Malmö, 4Department of Public Health and Caring Sciences, Uppsala, Sweden, 5Division of Endocrinology, Helsinki, 6Research Program for Diabetes and Obesity, University of Helsinki, Finland. Background and aims: LADA (latent autoimmune diabetes in adults) is sug- gested to be genetic mix of type 1 and type 2 diabetes, linked to HLA-DQB1 risk genotypes associated with autoimmunity as well as genes associated with type 2 diabetes, including TCF7L2 . Family history of diabetes (FHD) encom- passes both genetic and shared environmental factors and is a strong predic- tor of diabetes risk, although scarcely investigated in relation to LADA. Our aim was to investigate the risk of LADA in relation to family history of type 1 and type 2 diabetes. Materials and methods: We used data from a population based case-control study with incident cases of adult onset (≥35 years) diabetes, including 264 cases of LADA (Glutamic acid decarboxylase antibodies (GADA) positive (>10 IU/mL) with c-peptide >0.3 nmol/l, 796 cases of type 2 diabetes (GADA negative), together with 1047 controls without diabetes, randomly selected from the population. Self-reported information on diabetes in first and sec- ond degree relatives was collected and relatives with onset median); OR was estimated at 7.3 (95% CI 1.91-28.0) for FHD of T1D and at 12.4 (95% CI 5.5-27.9) for the combination of T1D and T2D in the family. No difference was seen for FHD of TD1 in female (OR 4.0; 95% CI 1.7-9.1) vs. male relatives (OR 4.4; 95% CI 1.9-10.2).The risk of T2D was associated with FHD of T2D (OR 2.9; 95% CI 2.3-3.6) but not FHD of T1D. In LADA pa- tients, FHD of T1D vs T2D was associated with higher GADA (197.8 vs 139.5 IU/mL, p=0.003) but lower c-peptide (0.55 vs 0.81 nmol/l, p=0.009) levels. Conclusion: The current data add support to the view that LADA is an ad- mixture of both T1D and T2D, also in terms of genetic risk. Supported by: Swedish Medical Research Council, FORTE, AFA, ALF a 174 Low birth weight is associated with an increased risk of latent autoimmune diabetes in adults (LADA) and type 2 diabetes: results from ESTRID a Swedish case-control study R. Hjort1, P.O. Carlsson2, M. Dorkhan3, L. Groop3, M. Martinell4, B. Rasouli1, T. Toumi5, S. Carlsson1; 1Institute of Environmental Medicine, Karolinska Institutet, Stockholm, 2Department of Medical Sciences, Uppsala University, 3Department of Clinical Sciences in Malmö, Lund University, Malmö, 4Department of Public Health and Caring Sciences, Uppsala University, Sweden, 5Department of Medicine, Helsinki University Central Hospital, Finland. Background and aims: A large body of research has recognized low birth weight as a risk factor for type 2 diabetes (T2D), hypothetically as an indica- tor of poor nutrition of the fetus, leading to insulin resistance. Genetic factors have also been implicated, e.g. it has been suggested that effects of paternal diabetes may be mediated by effects on intrauterine environment, manifested in low birth weight of the offspring. In contrast a link between high birth weight and type 1 diabetes (T1D) has been documented. Our aim was to in- vestigate, for the first time, the association between birth weight and LADA, a common diabetes form with features of both T1D and T2D. Materials and methods: We used data from ESTRID (Epidemiological Study of Risk factors for LADA and type 2 Diabetes), a Swedish population-based study. Eligible for the analysis were 116 incident LADA cases (≥ age 35 and Glutamic Acid Decarboxylase Antibodies (GADA) positive (>10 IU/mL) with c-peptide >0.3 nmol/l), 298 incident T2D cases (≥ age 35 and GADA negative) and 521 disease-free controls randomly sampled from the popula- tion. Information on birth weight and covariates was based on self-report. We present Odds Ratios (OR) and 95% Confidence Intervals (CI) calculated by logistic regression and adjusted for sex, age, current BMI (kg/m2) and family history of diabetes (FHD) in first and second degree relatives. Results: Low birth weight was associated with an increased risk for LADA as well as T2D; OR per kilogram reduction was estimated at 1.40 (CI; 1.03-1.90) and 1.39 (CI; 1.07-1.81) respectively and OR for subjects weighing less than 3 kg compared to more than 4 kg at birth was estimated at 2.06 (CI; 1.07-3.97) for LADA and 1.98 (CI; 1.16-3.40) for T2D. Further adjustments for educa- tion, physical activity, smoking and alcohol did not affect these estimates. Indications of an excess risk was seen both in those with FHD (OR 1.24, CI; 0.79-1.95) and without FHD (OR 1.59, CI; 1.04-2.43). A combination of low birth weight ( S 81 1 C OP 30 Integrative physiology 175 A novel mitochondrial mechanism controlling insulin secretion in obesity M. Liesa, S.B. Sereda, L. Stiles, L. Nocito, O.S. Shirihai; Boston University School of Medicine, USA. Background and aims: Reactive oxygen species can increase or decrease in- sulin secretion. However, it is unknown whether inhibiting an antioxidant system in islets can lead to improved insulin secretion without affecting their viability. Our goal is to identify whether modulating components related to heme metabolism, known to increase reactive oxygen species production but also antioxidant activity, can increase secretion without affecting viability. To this end, we studied the role of the mitochondrial transporter ATP-binding cassette B10 (ABCB10) regulating glucose stimulated insulin secretion in the context of diet-induced obesity, as ABCB10 is essential for proper heme me- tabolism and protection from oxidative stress. Materials and methods: Obesity in mice was induced by high fat diet feeding (45% Fat, Research Diets). Mice fed hypercaloric control diet (CD, calories matched with carbohydrates) or chow diet were used as controls. Glucose tolerance tests (1g/kg I.P.) after fasting were performed and both insulin (ELISA, ALPCO) and glucose (FreeLite glucometer) were measured. Stud- ies were carried out along BU IACUC and “Principles of laboratory animal care” (NIH). Mouse models are whole body ABCB10 +/- mice (n=10 back- crossed) and ABCB10-LoxP+/+ mice (pure) in C57Bl6J background. Islets were isolated 14-24 weeks after diet onset (weaning). Glucose stimulated in- sulin secretion was performed in modified DMEM by incubating islets in 2.8 mM glucose or 16.7 mM glucose for 30 minutes. Insulin in the media was measured by FRET (HTRF, CisBio). In the case of ABCB10-LoxP islets, they were transduced with Cre or Ds-Red control adenovirus at MOI 200 after isolation. Insulin was measured as in ABCB10 +/- islets. ABCB10 Real time PCR was performed using TaqMan probes. Results: High fat diet and control diet feeding for 5 months had the expected effects increasing weight, fasting glucose and insulin levels, when compared to chow diet feeding. No significant differences in these parameters were de- tected between wild type and ABCB10 +/- mice though. However, glucose tolerance tests showed that ABCB10 +/- mice were more tolerant than wild type mice and only after high fat diet feeding. This increase in glucose tol- erance was associated with higher blood insulin levels in response to glu- cose. In this regard, only high fat diet feeding resulted in ABCB10 +/- islets showing higher glucose stimulated insulin secretion when compared to wild type (40% increase). To address whether ABCB10 acutely regulated insulin secretion, we inactivated ABCB10 alleles after high fat diet feeding by Cre adenoviral expression in ABCB10 LoxP+/+ or LoxP+/- islets. Using this ap- proach, we reduced ABCB10 expression by 80-90% and 40-50% respectively. We found that only after high fat feeding, glucose stimulated insulin secretion was increased in ABCB10 knock out (LoxP+/+ - Cre; ~30%) and in LoxP+/- Cre (~15%), when compared to respective DsRed transduced controls. This increase in insulin secretion was also present in ABCB10 knock out islets from chow diet mice just exposed for 24 hours to high glucose (20mM) and NEFA (palmitate 0.4 mM). Conclusion: We report an unexpected role for the mitochondrial transporter ABCB10 regulating insulin secretion specifically in the context of high fat diet. We propose that lack of ABCB10 increases a reactive oxygen species- related high fat specific signal in islets that stimulates insulin secretion. In all, ABCB10 modulation could be a potential approach to improve insulin secretion specifically in the context of obesity. Supported by: Evans Center; BNORC-NIH/NIDDK 2P30DK046200 176 Circulating irisin is up-regulated by insulin infusion in obese, but not in lean humans, and is inversely associated with insulin sensitivity and respiratory exchange ratio M. Straczkowski1,2, A. Nikolajuk2, N. Matulewicz1, M. Stefanowicz1, M. Karczewska-Kupczewska1,3; 1Department of Metabolic Diseases, Medical University of Bialystok, 2Department of Prophylaxis of Metabolic Diseases, Institute of Animal Reproduction and Food Research, Polish Academy of Sciences, Olsztyn, 3Department of Metabolic Diseases, Institute of Animal Reproduction and Food Research, Polish Academy of Sciences, Olsztyn, Poland. Background and aims: Irisin is a recently discovered myokine, with the po- tential to induce brown-fat-like development of white adipose tissue and to increase energy expenditure. In humans, data on the relationships of circulat- ing irisin with insulin sensitivity and other metabolic parameters are incon- sistent. The aim of the present study was to assess the associations of serum irisin with insulin sensitivity and substrate oxidation, as well as the effect of insulin infusion on circulating irisin. Materials and methods: The study group consisted of 148 healthy subjects (115 males and 33 females, age between 18 and 35 years), 81 lean and 67 overweight or obese. Euglycemic hyperinsulinemic clamp and indirect calo- rimetry were performed. Serum irisin was measured in the baseline state and after the clamp. Results: Baseline irisin was not different between lean and overweight/obese group (p=0.44). Insulin infusion resulted in an increase in serum irisin in overweight/obese (p=0.006), but not in the lean group (p=0.73). In conse- quence, post-clamp irisin was higher in the overweight/obese (p=0.042). Baseline irisin was related to fasting NEFAs (r=0.26, p=0.008). Post-clamp irisin was positively related to BMI, waist circumference, triglycerides and resting energy expenditure (all p S 82 1 C VBM8 toolbox, and compared using t tests. Data were thresholded at p < 0.05, FDR corrected at cluster level. Results: WM but not GM density was higher in OLM versus OOM glob- ally in the brain. Exercise intervention increased WM density in both OLM and OOM subjects in the cerebellum. The intervention increased WM and decreased GM density in parietal superior regions in OLM and in cuneus and precuneus regions in OOM. When comparing OLM and OOM for the changes in WM density following intervention, cerebellum of the OLM showed higher increase of WM density compared to OOM (Figure 1). Conclusion: Maternal BMI during pregnancy influences brain atrophy in their offspring, especially the white matter. We show that exercise has an im- pact on brain morphology even in elderly women and in a brief time period. Being born to an obese mother compared to a lean or normal weight mother implies lower brain adaptability to exercise-induced changes in the WM den- sity. These changes were more pronounced in the cerebellum region and this could have been caused by axon caliber and myelination increase in a similar way to what happens in subjects learning a new motor skill. Clinical Trial Registration Number: NCT01931540 Supported by: DORIAN, EU FP7 Grant agreement N° 278603 178 Plasma FGF21 is released from the splanchnic circulation in response to exercise and regulated by insulin and glucagon: a human study P. Plomgaard1, J. Hansen1, J.O. Clemmesen2, N.H. Secher3, B.K. Pedersen4; 1Department of Clinical Biochemistry, 2Department of Hepatology, Rigshospitalet, 3Department of Anaesthesiology, 4Center of Inflammation and Metabolism, Rigshospitalet, Copenhagen, Denmark. Background and aims: FGF21 is a circulating protein that has beneficial metabolic effects. An increase in plasma FGF21 has been observed with fast- ing and exercise; probably secreted from the liver. However the regulation of plasma FGF21 in normal physiology is not completely understood. Aims: 1) to demonstrate that FGF21 is released from the liver in response to an acute exercise bout. 2) to investigate if glucagon and/or insulin regulates plasma FGF21 levels. Materials and methods: A: Ten healthy young males performed a 2 hour ex- ercise bout, with a hepatic vein and arterial catheter allowing drawing blood repeatedly and simultaneously before, during and after a 2 hour bicycling exercise bout at 60% VO2max. Hepatic blood flow was determined by use of indocyanine green. B: Ten subjects underwent 4 trails resting in supine posi- tion: 1) 1 hour of glucagon infusion 2) 1 hour of glucagon combined with a co-infusion of somatostatin 3) 1 hour of somatostatin infusion, 4) 1 hour of saline infusion (control). Blood samples were obtained every hour during the 8 hour experimental day. Plasma insulin, glucagon and FGF21 was measured by immunoassays. Results: The exercise study with the arterial-hepatic venous difference dem- onstrated a higher level of insulin and glucagon in the hepatic vein compared to the artery. During the exercise bout a decrease in plasma insulin and an increase in plasma glucagon were observed, which normalised in the recov- ery after the bicycling exercise. Plasma FGF21 gradually increased during the exercise bout to a 5-fold increase and rapidly returned to basal level into re- covery. Calculating the arterial- hepatic vein differences revealed a net release of FGF21 at rest before the bicycling commenced. The secretion of FGF21 in- creased gradually during the 2 hours of bicycling and returned to baseline in the recovery. The hepatic blood flow did not change during the experiment. In infusion study 1 (glucagon alone), plasma glucagon increased during the 1 hour infusion followed by compensatory increase in plasma insulin. In trial 2 (glucagon + somatostatin) glucagon increased similarly to trail 1, with no compensatory insulin increase. In trial 3 (somatostatin alone) both insulin and glucagon decreased, compared to trail 4 (saline), where no changes were observed. Interestingly, plasma FGF21 increased with a similar kinetics as observed with exercise, only when glucagon was infused and the compen- satory insulin response was blocked by somatostatin. A delayed increase in plasma FGF21 was observed in trail 3 (somatostatin alone). Conclusion: Here we present the first human data demonstrating by a direct measurement across the splanchnic organs that FGF21is released. The liver is the most likely organ to be responsible; however contributions from other organs as the spleen, intestines or pancreas cannot be excluded. Furthermore our data demonstrate that the splanchnic circulation releases exercise-in- duced FGF21 in humans. Exercise increased plasma glucagon concomitantly with a decrease in insulin and simulating this condition in resting healthy men, increased plasma FGF21 with a similar kinetics as observed with ex- ercise. These data add to the understanding of FGF21 in human physiology. Supported by: Augustinus Fonden 179 Selective expression of ROCK1 in the liver promotes insulin resistance and hepatic steatosis in diet-induced obese mice I.S. Lima1,2, S.-H. Lee3, M. Chung1, M.J. Kim1, M.P. Macedo2,4, Y.B. Kim1; 1Division of Endocrinology, Diabetes and Metabolism, Beth Israel Deaconess Medical Center - Harvard Medical School, Boston, USA, 2CEDOC - Chronic Diseases Research Center, Faculdade de Ciências Médicas – Universidade Nova de Lisboa, Portugal, 3Department of Internal Medicine - Division of Endocrinology and Metabolism, The Catholic University of Korea, Seoul, Republic of Korea, 4Portuguese Diabetes Association - Education and Research Center (APDP-ERC), Lisboa, Portugal. Background and aims: In the European Union approximately 29 million people suffer from a chronic liver condition. The prevalence of non-alcoholic fatty liver disease (NAFLD) is 2-44% in the European population and 42.6- 69.5% in people with type 2 diabetes. Furthermore, over 50% of adults in the European Union are overweight or obese. Obesity is a risk factor for NAFLD and is strongly associated with insulin resistance. Our previous data showed that liver-specific deletion of Rho-kinase 1 (ROCK1) caused a significant im- provement in insulin sensitivity and hepatosteatosis in obese mice induced by a high-fat diet. The current study was designed to further determine the physiological role of hepatic ROCK1 in regulating whole-body glucose and lipid metabolism. Materials and methods: Mice expressing a constitutively active (CA) mutant of ROCK1 in liver were studied. These mice started a high-fat diet (HFD) at 6 weeks of age for a period of 12 weeks. Insulin sensitivity and glucose toler- ance were assessed and body weight and glucose levels were also monitored. Hepatic and serum content in triglycerides and cholesterol was determined. Hematoxylin and eosin stain (H&E stain) of liver sections from control and CA-ROCK1 mice was performed. Gene expression of key molecules involved in lipid metabolism was also determined for control and CA-ROCK1 mice. Results: Liver-specific CA-ROCK1 mutant mice exhibited higher body weight 2 weeks after HFD feeding (21.6 ± 0.4 g N=10 vs. 23.6 ± 0.6 g N=10, p = 0.01) and this difference was increased by the period on HFD (33.8 ± 1.28 g N=10 vs. 39.4 ± 1.4 g N=10, p = 0.01). Blood glucose was also increased after 4 weeks of HFD (148.0 ± 3.7 mg/dL N=10 vs. 165.4 ± 7.1 mg/dL N=10, p = 0.05). These mice were insulin resistant, as revealed by the failure of blood glucose levels to decrease after in insulin injection (AUC 16058 ± 594.0 N=6 vs. 20581 ± 1102 N=10, p = 0.01), but normal glucose tolerant. These ef- fects were accompanied by hyperinsulinemia, increased hepatic triglycerides (430.2 ± 40.7 mg/dL N=9 vs. 674.1 ± 108.0 mg/dL N=8, p = 0.05) and serum cholesterol (94.2 ± 8.4 mg/dL N=10 vs. 131.4 ± 12.6 mg/dL N=9, p = 0.05) in the CA-ROCK1 mice. Histological analysis showed that hepatic steatosis by high-fat feeding was greatly increased in liver-specific CA-ROCK1 mutant mice compared with control mice. Moreover, activation of ROCK1 in liver caused an increase in gene expressions of key lipogenic enzymes, includ- ing FAS (fatty acid synthase) and ACC (Acetyl-CoA carboxylase). However, overexpression of hepatic CA-ROCK1 had no effect on gene expression in- volved in fatty acid oxidation and fatty acid uptake. Conclusion: Our data demonstrate that activation of hepatic ROCK1 is suf- ficient to cause insulin resistance and hepatic steatosis in diet-induced obese mice, suggesting an important role for hepatic ROCK1 in regulating fuel me- Diabetologia (2014) 57:[Suppl1]S1–S564 S 83 1 C tabolism. Thus, hepatic ROCK1 could be a molecular target for the treatment of obesity and obesity-related metabolic disorders such as NAFLD. Supported by: FCT SFRH/BD/71021/2010 180 Prediabetes and type 2 diabetes are associated with increased content of dipeptide carnosine in human skeletal muscle T. Kurdiova1, B. deCourten2, S. Vallova1, M. Balaz1, M. Vician3, M. Vlcek1, V. Belan4, I. Klimes1, D. Gasperikova1, W. Derave5, J. Ukropec1, B. Ukropcova1; 1Obesity section, Diabetes laboratory, Institute of Experimental Endocrinology, SAS, Bratislava, Slovakia, 2Monash University, Melbourne, Australia, 3Department of Surgery, Slovak Medical University, 4Radiology Clinic, University Hospital, Bratislava, Slovakia, 5Department of Movement and Sports Sciences, Ghent University, Belgium. Background and aims: Carnosine (β-alanine L-histidine) is dipeptide com- monly found in mammalian tissues and enriched in skeletal muscle. It has been shown to suppress many biochemical processes that accompany ageing and age related chronic diseases. Recent animal studies suggested the impor- tance of carnosine supplementation for glycemic control and prevention of type 2 diabetes. Our aim was to analyze skeletal muscle carnosine content in middle aged (45.1±1.2 years) lean healthy ( BMI 24.5±0.4 kg.m-²), over- weight/obese (BMI 29.3±0.8 kg.m-²), prediabetic (BMI 32.1±0.7 kg.m-²) and type 2 diabetic (BMI 30.7±0.9 kg.m-²) individuals (n=9/group). Materials and methods: Metabolic phenotyping included euglycemic hy- perinsulinemic clamp (insulin sensitivity), oral glucose tolerance test, MRI (abdominal adipose tissue content & distribution) and indirect calorimetry (resting energy expenditure). Physical activity was monitored with accel- erometers and validated questionnaire. Samples of m.vastus lateralis were obtained by needle biopsy. Skeletal muscle carnosine content (HPLC) and serum CN1 carnosinase activity were determined. Mitochondrial biogenesis related genes as well as relative muscle mitochondrial content were assessed by qPCR. Results: Skeletal muscle carnosine content progressively increased in pa- tients with prediabetes (by 30%, NS) and type 2 diabetes (by 39%, p S 84 1 C 182 Efficacy and safety of liraglutide versus placebo in subjects with type 2 diabetes and moderate renal impairment (LIRA-RENAL): a randomised trial G. Umpierrez1, S. Atkin2, S. Bain3, P. Rossing4, D. Scott5, M. Shamkhalova6, H. Bosch-Traberg7, A. Syrén7, M. Davies8; 1Emory University School of Medicine, Atlanta, USA, 2Weill Cornell Medical College Qatar, Doha, Qatar, 3Abertawe Bro Morgannwg University NHS Trust, Swansea, UK, 4Diabetes, Steno Research Center, Gentofte, Denmark, 5Clinical Research Development Associates, Rosedale, USA, 6Endocrinology Research Centre, Moscow, Russian Federation, 7Novo Nordisk A/S, Søborg, Denmark, 8University of Leicester, UK. Background and aims: Renal impairment in type 2 diabetes mellitus (T2DM) limits the available glucose lowering treatment options. The rationale for this trial was to establish the efficacy and safety of liraglutide 1.8 mg as add-on to existing glucose lowering drugs and/or insulin therapy in subjects with inadequately controlled T2DM and moderate renal impairment (estimated glomerular filtration rate [eGFR] 30-59 mL/min/1.73 m2; MDRD). Materials and methods: In this 26-week, double-blind, multicentre, mul- tinational, parallel group trial, adults with T2DM and moderate renal im- pairment, BMI of 20-45 kg/m2, HbA1c of 7.0-10.0% and on stable diabetes medication (unchanged medication and dose for ≥90 days prior to screening) were randomised 1:1 to receive either once-daily liraglutide 1.8 mg or pla- cebo. Liraglutide was initiated at 0.6 mg/day and incremented by 0.6 mg/day on a weekly basis to the target dose of 1.8 mg/day. The primary endpoint was change in HbA1c from baseline (BL) to Week 26 (Table). Results: 279 subjects were randomised (140 to liraglutide; 139 to placebo). All subjects, except for 2 in the placebo group, were exposed to trial medica- tion and included in the analysis. Liraglutide 1.8 mg showed superior gly- caemic control relative to placebo in subjects with moderate renal impair- ment with a low risk of hypoglycaemia and reduced body weight (Table). The most common adverse events (AEs) were GI side effects (liraglutide 35.7%, placebo 17.5%), mostly nausea and vomiting which resolved quickly. There was a higher incidence of AE leading to withdrawals in the liraglu- tide group (13.6%), compared to placebo (2.9%). No deterioration in renal function was observed (eGFR change from BL: -1% liraglutide; +1% placebo p=0.36). There was an increase from baseline in amylase and lipase levels in the liraglutide group that was not seen in the placebo group. One subject with elevated lipase (>3x ULN) and amylase (>2x ULN) at baseline was diagnosed with chronic pancreatitis on Day 11 of treatment with liraglutide. Conclusion: Liraglutide 1.8 mg showed superior HbA1c and weight reduc- tion with no unexpected safety or tolerability issues including no worsening of renal function in subjects with moderate renal impairment over 26 weeks. The efficacy, low incidence of hypoglycaemia and safety of liraglutide in sub- jects with T2DM and moderate renal impairment was demonstrated. Clinical Trial Registration Number: NCT01620489 Supported by: Novo Nordisk 183 Effect of saxagliptin on renal outcome O. Mosenzon1, D.L. Bhatt2, L.E. Litwak3, M. Shestakova4, K.K. Ray5, G. Leibowitz1, B. Hirshberg6, C. Wei6, A. Parker7, N. Iqbal8, B.M. Scirica2, R.C.W. Ma9, I. Raz1; 1The Diabetes Unit, Hadassah Hebrew University - Medical Center, Jerusalem, Israel, 2Brigham and Women’s Hospital, Harvard Medical School, Boston, USA, 3Hospital Italiano de Buenos Aires, Argentina, 4Endocrinology Research Center, Moscow, Russian Federation, 5St. Georges University of London, UK, 6AstraZeneca, Wilmington, 7AstraZeneca LP, Wilmington, 8BMS, Princeton, USA, 9The Chinese University of Hong Kong, Hong Kong. Background and aims: It was previously suggested that DPP-4 inhibitors may have a protective effect on diabetic kidney disease. Herein we studied the renal outcomes of 16,492 patients (pts) with type 2 diabetes, who were prospectively followed in the SAVOR TIMI 53 trial for a median of 2.1 years. Predefined renal outcomes (safety and efficacy) were: doubling of serum cre- atinine, the composite of initiation of chronic dialysis, renal transplant or serum creatinine >6.0 mg/dl, reduction from baseline of albumin/creatinine ratio (ACR) and categorical changes in ACR. Materials and methods: At baseline, pts were stratified into 3 groups accord- ing to renal function: normal or mild renal dysfunction (RD) [eGFR >50 mL/ min; N=13,916], moderate RD [eGFR 30-50 mL/min; N=2,240], or severe RD [eGFR S 85 1 C 184 Liraglutide 3.0 mg reduces severity of obstructive sleep apnoea and body weight in obese individuals with moderate or severe disease: SCALE sleep apnoea trial A. Blackman1, G. Foster2, G. Zammit3, R. Rosenberg4, L. Aronne5, T. Wadden6, B. Claudius7, C.B. Jensen7, E. Mignot8; 1Toronto Sleep Institute, MedSleep, Canada, 2School of Medicine, Temple University, Philadelphia, 3Clinilabs Sleep Disorders Institute, New York, 4NeuroTrials Research, Atlanta, 5Weill Cornell Medical College, New York, 6Perelman School of Medicine, University of Pennsylvania, Philadelphia, USA, 7Novo Nordisk A/S, Søborg, Denmark, 8School of Medicine, Stanford University, Palo Alto, USA. Background and aims: Obesity is strongly associated with obstructive sleep apnoea (OSA) and weight loss has been shown to reduce disease severity. This randomised, double-blind, parallel-group trial compared the effects of liraglutide 3.0 mg to placebo, both as adjunct to diet (500 kcal/day deficit) and exercise, on OSA severity and body weight. The primary endpoint was change in apnoea-hypopnoea index (AHI, number of apnoea/hypopnoea events per hour of sleep) after 32 weeks (Table). Materials and methods: Obese individuals without diabetes who had mod- erate (AHI 15-29.9 events/h) or severe (AHI ≥30 events/h) OSA and were unwilling/unable to use continuous positive airway pressure therapy were randomised 1:1 to liraglutide 3.0 mg or placebo for 32 weeks. Of 359 ran- domised individuals (mean baseline age 48.5 years, males 71.9%, AHI 49.2 events/h, severe OSA 67.1%, body weight 117.6 kg, BMI 39.1 kg/m2, HbA1c 5.7%), 276 (76.9%) completed the trial. Results: After 32 weeks, the reduction in AHI was significantly greater with liraglutide 3.0 mg than with placebo (-12.2 vs. -6.1 events/h, p=0.0150) (Ta- ble). Supporting the reduction in AHI, endpoints related to oxygen satura- tion, polysomnographic measures of sleep quantity and efficiency, and qual- ity of life also improved with liraglutide 3.0 mg, albeit statistically non-signifi- cantly compared to placebo. Liraglutide 3.0 mg produced significantly greater mean percentage weight loss compared to placebo (-5.7 vs. -1.6%, p10% weight loss targets after 32 weeks. Post-hoc analysis showed a relationship between weight loss and change in AHI. In addition, there were significantly greater reductions in neck circumference, HbA1c and systolic blood pressure (SBP) with lira- glutide 3.0 mg versus placebo. After 32 weeks, heart rate increased by about 2 beats/min with liraglutide 3.0 mg compared to placebo. The safety profile of liraglutide 3.0 mg was generally consistent with that previously seen with liraglutide in type 2 diabetes. Nausea and diarrhoea were the most common adverse events with liraglutide 3.0 mg (27% and 17% of individuals, respec- tively); most events were mild/moderate and transient. Conclusion: As an adjunct to diet and exercise, liraglutide 3.0 mg was gener- ally well tolerated and produced significantly greater reductions than placebo in AHI, body weight, SBP and HbA1c in obese individuals with moderate/ severe OSA. The results also indicate that weight loss improves OSA-related parameters. Clinical Trial Registration Number: NCT01557166 Supported by: Novo Nordisk 185 The incretin hormone GLP-1 is upregulated in critically ill ICU patients and regulates the metabolic response during acute inflammation: central role of IL-6 F. Kahles, C. Meyer, J. Möllmann, C. Lebherz, H.M. Findeisen, S. Diebold, A. Koch, F. Tacke, C. Trautwein, N. Marx, M. Lehrke; University Hospital Aachen, Germany. Background and aims: Hypoglycemia predicts adverse outcome in critically ill patients including those with sepsis or myocardial infarction. The underly- ing mechanism for inflammation-mediated hypoglycaemia are incompletely understood. Experimental data suggest that hyperinsulinemia might play a role here and that LPS induces insulin secretion in a glucose-dependent man- ner. Since the incretin hormone Glucagon-like-peptide 1 (GLP-1) also causes insulin secretion, we speculated that GLP-1 may be of relevance for inflam- mation dependent hypoglycaemia. Materials and methods: C57BL/6J, IL6 knockout (KO) and IL1 receptor KO mice were given i.p. injections of saline or lipopolysaccharides (LPS) (100μg/ kg), IL1β, IL6 and TNFα (all 4 μg/kg), respectively (n=6-12/ group). A total of 155 critically ill patients (112 with sepsis, 43 without sepsis) were studied prospectively upon admission to the medical intensive care unit (ICU) and compared to 134 healthy controls. Results: LPS upregulated GLP-1 secretion in C57BL/6J mice with a maximal 3,4 fold increase (p S 86 1 C 186 Saxagliptin in patients with prior heart failure - observations from the SAVOR-TIMI 53 trial B.M. Scirica1, E. Braunwald1, I. Raz2, O. Mosenzon2, A. Cahn2, J. Lopez-Sendon3, B. Hirshberg4, P.A. Pollack4, P. Steg5, A. Umez-Eronini1, K. Im1, D.L. Bhatt1; 1Medicine, Brigham and Women’s Hospital, Boston, USA, 2Medicine, Hadassah Medical Center, Jerusalem, Israel, 3Hospital Universitario La Paz, Madrid, Spain, 4AstraZeneca, Wilmington, USA, 5Medicine, Département Hospitalo-Universitaire FIRE, INSERM U-1148, Université Paris-Diderot , and Hôpital Bichat, Paris, France. Background and aims: Heart failure (HF) and T2DM frequently co-exist; however, there are limited data regarding the safety of a antihyperlycemic drugs in patients with T2DM and prior HF and thus clinical uncertainty over optimal glycemic control strategies. Materials and methods: The SAVOR-TIMI 53 trial evaluated the safety and efficacy of saxagliptin vs. placebo in 16,492 patients with T2DM, at high risk of cardiovascular disease, of whom patients with HF at baseline were a pre- specified sub-group. The 1° endpoint was CV death, MI, or ischemic stroke. The 2° endpoint included the 1° endpoint together with hospitalization for HF, unstable angina, or coronary revascularization. Median follow-up was 2.1 years. The 1° and 2° endpoints were adjudicated by a blinded events com- mittee. Results: 2,105 patients (12.8%) reported prior HF at baseline. Patients with prior HF were older and more likely to have established CV disease and renal impairment, with no difference in duration of DM or HbA1c. More patients with prior HF were on ASA, statins, B-blockers, and insulin, but fewer on metformin or sulfonylurea. There was no difference in the duration of DM and only a small difference (0.1%) in baseline HbA1c. Patients with prior HF were at double the risk for all CV events compared to patients with no prior HF. (Table) The relative effect of saxagliptin versus placebo was similar in patients with and without prior HF for the 1° and 2° endpoints. The increased risk of hospitalization for heart failure with saxagliptin was also similar re- gardless of prior HF, though the absolute risk difference with saxagliptin was greater in patients with prior HF. The incidence of hypoglycaemia, adverse events, and A1c reductions are presented in the Table. Conclusion: Despite being at substantially increased risk of CV complica- tions, saxagliptin neither increased nor decreased the risk of the 1° and 2° endpoints compared to placebo in patients with T2DM and prior HF. Simi- larly, there was no relative differential in the increased risk of hospitalization for HF with saxagliptin, though the absolute risk was greater. These data pro- vide important information regarding the safety and efficacy of an antihyper- glycemic agent in a traditionally difficult to treat population. Clinical Trial Registration Number: NCT01107886 Supported by: AstraZeneca and Bristol-Myers Squib OP 32 Complications: expect the unexpected 187 A randomised clinical trial on the efficacy of an adapted bowel preparation for diabetic patients undergoing a colonoscopy J.A. Flores-Le Roux1, M.A. Alvarez2, J.J. Chillarón1, A. Seoane2, D. Benaiges1, L. Carot2, S. Bacchiddu2, J. Pedro-Botet1; 1Endocrinology, 2Gastroenterology, Hospital del Mar, Barcelona, Spain. Background and aims: Need for colonoscopy is common among diabetic patients because of the high prevalence of gastrointestinal symptoms but they are more likely to have inadequately cleansed bowels which leads to a re- peat procedure that will increase patients’ exposure to unnecessary risk and costs. No studies have evaluated the importance of dietary recommendations before the procedure or the need for hypoglycaemic drug adjustments dur- ing preparation. A specific bowel preparation for patients with diabetes with special attention to dietary recommendations and modification of hypoglyce- mic medication could facilitate patient adherence and consequently improve bowel cleansing. We evaluated the efficacy and safety of an adapted protocol for colon cleansing in diabetic patients. Materials and methods: Single-blind experimental design, 80 persons with diabetes who were scheduled for colonoscopies were randomly assigned to either the experimental diabetic colon preparation (DMe) or the standard co- lon preparation (DMs). A group of 40 non-diabetic subjects was also includ- ed (noDM). DMe patients were instructed to follow a specific low-residues diet, with special attention to carbohydrate intake, without liquid diet the day before colonoscopy, and to modify the doses of hypoglycaemic medication prior to the procedure. DMs subjects received the standard recommenda- tions including general information on low-residues diet and liquid diet the day before the procedure with no counselling on hypoglycaemic medication adjustments. All patients received the same medication for bowel prepara- tion. Endoscopists blinded to the type of preparation scored the type of re- sidual stool and the percentage of bowel wall visualized for each segment of colon and for the overall examination using the Boston score. Hypoglycemic episodes during preparation were recorded. Results: 73 diabetic patients were randomized: 37 in DMe group and 36 in the DMs group. 40 noDM subjects were also included. NoDM patients were younger (56±15 vs 68±10 years, p S 87 1 C 188 Dissociation pattern in resting-state default mode network connectivity in type 2 diabetes patients Y. Cui, Y. Jiao, S. Ju, G.-J. Teng; Radiology, Southeast University, Nanjing, China. Background and aims: Type 2 diabetes mellitus (T2DM) has been shown to be associated with an increased risk of cognitive impairment. Patients with impaired cognition often show default-mode network (DMN) disruption. This study aimed to investigate the integrity of the DMN by using independ- ent component analysis (ICA) methods in patients with T2DM and to cor- relate the DMN functional connectivity (FC) changes with neurocognitive performance and clinical variables. Materials and methods: The current study was approved by the local ethics committee and written informed consent was obtained from all participants. Twenty-nine T2DM patients and thirty well-matched healthy controls were included in the study and underwent resting-state functional MRI (rs-fMRI) in a 3 Tesla unit. All participants underwent a detailed battery of neuropsy- chological tests. Clinical parameters such plasma glucose, HbA1c, insulin re- sistance, BMI and cholesterol levels were also collected. A group ICA method was used to extract the DMN of all participants. Two components were identified to be related to two sub-networks of the DMN, including the ante- rior and posterior parts of the DMN. Z-maps of the two sub-networks were compared between the two groups and correlated with the neurocognitive performance and clinical parameters by using Pearson correlation analysis. Results: Patients with T2DM showed significantly increased frontal con- nectivity around the medial prefrontal cortex (MPFC) in the anterior DMN (aDMN) and significantly decreased connectivity in the posterior cingulate cortex (PCC) and angular gyrus in the posterior DMN (pDMN) (Figure 1). The FC strength in aMDN was found to be negatively correlated with the score on complex-figure-test (r=-0.46) and positively correlated with the FPG (r=0.44). On the other hand, the FC strength in pDMN was negatively corre- lated with the disease duration (r=-0.55) and the time spent on trail-making- test (part B) (r=-0.68). These associations were independent of vascular risk factors and cerebral small vessel disease. Conclusion: The current study demonstrated the dissociation between an- terior and posterior DMN sub-networks in patients with T2DM. Our results highlight the important role of the DMN in the pathophysiology of T2DM- related cognitive impairment and suggest that abnormal DMN activity may be a trait associated with T2DM patients. Supported by: NSFC 81271739 189 Disrupted circadian arousal patterns comorbid with diabetes M. Kadono1, G. Hasegawa2, M. Fukui3, N. Nakamura3; 1Internal Medicine, Ayabe City Hospital, 2Endcrinology and Metabolism, Kyoto Second Redcross Hospital, 3Endcrinology and Metabolism, Kyoto Prefectural University of Medicine, Japan. Background and aims: Sleep-wake disturbances and disruptions of circadian rhythms are now recognized to be highly prevalent comorbidities in patients with medical illness. Likewise, it is supposed that circadian arousal rhythms were disrupted in parallel with metabolic and vascular vulnerability in dia- betes. We aimed to explore disrupted circadian arousal patterns comorbid with diabetes. Materials and methods: Ninety-one diabetic outpatients with BMI S 88 1 C events per hour during the total time of sleep, p=0.023]. A significant posi- tive correlation between microarousals and the AHI was detected (r=0.807, p S 89 1 C OP 33 Device utilisation and outcomes 193 Continuous intraperitoneal insulin infusion versus subcutaneous insulin for type 1 diabetes: a prospective, case-control trial proving non- inferiority H.J.G. Bilo1, P.R. van Dijk1, S.J.J. Logtenberg1, K.H. Groenier1, H. Feenstra2, R.O.B. Gans3, N. Kleefstra1; 1Diabetes centre, Isala clincs, Zwolle, 2Internal Medicine, Diaconessenhuis, Meppel, 3Internal Medicine, University Medical Center Groningen, Netherlands. Background and aims: Continuous intraperitoneal insulin infusion (CIPII) using an implantable pump is a last-resort treatment option for patients with type 1 diabetes mellitus (T1DM) who fail to reach acceptable glycaemic con- trol with intensified subcutaneous (SC) insulin regimens. Aim of this study was to compare the effects of long-term CIPII with SC insulin therapy in T1DM. Materials and methods: This is a 36 week, investigator initiated, prospec- tive, open-label matched-control study. Patients were eligible if they had been treated with either CIPII or SC insulin for >4 years and had a HbA1c of ≥53 mmol/mol. CIPII treated patients were matched to SC treated patients regarding age and gender. In order to account for inequality between treat- ment groups, as CIPII treated patients are a selected group of patients in need of a last-resort treatment, the primary endpoint was not a superiority but a non-inferiority assessment of the difference in HbA1c between both groups. A non-inferiority margin of -5.5 mmol/mol was predefined. Secondary out- comes were clinical and biochemical parameters, quality of life (SF-36 (scores 0-100)) and treatment satisfaction (DTSQ (scores 0-36)). Analysis were per- formed with ANCOVA, taking baseline differences into account. Results: During study, one patient withdrew consent. Subsequently 183 pa- tients, 36% male with a mean age of 50 years (SD 1) and diabetes duration of 26 years (SD 13), of which 39 were treated with CIPII and 144 with SC insulin therapy were analyzed. Age and gender were well matched. Results are presented in table 1. HbA1c remained stable within the CIPII group while it decreased with -1.0 mmol/mol (95%CI -1.9, -0.1) in the SC group. Using ANCOVA, the difference between treatment groups was -3.0 mmol/mol (95%CI -5.0, -1.0) and met the predefined non-inferiority criterion. Besides a difference in alanine aminotransferase (ALT) concentrations between groups of 3.6 U/L (95%CI 1.2, 6.0) being higher in the CIPII group, no other differ- ences were found. At baseline and at the end of the trial, SF-36 scores were lower among CIPII treated patients as compared patients treated with SC in- sulin. CIPII treated patients had lower SF-36 mental and physical component scores and higher treatment satisfaction as compared to patients treated with SC insulin at baseline and follow-up. Conclusion: CIPII therapy is non-inferior to SC insulin therapy with respect to glycaemic control in the treatment of poorly controlled T1DM patients. Besides a lower ALT among SC treated patients, there are no differences in clinical and biochemical parameters. Despite a lower quality of life, treatment satisfaction is higher among CIPII treated patients. This study underlines the effectiveness of long-term CIPII therapy as last resort treatment in T1DM. Clinical Trial Registration Number: NCT01621308 194 Three to four weeks of overnight closed loop insulin delivery during free living: analysis of randomised crossover studies in adults and adolescents with type 1 diabetes H. Thabit1, L. Leelarathna1, D. Elleri1, J.M. Allen1, A. Lubina-Solomon2, M. Stadler3, E. Walkinshaw2, A. Iqbal2, P. Choudhary3, M.E. Wilinska1, S.R. Heller2, S.A. Amiel3, M.L. Evans1, D.B. Dunger1, R. Hovorka1; 1University of Cambridge, 2Academic Unit of Diabetes, Endocrinology and Metabolism, Department of Human Metabolism, University of Sheffield, 3Diabetes Research Group, Kings College London, UK. Background and aims: Results from single-night laboratory-based closed loop studies demonstrated improved overnight glucose control and reduced risk of hypoglycaemia in subjects with type 1 diabetes. Assessments outside hospital settings over longer periods in free living conditions are underway. We combined data collected during free-living unsupervised randomised open-label crossover studies comparing overnight closed loop and sensor augmented pump therapy. Materials and methods: 40 participants with type 1 diabetes [24 adults re- cruited at three centres and 16 adolescents recruited at one centre (age 43±12 years, HbA1c 8.0±0.9%; mean±SD) (15.6±3.6 years, HbA1c 8.1±0.8%)] un- derwent training on study devices followed by two periods of sensor aug- mented pump therapy in combination with or without overnight closed-loop utilising a model predictive control algorithm to direct insulin delivery. The order of interventions was random; each period lasted four weeks in adults and three weeks in adolescents. Primary outcome was time when sensor glu- cose was in the target range between 3.9 and 8.0mmol/l. All analyses were by intention to treat. Results: Closed loop was started by participants on their own volition on 866 nights (89%). The proportion of time when sensor glucose was in target range (3.9-8.0mmol/l) between midnight and 08:00 was increased by 18.5% during closed-loop compared to sensor augmented therapy (P S 90 1 C 195 Impact of real-time continuous glucose monitoring system usage on endothelial function in adolescents with type 1 diabetes B. Głowińska-Olszewska, M. Jamiołkowksa, I. Jamiołkowska, W. Łuczyński, J. Tołwińska, A. Bossowski; Department of Pediatrics, Endocrinology, Diabetology with Cardiology Division, Medical University of Białystok, Poland. Background and aims: Chronic hyperglycaemia and recently defined “glu- cose variability” in the course of diabetes are established risk factors for car- diovascular disease. Children with type 1 diabetes (T1DM) are the high - risk group of accelerated early atherosclerosis development. Real-time continu- ous glucose monitoring system (RT-CGMS) provides with new possibilities for the detection of excessive glycaemic fluctuations and enables the patient to react quickly to decrease these adverse phenomena. The aim of the study was to assess the usefulness of RT-CGMS as an educational tool to find and reduce glycemic variability in order to improve the parameters of endothelial function in adolescents with T1DM. Materials and methods: Forty T1DM patients aged mean 14.6 years, diabe- tes duration: 7.4 years, mean HbA1c before the study: 9.35%, 19 boys and 21 girls were recruited. The study was based on one month continuous glucose sensors use combined with education of the patients and caregivers. Several parameters of glycemic variability (mean glucose, SD for the mean glucose, AUC>140 mg/dl, AUC S 91 1 C urinary concentrations of 8-iso-PGF2alfa , HbA1c, 24hr blood pressure, 3 days Continuous Glucose Monitoring sensor (CGM), self monitored blood glucose profiles. Results: AER decreased slightly in MDI group from a median value of 65 μg/ min (31-100 IQR) to 55 (33-155) after 3 years of follow-up. AER decreased significantly (p S 92 1 C OP 34 Novel approaches for beta cell protection 199 Specific overexpression of the calcium-sensor sorcin in pancreatic beta cells protects against ER-stress and diet-induced type 2 diabetes A.P. Marmugi, G.A. Rutter, I. Leclerc; Medicine, Imperial College London, UK. Background and aims: Pancreatic beta cells use electrical signals to couple changes in blood glucose concentration to insulin release via extracellular calcium (Ca2+) influx. Sorcin (SRI) is a Ca2+-binding protein whose overex- pression in cardiomyocytes rescues the abnormal contractile function of the diabetic heart and plays a role in terminating Ca2+-induced Ca2+-release. We explored here the role of sorcin in regulating Ca2+ fluxes in pancreatic beta cells and its ability to protect against glucolipotoxicity in vivo. Materials and methods: Two lines of transgenic mice were generated on a C56Bl/6 background permitting inducible overexpression of SRI cDNA with the TetOn-system specifically in beta cells. Animals bearing one (SRI-1) or ten (SRI-10) copies of the SRI transgene, and littermate controls (CTRL), were fed either a standard chow diet (SD) or a high fat diet (60% fat, HFD) and exposed to doxycycline in the drinking water (500mg/L) from 4 weeks onwards. Body weight measurements and fasting intraperitoneal glucose tol- erance tests (1g/kg, IPGTT) were performed at regular intervals. Quantita- tive RT-PCR and real-time calcium imaging with Fura-Red (excitation light at 420 and 480nm, 40x/1.4NA objective and a cMOS camera on board an Olympus IX-71 microscope) were performed on islets isolated from either transgenic animals or from wild-type mice transduced with adenoviruses overexpressing GFP-SRI or GFP. Results: Glucose tolerance during IPGTTs worsened considerably in CTRL males under HFD (120 min glycaemias (mmol/L): 12.3 ± 0.9 at 8-week-old, 22.1 ± 1.7 at 16-week-old, n=17) but not in CTRL males under SD (120 min glycaemia (mmol/L): 7.2 ± 0.2 at 8 weeks, 7.3 ± 0.5 at 16 weeks, n=10). Strik- ingly, and despite having similar body weights, male SRI-1 mice on HFD dis- played improved glucose tolerance at 16 weeks of age (areas under the curve (AUC; a.u.): 111.9 ± 8.8 for SRI-1, 136.2 ± 7.3 for CTRL, n=16, p S 93 1 C Results: The NPs were equilibrated in a dose-dependent manner in INS1 and beta-cells within 1-2 hours and localized to lysosomes as assessed by co- localization with LysoTracker staining. High concentration (250 µg/mL) of NPs induced significant cell death when UV-activated but was well-tolerated when not UV-activated, indicating successful activation of NPs in intact cells and allowing determination of a non-toxic treatment dose of 25 µg/mL NPs in cells. UV-activation of NPs in INS1 cells exposed for 18 hours to palmitate showed a significant rescue of lysosomal acidity and size compared to palmi- tate treatment alone or palmitate co-treated with NPs that were never UV-ac- tivated. Furthermore, palmitate-induced p62 and LC3-II accumulation were decreased following UV-activation of NPs, indicating downstream improve- ment of autophagic flux. Finally, UV-activated NPs were able to partially re- store fatty acid-induced decrease in glucose-stimulated insulin secretion in human islets, suggesting that rescuing lysosomal acidity and autophagic flux with acidic NPs can improve beta-cell function. Conclusion: The capacity of acidic NPs to restore lysosomal acidity, mor- phology and clearance supports an upstream role for impaired lysosomal dysfunction in the development of deregulated insulin secretion. Addition- ally, acidic NPs may have therapeutic value in diseases where a cellular defect in lysosomal acidity inhibits autophagic flux, such as the fatty acid-induced arrest of autophagic flux in the pancreatic beta-cell. 202 Reversible changes in pancreatic islet structure and function produced by elevated blood glucose M.F. Brereton1, M. Iberl1, K. Shimomura1, Q. Zhang2, P. Proks1, A.A. Adriaenssens3, I.I. Spiliotis1, W. Dace1, K.K. Mattis1, F.M. Gribble3, F. Reimann3, A. Clark2, P. Rorsman2, F.M. Ashcroft1; 1Department of Physiology, Anatomy and Genetics, University of Oxford, 2Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, 3Cambridge Institute For Medical Research, University of Cambridge, UK. Background and aims: Hyperglycaemia is common to all forms of diabetes and results from impaired insulin secretion and aberrant glucagon secretion due to changes in pancreatic islet cell function and/or mass. The extent to which hyperglycaemia per se underlies these alterations remains poorly un- derstood as previous studies have been restricted to in vitro culture systems, or animal models where diabetes is artificially induced by β-cell ablation and other confounding factors (e.g. obesity and insulin resistance) are common. We generated a mouse model in which diabetes could be rapidly and revers- ibly switched on and off to study the effect of hyperglycaemia on pancreatic islet structure and function. Materials and methods: An inducible human activating KATP channel muta- tion was expressed in β-cells of adult mice. This inhibited insulin secretion and resulted in rapid onset of diabetes. Islet morphology, ultrastructure and electrophysiological characteristics were studied. Results: 4-week exposure to hyperglycaemia was associated with a significant reduction in insulin-positive cells (% of islet area composed of insulin; con- trol: 86.5+/-0.3% vs. diabetic 27.8+/-1.7%). This reflected a significant reduc- tion in insulin mRNA and insulin protein levels in islets (P S 94 1 C interferon-gamma (IFN) modifies AS of >500 genes. Our recently published data indicate the presence of the Nova1 splicing factor in beta cells, but there is no information about its role on these cells. We have presently coupled specific knockdown (KD) of Nova1 with RNA sequencing to determine all splice variants and downstream pathways regulated by this protein in pan- creatic beta cells. Materials and methods: Nova1 expression was inhibited by 60% with the use of specific siRNAs. Three FACS-purified rat beta cell preparations (90- 95% pure) were RNA-sequenced under control conditions or after a 48-hour KD of Nova1. Samples were sequenced on an Illumina HiSeq 2000 sequencer and the data analysed using Tophat mapper and Flux Capacitor. The Cuf- flinks software suite was used to identify potentially novel transcripts. Ex- pression was considered changed if it fulfilled two criteria, namely p S 95 1 C also had a HbA1c target ≤6.5% and 40.7% of those with a loose SBP target also had loose HbA1c targets. At the 6 months follow-up the mean (±SD) SBP was 133.4±13.5 mmHg in the strict group (25% exceeding 140 mmHg), 135.7±13.3 mmHg in the medium group (25% exceeding 140 mmHg)and 139.0±15.0 in the loose group (25% exceeding 147 mmHg) (p S 96 1 C 209 Relationships between the risk of cardiovascular events in type 2 diabetes and both visit-to-visit variability and time-to-effect in systolic blood pressure T. Takao1, Y. Matsuyama2, H. Yanagisawa3, M. Kikuchi1, S. Kawazu1; 1Division of Diabetes and Metabolism, The Institute for Adult Diseases, Asahi Life Foundation, Tokyo, 2Department of Biostatistics, School of Public Health, University of Tokyo, 3Department of Public Health and Environmental Medicine, The Jikei University School of Medicine, Tokyo, Japan. Background and aims: Recent studies, mostly in treated hypertensive pa- tients, found a visit-to-visit BP variability is a risk factor for cardiovascular diseases (CVD) and organ damage independently of the mean BP. However, reports in patients with diabetes remain limited. Lack of a BP legacy effect on cardiovascular events has been reported in type 2 diabetes. However, data on the effects of early lowering of BP are still scarce. This study aimed to deter- mine whether systolic BP (SBP) variability is able to predict, independently of the mean SBP, the incidence of CVD events in patients with type 2 diabetes, and to analyze the time-to-effect relationship between SBP and the risk of these events. Materials and methods: A total of 652 (538 men, 114 women) patients with type 2 diabetes who first visited our hospital between 1995 and 1996, with at least 1 hospital visit per year and no history of CVD, and undergone 4 or more SBP determinations, were retrospectively enrolled. Patients were fol- lowed through June 2012. SD or CV was used as a measure of SBP variability. Risk of CVD events was evaluated by multivariate Cox proportional hazard models. SBP was analyzed as a mean value and a time-dependent covariate using the last observation carried forward or moving-mean during 1 year to 17 years preceding the events. For all analyses, SAS (version 9.3) was used. Results: The mean values at baseline were age 55.7 years, duration of dia- betes 5.6 years, BMI 23.4 kg/m2, BP 133.7/77.9 mmHg, HbA1c 8.0%, Total- cholesterol (TC) 209.3 mg/dl, and HDL-cholesterol (HDLC) 49.8 mg/dl. 275 patients were current smokers. The median follow-up period from first visit was 11.5 years, and the total number of SBP measurements was 53,949 (per- patient median: 78). By the end of follow-up, CVD events occurred in 71 patients. Hazard ratios for the incidence of CVD unadjusted, adjusted for age and sex, and for age, sex, duration of diabetes, current smoker, mean SBP, mean HbA1c, mean TC/HDLC, mean BMI, and the number of SBP meas- urements (ln-transformed), increased across tertiles of SD and CV of SBP (Table). The incidence of CVD was significantly associated with SBP during the preceding 2 to 8 years, with the highest significance during the preceding 3 years, and borderline significance during the preceding 1 and the preceding 9 to over 17 years. Conclusion: Visit-to-visit SBP variability was able to independently predict CVD events in patients with type 2 diabetes. Increased SBP during the pre- ceding 3 years resulted in the highest risk of CVD; therefore, to prevent CVD, SBP management should focus on stable and well-timed control. 210 Re-evaluation of patients with type 2 diabetes from the RIACE cohort using the 8th Joint National Committee cut-offs for blood pressure A. Solini1, G. Zoppini2, C. Fondelli3, E. Orsi4, M. Arosio5, R. Trevisan6, M. Vedovato7, F. Cavalot8, O. Lamacchia9, R. Buzzetti10, G. Penno1, G. Pugliese11; 1Department of Clinical and Experimental Medicine, University of Pisa, 2Division of Endocrinology and Metabolic Diseases, University of Verona, 3Diabetes Unit, University of Siena, 4Diabetes Unit, Fondazione IRCCS Cà Granda - Ospedale Maggiore Policlinico, Milan, 5Endocrinology Unit, San Giuseppe Hospital, Milan, 6Endocrinology and Diabetes Unit, Ospedali Riuniti, Bergamo, 7Department of Clinical and Experimental Medicine, University of Padua, 8Department of Clinical and Biological Sciences, University of Turin, Orbassano, 9Department of Medical Sciences, University of Foggia, 10Department Experimental Medicine, La Sapienza University, Rome, 11Department of Clinical and Molecular Medicine, La Sapienza University, Rome, Italy. Background and aims: The 8th Joint National Committee (JNC8) report has raised the recommended blood pressure (BP) threshold for drug therapy in diabetic subjects from 130/80 (JNC7) to 140/90 mmHg. This study was aimed at re-evaluating with the new BP cut-offs the prevalence of hypertension, an- ti-hypertensive treatment, and achievement of target BP in patients with type 2 diabetes from the Renal Insufficiency And Cardiovascular Events (RIACE) Italian Multicenter Study. Materials and methods: The RIACE cohort consists of 15,773 patients, con- secutively visiting 19 Diabetes Clinics throughout Italy in years 2007-2008. Exclusion criteria were dialysis or renal transplantation. BP was measured with a mercury sphygmomanometer on the right arm after at least 5 min of rest. Major acute cardiovascular disease (CVD) events were adjudicated based on hospital discharge records. Results: Using the JNC7 BP targets, 6,854 patients (43.5%) had systolic BP S 97 1 C OP 36 Molecular mechanisms of insulin action in vitro 211 The regulation of Wnt signalling pathways by insulin and NEFAs in skeletal muscle and adipose tissue of healthy humans M. Karczewska-Kupczewska1,2, A. Nikolajuk2, M. Stefanowicz1, N. Matulewicz1, M. Straczkowski1,2; 1Department of Metabolic Diseases, Medical University of Bialystok, 2Department of Prophylaxis of Metabolic Diseases, Institute of Animal Reproduction and Food Research, Polish Academy of Sciences, Olsztyn, Poland. Background and aims: The Wnt signaling plays an important role in em- bryonic development, cell proliferation, cell cycle, differentiation, apoptosis and tissue homeostasis. The Wnt ligands bind to frizzled (Fzd) receptors and activate canonical (β-catenin dependent) and non-canonical (β-catenin in- dependent) pathways. Dysregulation of Wnt signal transduction is associated with many pathophysiological states including metabolic disorders. The aim of the present study was to assess skeletal muscle and adipose tissue expres- sion of genes associated with Wnt signaling pathways in young healthy popu- lation, their regulation by hyperinsulinemia and serum NEFAs elevation and their relationship with insulin sensitivity. Materials and methods: We studied 20 healthy male subjects (mean age 25.20±3.15 years, mean BMI 26.47±4.64 kg/m2). The biopsies of vastus lat- eralis muscle and subcutaneous adipose tissue were performed at baseline and after 6 hours of euglycemic hyperinsulinemic clamp with or without In- tralipid/heparin infusion. The participants were divided into subgroups of high (high-IS) and low insulin sensitivity (low-IS). Results: The muscle mRNA expression of β-catenin, Fzd coreceptor-LDL receptor-related protein 6 (Lrp6), Dishevelled 2 (Dsh2) and adipose tissue mRNA expression of Lrp6 were higher in low-IS, whereas adipose tissue mR- NAof Dsh2 and glycogen synthase kinase 3β (GSK3β) were lower in low-IS (all p S 98 1 C 214 GLP-1 improves the reduction of glucose uptake through SIRT-1 in skeletal muscle cells under palmitate induced-insulin resistance T. Kim1, J. Jeon2, E. Ha2, S.-E. Choi3, S.-Y. Ahn2, S. Han2, H. Kim2, D. Kim2, Y. Kang3, K.-W. Lee2; 1Internal Medicine, Kwandong University, Gangneung, 2Endocrinology and Metabolism, 3Institute of Medical Science, Ajou University School of Medicine, Suwon, Republic of Korea Background and aims: Insulin resistance is a major pathogenesis of type 2 diabetes mellitus. Recently, the target molecules for diabetes treatment have been identified. Glucagon like peptide-1 (GLP-1) was developed as an anti- diabetic agent and is known to have a potent glucose-dependent insulino- tropic action on the pancreas as well as extrapancreatic actions. SIRT1, as an NAD-dependent deacetylase, has been demonstrated in the role of protecting ageing-related diseases and beneficial effects of metabolic homeostasis. We investigated whether the actions of GLP-1 were mediated by SIRT-1 activa- tion in skeletal muscle cells under palmitate induced-insulin resistance. Materials and methods: After confirming the action of GLP-1 to palmitate- induced insulin resistance, we determined 2-[N-(7-nitrobenz-2-oxa-1,3- diazol-4-yl)amino]-2-deoxy-D-glucose (2-NDBG) uptake, GLUT4 mRNA, and the phosphorylation levels of protein related to insulin signal pathway (IRS-1, Akt) in human skeletal muscle myotubes (HSMM) exposed to pal- mitate (200 μM, for 24 hours) and compared to those in HSMM exposed to palmitate and GLP-1 (200 nM) simultaneously. To elucidate whether SIRT- 1 contributed to GLP-1 action in the palmitate-induced insulin resistance, we compared the levels of proliferator-activated receptor-γ-co-activator 1α (PGC1α) deacetylation in HSMM exposed to palmitate and GLP-1 (200 ng/ ml) for 24 hours. Moreover, when exposed to SIRT1 inhibitor (EX527) in the same conditions, we demonstrated the changes of GLUT4 mRNA expression and insulin signaling. Results: GLP-1 restored the reduction of glucose uptake, GLUT4 mRNA levels, and GLUT4 promoter activity by palmitate in human skeletal muscle myotubes. GLP-1 reduced acetylation of PGC1α in HSMMs over expressed PGC1α. This result suggested that GLP-1 activated a deacetylase, such as SIRT-1. SIRT1 inhibitor (EX527) suppressed GLUT4 mRNA expression by GLP-1 in HSMMs. SIRT1 inhibitor (EX527) prevented phosphoryla- tion (IRS-1, Akt) of protein related to insulin signal pathways by GLP-1 in HSMMs. This suggested the expression of GLUT4 mRNA and the activation of insulin signaling by GLP-1 were associated with SIRT1. Conclusion: Our data suggests GLP-1 improved palmitate-induced insulin resistance in HSMM by the activation of SIRT-1. 215 Overexpression of the glucose transporter 1 in renal mesangial cells protects against cellular stress via Nrf2/NQO1 T. Fleming1, E. Schleicher2, H.-U. Häring2, P.P. Nawroth1; 1Internal Medicine I and Clinical Chemistry, University of Heidelberg, Heidelberg, Germany, 2Department of Internal Medicine, University of Tübingen, Germany. Background and aims: Diabetic patients have been shown to be susceptible to the formation and accumulation of triose phosphates and the reactive me- tabolite methylglyoxal (MG), formed by non-enzymatic degradation on these intermediates. These findings suggest that it is not the condition of hyper- glycaemia, but how glucose is handled, which underlies the development of diabetic complications. In this project, this was addressed by characterizing rat mesangial cells which stably overexpress the glucose transporter-1 pro- tein, the rate-limiting step for the movement of glucose into the cell. In such cells, glucose uptake and utilization is increased regardless of other regulator factors. Materials and methods: The phenotype of rat mesangial cells over-express- ing glucose transporter-1 (GLUT1) and the respective control cells, stably expressing lac operon (LacZ) were characterized with respect to growth and energy metabolism under basal condition (11mM glucose). Results: It was found that despite a significantly higher glucose uptake, rat mesangial cells over-expressing glucose transporter-1 (GLUT1) had a re- duced glycolytic rate. They also had reduced levels of metabolic cofactors and intermediates. They were also found to have significantly lower number of hyper-polarized mitochondria, suggestive of altered respiratory function. This was confirmed by the decreased rate of oxygen consumption. Mitochon- drial superoxide production was shown to be decreased in the GLUT1 cells; however, generally oxidative stress within these cells was increased, sugges- tive of increased cellular stress as a consequence of increased glucose uptake. Preliminary screening of antioxidant defence enzymes showed that NAD(P) H dehydrogenase(quinone) 1 (NQO1), was significantly increased in terms of expression and activity as well as its regulator, Nrf2. The observed increase in antioxidant defences was confirmed by the increase in total antioxidant capacity, and increased tolerance to hydrogen peroxide and MG. Conclusion: This data indicates that under conditions of high glucose uptake, antioxidant defences are increased to help protect the cell from the increased cellular stress The observed increases in defences system whilst protective are ultimately detrimental to the cell. Supported by: DFG (BI-1281/3-1 & NA 138 /7-1), DZD/BMBF, Dietmar- Hopp-Stiftung 216 Extracellular vesicles released by hypoxic adipocytes impair insulin signalling and glucose uptake in adipocytes S. Mora1, J. Mleczko1, J. Falcon-Perez2; 1Cellular and Molecular Physiology, University of Liverpool, UK, 2CIC BioGUNE, Bilbao, Spain. Background and aims: Exosomes and microvesicles are extracellular vesicles (EVs) shed by many cell types that mediate cell to cell communication. EVs contain proteins and RNA species that modulate biochemical responses in target cells. Adipocytes have been shown to produce extracellular vesicles. During obesity adipocytes are subject to hypoxia. The aim of our study was to investigate whether EVs released by hypoxic adipocytes have autocrine ef- fects in regulating insulin action and glucose uptake. Materials and methods: We isolated EVs from the conditioned media of differentiated 3T3L1 adipocytes cultured under normoxia (control cells) or from cells exposed to hypoxia (1% O2) for 24hrs. EVs were purified from the culturing media using centrifugation techniques and characterized using bio- physical and biochemical methods, including nanoparticle tracking analysis, cryoelectron microscopy, and immunoblotting techniques. EVs were then tested for their ability to modulate insulin action and insulin-stimulated glu- cose uptake in 3T3L1 adipocytes. Independent sets of differentiated 3T3L1 adipocytes were either left untreated or treated with EVs purified from con- trol or hypoxic adipocytes for 24hr and subsequently their response to insulin stimulation was evaluated by western blotting and glucose uptake assays. Results: We found that 3T3L1 adipocytes release EVs of heterogeneous sizes as determined by cryoelectron microscopy and nanoparticle tracking anal- ysis. Hypoxia increased the release of EVs by adipocytes. Immunoblotting analysis showed that EV preparations were enriched in exosomal markers and markedly devoid of other organelle markers. Adipocytes treated with EVs obtained from hypoxic cells displayed a reduced insulin-mediated acti- vation of glucose uptake compared to those left untreated or treated with EVs derived from control adipocytes. No differences in the expression levels of the glucose transporter Glut4 were observed in cells treated with EVs from hy- poxic adipocytes compared to those left untreated or treated with EVs from control cells. However, cells treated with EVs released from hypoxic cells ex- hibited reduced insulin-mediated activation of phosphatidylinositol 3-kinase as seen by a reduced phosphorylation of the downstream kinase AKT, with- out alterations in the total AKT levels. Conclusion: Our data suggest that EVs released by hypoxic adipocytes con- tribute to insulin resistance, as they impair insulin signalling and glucose up- take in adipose cells. Diabetologia (2014) 57:[Suppl1]S1–S564 S 99 1 C OP 37 Metformin: new insights into an old drug 217 Mechanism of increase in plasma intact GLP-1 by metformin in type 2 diabetes: stimulation of GLP-1 secretion or reduction in plasma DPP-4 activity? T. Wu1, M.J. Bound1, C.F. Deacon2, K.L. Jones1, M. Horowitz1, C.K. Rayner1; 1Discipline of Medicine, University of Adelaide, Australia, 2Department of Biomedical Science, University of Copenhagen, Denmark. Background and aims: Metformin increases plasma intact glucagon-like peptide-1 (GLP-1) concentrations. However, it remains uncertain whether this reflects an enhancement of GLP-1 secretion and/or a reduction in plasma dipeptidyl peptidase-4 (DPP-4) activity. We, therefore, evaluated the effects of metformin on plasma DPP-4 activity, and total and intact GLP-1 concen- trations, before and during an intraduodenal (ID) glucose infusion in type 2 diabetes. Materials and methods: We retrospectively assayed plasma DPP-4 activity in 12 Caucasian males with diet-controlled type 2 diabetes (HbA1c: 6.5 ± 0.1% (47.6 ± 1.4 mmol/mol); duration of known diabetes: 3.5 ± 0.9 years; age: 63.7 ± 1.9 years; BMI: 29.9 ± 1.2 kg/m2), treated with metformin 850 mg bd or pla- cebo each for 7 days in a randomised crossover design, with 14 days ‘washout’ between. On days 5 or 8 of each treatment (6 patients each), subjects took their morning dose of metformin or placebo after an overnight fast, followed 30 min later by an ID infusion of glucose (60 g over 120 min, ie. 2 kcal/min). Results: Plasma fasting DPP-4 activity was less during metformin treatment than placebo (18.4 ± 1.6 vs. 20.5 ± 1.7 U/L, P = 0.014). During ID glucose infusion (t = 0-120 min), there was a slight reduction in plasma DPP-4 activ- ity on both days (P < 0.001), without any difference between metformin and placebo. Both total and intact GLP-1 concentrations were greater after met- formin than placebo (P = 0.014 and 0.006, respectively). During ID glucose infusion, the increment in intact GLP-1 (ie. difference in the areas under the curve (AUC)) after metformin versus placebo was directly related to the AUC for total GLP-1 (r = 0.61, P = 0.028), and tended to be inversely related to plasma DPP-4 activity (r = -0.48, P = 0.098). Conclusion: The increase in plasma intact GLP-1 concentrations by met- formin is, at least in part, attributable to stimulation of GLP-1 secretion - a reduction in soluble DPP-4 activity may also make a modest contribution. Clinical Trial Registration Number: ACTRN12608000433303 Supported by: MSD funded investigator-initiated study 218 Intestinal glucose uptake is modulated by metformin J.B.L. Hoekstra, L. Bähler, K. Stroek, H. Verberne, F. Holleman; Academic Medical Center, Amsterdam, Netherlands. Background and aims: Obesity is the result of an imbalance between en- ergy intake and energy expenditure. Recently brown adipose tissue (BAT) has gained a lot of interest because of its capacity to convert calories into heat. BAT shows an intense 18F-Fluorodeoxyglucose-(FDG)-uptake on Positron- Emission-Tomography-(PET)-CT scans. Physiological 18F-FDG accumula- tion is also frequently observed in the intestine, however, the possibility of the intestine as an energy dissipating tissue has not yet been explored. Metformin is one of the few drugs in the treatment of diabetes mellitus(DM) that is as- sociated with moderate weight loss. Interestingly, patients using metformin are known to have increased 18F-FDG-uptake in the colon. The aim of this study was to retrospectively assess determinants of 18F-FDG-uptake in the intestine in patients who underwent a primary diagnostic 18F-FDG-PET-CT scan. Materials and methods: Consecutive 18F-FDG-PET-CT scans performed between January 1st and April 31st 2011 were analysed. 18F-FDG intestinal uptake was visually assessed using the 4-point scale described by Gontier et al. which uses liver uptake as a reference. (1; lower, 2; similar, 3; moderately higher than hepatic activity and 4; intense and diffuse uptake). Differences in continuous variables between the 4 grades were assessed using the Kruskal- Wallis test. Differences in categorical variables between the 4 groups were assessed using the Fisher-Freeman-Halton exact test. All variables were used as covariate in a forward logistic regression model (e.g. grade 1 & 2 vs 3&4). Results: Available for analysis were 270 18F-FDG-PET-CT scans of 270 pa- tients. The majority of patients was female (51.9%) with a median age of 61 [52-71] years, a BMI of 25.0[22.5-28.7] kg/m2. Approximately 20% of the subjects smoked (smoking >1 daily). Most patients had a grade 2 (44%) or grade 3 (39%) uptake of 18F-FDG in the intestines, far less subjects had a grade 1 (9%) or grade 4 (8%) uptake of 18F-FDG. There were no significant differences between the four grading groups with regard to age, sex, BMI or smoking status. Type 1 DM was observed in 2 % of the subjects and 37% of the patients had type 2 DM. Of the DM patients, 87.2% used insulin, 87.2% used metformin and 82.1% used sulfonylurea derivatives (SU). There was a positive trend over the 4 grades for the use of insulin, SU and metformin but also for diuretic use. The use of insulin (p S 100 1 C flurodeoxyglucose (FDG) PET during euglycemic hyperinsulinemic clamp. Magnetic resonance imaging (MRI) was used as an anatomical reference. Results: Both active treatments improved glycemic control (p S 101 1 C 222 Occurrence of clinical remission in adult-onset type 1 diabetes is associated with decreased risk of microvascular complications P. Niedzwiecki, S. Pilacinski, A. Adamska, D. Zozulinska-Ziolkiewicz; Clinic of Internal Medicine and Diabetology, Poznan University of Medical Sciences, Poland. Background and aims: Partial clinical remission is common in the initial course of type 1 diabetes. In this phase of disease substantial insulin secretion contributes to good metabolic control and low incidence of acute complica- tions. The aim of the study was to determine the association between pres- ence of partial remission and occurrence of chronic complications of type 1 diabetes. Materials and methods: 240 consecutive patients (77 women, 143 men), aged 26 (IQR: 22-31) hospitalized with newly diagnosed type 1 diabetes in the Department of Internal Medicine and Diabetology in 2004-2007 were asked to participate. Of these, 220 were included in the study. 133 of patients completed prospective follow-up. Finally, data of 81 patients (24 women, 57 men), aged 33 (IQR 29-38) were included into endpoint analysis. Clinical remission was defined as time in which all of the following criteria were met: HbA1c below 6.5 % , dose of exogenous insulin below 0.3 U / kg body weight and serum C-peptide concentration above 0.5 ng / ml. Patients were divided into those who were in remission at any time during follow-up (remitters) and non-remitters. At 7 years of follow-up occurrence of chronic microvas- cular complications of diabetes (retinopathy, diabetic kidney disease and neuropathy) was evaluated. Results: In non-remitters group higher incidence of at least one microvascu- lar complication (46.4 vs. 7.6 %, p < 0.0001), higher incidence of retinopathy (42.8 vs. 5.7 %, p < 0.0001), and neuropathy (21.4 vs 1.9 %, p = 0.006) was found. In univariate logistic regression, significant association was found between absence of remission and occurrence of at least one microvascular complication (OR: 10.6, 95% CI : 2.94-38.22, p = 0.002). In the Cox propor- tional hazards regression model that included clinically significant param- eters at diagnosis (presence of ketoacidosis, cigarette smoking and HbA1c value) as covariates, absence of remission was associated with occurrence of chronic complications of diabetes at 7 years [HR: 3.65 (95% CI 1.23-4.56), p=0.04]. Conclusion: Occurrence of clinical remission of diabetes is associated with reduced risk of chronic microvascular complications at 7-year follow-up. Supported by: The Polish Diabetes Association 223 Proinsulin in early atherosclerosis: causal factor or bystander? R.J. Strawbridge1, A. Silveira1, P. Gustafsson1, K. Gertow1, D. Baldassarre2,3, F. Veglia3, S. Humphries4, U. de Faire5, E. Tremoli2,3, A. Hamsten1, the IMPROVE study group; 1Department of Medicine Solna, Karolinska Institutet, Stockholm, Sweden, 2Dipartimento di Scienze Farmacologiche e Biomolecolari, Universita di Milano, 3IRCCS Centro Cardiologico Monzino, Milan, Italy, 4Centre for Cardiovascular Genetics, University College London, UK, 5Department of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden. Background and aims: Increased proinsulin levels (relative to insulin levels) have been demonstrated to predict future incidence of both Type 2 Diabetes (T2D) and cardiovascular disease (CVD), independently of classical risk fac- tors, although the mechanisms of these relationships are poorly understood. A genome-wide association meta-analysis study conducted by the MAGIC consortium, and including this group, has identified a number of loci and genetic variants associated with increased circulating proinsulin levels. This project aims to determine whether proinsulin and/or proinsulin-associated pathways play a causal role in determining severity and rate of progression of early, subclinical atherosclerosis. Materials and methods: The IMPROVE study (consisting of 3711 subjects, of whom 917 are T2D subjects) without clinical signs of CVD at enrolment, who have been carefully phenotyped and monitored with repeated state- of-the-art high-resolution carotid ultrasound examinations. Biochemical phenotyping includes measurement (by Elisa) of circulating insulin and pro- insulin levels. All subjects have been genotyped on the Illumina CardioMe- tabo200K platform, providing good coverage of 5 of the 9 loci associated with proinsulin levels. Additional genotyping to cover the remaining 4 loci has been performed. Standard statistical methods and a Mendelian randomisa- tion approach will be used to examine the influence and causality of proinsu- lin on early atherosclerosis. Results: Proinsulin levels demonstrated associations with progression of (but not baseline) carotid intima media thickness (cIMT). In contrast, proinsulin- associated genetic variants influenced baseline but not progression cIMT measures, independently of established CVD risk factors and proinsulin lev- els. Novel associations between a) novel genetic variants and proinsulin and b) proinsulin-associated SNPs and cIMT have been observed, however these require replication. Conclusion: Proinsulin does not appear to be a causal factor in early sub- clinical atheroscelerosis, however proinsulin-associated mechanisms are im- plicated. Supported by: the European Commission and the European Union Framework Programme 7 224 Risk of cardiovascular disease events: the impact of diabetes and anti-diabetic drugs: a nested case-control study M. Gejl1,2, J. Starup-Linde2, J. Scheel-Thomsen3, J. Rungby4, S. Gregersen2, P. Vestergaard5; 1Department of Biomedicine – Pharmacology, Aarhus University, 2Department of Endocrinology and Internal Medicine (MEA), Aarhus University Hospital, 3Department of Neurology, Aalborg University Hospital, Aalborg, 4Diabetes Research Division, Dept. of Medicine, Copenhagen University Hospital Gentofte, Copenhagen, 5Department of Clinical Medicine and Endocrinology, Aalborg University Hospital, Denmark. Background and aims: Diabetes Mellitus (DM) is associated with an in- creased risk of cardiovascular disease. We investigated the effects of antidia- betic drugs on the composite endpoint (CE) of ischemic heart disease, heart failure and stroke incidence in patients with DM. Materials and methods: We conducted a nested case-control study. Cases were patients with type 1 or type 2 DM who subsequently suffered from CE; controls were DM patients with no history of CE after DM diagnosis. Using the Danish National Hospital Discharge Register, we included DM patients with information on date of DM diagnosis, date of CE, and comorbidities. From the Central Region of Jutland, Denmark, medication use and bio- chemical parameters (LDL, HDL, total cholesterol, triglyceride, HbA1c, and creatinine) were collected. Logistic regression analyses were conducted. The analyses were mutually adjusted for comorbidities, pharmaceutical use, and the biochemical parameters. Results: The data included 14,454 patients with DM who contributed a to- tal of 115,225 person-years. 14.8 percent (2,257) suffered from a subsequent CE. CE prior to DM diagnosis (OR= 14.83, 95%CI: 12.63-17.41), atrial fi- brillation (OR= 1.99, 95%CI: 1.57-2.51), and age at diabetes diagnosis (OR= 1.01, 95%CI: 1.007-1.014) all significantly increased the risk of subsequent CE. Patients with type 2 DM had a higher risk compared to type 1 DM pa- tients (OR= 3.05, 95%CI: 2.57-3.63). The diabetes complications retinopathy (OR= 1.70, 95%CI: 1.30-2.23), nephropathy (OR= 1.58, 95%CI: 1.28-1.95), neuropathy (OR= 1.99, 95%CI: 1.57-2.53) and peripheral artery disease (OR= 2.06, 95%CI: 1.68-2.56) increased the risk of CE so did male gender (OR= 1.12, 95%CI: 1.00-1.25), and hypertension (OR= 1.39, 95%CI: 1.20- 1.62). Treatment with statins reduced the risk of CE significantly (OR= 0.65, 95%CI: 0.57-0.73) when looking at the DM group as a whole, when grouped only significant in subjects with type 2 DM (OR= 0.59, 95%CI: 0.51-0.67). Biguanides (OR= 0.54 95% CI; 0.44-0.67) and liraglutide (OR= 0.41 95% CI; 0.32-0.52) (type 2 DM) both significantly decreased the risk of CE. DPP-4 inhibitors and β-cell stimulating agents had a neutral effect. In the subgroup analysis, a total of 1,351 DM patients (10,229 person-years), had information on biochemical values and were included. 14.9 percent (201) suffered from a subsequent CE. When results were adjusted for biochemistry, liraglutide (OR= 0.25, 95%CI: 0.11-0.56) was the only antidiabetic drug that retained a significant reduction on the risk of CE. Conclusion: We have shown an association between use of biguanides and liraglutide and a reduced risk of CE in patients with type 2 DM. The effect of liraglutide was not tied to patient biochemical values, e.g. cholesterol or glucose-control (HbA1c). Supported by: Novo Nordisk Diabetologia (2014) 57:[Suppl1]S1–S564 S 102 1 C OP 39 Inflammation in obesity and type 2 diabetes 225 Activation of oestrogen receptor alpha in macrophages controls high-fat diet-induced inflammation of adipose tissue and prevents obesity and glucose intolerance E. Riant1, S. Handgraaf1, A. Zakaroff1, D. Teixeira2, R. Burcelin1, J.-F. Arnal1, A. Bouloumié1, P. Gourdy1,3; 1Institut des Maladies Métaboliques et Cardiovasculaires, INSERM U1048, Toulouse, France, 2Department of Biochemistry, Faculty of Medecine, Porto, Portugal, 3Diabetology Department, Toulouse, France. Background and aims: Estrogens have been recognized as key regulators of body composition and glucose homeostasis. However, although both clinical and experimental data clearly evidenced the crucial role of estrogen receptor alpha (ERa), the mechanisms involved in the protective actions of estrogens against obesity and diabetes remain obscure. For instance, although estrogens are known to modulate inflammatory responses through ERa dependent ef- fects, it is still uncertain wether specific actions of these sex steroid hormones on the stroma-vascular fraction (SVF) of adipose tissue could contribute to their benefits on body composition and glucose metabolism. In the present study, we aimed to determine 1) the influence of estrogens on the adaptation of SVF cells in response to a nutritional stress; 2) the contribution of ERa- expressing myeloid cells to the prevention of obesity by estrogens. Materials and methods: In a first set of experiment, 3 groups of C57Bl/6 female mice were subjected to a chow or a high-fat (HFD) diet for 12 weeks: ovariectomized (estrogen deficiency), sham-operated (endogenous estro- gens) and ovariectomized treated with 17b-estradiol (E2, 80µg/kg/d, sc). Weight gain, body composition and glucose tolerance were monitored and both isolated adipocytes and SVF from visceral adipose tissue (VAT) were analyzed by flow cytometry and RT-qPCR. Then, metabolic phenotype and SVF characteristics were studied in HFD-fed mice with specific invalidation of ERa in myeloid cells (ERa-LysM-Cre+ mice). Results: Both endogenous estrogens and E2 administration prevented HFD- induced obesity and glucose intolerance, as well as adipocyte hypertrophy. HFD increased the number of CD45- cells in VAT SVF, irrespective of es- trogen status, but VAT infiltration by immune cells (CD45+) was strongly reduced by endogenous estrogens and E2, with a significant decrease in macrophage content (-65% CD45+/F4/80+ cells) and a predominance of M2 macrophages, as confirmed by mRNA analysis (+13-fold Ym1; -4.5-fold iNOS in E2-treated as compared to ovariectomized mice). VAT infiltration by B and TCD4+ and TCD8+ lymphocytes was also reduced by estrogens, but the pool of regulatory T cells was preserved. Demonstrating the role of ERa-LysM-Cre+ mice were characterized by a significant increase of HFD- induced inflammation of VAT and developed exacerbated adiposity and glu- cose intolerance. Conclusion: These data demonstrate that estrogens limit HFD-induced in- flammation of VAT and suggest that macrophages of the SVF largely con- tribute to their protective action against obesity and though the activation of ERa. 226 Heme oxygenase-1 drives metaflammation and insulin resistance in mouse and man A. Jais1, E. Einwallner1, O. Sharif2, M. Bilban1, C. Duvigneau3, C. Lumeng4, W. Patsch5, O. Wagner1, J.A. Pospisilik6, H. Esterbauer1; 1Department of Laboratory Medicine, Medical University of Vienna, 2Dept. of Medicine 1, Laboratory of Infection Biology, Medical University of Vienna, 3Institute of Medical Biochemistry, University of Veterinary Medicine, Vienna, Austria, 4University of Michigan, Ann Arbor, USA, 5Paracelsus Medical University, Salzburg, Austria, 6MPI for Max Planck Institute of Immunobiology and Epigenetics & Epigenetics, Freiburg, Germany. Background and aims: Obesity and diabetes affect more than half a billion individuals worldwide. Interestingly, the two conditions do not always co- incide and the molecular determinants of ‘healthy’ versus ‘unhealthy’ obe- sity remain ill-defined. Chronic metabolic inflammation (metaflammation) is believed to be pivotal. Here, we tested a hypothesized anti-inflammatory role for heme oxygenase-1 (HO-1) in the development of human and mouse metabolic disease. Heme oxygenases catalyze the oxidative degradation of heme, a potentially harmful pro-oxidant, into biologically active products: biliverdin IXα, carbon monoxide(CO) and ferrous iron. In addition to its role in heme catabolism, HO-1 plays important roles in various pathophysiologi- cal states associated with cellular stress. Materials and methods: To address the metabolic role of HO-1 in vivo in macrophages, we have generated a myeloid-cell specific conditional HO-1 knockout mouse model (‘MacHO’) and examined the role of HO-1 in the development of obesity-associated adipose tissue inflammation and insulin resistance. We fed either low-fat (10% calories derived from fat) or high-fat (60% of calories derived from fat) diets to male MacHO and littermate con- trol mice for a total of 21 weeks. Results: A significant difference in body-weight between the high fat diet groups was observed after 8 weeks of HFD feeding. Both insulin sensitivity and glucose tolerance were improved in MacHO mice fed the HFD compared with littermate control mice. Similarly, obese MacHO mice showed signifi- cantly higher insulin sensitivity as measured by insulin tolerance test when compared to HFD-fed control-mice, whereas insulin sensitivity was com- parable between both groups on control diet. Histological analysis of HFD WAT sections showed a higher density of smaller adipocytes and a signifi- cant reduction in macrophage cell numbers, especially of the Cd11c+ pro- inflammatory macrophage subpopulation. Furthermore, we found that mice lacking myeloid Hmox1 exhibited a strong reduction in hepatic accumula- tion of triglycerides and cholesterol. Histological analysis showed decreased vacuole formation and a reduced number of lipid droplets, consistent with amelioration of hepatic steatosis. In line, in highly matched biopsies from ‘healthy’ versus insulin resistant obese subjects we find HO-1 to be amongst the strongest positive predictors of metabolic disease in humans. Conclusion: Intriguingly, cellular assays show that HO-1 is defining pre- stimulation thresholds for inflammatory skewing and NF-kB amplification in macrophages. These findings identify HO-1 inhibition as a potential thera- peutic strategy for metabolic disease. Supported by: WWTF, MPG, ÖDG, FWF 227 Adipose stem cells contribute to inflammation and insulin resistance through both TH-17 polarisation and increased pro-inflammatory cytokine secretion A. Eljaafari1,2, M. Robert1,3, B. Osta4, M. Chehimi1, C. Durand1, S. Chanon1, E. Lefai1, H. Vidal1, L. Pirola1; 1Unité 1060-Carmen, INSERM, Pierre Benite, France, 2Recherche Clinique, Centre Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite, 3Gastro-Enterology Surgery, 4Unité Immunogénomique et Inflammation, EA4130, Hopital E. Herriot, Lyon, France. Background and aims: In obesity, white adipose tissue (WAT)-infiltrating immune cells induce a state of low grade inflammation. Among those cells, Th-17 cells have been identified as contributors of WAT inflammation. How- ever, the mechanisms by which Th-17 are activated remain to be elucidated. Therefore, to gain further insight into the inflammatory processes taking place into WAT during obesity, we investigated whether adipose tissue-de- rived stem cells (ASC) could promote Th-17 cell activation. Materials and methods: Human ASC were isolated from the stromal vas- cular fraction of residual WAT obtained during supra-mesocolic abdominal surgical procedures. Mononuclear cells (MNC) were prepared from blood donors, or from three surgical patients, as autologous controls. Co-cultures were performed with MNC and graded doses of ASC. Phyto-hemagglutinin A (PHA) was used for MNC stimulation. Cytokine secretion and mRNA expression were measured by ELISA and quantitative PCR, respectively. Binding of STAT3 and STAT5 on the IL-17 genomic locus was measured by chromatin immunoprecipitation (ChIP) in co-cultured cells. Finally, adi- pogenesis was induced in the presence or absence of conditioning medium collected from ASC-MNC co-cultures. Insulin-sensitivity and adipocyte dif- ferentiation were measured through AKT phosphorylation, and expression of mRNA markers, respectively. Results: Co-cultures of ASC and PHA-activated MNC resulted in a dose- dependent increase of IL-17A secretion (p S 103 1 C played an important role in ASC-mediated enhancement of IL-17A produc- tion, as co-cultures in transwell systems drastically decreased IL-17A produc- tion (pA mutation substituted asparagine at residue 325 in place of isoleucine (A20I325N) that lies in a highly conserved beta-hairpin loop in A20s ovarian tumour (OTU) domain, away from the C103 catalytic residue. Despite the presence of increased circulating granulocytes and CD44high T- cells, adult mice appear grossly normal and healthy; we therefore tested the effect of the mutation on islet biology. Results: Here we demonstrate that pancreatic islets harbouring the A20I325N mutation to exhibit cell-intrinsic increased susceptibility to inflammatory triggers, with loss of metabolic function. When challenged with mild inflam- matory stress by syngeneic transplantation of A20I325N islets into wildtype (WT) hosts, A20I325N islet grafts showed an abnormal production of inflam- matory factors (e.g. CXCl10, CXCL1) with significant neutrophil infiltra- tion. Furthermore, mutant islets exhibited severe glucose intolerance in an i.p.GTT, and an i.v.GTT revealed that beta-cell function was impaired. When directly challenged with an immunological insult in the form of allogeneic transplantation, A20I325N islets were hyper-inflammatory and more rapidly destroyed than WT allogeneic islets. Immunoblot analysis of cells harbor- ing the A20I325N mutation show exaggerated poly-ubiquitination of RIP1 with exacerbated NF-κB and JNK/AP-1 signaling following TNF-stimulation. In- vitro transfection studies using NF-κB and AP-1 luciferase reporter plasmids show WT A20 to inhibit luciferase production at the promoter level. Thus, A20 is critical for controlling NF-κB but also JNK activation; and loss of OTU function results in tissue hypersensitivity to inflammatory triggers and meta- bolic impairment. Ectopic expression of WT A20 in A20I325N islets restored control of NF-κB and JNK/AP1 pathways and rescued A20I325N islets from hyper-inflammation. Whole body A20I325N mutant mice show normal glucose homeostasis, however, in a diet induced obesity model (45 kcal % fat), mice homozygous for the point mutation (A20I325N/I325N) develop more severe glu- cose intolerance, compared to littermate controls and mice heterozygous for the point mutation (A20I325N/+). However, when A20I325N/+ mice are subjected to a low-dose lipopolysaccharide (LPS; i.p.) challenge, they also developed marked glucose intolerance compared to littermate WT controls. Further- more, glucose intolerance was coupled with islet-specific immune infiltration of the pancreas. Conclusion: These data identify A20 as a diabetes protective gene. A20 negative feed-back on NF-κB and JNK/AP-1 pathways maintain islet inflam- matory and metabolic homeostasis under vigorous inflammatory states. In contrast, functional impairment by SNPs can uncouple A20s protective effect and lead to diabetogenic outcomes. 230 Anti-CD3 / anti-CXCL10 antibody combination therapy reverts type 1 diabetes in two mouse models S. Lasch, P. Müller, M. Bayer, J.M. Pfeilschifter, E. Hintermann, U. Christen; Goethe University Hospital, Frankfurt am Main, Germany. Background and aims: Therapy of type 1 diabetes (T1D) with anti-CD3 monoclonal antibodies (a-CD3) results in a blockade of the autoimmune process in both animal models and patients with T1D patients. Treated pa- tients show a reduced insulin need and elevated C-peptide levels. Unfortu- nately, this effect is only temporal and patients revert within a few years to the status they had at the beginning of the therapy. In animal models several combination therapies using a-CD3 and treatments, such as rapamycin, nasal insulin, vitamin D, FTY 720, or Lactococcus lactis have been successful in Diabetologia (2014) 57:[Suppl1]S1–S564 S 104 1 C blocking T1D. We used a different approach and aimed at a blockade to cel- lular re-entry into the islets of Langerhans after a-CD3 treatment by blocking the key chemokine CXCL10. Materials and methods: We used the well-established RIP-LCMV-GP mouse model of T1D. As a target autoantigen in the β-cells, such mice express the glycoprotein (GP) of the lymphocytic choriomeningitis virus (LCMV) under control of the rat insulin promoter (RIP). RIP-LCMV-GP mice only develop T1D only after infection with LCMV. For the present study we used a com- bination therapy (CT) and treated diabetic RIP-LCMV-GP mice first with a-CD3 (3 injections in 3 days) followed by administration of a neutralizing anti-CXCL10 monoclonal antibody (a-CXCL10) (8 injections over 3 weeks). Results: The CT reverted T1D in RIP-LCMV-GP mice significantly (CT: 65% reversion; p=0,0078 / Control: 20% reversion). Thereby, the efficacy of the monotherapies was improved (a-CD3 alone: 47% reversion; p=0,1168 / a- CXCL10 alone: 36% reversion; p=0,3735). Flow cytometry and histological experiments demonstrate a marked reduction of CD4 and CD8 T cells in the pancreas of mice treated with the CT compared to mono-therapies at day 31 after infection. Importantly, the presence of islet antigen (LCMV-GP)- specific CD8 T cells was reduced dramatically. This effect is long-lasting since CT-treated islets show almost no infiltrates up to 181 days after infection. In order to confirm our data in a different model, we conducted a pilot study with NOD mice, in which CT cured 50% of diabetic mice whereas no mice show diabetes reversion with a-CD3 or without therapy. Conclusion: In summary our data suggest that a combination therapy of an- ti-CD3 and anti-CXCL10 antibodies results in a significant reduction of T1D in two independent mouse models. The neutralization of the key chemokine CXCL10 prevented a re-entry of aggressive T cells into the pancreas and thereby abrogated the destruction of beta-cells. Such a combination therapy might constitute a novel treatment for patients suffering from T1D. Supported by: TRIP / Goethe University Hospital 231 MiR-125a-5p is up-regulated in un-functional CD4+FOXP3+ T regulatory cells deriving from pancreatic lymph nodes of patients with type 1 diabetes and targets C-C chemokine receptor type 2 G. Ventriglia1, G. Sebastiani1, A. Stabilini2, F. Mancarella1, L. Nigi1, T. Brusko3, M. Battaglia2, F. Dotta1; 1University of Siena and Umberto di Mario Foundation, 2Diabetes Research Institute, San Raffaele Scientific Institute, Milano, Italy, 3University of Florida, Gainesville, USA. Background and aims: We previously showed that CD4+CD25brightCD127- T regulatory (Treg) cells deriving from pancreatic lymph node (PLN) of pa- tients with type 1 diabetes (T1D) are epigenetically imprinted to be Treg, but for unknown reasons do not function as such in vitro. Importantly, this functional defect is present only in Treg cells residing in the PLN of T1D patients and not in those circulating in their peripheral blood (PB). Some microRNAs (miRNAs), an abundant class of small non-coding RNAs that regulate gene expression by affecting the degradation and translation of target mRNAs, are differentially expressed in Treg cells isolated from PLN as com- pared to those isolated from PB of T1D patients. Among the differentially expressed miRNAs, we found that miR-125a-5p is specifically up-regulated in Treg cells isolated from PLN of T1D patients as compared to those isolated from PB of the same donors and to those isolated from non-diabetic donors. This differential expression was not identified in T conventional cells of T1D patients, uncovering both a cell (i.e., Treg cells) and a disease-specific (i.e., T1D) miRNA expression. Thus, we aimed at further investigating the role of miR-125a-5p in Treg cells. Materials and methods: CD4+CD25brightCD127- Treg cells were FACS sorted from PLN and PB of patients with T1D and from non-diabetic donors. 200 sorted cells were analyzed and miRNA expression was evaluated by RT- PCR using Taqman miRNAs single assay. Data analysis was performed using 2-ddCt method. Bioinformatic miRNAs target prediction was performed us- ing Targetscan 6.2. Dual luciferase reporter assay was performed to experi- mentally validate the predicted target gene. CCR2 expression was determined by flow cytometry. Results: Computational target gene analysis of miR-125a-5p was performed. Target prediction analysis retrieved some potential interesting genes involved in Treg-cell biology, including: C-C chemokine receptor type 2 (CCR2), inter- leukin 6 receptor (IL6R), interferon regulatory factor 4 (IRF4). Among these putative targets we focused on CCR2 to verify whether this chemokine recep- tor is key in the observed Treg-cell dysfunction. Firstly, this prediction was validated by renilla luciferase assay: a decreased luciferase activity following miR-125a-5p overexpression was observed and this reduction was restored upon miRNA predicted target site mutations. Secondly, CCR2 expression on Treg cells was tested by flow cytometry and frequency of CCR2+ Treg cells was correlated with miR-125a-5p RT-PCR levels. Interestingly, lower expres- sion of CCR2+ Treg cells correlated with higher expression of miR-125a-5p, suggesting an inverse correlation. Conclusion: A disease- and tissue-specific miRNA, miR-125a-5p, has been identified in Treg cells isolated from PLN of patients with T1D. miR-125a-5p targets CCR2, thus leading to reduced CCR2 expression. Reduced expression of CCR2 may induce migration dysfunction in Treg cells residing in the PLN of T1D patients. Therefore these Treg cells could be unable to reach the target organ where they are supposed to control and possibly block the autoreactive immune response. Further in vitro and in vivo studies are ongoing to confirm this hypothesis. Supported by: JDRFi nPOD Grant 232 Cholecalciferol increases suppressor function in regulatory CD4+CD25+Foxp3+ T-cells from patients with new onset type 1 diabetes: a randomised controlled trial G. Treiber1, B. Prietl1, E. Fröhlich-Reiterer2, E. Höller1, A. Ribitsch1, M. Fritsch3, B. Rami-Merhar3, C. Steigleder-Schweiger4, M. Borkenstein2, T.R. Pieber1; 1Devision of Endocrinology and Metabolism, 2Department of Pediatrics and Adolescent Medicine, Medical University Graz, Austria, 3Department of Pediatrics, Medical University Vienna, 4Paracelsus Medical University, Salzburg, Austria. Background and aims: Vitamin D has an important role in increasing the effects of innate immune processes while restraining the adaptive immune system, leading to improved outcomes in autoimmune diseases. Vitamin D deficiency is one environmental factor to influence the risk of type 1 diabetes (T1D) and in vitro data reinforce the role of vitamin D in the pathogenesis of T1D. In this study we evaluate the effects of Cholecalciferol on function of regulatory T-cells and residual b-cell function in patients with recent onset T1D. Materials and methods: In this randomized, double-blinded, placebo con- trolled trial thirty patients (age 15 [10-16], 77% male) with new-onset T1D were assigned to receive oral therapy of cholecalciferol (70IU/kg bodyweight/ day) or placebo for 12 months. Cellular frequencies were determined by FACS-analysis and functional tests were assessed with ex vivo suppression co-cultures. Mixed meal tolerance tests were performed to assess b- cell func- tion. Results: Suppressive capacity of regulatory T cells (Treg) increased with Cholecalciferol after 3, 6 and 12 month (p=0.0005) and delta suppression capacity between Cholecalciferol (28.1 [14.9 to 43.3]%) and placebo (-20.9 [-30.1 to -11.3]%) was significant (p=0.002). Treg frequency did not dif- fer after 12 month between the two groups (5.75±1.48% vs 5.38±1.73%, p=0.187). HbA1c and insulin dose were similar between the groups, fast- ing C-peptide appeared to decrease slower (p=0.078) in the Cholecalciferol group (0.62±0.18 to 0.59±0.34 ng/ml) than in placebo group (0.65±0.27 to 0.40±0.25 ng/ml). Serum calcium and parathormone stayed within the nor- mal range. Conclusion: Cholecalciferol, the inactive form of vitamin D used as adjunc- tive immunomodulatory therapy with insulin is associated with an improve- ment in suppressor function of regulatory T-cells in patients with new-onset T1D. Assessment of vitamin D status and a targeted supplementation may be indicated after the diagnosis of T1D. Clinical Trial Registration Number: NCT01390480 Supported by: EFSD/MSD Diabetologia (2014) 57:[Suppl1]S1–S564 S 105 1 C OP 41 The human methylome in diabetes 233 Novel epigenetic markers of type 2 diabetes in the liver S. Cauchi1,2, R. Caiazzo2,3, B. Pichard1,2, A.-S. Régent1,2, M. Canouil1,2, S. Lobbens1,2, L. Yengo1,2, A. Leloire1,2, V. Raverdy2,3, O. Poulain-Godefroy1,2, F. Pattou2,3, P. Froguel1,2; 1Genomics and Metabolic Diseases, UMR CNRS 8199, 2European Genomic Insitute for Diabetes (EGID), 3Inserm UMR 859, Lille, France. Background and aims: The liver plays an important role in the regulation of glucose homeostasis. Epigenetic markers in this organ could contribute to the pathophysiology and heritability of type 2 diabetes (T2D). Our aims were to define DNA methylation patterns associated with T2D risk in the liver, cor- relate these epigenetic factors with gene expression pathways and decipher intermediate metabolic traits. Materials and methods: We analyzed the genome-wide methylation patterns in the liver tissue of 96 T2D cases and 96 matched normoglycemic controls using Infinium Human Methylation 450 BeadChips (Illumina). A beta-mix- ture quantile normalization method was applied for correcting probe design bias. Genome-wide expression profiles were also assessed in a subset of 24 T2D cases and 24 matched normoglycemic controls using HumanHT-12 v4 Expression BeadChips (Illumina). Results: We identified 158 sites differentially methylated between T2D cases and controls after Bonferroni correction for multiple comparisons (p S 106 1 C array, (ii) targeted pyrosequencing of published methylation variants: IGF2, PAX8, ZFYVE28, EXD3, BOLA3 • Validation using an external dataset Results: The in utero exposed group had an excess of underweight individu- als (49%) compared to postnatal exposed (32%) or unexposed (30%); more postnatal exposed offspring (26%) were overweight compared to in utero exposed (13%) or unexposed (14%) (proportional odds model df=2, devi- ance ratio 3.26, p=0.039). Of the underweight, in utero exposed had a higher mean glucose 120minutes post-glucose challenge (5.8mmol/l) compared to postnatal exposed or unexposed (both 4.8mmol/l) (ANOVA p S 107 1 C firm that Nur77 transcriptionally activates FGF21 expression, ChIP assay, EMSA and mutagenesis analysis were employed. Results: We found that fasting induced hepatic Nur77 and FGF21 expres- sion. Glucagon and forskolin increased Nur77 and FGF21 expression in vivo and in vitro, respectively. Ad-Nur77 increased FGF21 production in cultured hepatocytes. Hepatic overexpression of Nur77 via tail vein injection of Ad- Nur77 increased FGF21 mRNA and protein levels. Moreover, siRNA-Nur77 abolished the effect of forskolin on FGF21 expression. The results from ChIP assay, EMSA and mutagenesis analysis showed that Nur77 bound to puta- tive NBRE of FGF21 promoter in cultured hepatocytes and fasting induced Nur77 binding to FGF21 promoter in vivo. Conclusion: This study shows that Nur77 mediates fasting-induced hepatic FGF21 production. The present study suggests an alternative mechanism by which hepatic FGF21 transcription is mediated under fasting conditions. Supported by: NRF-2012R1A2A1A03670452 239 Overexpression JAZF1 protected ApoE-/- mice from atherosclerosis by inhibiting hepatic cholesterol synthesis H. Wang1,2, L. Li1,2, P. Yan3, G. Yang3; 1the Key Laboratory of Laboratory Medical Diagnostics in Ministry of Education, Chongqing, China, 2Department of Clinical Biochemistry, College of Laboratory Medicine, Chongqing Medical University, 3Department of Endocrinology, the Second Affiliated Hospital, Chongqing Medical University, China. Background and aims: Genome wide association studies have suggested an association of Juxtaposed with another zinc finger gene 1(JAZF1) with type2 diabetes mellitus(T2DM). As an inhibitor of the TAK1/TR4 signaling pathway, JAZF1 has been shown to be involved in gluconeogenesis, lipid me- tabolism and insulin sensitivity. However, its role in insulin resistance and atherosclerosis in vivo remains unknown. The present study was designed to investigate in vivo the impact of JAZF1 on insulin resistance-associated dyslipidemia and atherosclerosis. Materials and methods: We established Adenovirus-mediated JAZF1 over- expression to characterize the role of JAZF1 in the regulation of lipid metabo- lism and the development of atherosclerosis in normal chow- or high fat diet (HFD)-fed ApoE-/- mice .We use euglycaemic-hyperinsulinaemic clamping to examine Insulin sensitivity, Cholesterol de novo synthesis was measured by intraperitoneal acetate injection and atherosclerotic plaques were quantified by histological analysis ,A dual- luciferase reporter assay was used to assess the ability of JAZF1 to regulate HMGCR transcriptional activity, mRNA and protein expressions were measured by qRT-PCR and Western blot, respec- tively. Results: we showed that JAZF1 overexpression improved HFD-induced he- patic insulin resistance in C57BL/6J mice. In HFD-fed ApoE-/- mice, JAZF1 overexpression decreased serum cholesterol levels and hepatic cholesterol synthesis by inhibiting CREB dependent 3hydroxy3methylglutaryl coen- zyme A reductase (HMGCR) promoter transcriptional activity and JAZF1 overexpression had significantly reduced aortic and aortic sinus en face and cross- sectional plaque areas in HFD-fed ApoE-/- mice. Conclusion: we report the first evidence that JAZF1 overexpression protect- ed against development of atherosclerosis and insulin resistance induced by a HFD in apoE-/- mice and in C57BL/6J mice. Furthermore, we provide mo- lecular explanation by which JAZF1 can regulate hepatic cholesterol synthe- sis and HMGCR transcriptional activity via CREB -dependent mechanisms. Our data delineate that JAZF1 overexpression may represent a promising strategy for the treatment of insulin resistance-associated dyslipidemia and atherosclerosis. Supported by: NNSFC (81270913,81300702,81300670) 240 Defective autophagy results in reduced glycogen breakdown in the liver of IUGR newborn Wistar rats. Consequences for glucose homeostasis E. Fernandez-Millan1, J. de Toro-Martin2, E. Lizarraga-Mollinedo1,2, F. Escriva1,2, C. Alvarez1,2; 1CIBERDEM, ISCIII, 2Biochemistry and Molecular Biology II, School of Pharmacy, UCM, Madrid, Spain. Background and aims: Glycogen autophagy is a highly regulated process that represents a mechanism of glucose homeostasis under conditions of de- mand for the production of this sugar as in the liver of newborn animals. The abrupt interruption of the maternal glucose supply with birth, leads to the development of hypoglycaemia in the newborn. This promotes glucagon secretion and triggers the normal mechanisms for the newborn to adapt to the postnatal environment. Glucagon acts through the cAMP/protein Kinase A pathways and induces glycogen autophagy. We previously showed that in- trauterine growth restriction (IUGR) induced massive glycogen accumula- tion in the liver of fetuses and neonates. Interestingly, IUGR neonates were slightly hypoglycaemic and showed hypoglucagonaemia and glucagon resist- ance but insulin hypersensitivity. In this study we set out to examine the au- tophagic response in the liver of IUGR newborns and whether this process might be linked to the development of insulin resistance associated to IUGR. Materials and methods: Newborn rats naturally delivered from pregnant females fed ad libitum or 65% food-restricted from the last third of gesta- tion, were kept separated from their mothers and used 3 or 6h after birth. Liver glycogen content was quantified by a KOH-ethanol precipitation and acid hydrolysis method. The protein levels of the enzymes involved in glu- cose metabolism, the markers of autophagy and ER stress were determined by western blot in hepatic homogenizes. Ultraestructural analysis of the neo- natal liver was performed by electron microscopy. Results: At birth (0h), liver glycogen values of control (C) and IUGR rats were 23.8 ± 3.2 and 57.4 ± 7.3 mg/g (p S 108 1 C OP 43 GLP-1 analogues: novel formulations 241 Benefits of a fixed-ratio formulation of once-daily insulin glargine/ lixisenatide (LixiLan) vs glargine in type 2 diabetes inadequately controlled on metformin J. Rosenstock1, M. Diamant2, L. Silvestre3, E. Souhami3, T. Zhou4, V. Fonseca5; 1Dallas Diabetes and Endocrine Center, USA, 2VU University Medical Center, Amsterdam, Netherlands, 3Sanofi, Paris, France, 4Sanofi, Bridgewater, USA, 5Tulane University Health Sciences Center, New Orleans, USA. Background and aims: LixiLan is a fixed-ratio formulation combining insu- lin glargine (GLARG) with lixisenatide (Lixi) in a single pen device currently in development for the management of type 2 diabetes mellitus (T2DM). This randomized, open-label study assessed the efficacy and safety of Lixi- Lan (GLARG 2 U / Lixi 1 µg) vs GLARG in insulin-naïve T2DM on top of metformin. Materials and methods: Patients (mean baseline A1C 8.0%, BMI 32.1 kg/m 2, diabetes duration 6.7 yrs) were randomized to LixiLan (n=161) or GLARG (n=162) for 24 weeks. Primary objective was non-inferiority in A1C reduc- tion; if non-inferiority met, statistical superiority was tested. Results: At 24 weeks, mean A1c was reduced to 6.3% and 6.5% with Lixi- Lan and GLARG respectively, establishing statistical superiority of LixiLan (LS mean difference: -0.17% [-0.312% to -0.037%; p=0.0130]), and 84% and 78% achieved A1C 10% to ≤12%, age 18-80 years, BMI 25-45 kg/m2, and on stable (≥3 months) diet and exercise and/or monotherapy or any combination of met- formin, sulfonylurea, and thiazolidinedione. Treatment was initiated by placing a 3-month ITCA 650 mini-pump delivering 20 mcg/day, which was then replaced by a 6-month ITCA 650 mini-pump delivering 60 mcg/day for 26 weeks. Pre-study oral antidiabetic agents (OADs) were maintained un- changed for the 39 week of treatment. The primary endpoint was change in A1C from baseline to week 39. Results: At the time of this initial interim analysis, 50, 39, and 25 of the 60 sub- jects enrolled had completed 13, 19, and 26 weeks of treatment; respectively. Mean baseline characteristics for the entire cohort (n=60) were A1C 10.7%, age 52.1 yrs, BMI 32.1 kg/m2, duration of diabetes 8.9 yrs, OAD use 69%. Mean reductions of A1C at Weeks 13 (n=50), 19 (n=39), and 26 (n=25) were -2.5%, -2.9%, and -3.2%, respectively. A1C reductions ≥2% were achieved by 78% of subjects who completed at least 13 weeks of treatment; 50% achieved >3% and 22% achieved ≥4% reductions. A1C targets of S 109 1 C (36%). The likelihood of achieving all four pre-prandial BG values within the recommended range (≥3.9 to ≤7.2mmol/L) was significantly greater with IDegLira (DUAL I 48%; DUAL II 44%) compared with IDeg (DUAL I 41%; DUAL II 27%) or Lira (32%). At EOT, the proportion of patients with all 9 BG values within the range ≥3.9 to S 110 1 C OP 44 Diabetes and cancer 245 Cancer occurrence in type 1 diabetes patients: a 4-country study with 8800 cancer cases in 3.7 mio person-years B. Carstensen1, S. Friis2, J. Harding3, D.J. Magliano3, R. Sund4, I. Keskimäki5, R. Ljung6, A.-M. Svensson7, S. Wild8, S. Gudbjörnsdottir7; 1Steno Diabetes Center, Gentofte, 2Danish Cancer Society, Copenhagen, Denmark, 3BakerIDI Heart and Diabetes Institute, Melbourne, Australia, 4University of Helsinki, 5National Institute for Health and Welfare, Helsinki, Finland, 6Karolinska Insititutet, Institute of Environmental Medicine, Stockholm, 7University of Gothenburg, Sweden, 8University of Edinburgh, UK. Background and aims: Diabetes (DM) patients carry an excess risk of can- cer in the order of 20-25%; this is mainly derived from follow-up of type 2 patients (T2D). The excess risk of cancer among type 1 (T1D) patients is described here as it is anticipated to be different from T2D patients. Materials and methods: T1D patients from four countries with nation-wide diabetes registers: Australia (1997-2008), Denmark (1995{-2009), Finland (1972-2010), and Sweden (1987-2011) were followed for cancer occurrence. T1D was defined by diagnosis of DM before age 30. Cancer incidence rates were compared to population cancer incidence rates from national cancer registries. We used Poisson-models for rates, adjusting for age and date of follow-up, and date of birth. We estimated the overall rate ratio (RR) for all T1D patients and the effect of time since DM diagnosis. Results: There were a total of 8,807 cancers among T1D patients during 3.7 million person-years of follow-up with median age at cancer diagnosis 51.1 (IQR: 43.5,59.5). Overall, we found an RR of any type of cancer of 1.00 (95% CI: 0.97-1.03) among men and 1.05 (95% CI: 1.02-1.08) among women. The highest RRs were found for colorectal cancer, (RR=1.13 (M), 1.14 (F)), liver cancer (RR=2.14 (M), 1.50 (F)), pancreas cancer (RR=1.74 (M), 1.31 (F)), endometrial cancer (RR=1.4 kidney cancer (RR=1.28 (M), 1.44 (F)), and thy- roid cancer (RR=1.29 (M), 1.46 (F)), all significant. We found a strong effect of diabetes duration, with an RR of 2.5 during the first year, decreasing to 1.2 (M) and 1.1 (F), after 2-5 years. Conclusion: Some of the observed excess risk may be explained by risk fac- tors for cancer being more frequent in T1D patients (obesity), however this effect is presumably smaller than in T2D patients, and hence consistent with the smaller excess risk. The long-term RR (>5 years of DM) is less than 1.2, which means that some small effect of exogenous insulin cannot be excluded, but the study is also consistent with an assumption of no such effect. Supported by: EFSD 246 Time-varying incidence of cancer after incident type 2 diabetes: differences by obesity-versus non-obesity-related cancers A. Renehan1, E. Badrick2, M. Sperrin2, M. Carr2, T. Moran3, I. Buchan2; 1Institute of Cancer Sciences, University of Manchester, 2Institute of Population Health, University of Manchester, 3NHS, Northwest Cancer Intelligence Service, Manchester, UK. Background and aims: People with Type 2 diabetes (T2D) are at increased incident risk of several cancer types, many of which are also causally linked with excess body weight. To explore these relationships further, we deter- mined the incidence of cancer in people with incident T2D during different time windows following diabetes diagnosis, stratified as obesity- and non- obesity-related cancers. Materials and methods: We used the Salford Integrated record database linked with the Northwest cancer Intelligence service (UK) for data from 1995-2010. We identified incident cohorts with (N = 10,328) and without T2D, who were matched (1: 2) by age, sex and index year. Following exclusion of previous can- cers, first site-specific cancers were identified prospectively in both cohorts. Obesity-related cancers were post-menopausal breast, colorectal, endometrial, kidney, gallbladder, pancreas, liver, and ovarian. Risk estimates were expressed as hazard ratios (HRs) and their 95% confidence intervals (CIs). Results: Diabetic patients had a significantly higher BMI than matched con- trols (mean 30.9 (SD6.7) vs 27.3 (SD5.4)), and were more likely to be ever smokers (60.0% vs 52.63%). Within 6 months following diabetes onset, par- ticipants with T2D were at increased risk of obesity-related (HR 1.54, 95% CI: 0.98, 2.40) but not non-obesity-related cancers. From 6 months to 10 years, risk associations for obesity- and non-obesity-related cancers remained close to null. After 10 years following diabetes onset, participants with T2D were at increased risk of non-obesity-related (HR 1.49, 95% CI: 1.13, 1.97) but not obesity-related cancers. These observations were robust to adjustment for smoking. Conclusion: We confirmed that people with incident T2D are at increased risk of certain cancers; the risk for obesity-related cancers is particularly el- evated near to the time of diabetes onset, and is likely to present detection time bias. Paradoxically, we noted elevated risk of non-obesity-related can- cers among people with T2D with greater than 10 years disease duration, emphasizing the importance of assessing diabetes-cancer associations using time-varying approaches. Supported by: EFSD 247 Diabetes and cancer risk in a population-based study with 20 years of follow-up: the Rotterdam Study K.M.J. De Bruijn1, R. Ruiter2, T.E. de Keyser2, A. Hofman2, B.H. Stricker2, C.H.J. van Eijck1; 1Surgery, Erasmus University Medical Center, 2Epidemiology, Erasmus University Medical Center, Rotterdam, Netherlands. Background and aims: Type 2 diabetes is associated with an increased risk of cancer. Most observational studies on this topic analyze diabetes as a di- chotomous risk factor without adjusting for diabetes Duration before cancer occurrence. The objective of this study was to investigate the association be- tween diabetes duration and cancer risk in more detail. Materials and methods: In this prospective cohort study, incident diabetes was diagnosed on the basis of clinical information and use of glucose-lower- ing medication. Details on incident cancers were obtained via general practi- tioners and linkage to registers for pathology. Time-dependent Cox propor- tional hazard models were used. Additionally, a five year latency period was taken into account. Results: The study comprised 10,746 individuals. Diabetes was associated with an increased overall risk of incident cancers (HR 1.22, 95%CI 1.07-1.39). Distinguished by type of cancer, diabetes was associated with an increased risk of breast (HR 1.50, 95%CI 1.05-2.15) and pancreatic cancer (HR 2.93, 95% CI 1.75-4.89), but with a decreased risk of prostate cancer (HR 0.59, 95%CI 0.38-0.93). As a diabetes diagnosis of less than three months before a cancer diagnosis was associated with strongly increased risks for all cancers (HR 3.30, 95%CI 2.50-4.32) and pancreatic cancer (HR 28.74, 95%CI 6.32-130.58), at least a part of the association might be explained by detection bias or by pro- topathic bias. After adjusting for a latency period of five years, the association regarding the decreased risk of prostate cancer remained statistically signifi- cant, whereas the risk of rectal cancer increased (HR 2.26, 95%CI 1.04-4.95). Conclusion: Although the magnitude of the association between diabetes and increased risk of cancer seems to be inflated by detection- or protopathic bias, an independent association remains. Future studies investigating this association should adjust for diabetes duration and a plausible etiological risk window. Diabetologia (2014) 57:[Suppl1]S1–S564 S 111 1 C 248 Cancer therapy alters whole body metabolism later in life: a potential role for epigenetic mechanisms V. Nylander1, D. Simar2, M. Aznar3, L. Specht3, J.R. Zierath1, R. Barres1; 1The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Denmark, 2School of Medical Sciences, University of New South Wales, Sydney, Australia, 3University of Copenhagen, Denmark. Background and aims: Cancer survivors are at increased risk of acquiring metabolic dysfunction such as glucose intolerance and Type 2 Diabetes, years after end of cancer treatment. Epidemiological studies identified total body irradiation as the highest risk factor. Given the importance of muscle tissue in whole body glucose metabolism, we focused on skeletal muscle progenitors cells. We hypothesized that irradiation causes epigenetic changes in this cell type that could account for the delayed, deleterious effects on metabolic function. Aim: To determine if irradiation alters the DNA methylome of muscle pro- genitor cells, leading to altered differentiation potential and metabolic dys- function. Materials and methods: Male C57Bl/6 mice were irradiated (IRR) or not (CTL) with one dose of total body irradiation (6 Gy). Mice were allowed to recover from acute irradiation effects for 5 weeks before randomisation into groups fed CHOW or high fat diet (HFD, 60 % energy intake from fat). Meta- bolic characterization of mice was done by MR scan, glucose tolerance test and indirect calorimetry. Muscle progenitor cells were isolated by FACS sort- ing to investigate for deficiencies in intracellular pathways controlling me- tabolism or cell differentiation. In vitro experiments were performed using L6 rat myoblasts, which were irradiated with one dose of 1, 3 or 6 Gy. Cells were allowed to recover from irradiation for 4 weeks, before differentiation was induced and insulin signalling and DNA methylation was investigated. Statistical analysis used was one-way and two-way ANOVA for MR and body weight, and repeated measures analysis for indirect calorimetry data and glu- cose tolerance tests. Results: After recovery from irradiation, calorie intake, oxygen consumption and energy expenditure were similar in both groups. Prolonged HFD induced impaired glucose homeostasis in IRR-HFD, but not in IRR-CHOW. This was accompanied by increased fasting insulin levels in IRR-HFD (50%, p= 0.033), whereas fasting insulin was decreased in IRR-CHOW (20%, p= 0.022), as compared to respective controls. Muscle progenitor cells collected 17 weeks post irradiation and cultured in vitro exhibited decreased growth rate (30%, p=0.03) compared to cells collected from CTL animals. Muscle cells collected from IRR animals showed altered signalling pathways controlling cell divi- sion and differentiation. Irradiation of rat muscle cell line phenocopied the impaired activation of signalling and exhibited an altered differentiation and DNA methylation profile. Conclusion: Our study confirms that irradiation induces a long-term al- teration of glucose metabolism and suggests that irradiation stably alters the epigenome and the differentiation potential of muscle progenitor cells. Irradiation-induced epigenetic modification could be a mechanism by which cancer survivors develop metabolic complications later in life. Supported by: EFSD OP 45 Diabetic foot: mechanisms of wound healing 249 Overexpression of cutaneous mitochondrial superoxide dismutase in recently diagnosed type 2 diabetic patients D. Ziegler1,2, A. Strom1, S. Püttgen1, J. Brüggemann1, I. Ziegler1, B. Ringel1, M. Roden1,2, German Diabetes Study Group; 1Institute for clinical Diabetology, German Diabetes Center at Heinrich Heine University, 2Department of Endocrinology and Diabetology, University Hospital, Düsseldorf, Germany. Background and aims: Oxidative stress resulting from enhanced free-rad- ical formation and immune-mediated processes has been implicated in the pathogenesis of diabetic neuropathy. Since manganese superoxide dismutase 2 (SOD2) is responsible for superoxide detoxification in mitochondria, we hypothesized that cutaneous SOD2 could be overexpressed in recently diag- nosed diabetes. Materials and methods: We assessed skin biopsies and nerve function in 71 participants of the German Diabetes Study (GDS) with recently diagnosed type 2 diabetes (age: 54.5±0.9 [SEM] years; male: 67.6%; BMI: 32.2±0.7 kg/ m², diabetes duration: 1.0±0.1 years; HbA1c: 6.5±0.1%) and 52 healthy con- trol subjects (age: 55.2±1.2 years; male: 53.8%, 25.4±0.5 kg/m²). Intraepi- dermal nerve fibre density (IENFD) was assessed by immunohistochemis- try in 3-mm punch biopsies from the distal leg. Subepidermal SOD2 area was determined by immunofluorescence using a polyclonal SOD2 antibody. CD31 immunohistochemistry was used to quantify subepidermal endothe- lial cell area. Peripheral nerve function was assessed by motor and sensory nerve conduction velocity (MNCV, SNCV), vibration perception thresholds (VPT), and thermal detection thresholds (TDT), while cardiac autonomic nerve function was assessed by heart rate variability (HRV) in the time and frequency domains. All comparisons were adjusted for sex, age, and BMI. Results: IENFD was lower by 24% in diabetic subjects vs controls (6.9±0.3 vs 9.1±0.4 fibres/mm; P S 112 1 C depending on the members of the Notch system involved. Diabetic wounds are characterized by impaired coordination of several cellular processes such as angiogenesis and cell differentiation. We have studied the potential role of Notch signaling in diabetic wound healing. Materials and methods: Human dermal fibroblasts (HDF) and human der- mal endothelial cells were used for in vitro studies and several animal models of diabetes (db/db mice and streptozotocin-induced diabetic mice) were used for in vivo studies. The functional consequence of the notch system modula- tion was studied in vitro by assessment of the migration of HDF and by an- giogenesis assay. Notch pathway inhibition was induced either by γ-secretase inhibitors or by specific siRNA silencing of the Notch receptors (1-4). Using cre-lox system we have generated mice that lack Notch 1 in the skin. Wound healing rate was evaluated both in db/db mice and in skin-specific Notch1 knock-out mice in which diabetes was induced by streptozotocin. Results: Notch signaling was activated in the skin of several animal models of diabetes and biopsies from patients with diabetic wounds. Hyperglycemia ac- tivated Notch pathway and had repressive effect on fibroblasts migration and angiogenesis. Mechanistically, we found that hyperglycaemia enhances delta4 expression in a Notch1-dependent manner and this positive delta4-Notch1 feedback loop contributed to the impaired wound healing in diabetes. This was confirmed in vivo where inhibition of Notch signaling with γ-secretase inhibitors improved wound healing rate just in diabetic (db/db mice) but not in control non-diabetic animals. Using loss-of-function genetic approaches we demonstrated both at the cellular level (fibroblasts, endothelial cells) as well as in an animal model that the Notch1 activation was the key player of the repressive effects of Notch on wound healing in diabetes, which was con- firmed in the biopsies from patients with diabetic foot ulcers. Conclusion: Glucose activates a positive feedback loop (delta4-Notch1) that contributes to the deleterious wound healing in diabetes. Supported by: VR, Family Erling-Persson Foundation, SLL, Tore Nilssons Foun- dation 251 SP regulates macrophage function in diabetic wound healing E.C. Leal1,2, A. Tellechea1,2, S. Kuchibhotla2, M.E. Auster2, F.W. LoGerfo2, L. Pradhan-Nabzdyk2, E. Carvalho1, A. Veves2; 1Center for Neurosciences and Cell Biology, Coimbra, Portugal, 2Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, USA. Background and aims: Chronic diabetic foot ulcerations develop in areas affected by diabetic neuropathy. As consequence, neuropeptides, such as sub- stance P (SP), known to modulate inflammation, showed to be important in wound healing. However, little it is known about the effect of SP on mac- rophages activation. We aimed to study the effect of SP on macrophage func- tion in diabetic wound healing progression. Materials and methods: We used wild type (WT) mice and two types of ge- netically modified mice: one deficient of the NK1 receptor of SP (NK1RKO) and one deficient of the TAC1 gene that encodes for SP and other takinins (TAC1KO). Also, we treated the wound with SP or CJ012,255 (CJ), an in- hibitor of SP receptor. Diabetes was induced by sreptozotocin intraperito- neal injection, 50 mg/Kg, 5 days. The animals were kept 8 weeks diabetic prior wound healing experiment. M1 and M2 macrophages were identified by using immunohistochemistry at several phases of wound healing: baseline (Day-0), Day-3 and Day-10 post-wounding. MCP-1, IL-a and KC was quanti- fied by q-RT-PCR. Results: The M1/M2 ratio was increased at Day-0 in WT-diabetic mice and nondiabetic and diabetic NK1R and TAC1 mice when compared to WT- nondiabetic mice. In WT-nondiabetic mice, the ratio increased at Day-3 but returned to normal levels by Day-10. In contrast, in WT-diabetic and non- diabetic and diabetic NK1R and TAC1 mice at Day-0, Day-3 and Day-10, the M1/M2 ratio remains high, suggesting a persistent inflammation. At Day-10, the M1/M2 ratio was increased in CJ- treated WT-nondiabetic and WT-dia- betic mice. In contrast, SP treatment reduced the diabetes-induced increase in M1/M2 ratio. Similar results were observed in the skin gene expression of the monocyte chemoattractant 1 (MCP-1) that recruits monocytes in areas of inflammation where they are converted to macrophages. At Day-0, MCP-1 was increased in the TAC1KO mice. MCP-1 expression increased at Day-3 in all non-diabetic and diabetic mice. At day-10, all mice groups had a de- creased MCP-1 expression but, nonetheless, it was increased in WT-diabetic. SP treatment increased MCP-1 expression at Day-3 in both WT-nondiabetic and diabetic wounds. At Day-10, SP treatment reduced MCP-1 expression in both WT-nondiabetic and diabetic mice while CJ-treated mice had high- er expression when compared to SP-treated mice. Moreover, diabetic mice showed an increase in IL-6 expression at baseline while SP induced a further increase in IL-6 gene expression in WT-nondiabetic mice at Day-3. NK1RKO and TAC1KO mice showed an increase in IL-6 expression at baseline. KC gene expression was also increased in WT diabetic mice at baseline and this increase persisted at Day-10. SP induced an acute increase at Day-3 that was followed by a reduction to normal levels at Day-10. NK1RKO and TAC1KO mice had also higher KC gene expression at baseline and NK1RKO mice showed a higher KC expression at Day-3. Conclusion: In conclusion, the reduction in SP availability results in a pro- inflammatory activation of skin macrophages before wounding. SP also plays a major role in shifting macrophages to the M2 activation promoting wound healing. Furthermore, SP induced an increase in the acute inflammatory phase of wound healing during the early stages of wound healing but con- siderably reduces it in later stages, which is essential for a good skin repair. Supported by: NIH 1R01NS066205, NIH 1R01DK076937, NIH 1R01NS046710, EFSD, FCT-PTDC/2010 252 Living human keratinocytes as a therapeutic option to improve diabetic ulceration D. Maucher, C. Schürmann, I. Goren, S. Frank; Pharmazentrum Frankfurt, Germany. Background and aims: Diabetes-associated foot ulcerations (DFU) repre- sent a serious clinical complication of wound healing. To date, pharmaco- logical approaches to improve diabetic wound healing disorders still remain poor and limited. The lack of knowledge for underlying molecular and cel- lular mechanisms of impaired wound healing particularly adds to this un- satisfactory clinical condition. Here we used human keratinocytes (BAWD, bio-active wound dressing) as a dynamic approach to deliver a living source of multiple wound-induced mediators for treatment of highly disturbed heal- ing conditions in severely diabetic mice. Materials and methods: BAWD consists of human keratinocytes cell line (KCBl1). The cells were cultured on a hyaluronic acid matrix. The KCBl 1 cell line represents well-characterized, safe, and highly proliferative keratinocytes, which have been initially isolated from human foreskin. For wound healing studies, six full-thickness wounds (5 mm in diameter, 3-4mm apart) were made on the back of severely diabetic C57BL/6J-db/db mice. Upon injury, mice received a coverage of skin wounds using hyaluronic acid matrix alone (vehicle) or living keratinocytes (BAWD, therapeutic). Skin biopsy specimens were obtained from the animals at day 3, 7 and 10 after injury. Protein and RNA isolation was performed and wound biopsies were fixed for histology. The transcriptome of hyaluronic acid- and BAWD-covered wound tissue was analyzed by direct mRNA sequencing. Paraffin sections (4µm) were stained for collagen deposition (azan trichrome) or incubated using antisera against CD31, VEGF or myosin heavy chain 7. Results: BAWD strongly improved diabetes-disturbed wound healing. BAWD-covered wounds showed a very robust formation of granulation tis- sue in the presence of hyperglycaemia. New tissue was characterized by high expression levels of angiogenic VEGF. Induced expression of the endothelial marker Tie-2 and a high density of newly formed blood vessels defined newly formed tissue. Moreover, gene expression analyses of BAWD-improved wound tissue showed a keratinocyte-mediated induction of a large set of genes, which are known to be specifically involved in skeletal and heart mus- cle development, differentiation and function. Improved wound tissue upon BAWD-treatment also showed combined expression levels of the mesenchy- mal stem cell markers Sca-1 and CD29. Conclusion: BAWD markedly improved wound healing in diabetic mice by accelerating the formation of new tissue at sites of diabetes-impaired wounds repair. Supported by: GRK 1172 Diabetologia (2014) 57:[Suppl1]S1–S564 S 113 1 C OP 46 New approaches and clinical outcomes of islet or pancreas transplantation 253 Macro-encapsulated human embryonic stem cell-derived implants meet characteristics of clinical human islet cell grafts E. Motté1, E. Szepessy1, K. Suenens1, G. Stangé1, M. Bomans2, D. Jacobs-Tulleneers-Thevissen1, Z. Ling1, E. Kroon3, D. Pipeleers1; 1Diabetes Research Center, 2Beta-Cell nv, Jette, Belgium, 3Viacyte Inc., San Diego, USA. Background and aims: Shortage of good quality human pancreases for use in organ and islet cell transplantation has led to development of large-scale laboratory sources that generate insulin-producing implants. Pancreatic en- docrine cells can be formed from human embryonic stem (huES) cells fol- lowing in vitro derivation to pancreatic endoderm (PE) and further differen- tiation in immune-incompetent mice. This study examines to which extent huES-generated implants meet characteristics of human islet cells as used in clinical transplantation. Materials and methods: Free, alginate-micro- and Theracyte-macro-encap- sulated huES-derived PE-grafts of similar size were implanted in mice with severe-combined-immune-deficiency. Their in vivo differentiation was com- pared through plasma human C-peptide levels and cellular analysis of im- plants at posttransplant week 20. Implants retrieved from Theracyte-macro- devices were analyzed for glucose-regulated hormone synthesis and release. Data were compared with those collected for human islet cells. Results: Endocrine enrichment was higher in encapsulated than free hu-ES- implants or in clinical-grade human islet cell grafts, with enrichment in single hormone-positive alpha, beta and delta cells in macro-devices and alpha cells in microcapsules. Macro-huES-implants resulted in higher plasma human C- peptide than free human islet cell transplants with similar cell number. They exhibited equally rapid insulin- and glucagon-secretory responses to increas- ing and decreasing glucose concentrations during perifusion. Their insulin secretory amplitude to glucose was lower, in part attributable to a lower cellu- lar hormone content; this was associated with lower glucose-induced insulin biosynthesis, but not with lower glucagon-induced stimulation of release. Conclusion: Macro-encapsulated huES-derived implants can reach an endo- crine composition and function with therapeutic potential. Their beta cells exhibit rapid glucose responsiveness with lower amplitude, compatible with an immature functional state. Comparative data with clinical-grade human islet cell grafts set references for further development and clinical translation. Supported by: EC-FP7, JDRF encapsulation consortium, Flemish Government, FWO 254 Impaired processing of human pro-islet amyloid polypeptide (proIAPP) promotes early islet graft failure J.A. Courtade, P.C. Orban, C.B. Verchere; Pathology and Laboratory Medicine, The University of British Columbia, Vancouver, Canada. Background and aims: Islet transplantation is a promising therapy for type 1 diabetes. Long-term survival of islet grafts may be compromised not only by immune rejection but also other factors including the formation of islet amyloid from islet amyloid polypeptide (IAPP). IAPP aggregation has been linked to beta-cell dysfunction, apoptosis, and islet inflammation but the underlying mechanism of islet amyloid formation remains elusive. Elevated circulating proinsulin:insulin ratios have been detected in type 2 diabetes and in type 1 diabetic recipients of islet transplants. Processing of the IAPP precursor, proIAPP, may be similarly compromised in islet transplants and type 2 diabetes, since like proinsulin, proIAPP is processed by the beta cell prohormone convertases PC1/3 and PC2. We hypothesize that accumulation of proIAPP intermediates leads to amyloid formation and beta-cell dysfunc- tion and assessed the impact of impaired proIAPP processing in transplanted islets. Materials and methods: NOD.SCID mice (age 8-10 weeks) were made dia- betic with streptozotocin and transplanted with islets from mice with beta- cell expression of human proIAPP and lacking PC2 (hIAPPTg/0; PC2-/-) or controls (hIAPPTg/0;PC2+/-, hIAPP0/0;PC2-/-, hIAPP0/0;PC2+/-). Blood glucose was monitored weekly in recipients for 16 weeks or until graft failure (re- turn of hyperglycaemia). Islet graft sections were stained for insulin, amyloid deposition (thioflavin S) and TUNEL positivity. Results: Diabetes (blood glucose >16.9 mM) returned in almost all (90%) of recipients of hIAPPTg/0;PC2-/- islets within 16 weeks post-transplant, whereas recipients with normal human proIAPP processing (hIAPPTg/0;PC2+/-) re- mained normoglycemic for the entire 16 weeks. Recipients of grafts express- ing native rodent IAPP, with or without PC2, similarly maintained normogly- cemia. Despite early graft failure in recipients of islets with impaired human proIAPP processing, beta-cell mass remained unchanged and amyloid depo- sition was lower compared to human proIAPP-expressing mice with normal processing. Islet grafts had few TUNEL-positive beta cells and no difference in the number of TUNEL-positive cells among groups, suggesting that im- paired human proIAPP processing compromises beta-cell function rather than promoting beta-cell apoptosis. Conclusion: Impaired human proIAPP processing in islet transplants induc- es early graft failure without affecting beta-cell mass or increasing amyloid formation. These data suggest that impaired proIAPP processing promotes dysglycemia, possibly via the formation of early (pro)IAPP aggregates that in- duce beta-cell dysfunction. Our work implicates IAPP precursors as potential targets to prolong islet graft survival. Supported by: CIHR Transplant Training Program 255 In type 1 diabetic patients pancreas transplanted simultaneously with kidney preserves long-term kidney graft ultrastructure better than kidney transplantation alone J.P.H. Lindahl1, F. Reinholt2, I.A. Eide1, A. Hartmann1, O. Øyen1, T. Jenssen1,3; 1Department of Transplant Medicine, 2Department of Pathology, Oslo University Hospital, Norway, 3University of Tromsø, Norway. Background and aims: In type 1 diabetic patients with end-stage renal dis- ease (ESRD) we aimed to determine whether long-term normoglycaemia, as achieved by successful simultaneous pancreas and kidney (SPK) transplan- tation, would preserve kidney graft structure, determined by morphometry and electron microscopy (EM), and function compared with live donor kid- ney (LDK) transplantation alone. Materials and methods: Estimated GFR (eGFR) was calculated in 25 recipi- ents with SPK and 17 recipients with LDK in a stable phase 3 months after transplantation and annually during follow-up. Ultrasound guided kidney graft biopsies were obtained from all participants at follow-up. All biopsies were processed for light microscopy to measure glomerular volume, and EM to obtain data on glomerular basement membrane (GBM) width, mesangial fractional volume and podocyte coverage of the capillary basement mem- brane. Results: SPK and LDK recipients were similar in age and diabetes duration at engraftment. The median duration of follow-up was 10.1 years. HbA1c was 5.6±0.4 % and 8.3±1.7 % in the SPK and LDK group, respectively (p S 114 1 C 256 Coronary artery disease in type 1 diabetic patients long-term after simultaneous pancreas and kidney transplantation compared with kidney transplantation alone T. Jenssen1,2, J.P.H. Lindahl1, A. Hartmann1, K. Endresen3, A. Günther4; 1Department of Transplant Medicine, Oslo University Hospital, 2University of Tromsø, 3Department of Cardiology, 4Department of Radiology, Oslo University Hospital, Norway. Background and aims: Improved long-term glycaemic control protects against development of cardiovascular disease and death in type 1 diabetic patients. We aimed to determine whether long-term normoglycaemia, as achieved by successful pancreas and kidney transplantation (SPK), would change the development of coronary artery disease (CAD) compared with successful live donor kidney transplantation alone (LDK). Materials and methods: Twenty type 1 diabetic patients who had received SPK grafts were compared with 11 recipients of LDK. All patients included underwent baseline (pre-transplant) and follow-up coronary angiography. Computed tomography (CT) examination was performed at follow-up to ob- tain coronary artery calcium scoring (CACS). Logistic regression was used to evaluate association between CACS and CAD. Results: Recipients of SPK grafts and LDK were similar in terms of age and diabetes duration at engraftment. The median duration of follow-up was 9.2 years. HbA1C was 5.6±0.4 % and 8.4±2.0 % in the SPK and LDK group, re- spectively (p0.3% and the top 10% >0.6%. The lower 25% had little or no progression. When comparing the highest vs low- est quartile, predictors of rapid progression at baseline were younger age at diagnosis (mean 58.2 vs 67years, p S 115 1 C group characteristics were determined and the prevalence of microvascular complications over time was investigated. Materials and methods: 5423 T2DM patients with at least two HbA1c fol- low-up measurements were selected out of 9849 T2DM patients from the primary Diabetes Care System cohort. The mean follow-up period was 5.7 years (range 2 to 9 years). Latent Class Growth Modelling (LCGM) was per- formed to identify subgroups of patients with distinct trajectories of HbA1c levels. Multinomial logistic regression analyses were conducted to determine subgroup characteristics. Associations of different subgroups with retinopa- thy, microalbuminuria and medication use during follow-up were studied by constructing plots and by graphically comparing differences between sub- groups. Results: Four subgroups with distinct trajectories of HbA1c levels were identified (Figure). The first and largest subgroup (83%) maintained good glycemic control over time (HbA1c ≤ 53 mmol/mol), the second subgroup (8%) initially showed severe hyperglycaemia, but reached the recommended HbA1c target within 2 years. Patients within this subgroup had significantly higher baseline HbA1c levels but were otherwise similar to the good glycemic control subgroup. The third subgroup (5%) showed hyperglycaemia and a de- layed response without reaching the recommended HbA1c target. The fourth subgroup (3.0%) showed deteriorating hyperglycaemia over time. Patients within the last two subgroups were significantly younger, had higher baseline HbA1c levels and had a longer diabetes duration at baseline. These subgroups also showed a higher prevalence of retinopathy and microalbuminuria com- pared to the good glycemic control subgroup. Conclusion: This is the first study identifying subgroups with distinct tra- jectories of HbA1c levels in a T2DM cohort. More than 90% of the patients reached and maintained good glycemic control. Two subgroups showed a more unfavorable course of glycemic control. These T2DM patients were younger, had higher HbA1c levels and a longer diabetes duration at baseline. 259 Statistical power enhancement in diabetic drug trial by selectively sampling participants from the tails of biomarker distributions: a DIRECT study A. Ashfaq1, N.A. Pasdar1, A. Kurbasic1, D. Shungin1, E.R. Pearson2, P.W. Franks1; 1Lund University Diabetes Research Centre, Department of Clinical Sciences, Lund University, Malmö, Sweden, 2Division of Cardiovascular & Diabetes Medicine Medical Research Institute, University of Dundee, UK. Background and aims: Within the IMI DIRECT Study we are seeking to identify biomarkers that modify treatment response, primarily using obser- vational studies to generate hypotheses that will be validated in clinical trials. Here, we compared statistical power to detect gene-metformin interactions in two types of trials: approach ‘A’ represents interactions evaluated in an exist- ing clinical trial that involves randomly selected participants; approach ‘B’ involves selecting participants on the basis of discordant levels of genetic risk score (GRS) comprised of 20 variants that modify treatment effects (geno- type-based recall: GBR). Materials and methods: Model assumptions were based on published data from the Diabetes Prevention Program, a randomized controlled trial of met- formin treatment vs. placebo. We simulated GRS-metformin interactions, as- suming the 20 GRS variants convey comparable gene-metformin interaction effects to those reported for the MATE1 locus. The effects of the metformin and placebo interventions were estimated assuming randomised treatment allocation. To simulate approach ‘A’, participants were randomly sampled from a larger population-sampling frame (N=10,000-50,000), as might be the case in a conventional clinical trial. To simulate approach ‘B’, participants were purposefully sampled at random and from the upper tail of GRS distri- bution from the same sampling frames. Time to diabetes event was calculated using the proportional hazards function. Sample size calculations were per- formed using 1000 iterations per simulation in the R software package. Results: Figure 1 shows representative results from these analyses: to obtain 80% power to detect the specified gene x metformin interaction effect, the GBR approach would require 300-400 participants, whereas the random- sampling approach would require roughly 800 participants for rare alleles (Fig. 1a). For common alleles a sample size of ~250 would suffice with the GBR approach compared with sample size of ~650 for the random sampling approach. Similarly, where interaction effects are smaller in magnitude, the difference in sample-size requirements for the random vs. the GBR sampling approach increases substantially. The size of the sampling-frame also affects power (Fig. 1b). Conclusion: The GBR approach for validating observations of gene-drug interaction offers a potentially powerful alternative to conventional clinical trials, which is especially appealing when gene variants are rare and pheno- typing is expensive. Supported by: EC and EFPIA. Diabetologia (2014) 57:[Suppl1]S1–S564 S 116 1 C 260 Antidepressant medication use and trajectories of fasting plasma glucose and HbA1c levels: a 9-year longitudinal study of the DESIR cohort M. Azevedo Da Silva, A. Dugravot, B. Balkau, A. Singh-Manoux, H. Nabi; CESP - U.1018, INSERM, Villejuif, France. Background and aims: Some research has suggested that antidepressant medication use (AMU) may increase the type 2 diabetes (T2D) risk. However the causal mechanism of this association remains unclear. To study the bio- logical plausibility of this relationship, we examined the association between AMU and changes in fasting plasma glucose (FPG) and in HbA1c levels over time. Materials and methods: Participants were 4869 men (49.2%) and women, free of T2D and aged 30-65 years at baseline, followed for 9 years between 1994 and 2005 in the French D.E.S.I.R. cohort study. AMU, FPG and HbA1c were assessed concurrently at 4 medical examinations in phases 1 (1994- 1996), 2 (1997-1999), 3 (2000-2002) and 4 (2003-2005). Linear mixed models were used to examine the longitudinal associations of AMU with FPG and HbA1c levels. Results: FPG and HbA1c levels increased over follow-up. In model adjusted for sociodemographic characteristics, health-related behaviours, medical conditions and other medications, there was no difference in mean FPG and HbA1c levels (β=-0.001 p=0.974, β=0.012 p=0.662, respectively) among an- tidepressant medication users and non-users at baseline (1994-1996). The interaction term between time and AMU, suggested no increase in FPG and HbA1c levels in antidepressant users comparatively to non-used over 9-year of follow-up (β=-0.004 p=0.535, β=-0.002 p=0.548 respectively). Similar pat- terns of associations were observed when the type and the cumulative use of antidepressants over follow-up were considered. Conclusion: Our results suggest that AMU is not associated with an increase in FPG and HbA1c levels over 9-year of follow-up. This suggests that the nature of the association between AMU and T2D risk, observed in previous studies, may not be causal. Supported by: The Ministry of Research (France) OP 48 Novel targets for anti-inflammatory therapies 261 Exogenous adiponectin administration through a subcutaneous minipump reverses high-fat diet-induced impairment of adipose tissue metabolism P. Matafome1, T. Rodrigues1, A. Pereira1, L. Letra1, H. Azevedo2, A. Paixão1, M. Silvério1, A. Almeida2, C. Sena1, R. Seiça1; 1Faculty of Medicine, University of Coimbra, 2Hitag, Biocant, Cantanhede, Portugal. Background and aims: Despite adiponectin administration is thought to be a future strategy to obese or type 2 diabetic patients, it still is a myth, due to its expensive costs and absence of studies demonstrating the effective- ness of chronic exogenous adiponectin. Our aim was to develop a method to produce adiponectin for long-term adiponectin administration and test its usefulness in improving metabolic profile and adipose tissue metabolism in high-fat diet fed rats. Materials and methods: Adiponectin (98 μg/day) was administered during 28 days through a subcutaneous minipump with continued release to normal Wistar rats fed a high-fat diet. Several systemic markers of dysmetabolism and pathways of insulin signaling and lipid storage in epididymal and subcu- taneous adipose tissue were assessed. Results: Exogenous adiponectin was able to decrease body weight, fasting glycaemia, HbA1c and cholesterol levels (total and non-HDL). In adipose tis- sue, adiponectin reverted high-fat diet-induced impairment of insulin sign- aling and decreased IκBα and PPARγ levels, which were mainly observed in epididymal adipose tissue. High-fat fed rats showed little activation of lipoly- sis during fasting in epididymal adipose tissue, despite no alterations were observed in the total amount of the protein. This was only partially reverted by adiponectin. Conclusion: Long-term adiponectin administration through a subcutaneous minipump was able to improve pathways of insulin signaling and lipid stor- age in adipose tissue after the consumption of a high-fat diet. Adiponectin was also able to improve the metabolic prolife, probably as a result of im- proved adipose tissue metabolism. Supported by: FCT 262 Immuno-neutralisation of extra-cellular nicotinamide phosphoribosyltrasferase as a therapeutic strategy for treatment of type 2 diabetes P.W. Caton1, J. Kieswich2, M. Silvestre1, M.M. Yaqoob2; 1Centre for Diabetes, Blizard Institute, Queen Mary London, 2Translational Medicine and Therapeutics, William Harvey Research Institute, Queen Mary London, UK. Background and aims: The protein extra-cellular nicotinamide phosphori- bosyltransferase (eNAMPT; also termed visfatin/PBEF) is elevated in serum of type 2 diabetes (T2D) patients and raised eNAMPT levels are reported to strongly correlate with declining beta-cell function. Moreover, eNAMPT re- portedly exerts pro-inflammatory effects suggesting a potential pathophysi- ological role for eNAMPT in T2D. However, the precise functional relation- ship between elevated serum eNAMPT and the pathophysiology of T2D has not been determined. Interestingly, eNAMPT exists in serum as a monomer and a dimer, suggesting a potential structure/function relationship, although the precise function and specific changes in each form in T2D have not been determined. Materials and methods: To investigate the role of eNAMPT in T2D, C57Bl/6 mice were fed a high-fat diet (HFD; 60% fat) for 10 weeks. In weeks 9 - 10, mice were administered polyclonal anti-NAMPT antibody (2.5 ug/ml; 2 dos- es/week) or saline equivalent. After week 10, mice were fasted for 16 h prior to glucose tolerance testing (2 g/kg body weight glucose) or sacrificed in the fed state for tissue (islets, liver and white adipose tissue; qPCR and western blot) and serum analysis Results: Serum eNAMPT monomer levels were elevated in HFD mice, whilst dimer levels were unchanged, as assessed by Native-PAGE immunoblot. Immuno-neutralization of eNAMPT corrected HFD-mediated fasting and fed hyperglycaemia, hyperinsulinemia and impaired glucose tolerance. These Diabetologia (2014) 57:[Suppl1]S1–S564 S 117 1 C effects were mediated through improvements in beta-cell health (increased ex vivo and in vivo islet glucose-stimulated insulin secretion and increased islet size) and correction of hepatic insulin resistance (reduced gluconeogen- ic gene expression, increased insulin signalling). Consistent with a pro-in- flammatory function of eNAMPT, improvements in glycaemia and beta-cell health following eNAMPT-immuno-neutralization were linked to lowering of HFD-mediated increases in gene expression of islet and white adipose tis- sue pro-inflammatory cytokines, chemokines and immune cell markers, as well as decreased mRNA levels of islet pro-apoptotic markers. Conclusion: Elevated serum eNAMPT monomer levels may contribute to pathophysiology of T2D, through pro-inflammatory effects. Moreover, selec- tive neutralization or inhibition of eNAMPT monomer represents a potential therapeutic target for treatment of T2D. Supported by: PWC is a recipient of an EFSD/Lilly Research Fellowship 263 TLR4 inhibition blocks cytokine production and restores beta cell survival in human pancreatic islets W. He1, C. Schuster-Klein2, B. Guardiola-Lemaître2, K. Maedler1; 1Centre for Biomolecular Interactions, University of Bremen, Germany, 2ADIR-Servier Group, Suresnes, France. Background and aims: Inflammatory signals are strong mediators of meta- bolic failure in fat, liver, brain, muscle as well as in pancreatic islets. Toll-like receptor-4 (TLR-4) signaling is one of the major pro-inflammatory pathways, whose ligands as well as downstream products, i.e. cytokines and chemokines, are increased systemically in patients with T2D as well as in at-risk individu- als. TLR4 knockout mice are protected from the metabolic consequences of a high fat diet. In the present study we investigated the consequences of TLR4 activation in human islets and whether pharmacological inhibition of TLR4 restores beta-cell survival in a diabetogenic milieu. Materials and methods: Isolated human islets were exposed to the TLR4 li- gand Lipopolysaccharide (LPS) in the presence or absence of TLR4 receptor small molecule antagonist, TAK-242 (Resatorvid). The effect of TLR4 inhibi- tion was compared with those achieved by the depletion of islet resident mac- rophages by treatment of human islets with clodronate liposomes. Expression and secretion of pro-inflammatory cytokines/chemokines were evaluated by RT-PCR and ultrasensitive ELISA. TLR4 downstream signaling activation and apoptosis were analyzed by Western Blotting. Results: Treatment of isolated human islets with LPS elevated mRNA levels of pro-inflammatory cytokines/chemokines, including IL-1alpha, IL-1beta, IL- 6, TNFalpha, CCL2 and IL-8. LPS induced IL-1beta secretion was confirmed by ELISA. TAK-242 completely blocked LPS-induced cytokine/chemokine expression and IL-1β secretion, while clodronate treatment only inhibited IL-1alpha/beta but not IL-6, IL-8, TNFalpha and CCL2 expression. TAK-242 could completely antagonize the effects of LPS on TLR-4 downstream signal- ing and apoptosis in human islets, it inhibited phosphorylation of inhibitor of nuclear factor kappa B alpha (IkBα) and c-Jun N-terminal kinase (JNK) as well as caspase-3 cleavage. Conclusion: Our results show that TLR4 activation in human islets leads to islet inflammation and beta-cell apoptosis, which results in beta-cell failure, while the TLR4 inhibitor TAK-242 inhibits cytokine production and restores beta-cell survival. Supported by: the European Research Council 264 Intervention with Caspase-1 inhibitor attenuates the metabolic syndrome and prevents non alcoholic-steatohepatitis (NASH) in high fat diet-fed LDLR-/-.Leiden mice P.Y. Wielinga, A.M. van den Hoek, K. Salic, W. Liang, T. Kooistra, R. Kleemann; Metabolic Health Research, Netherlands Organization for Applied Scientific Research (TNO), Leiden, Netherlands. Background and aims: Non-alcoholic steatohepatitis (NASH) is a serious liver pathology which develops as a complication of the metabolic syndrome. Recently, the inflammasome was proposed to be involved in the development of NASH. Here we investigate whether a chemical inhibitor of caspase-1 in already manifest metabolic syndrome would prevent the development to NASH. Materials and methods: Male LDLR-/-.Leiden mice were fed a high fat diet (HFD; group 1) or low fat diet (LFD; group 2) for 21 weeks. In a third group, intervention with caspase-1 inhibitor Ac-YVAD-CMK (40 mg/kg daily) was started after 9 weeks of HFD (manifest metabolic syndrome) and continued until 21 weeks. Results: HFD treated mice developed obesity and insulin resistance after al- ready 9 weeks of HFD feeding. Intervention with caspase-1 inhibitor attenu- ated a further development of insulin resistance, and reduced body weight gain as well as adipose tissue inflammation compared to HFD. Histopatho- logical analysis of the livers clearly demonstrated prevention of NASH de- velopment with caspase-1 inhibitor : livers were less steatotic and neutrophil infiltration was diminished. Additionally, hepatic fibrosis quantified by sirius red staining and acta2 and col1a1 gene expression as observed in HFD treat- ed mice, was completely prevented. Conclusion: Intervention with a caspase-1 inhibitor in already established disease improved hallmarks of the metabolic syndrome and prevented the development of NASH. Our data further support the importance of cas- pase-1/inflammasome in the development of NASH and demonstrate that therapeutic intervention in the already ongoing disease process is feasible. Diabetologia (2014) 57:[Suppl1]S1–S564 S 118 1 C PS 001 Pragmatic prediction and prevention of type 2 diabetes 265 Family history of diabetes is a strong predictor of diabetes, hypertension and metabolic syndrome in Sri Lankan adults P. Ranasinghe1, R. Jayawardena2, P. Katulanda3; 1Department of Pharmacology, Faculty of Medicine, Colombo, Sri Lanka, 2Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia, 3Department of Clinical Medicine, Faculty of Medicine, Colombo, Sri Lanka. Background and aims: Diabetes mellitus has become an important health concern in South Asia. Family history (FH) is a common non-modifiable risk factor for most of the chronic non-communicable diseases, as it is a collective reflection of the genetic susceptibility, shared environments and behaviors. Information on FH may serve as a unique and useful tool for public health and preventive medicine. The increased genetic predisposition amongst South Asians probably makes FH more important in risk assessment than in other ethnic groups. The present study aims to describe the influence of FH on diabetes prevalence and associated metabolic risk factors in a nationally- representative sample of South Asian adults from Sri Lanka. Materials and methods: A cross-sectional community based national survey was conducted in 7 of the 9 provinces in Sri Lanka. Five thousand adults were recruited for the study using a multi-stage stratified cluster sampling tech- nique. An interviewer administered questionnaire was used to collect data, which included; age, gender, area of residence, ethnicity, level of education, household monthly income, duration of diabetes, FH, height, weight, waist circumference, hip circumference and blood pressure. FH was evaluated at three levels, a) parents, b) grandparents (paternal and maternal) and c) sib- lings. Presence of diabetes in children was also evaluated. A binary-logistic regression analysis controlling for confounders (age, gender, body mass index and physical activity) was performed in all patients with ‘presence of diabetes’ as the dichotomous dependent variable and using FH in father, mother, ma- ternal grandmother/grandfather, paternal grandmother/grandfather, siblings and children as binary independent variables. A p value 15. The AUC for T2D was 0.766, p < 0.001, whereas for AGT the area under the curve was 0.663, p < 0.001. Using the optimal cut-off value of 10 with the largest area under the curve to identify previously undiagnosed pa- tients with T2D, AGT resulted in a sensitivity of 91%, 67% and specificity of Diabetologia (2014) 57:[Suppl1]S1–S564 S 119 1 C 52%, 59%. To achieve this, only 50% of the total population would need to be screened. Increasing the cutoff value of the score to 15 changed the sensitivity to 38%, 25% and specificity to 86%, 90% respectively. Apart from total cho- lesterol and LDL cholesterol, several risk factors for cardiovascular disease: age (p < 0.001), BMI (p < 0.001), WC (p < 0.001), systolic and diastolic blood pressure (p < 0.001), fasting and 2h plasma glucose (p < 0.001), triglycerides (p = 0.016), HDL cholesterol (p = 0.003) had a direct association with the FINDRISK values. Conclusion: FINDRISC proved to be simple and effective test to screen sub- jects at high risk for T2D. With the optimal cut-off level, the FINDRISC iden- tified 91% undetected T2D and 67% AGT with relatively high PPV 48% for AGT, but only 11% of PPV for new T2D. 268 Trends in leisure time physical activity in Danish adults with diabetes S. Molsted1, N.F. Johnsen2, O. Snorgaard3; 1Department of Cardiology, Nephrology & Endocrinology, Nordsjællands Hospital, Hillerød, 2National Institute of Public Health, University of Southern Denmark, 3Department of Endocrinology, Hvidovre Hospital, Copenhagen, Denmark. Background and aims: During the last decades there has been an increased focus on physical activity and exercise training in treatment of diabetes. The primary aim of this study was to compare leisure time physical activity (PA) reported in 2000, 2005 and 2010 by Danish subjects with diabetes. Further- more, we analysed the subjects’ smoking, alcohol consumption, and body mass index (BMI). Materials and methods: Data comprised level of leisure time PA in four categories: physical inactivity, moderate activity, medium activity, and high activity; smoking; alcohol consumption; and BMI provided by The Danish Health and Morbidity Surveys, nationwide surveys (n=~15,000) from the general population. Subjects older than 45 years at the time of the surveys were included from cross-sectional analyses from 2000, 2005 and 2010. Results: The diabetes prevalence was 4.5% (n=386) in 2000, 6.0% (n=489) in 2005, and 7.5% (n=622) in 2010. In subjects with diabetes, percentages of inactive women decreased from 42.2% to 21.8% (p S 120 1 C 271 Sex-specific differences in prevention of type 2 diabetes mellitus: a systematic review and meta-analysis J. Harreiter1, A. Glechner2, G. Gartlehner2,3, A. Kautzky1, S. Rohleder1, M. Van Noord2, A. Kaminski-Hartenthaler2, J. Tuomilehto4,5, A. Kautzky-Willer1; 1Department of Medicine III, Divison of Endocrinology & Metabolism, Gender Medicine Unit, Medical University Vienna, 2Department for Evidence-based Medicine and Clinical Epidemiology, Danube University, Krems, Austria, 3Research Triangle Institute (RTI) International, North Carolina, USA, 4Diabetes Prevention Unit, National Institute for Health and Welfare, Helsinki, Finland, 5King Abdulaziz University, Jeddah, Saudi Arabia. Background and aims: Type 2 diabetes mellitus (T2DM) is becoming an increasing challenge worldwide. Lifestyle interventions and several glucose lowering medications have been proved to be effective in preventing the progression to T2DM in people with prediabetes. The ADA/EASD position statement 2012 claims a patient centred approach, in which sex and gender aspects are relevant aspects. Sex-specific differences of the efficacy and safety of diabetes prevention strategies have not been studied so far. Materials and methods: A systematic search of PubMed, Cochrane, Embase, CINAHL Web of Science, and reference lists of review articles from 1980 to June 2013 was performed. Literature was dually reviewed and risk of bias was rated for each study. Only randomised controlled trials with at least 1-year follow-up that compared active lifestyle interventions with usual care and pharmacotherapy trials with glucose lowering drugs that compared active treatment with placebo were included. Random effects meta-analysis of out- comes of interest was performed. Results: We identified 2548 relevant abstracts; 309 full-text articles were re- trieved for further examination. Eighteen RCTs (44articles) met the eligibility criteria. Only 3 published sex-specific results; of 9 RCT´s unpublished sex- specific data was supplied upon request. A similar efficacy of risk or harms of lifestyle interventions and pharmacological therapies between men and women was found. Lifestyle interventions reduced the risk to develop T2DM in prediabetic people in both sexes. After 1 year lifestyle intervention group had a lower risk than patients in control group (RR:0.60; 95% CI:0.35-1.05; 4RCTs,888patients). A 37% lower risk of progressing to T2DM for pre-dia- betic people receiving an average of 3 years of lifestyle intervention compared with those under usual care (RR:0.63; 95 CI:0.51-0.79; 5RCTs,1555patients) was found. Similar preventive effects of lifestyle interventions were found in men and women after 1 and 3 years. (p=0.61, p= 0.20). Only few studies pro- vided sex-stratified results of pharmacological therapies in T2DM preven- tion. Long-term evidence on potential sex differences of diabetes-associated comorbidity and mortality is missing. Conclusion: The evidence indicates that intensified lifestyle intervention in structured programs is equally effective in both sexes. Therefore individually designed interventions to increase physical activity and healthy nutrition by behavioral modification for effective diabetes prevention should be offered for both sexes. More studies to reveal sex and gender-specific differences are necessary to follow a more patient centred approach. Clinical Trial Registration Number: PROSPERO 2012:CRD42012003102 Supported by: BGM 12013 272 Beneficial effect of pitavastatin on the incidence of diabetes in women was not associated with age: sub-analysis of J-PREDICT T. Shiba1, K. Sakamoto1, C. Ito2, T. Yamazaki3, J. Kishimoto4, M. Noda5, Y. Terauchi6, M. Odawara7, H. Kitazato8, K. Maemura9, K. Tobe10, Y. Iwamoto11, Y. Akanuma12, T. Kadowaki13, the J-PREDICT study investigators; 1Toho University, Tokyo, 2Grand Tower Medical Court, Hiroshima, 3The University of Tokyo Hospital, 4Kyushu University Hospital, Fukuoka, 5National Center for Global Health and Medicine, Tokyo, 6Yokohama City University Graduate School of Medicine, Kanagawa, 7Tokyo Medical University, 8Omori Red Cross Hospital, Tokyo, 9Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, 10University of Toyama, 11Tokyo Women‘s Medical University, 12The Institute for Adult Diseases, Asahi Life Foundation, Tokyo, 13The University of Tokyo, Japan. Background and aims: Diabetes mellitus (DM) increases the risk of coro- nary heart disease (CHD) far more in women than men. If statin therapy can prevent or delay the development of diabetes, it is helpful for prevent- ing CHD, especially in women. Thus, we conducted a sex-specific analysis in J-PREDICT comparing the effect of pitavastatin (PIT) on the incidence of diabetes in woman versus men. Materials and methods: J-PREDICT study was a prospective randomized, open-label, blinded-endpoint trial evaluating the effect of PIT on the inci- dence of diabetes (n=1,269; 635 Control [lifestyle modification alone], 634 PIT [PIT 1-2 mg + lifestyle modification]) in Japanese subjects with impaired glucose tolerance (IGT). The primary outcome was incidence of DM defined as 2-h plasma glucose of ≥200 mg/dl or fasting plasma glucose of ≥126 mg/dl measured at least once in 75g OGTT performed every six months. This study was performed based on the full analysis set (total, n=1090; men, n=683; women, n=407). Results: Women were older than men and lower in body mass index. Women and men had similar levels of glycemic and insulin-related parameters. The incidence of diabetes in women was lower by 32% (HR, 0.68; 95%CI, 0.49- 0.93; p=0.02) in the PIT group compared with control group, but not sig- nificantly different between the group in men (HR, 0.97; 95%CI, 0.78-1.21; p=0.78). The analyses using stratified log-lank test and Cox proportional haz- ard model also brought a positive PIT effect only in women. Trends towards higher incidence of diabetes with increasing age were observed only in men. The beneficial effect of PIT was most in a group of 55 to 64 years in both genders, and only significant in women (p=0.02). Thus, incidence of diabetes in women was not associated with age ( S 121 1 C time clinical outcomes using the UKPDS Outcomes Model. Multiple simula- tions examined the impact of a hypothetical intervention that delayed diabe- tes onset by 1, 3, 5 or 7 years on costs and (quality-adjusted) life expectancy. For simplicity, we assumed that all patients would experience the same period of delay regardless of their characteristics or previous medical history. All future costs and effects were discounted at 3.5% (UK) and 3.0% (US). Results: Subjects were mean (1SD) age 66.3 (6.9) years, HbA1c 43 (7) mmol/ mol, body mass index 31.5 (5.7) kg/m2, and 51% were male. The longer dia- betes onset was delayed the greater was the increase in life expectancy and quality-adjusted life years (QALYs) and the decrease in the costs of diabetic complications. When diabetes onset was delayed by 7 years, undiscounted life expectancy measured from the simulation baseline increased from 13.8 to 15.4 years; discounted QALYs increased from 7.6 (US 8.0) to 8.4 (US 8.8) years, with a corresponding decrease in the discounted costs of complications from £16,741 (US $140,967) to £16,377 (US $140,369). Assuming an annual cost for the hypothetical intervention of £1,000 (US $1,600), the estimated cost per QALY gained when onset was delayed by one year was £7,240 (US $11,546), falling to £6,909 (US $10,251) when onset was delayed by 7 years. If the annual cost for the hypothetical intervention was £3,000 (US $4,800), the estimated cost per QALY gained when onset was delayed by one year was £22,691 (US $34,870), falling to £21,302 (US $32,239) when onset was delayed by 7 years. Conclusion: In our analysis, interventions that can delay the onset of dia- betes substantially increased remaining life expectancy and quality-adjusted survival and slightly reduced remaining lifetime complication costs. At an- nual costs of between £1,000 and £3,000 (US $1,600 and $4,800), the hypo- thetical intervention yielded a cost per QALY gained that remained broadly within the ranges considered cost-effective in the UK and US. Further work is required to explore heterogeneity in the patient population and parameter uncertainty in the modelling. Developing such interventions continues to be a high priority, especially given the continued growth in the global popula- tion with type 2 diabetes. Supported by: Novartis PS 002 Aetiological epidemiological studies of type 2 diabetes 274 Blood glucose levels at baseline and incidence of type 2 diabetes: a prospective cohort study of 0.5 million adults in the China Kadoorie Biobank H. Du1, L. Li2,3, I. Millwood1, F. Bragg1, L. Yang1, Y. Chen1, Y. Guo2, Z. Bian2, J. Chen4, R. Collins1, R. Peto1, Z. Chen1; 1CTSU, University of Oxford, UK, 2Chinese Academy of Medical Sciences, 3School of Public Health, Peking University, 4China National Center for Food Safety Risk Assessment, Beijing, China. Background and aims: Random blood glucose (RBG) levels may be used to screen for diabetes, however the predictive value of RBG levels within the normoglycemic range on future diabetes risk has not been well studied, par- ticularly among Chinese population. Materials and methods: The study population includes 496,720 individuals enrolled into the China Kadoorie Biobank between 2004-8 from 10 diverse localities across China without prior physician diagnosed diabetes. At base- line, RBG was measured by venous blood spot test and those with RBG ≥ 11.1 mmol/L and those fasting blood glucose ≥ 7.0 mmol/L were identified as newly detected diabetes (2.8%) and excluded from the current analysis. Data on type 2 diabetes incidence was collected through electronic linkage with mortality and morbidity registries as well as with the national health insurance system. RBG was related to incidence of type 2 diabetes using Cox proportional Hazard models, with adjustment for potential confounders. Results: The overall mean age of participants (n=474,423) was 51 years, mean BMI was 23.6 kg/m2, and mean RBG was 5.7 mmol/L. During the 7-year follow-up, 4194 incident cases of type 2 diabetes were identified. A dose-re- sponse relationship was observed between baseline RBG and risk of diabetes. Compared to those with RBG < 6.0 mmol/L, those with 6.0 ≤ RBG < 6.9 mmol/L had a diabetes hazard ratio (HR) 2.8 (95% CI: 2.7-2.9), those with 7.0 ≤ RBG < 7.8 mmol/L had HR 3.8 (3.5-4.1) and those with RBG ≥ 7.8 (i.e. pre-diabetes) had HR 6.8 (6.3-7.4). The association of RBG with diabetes risk was slightly stronger in women than in men, and in younger than in older people. Stratifying the analyses by hours since last meal (< 2 hours, 2-4 hours, 4-6 hours, or ≥ 6 hours) revealed a stronger association among those who had fasted for a longer period (HR of RGB ≥ 7.8 mmol/L 12.0 [5.3-27.2]. Conclusion: In this group of adult Chinese, higher RBG levels, even those within the normoglycemic range, constitute an independent risk factor for type 2 diabetes. Such levels should probably be taken into account in identify- ing people at increased risk for diabetes. Supported by: Wellcome Trust; MRC; British Heart Foundation; Cancer Re- search UK; ChinaTRD 275 Stressful life events and the metabolic syndrome: the Hoorn Study F. Rutters1, S. Pilz2, A.D. Koopman1, S.P. Rauh1, S.J. Te Velde1, C.D.A. Stehouwer3, P. Elders4, G. Nijpels4, J.M. Dekker1; 1Department of Epidemiology and Biostatistics, VUmc, Amsterdam, Netherlands, 2Department of Internal Medicine, Medical University of Graz, Austria, 3Maastricht University Medical Centre, Department of Internal Medicine, Netherlands, 4Department of General Practice, VUmc, Amsterdam, Netherlands. Background and aims: The amount of psychosocial stress may play a role in the development of the Metabolic Syndrome. The aim of our current study was to evaluate whether in a population-based cohort of older men and women the number of stressful life events is associated with the Metabolic Syndrome incidence, and whether any such relationship is mediated by be- havioural factors. Materials and methods: The association between the number of stressful life events experienced at baseline and Metabolic Syndrome incidence after 6 years of follow up, was assessed in the Hoorn study. Subjects with the Meta- bolic Syndrome at baseline, defined according to the Adult Treatment Panel III (NCEP), were excluded. Results: We included 1099 participants (47% male; age 60.1±7 years). During 6.5 years of follow-up, 238 subjects (21.7%) developed the Metabolic Syn- drome. Using logistic regression, a positive association was observed between Diabetologia (2014) 57:[Suppl1]S1–S564 S 122 1 C Metabolic Syndrome incidence at follow-up and the number of stressful life events [OR 1.11 (1-1.25)]. However, we also observed effect modification for education level (P value interaction, p=0.01). In the low-education group, the model adjusted for age and sex showed a significant association between Metabolic Syndrome incidence and the amount of stressful life events [OR 1.20 (1.03-1.39)]. No such significant association was observed in the middle/ high-education group [OR 1.05 (0.83-1.24)]. Additionally, a chi-square anal- ysis showed a linear-by-linear association between the number of Metabolic Syndrome abnormalities and the amount of stressful life events (p S 123 1 C eGFR was negative correlated with age (r=-0.3; p 5/hour in an overnight polysomnography were excluded from analyses. Multiple linear regression analyses were performed to explore the associations between sleep and glucose metabolism parameters. Variables with skewed distribution were log-transformed before entering into analyses. Results: A total of 131 children and adolescents were studied. Among them, 13 had moderate-to-severe OSA, leaving 118 subjects eligible for the pre- sent analysis (mean age ± SD: 13.1 ± 3.3 years; male: 44.9%). After adjust- ment for gender, Tanner stage, BMI z-score, and OAHI, sleep duration and sleep efficiency measured by actigraphy were negatively associated with log- transformed 2-hour plasma glucose levels (β ± se: -0.042 ± 0.022, p = 0.054 and -0.009 ± 0.004, p = 0.033, respectively) and positively associated with log-transformed ISSI-2 (β ± se: 0.102 ± 0.049, p = 0.038 and 0.023 ± 0.009, p = 0.014, respectively). There were no significant associations of log-trans- formed ISOGTT (p > 0.20) with sleep duration and sleep efficiency as meas- ured by actigraphy. Moreover, no associations were found between subjective sleep duration and sleep efficiency as measured by sleep diary and all glucose metabolism parameters as mentioned above (p > 0.20) (Table 1). Conclusion: Short sleep duration and low sleep efficiency as measured by actigraphy are independent risk factors for worse glucose tolerance and β-cell function in children and adolescents. Supported by: RGC of Hong Kong SAR (Project no. CUHK 4465/06M) 280 Glycated albumin is a useful indicator for screening impending diabetes and predicting beta cell dysfunction in the pre-diabetic condition S.M. Hong1, S.Y. Rhee1, J. Kim2, Y.C. Hwang1, C.H. Jung3, K.J. Kim4, W.S. Jeon5, S.M. Jin6, B.W. Lee4, J.T. Woo1; 1Department of Endocrinology and Metabolism, Kyung Hee University School of Medicine, 2Department of Internal Medicine, Dongsuwon General Hospital, 3Department of Internal Medicine, Asan Medical Center, University of Ulsan College of Medicine, 4Department of Internal Medicine, Severance Hospital, University of Yonsei College of Medicine, 5Department of Internal Medicine, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 6Department of Internal Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea. Background and aims: Pre-diabetes is known as a pre-clinical stage of in- creased risk for overt diabetes mellitus (DM) and cardiovascular disease. Because glycated albumin (GA) has been suggested to have more potential for assessing insulin secretory dysfunction and glycemic fluctuation than HbA1c, we studied the clinical significance of GA in this stage. Materials and methods: We enrolled the 1379 anti-diabetic drug naïve subjects in retrospective, multi-center, cross-sectional manner. According to the 75-g OGTT, the subjects were classified as normal glucose tolerance (NGT), isolated IFG (i-IFG), isolated IGT (i-IGT), combined glucose intoler- ance (CGI) and DM subgroup. We analyzed clinical characteristics of these 5 groups including GA, insulin sensitivity (HOMA2%S), and insulin secretion (HOMA2%B) index. Results: Mean GA was 11.6±1.4, 12.3±1.8, 12.3±1.9, 13.0±1.9, 18.8±7.9 in NGT (n=295, 21.4%), i-IFG (n=257, 18.6%), i-IGT (n=103, 7.4%), CGI (n=257, 18.6%), and DM (n=466, 34%) subgroup. After adjusting covariates, adjusted mean of GA was 12.2±0.1, 12.2±0.2, 13.1±0.1 in i-IFG, i-IGT, and CGI subgroup (p S 124 1 C 281 A population approach to modelling the oral glucose tolerance test in normal and impaired glucose tolerant states M.J. Theodorakis, G.P. Chrousos; Endocrinology, Metabolism and Diabetes Unit, University of Athens, Greece. Background and aims: The conventional approach to analyzing oral glucose tolerance testing (OGTT) data requires model identification in each individ- ual separately (standard two stage), ignoring knowledge about the population as a whole. However, the OGTT is sparsely sampled and individual estimates are often not resolvable from the available data. Materials and methods: We applied a population approach, nonlinear mixed effects modeling, to plasma glucose, insulin and C-peptide data obtained from a 120-minute OGTT undertaken by 106 subjects forming five groups with varying glucose tolerance. This method provides estimates of popula- tion means, variances and covariances of model parameters, empirical Bayes estimates of individual parameter values and measures of intra-individual (within-subject) and inter-individual (between-subject) variability. The lat- est version of the oral glucose minimal model was used to evaluate insulin sensitivity and a combined model approach was used to assess β-cell secre- tion. These models allowed for the reconstruction of insulin secretion and glucose absorption profiles and gave population indexes of insulin sensitivity (SI=6.51±;1.20×10 -4 min-1•µU-1•ml), fractional hepatic extraction of insulin (F=0.522×0.291) and fractional insulin clearance (ki = 0.258×0.151 min -1). Results: Individuals with type 2 diabetes (T2DM) had significantly higher HOMA-IR, HbA1c and triglyceride levels compared to subjects with normal glucose tolerance (NGT). They also had increased fasting plasma insulin lev- els, significantly reduced insulinogenic index from 0 to 20 minutes after oral glucose, and increased fasting glucagon levels compared to NGT subjects. All had similar fasting plasma biochemical values and were without medical histories of gastroparesis, kidney disease or microvascular disease. The popu- lation model, including data from all groups of individuals, was implemented in NONMEM. When empirical post hoc Bayes estimates obtained from the population approach were used to determine individually predicted values, these points collapsed to the line of unity, showing no systematic deviations from the observed data. NGT subjects were found to have the highest median values for SI, as well as the most variability. Mean estimates and standard er- rors for SI (104min -1•µU-1•ml) in the different groups were: Group 1 (normal fasting and 2-hr plasma glucose; 17.7±;1.8), Group 2 (normal fasting / im- paired 2-hr plasma glucose; 9.95±;2.0), Group 3 (impaired fasting and 2-hr plasma glucose; 5.67±;1.0), Group 4 (normal fasting / diabetic 2-hr plasma glucose; 3.82±;1.1) and Group 5 (diabetic fasting and 2-hr plasma glucose; 9.39±;3.8). Individual insulin secretion profiles were also reconstructed from empirical Bayes estimates. Conclusion: Whereas the traditional approach to parameter estimation failed to recover estimates in more than one third of the population, the pop- ulation approach provided individual estimates in all subjects. Examination of the empirical Bayes estimates showed that individual parameter estimates were able to differentiate well between individuals at glucose tolerant states ranging from euglycemia to overt type 2 diabetes. Our findings suggest popu- lation analysis is a powerful tool to obtain accurate assessments of indexes of insulin sensitivity and β-cell function from the OGTT, especially in epide- miological studies with large numbers of sparsely sampled subjects. Supported by: Kapodistrias Award, University of Athens 282 Family history is not associated with lifestyle, clinical, or anthropometric factors in newly diagnosed type 2 diabetes mellitus patients: the DD2 study E. Svensson1, K. Berensci1, S. Sander1, A. Mor1, J. Rungby2, J.S. Nielsen3, S. Friborg4, I. Brandslund5, J.S. Christiansen6, A. Vaag7, H. Beck-Nielsen3, H.T. Sørensen1, R.W. Thomsen1; 1Department of Clinical Epidemiology, 2Department of Pharmacolgy, Aarhus University, 3Diabetes Research Centre, Department of Endocrinology, Odense University Hospital, 4Department of Endocrinology, Odense University Hospital, 5Department of Biochemistry, Lillebaelt Hospital, Vejle, 6Department of Internal Medicine and Endocrinology, Aarhus University Hospital, 7Department of Endocrinology, Rigshospitalet and University of Copenhagen, Denmark. Background and aims: A positive family history of type 2 diabetes mellitus (T2D) increases the risk for developing T2D approximately two-fold, likely due to both genetic and lifestyle factors. It is unknown how having a fam- ily history of T2D is related to demographic, clinical-, lifestyle and anthro- pometric factors at T2D debut, thus we aimed to examine this in a cross- sectional study. Materials and methods: All participants in the Danish Centre for Strategic Research in Type 2 Diabetes (DD2) cohort responded to a questionnaire, in which detailed family history of diabetes (grandparents, parents, siblings and children) was assessed. We examined the prevalence and relative risk, and corresponding 95% confidence interval (CI) of having one first-degree rela- tive vs no first-degree relative with history of T2D according to different de- mographic, lifestyle, clinical and anthropometric factors at T2D debut. Results: Of 2,718 T2D patients, 1,191 (44%) had one or more first-degree family relatives with diabetes, including 377 (14%) with two or more first- degree relatives. Of the patients with positive family history, 20% had affected mothers, while 17% had affected fathers. Among the newly diagnosed T2D patients, a family history of T2D was to a lesser extent seen in men compared with women (adjusted relative risk (aRR) 0.84, 95% CI 0.78-0.90). Individuals aged less than 40 years at their T2D debut were more likely to have a family history of diabetes as compared with patients aged more than or equal to 60 years at T2D debut (aRR 1.34, 95% CI 0.94-1.91). In contrast, family history of T2D was evenly distributed according to presence or absence of central obesity, large weight gain since age 20, physical activity, alcohol consumption and Charlson comorbidity score. Conclusion: Almost half (44%) of newly diagnosed T2D patients have a first degree family relative with diabetes. Our results confirm that T2D heredity is stronger on the maternal side. Having a family history of diabetes is thus associated with female gender and younger age at T2D debut, whereas fam- ily history does not appear to be associated with presence of comorbidity or specific lifestyle-, clinical- and anthropometric factors at debut. Supported by: Danish Agency for Science, Danish Health and Medicines Au- thority, DDA 283 Determinants of skin autofluorescence in the general population and in type 2 diabetes B.H.R. Wolffenbuttel1, R. Graaff1, H.L. Lutgers1, K. Eny2, A.D. Paterson2, S.N. Slagter1, J.V. van Vliet-Ostaptchouk1, M.M. van der Klauw1; 1Endocrinology, University Medical Center Groningen, Netherlands, 2Program in Genetics & Genomic Biology, Hospital for Sick Children, Toronto, Canada. Background and aims: Skin autofluorescence (SAF) is a non-invasive mark- er of advanced glycation end products and is associated with long-term dia- betes complications and cardiovascular morbidity and death. SAF increases with age, and is higher among individuals with diabetes. Materials and methods: We assessed the determinants of SAF in people with and without type 2 diabetes (T2DM) in the LifeLines Cohort study, a ran- dom population-based sample of inhabitants of three Northern provinces of the Netherlands. For this cross-sectional analysis, we included subjects 18- 90 years of age, who had both genetic data available and SAF measurement collected between January 2008 and March 2011. We excluded subjects who were known to have type 1 diabetes and those with serum creatinine >140 mcmol/l , leaving 9025 individuals (of whom n=314, 3.5% with T2DM) for analysis. Skin autofluorescence (SAF) was measured non-invasively with the AGE Reader. Diabetologia (2014) 57:[Suppl1]S1–S564 S 125 1 C Results: Mean (±SD) age was 49 (±11) yrs for the non-diabetic participants and 59 (±11) yrs for the T2DM group, while mean SAF was 2.04 (±0.44) arbitrary units (AU) vs 2.45±0.59 AU. Linear regression showed that age, sex, body mass index, HbA1c, glomerular filtration rate (GFR), smoking, and a genetic polymorphism of N-acetyltransferase 2 (NAT2) significantly influ- enced SAF in the general population (all p S 126 1 C the different methodologies employed. Tissues from a total of 101 cases were then examined and scored for several indices of viral infection including VP1 immunopositivity, class I MHC hyperexpression, positive in situ hybridisa- tion signals and the presence of viral sequences. Among 28 controls, a total of 54 indices of viral infection were assessed independently and 7 were reported as positive (13%). In 52 type 1 diabetes cases (including those with residual insulin-containing islets and those without) 99 indices were scored and 51 were positive (51.5%). Among 21 autoantibody positive cases, 43 indices were assessed and 16 were positive (37.2%). PCR amplification and nucleic acid sequencing analysis of RNA extracted from cultured spleen cells or PBMCs revealed the presence of several enterovirus serotypes and the positivity cor- related with VP1 staining in sections of tissue from the same samples. Conclusion: The results support the conclusion that enteroviral infection oc- curs at much higher frequency in T1D pancreas than in controls. Virus can be found both in blood samples and in other tissues (e.g. pancreas, spleen). Similar evidence of persisting virus was also found in AAb+ cases, supporting the hypothesis that enteroviruses may play a role early in the development of T1D. Supported by: EU Framework 7 (PEVNET); JDRF nPOD-V 286 Increased antibodies against other autoimmune diseases in first degree relatives of patients with type 1 diabetes M. Szelachowska, K. Siewko, A. Zielinska, R. Maciulewski, A. Poplawska-Kita, D. Lipinska, M. Gorska; Department of Endocrinology, Diabetology and Internal Medicine, Medical University, Bialystok, Poland. Background and aims: It is well known that T1D is associated with other autoimmune diseases. The aim of this study was to compare the prevalence of various auto-antibodies in first-degree relatives of patients with T1D and healthy individuals with negative family history of diabetes. Materials and methods: The group studied consisted of 90 relatives and 60 healthy individuals. Serum concentrations of antibodies to anti-21-hy- droxylase (21-OH-Abs), anti-gastric parietal cell antibodies (GPC-Abs), anti-thyroglobulin antibodies (TG-Abs), anti-thyroid peroxidase antibodies (TPO-Abs) and anti-TSH receptor antibodies (TSHR-Abs) were measured by commercial radioimmunoassay. Results: Positive antibodies against pancreatic islet antigens were found in 34.4% of the relatives (IAA in 23.3%, GADA in 16.7% and IA-2A in 2.2%) and in none of the controls. Other antibodies (mainly TPO-Abs, TSHR-Abs and GPC-Abs) were detected in 40% of all relatives and in 93.5% of these with positive anti-islet antibodies. Median levels of 21-OH-Abs, GPC-Abs, TPO- Abs and TSHR-Abs were significantly higher in the relatives, in particular these with positive anti-islet antibodies, as compared with the group of rela- tives with no anti-islet antibodies and the controls. A positive correlation be- tween IAA and TPO-Abs levels was noted in the whole group of relatives, as well as in a subgroup with anti-islet antibodies (r=0.549, p S 127 1 C Results: Overall, GV parameters (CGM: interquartile rangeday, SDday and %glycemiaday >7.7 mmol/l; SMBG: SDday and rangeday) were inversely cor- related with FBM (r=-0.6 to -0.7, p≤0.001). All patients except 1 in remis- sion and 4/21 FDR had a FBM ≤percentile 10 (P10) of healthy controls; 2 of these 4 FDR developed T1D within 9 months. Among the 17 FDR with FBM >P10, 2 exhibited elevated GV parameters. A stronger correlation was found between GV parameters and M values (r=-0.8, pP33). In autoAb+ FDR, M ≤P33 of controls detected presence or development of dysglycemia within 1 year with 88% sensitivity and 93% specificity. Especially when considering multiple parameters, CGM was more sensitive than SMBG in detecting individuals who later developed IGT or T1D. Conclusion: In autoAb+ FDR low glucose disposal rate, also reflecting insulin resistance, outperforms FBM in diagnosing or predicting dysglycemia and is closely related to elevated CGM parameters. GV could provide non-invasive outcome measures in clinical trials. Supported by: JDRF (USA), EU FP-7 (Brussels), FWO (Brussels) 289 Baseline and five year treatment characteristics of adult onset type 1 diabetes D. Hesse, A.A. Nielsen, M. Ridderstråle; Steno Diabetes Center, Gentofte, Denmark. Background and aims: Much less is known about adult onset type 1 diabetes (AOT1D) than childhood or adolescent onset T1D. We sought to character- ize the initial five years of treatment of AOT1D. Materials and methods: Subjects with AOT1D between 2001 and 2012 were identified in the electronic medical records of our diabetes center. Baseline demographics and diabetes related parameters for up to five years after on- set were registered. Patients with latent autoimmune diabetes of the adult (LADA) were not included. Data are shown as mean ± standard deviation. Non-parametric tests were used throughout and a p-value S 128 1 C with DKA (5.1 weeks, SD 6.2) and those without (6.5 weeks, SD 18.8) was comparable (p=0.42). Positive autoantibody (AAb) status was not associated with symptom type, number or duration or DKA at presentation. GAD AAb titre had no effect on frequency of presentation with DKA. 8.2% of those pre- senting with DKA had at least one other autoimmune condition versus 4.5% of those with no DKA (p S 129 1 C PS 004 Epidemiology of obesity and ectopic fat 293 Women from South Asia, Middle East and Africa at increased risk of postpartum weight retention and type 2 diabetes: a multi-ethnic population based cohort study in Oslo, Norway C.W. Waage1,2, R.S. Falk3, C. Sommer1,4, K. Mørkrid1,2, K.R. Richardsen5,6, A. Bærug7, N. Shakeel2, K.I. Birkeland1,4, A.K. Jenum2,6; 1Department of Endocrinology, Morbid Obesity and Preventive Medicine, Oslo University Hospital, 2Institute of Health and Society, Department of General Practice, Faculty of Medicine, University of Oslo, 3Department of Biostatistics, Epidemiology and Health economics, Oslo University Hospital, 4Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, 5Norwegian Resource Centre for Womens Health, Department for Womens and Childrens Health, Oslo University Hospital, 6Faculty of Health Sciences, Oslo and Akershus University College of Applied Sciences, 7Norwegian Resource Centre for Breastfeeding, Oslo University Hospital, Norway. Background and aims: Pregnancy has been considered a critical period for development of overweight and its complications including type 2 diabetes, both in the mother and through diverse mechanisms including epigenetic modifications, also in her child. Immigrant women of Asian and African origin living in Europe are at higher risk for obesity, type 2 diabetes and car- diovascular diseases than the native European population. Women with ges- tational weight gain above recommended levels from American Institute of Medicine are at increased risk of postpartum weight retention and thereby of later obesity and related co-morbidities. The aim of the study was to explore ethnic differences in postpartum weight retention three months postpartum in a population-based, multi-ethnic study of pregnant women. Materials and methods: A multi-ethnic population based cohort study from Oslo, Norway of 823 healthy pregnant women, 59 % with ethnic minority background, included from 2008-2010. A total of 642 (78% of 823) were fol- lowed till three months postpartum. A multiple linear regression analyses were performed to model the relationship between postpartum weight reten- tion and ethnicity. Results: Unadjusted mean postpartum weight retention was 2.3 kg (95% con- fidence interval 1.7-2.9) for women from Western Europe and ranged from 3.7-6.3 kg among ethnic minority groups. The proportion of women in the highest quintile (postpartum weight retention 8.5-24.4 kg) differed by eth- nicity; 12% among Western European, 9% among East Asians, 25% among South Asians, 28% among Middle Eastern, 29% among East European and 41% among African women (p S 130 1 C 296 Visceral fat is associated with hyperglycaemia after renal transplantation M. von Düring1, T. Jenssen1,2, J. Bollerslev3, K. Godang3, A. Aasberg4, A. Hartmann1; 1Department of Organ Transplantation, Section of Nephrology, University of Oslo, Rikshospitalet, 2Institute of Clinical Medicine, Faculty of Health Science, University of Tromsø, 3Section of Specialized Endocrinology, 4Department of Pharmaceutical Biosciences, School of Pharmacy, University of Oslo, Rikshospitalet, Norway. Background and aims: Hyperglycaemia and new onset diabetes after trans- plantation (NODAT) are common complications in renal transplant recipi- ents and are associated with impaired long-term survival. Studies in the non- transplant population suggest that visceral fat facilitates development of type 2 diabetes. The role of visceral fat for development of NODAT is not known. The aim in this study was to elucidate this relationship between visceral fat content and hyperglycaemia in renal transplant patients. Materials and methods: We studied 159 renal transplant patients without a prior diagnosis of diabetes. All of them underwent oral glucose-tolerance tests (OGTTs) in a stable phase 10 weeks after transplantation. Visceral fat content was analyzed by a newly validated software (CoreScan) applied after total body composition DXA-scans using Lunar Prodigy, software version 14.10. Results: The amount of visceral fat (median 1.0 kg, interquartile range = IQR 0.4 - 1.9 kg) was highest in patients with NODAT (median 2.2 kg, IQR 1.2- 2.7). There was a significant difference in the amount of visceral fat between the categories of glucose tolerance; NODAT, impaired glucose tolerance (me- dian 1.2 kg), impaired fasting glucose (median 1.0 kg) and normal glucose tolerance (median 0.8 kg) (Kruskal-Wallis ANOVA, p = 0.003). The percent- age visceral fat of the total fat mass was 97% higher in patients with NODAT (median 7.7%, IQR 6.0-8.6%) compared with NGT patients (median 3.9%, IQR 2.0-5.8%) (p < 0.001). Percentage visceral fat of total fat mass was also a better predictor of both fasting (fPG, R2 = 0.116, p < 0.001) and 2-hour plasma glucose (2hPG, R2 = 0.082, p < 0.001), compared to total body fat per se or BMI in a multiple regression analysis. Conclusion: Visceral fat is adversely associated with glucose metabolism in renal transplant recipients. Whether reduction of visceral fat mass may pre- vent development of NODAT needs further studies. 297 Increased risk for diabetes development in non-diabetic Korean subjects with hypertriglyceridaemic waist phenotype E.-J. Rhee, M. Lee, H. Park, W. Jeon, S. Park, C.-Y. Park, W.-Y. Lee, K.-W. Oh, S.-W. Park; Endocrinology, Department of Endocrinology and Metabolism, Kangbuk Samsung Hospital, Sungkyunkwan University, Seoul, Republic of Korea. Background and aims: Hypertriglyceridemic waist (HTGW) phenotype is a simple and inexpensive screening parameter to identify people at increased risk for cardiovascular disease. We evaluated whether the HTGW phenotype predicts diabetes in Korean urban adults. Materials and methods: 2,900 non-diabetic subjects (mean age 44.3 years) including 2078 male (71.7%) and 822 female (28.3%), who did annual medi- cal check-up in a university hospital for four consecutive years were recruited. The subjects were divided into four groups according to the baseline serum triglyceride (TG) and waist circumference (WC); normal WC-normal TG (NWNT), normal WC-high TG (NWHT), high WC-normal TG (HWNT), high WC-high TG (HWHT). High serum TG was defined as TG≥150 mg/ dL and high WC was defined as WC≥90 cm for men and ≥85cm for women. New cases of diabetes were determined according to the self-questionnaire of the participants and fasting plasma≥126 mg/dL. Cox proportional hazard model analysis was used to assess the cumulative incidence of diabetes ac- cording to baseline HTGW phenotypes. Results: 101 (3.5%) new diabetes cases were diagnosed during four years of follow-up period. The subjects in HWHT group had the highest incidence of diabetes (8.3%) compared with the NWNT group (2.2%). The adjusted hazard ratio (aHR) for developing diabetes in the presence of HWHT phe- notype at baseline was 4.11 (95% confidence interval [CI] = 2.40-7.06) after adjusting for age, and 2.44 (95% CI = 1.38-4.33) after adjusting for age, sex, total cholesterol and systolic pressure, when NWNT group was considered as the reference group. Conclusion: This study demonstrated that HWHT showed the highest risk for diabetes development during four years of follow-up period. Recognizing HWHT type would be useful to identify individuals at high-risk of diabe- tes and, which is of great significance in reducing the incidence of diabetes among Korean urban adults. 298 Metabolically healthy obesity, presence or absence of fatty liver, and risk of type 2 diabetes in Japanese individuals Y. Heianza1,2, Y. Arase2, K. Fujihara1, S. Hsieh2, K. Saito1,2, O. Hanyu1, S. Kodama1,2, S. Hara2, H. Sone1,2; 1Department of Internal Medicine, Niigata University Faculty of Medicine, 2Health Management Center, Toranomon Hospital, Tokyo, Japan. Background and aims: Whether a “metabolically healthy obese (MHO) phe- notype” that does not include typical obesity-related metabolic abnormali- ties is really a benign or malignant state has not been elucidated in terms of predicting the progression to type 2 diabetes. We investigated whether the MHO phenotype was associated with an increased risk of the development of diabetes. If so, we aimed to determine what factors could explain this finding. Materials and methods: Studied were 8090 Japanese individuals without dia- betes (diabetes indicated by the ADA criteria). Metabolic health status was assessed by common clinical markers: blood pressure, triglycerides, HDL- cholesterol, and fasting glucose concentrations. Cut-off value for obesity (O) or normal weight (NW) was a BMI of 25.0 kg/m2. Participants were catego- rized at the baseline examination into 4 phenotypes: 1) metabolically healthy and normal weight (MHNW), 2) metabolically healthy and overweight or obese (MHO), 3) metabolically abnormal and normal weight (MANW), or 4) metabolically abnormal and overweight or obese (MAO). Results: The 5-year incidence rate of diabetes was 1.2% (n=58/4749) in MHNW individuals, 2.8% (n=20/719) in MHO individuals, 6.0% (n=102/1709) in MANW individuals, and 10.3% (n=94/913) in MAO indi- viduals. Although the MHO individuals had no or 1 metabolic factor, 47.8% had ultrasonographic fatty liver (FL). The MHO group had a significantly increased risk of diabetes compared to the MHNW group (multivariate-ad- justed OR, 2.23 (95% CI 1.33, 3.75)), but this risk was attenuated after adjust- ment for FL. We then assessed the combined effect of obese phenotypes and FL on the development of diabetes. MHNW/non-FL group had the lowest incidence rate of diabetes (0.9%), and MHO/non-FL group also had a simi- larly low incidence rate of diabetes (1.1%). On the other hand, the MHO/FL group had an elevated incidence rate (4.7%). Incidence rate of diabetes was markedly high at 8.5% in the MANW/FL group and 12.6% in the MAO/FL group. Compared to the MHNW/non-FL group, the risk of diabetes in the MHO/non-FL group was not significantly elevated (OR 1.01 (0.35, 2.88)). However, the MHO/FL and MHNW/FL groups had similarly elevated risks of diabetes (OR 4.09 (95% CI 2.20, 7.60) and 3.16 (1.78, 5.62), respectively). A prediction model with age, sex, and obese phenotypes had an area under the ROC curve (AUCROC) of 0.754 for the development of diabetes. The AUCROC was slightly but significantly (p S 131 1 C 299 Presence and severity of obstructive sleep apnoea is independently associated with glycometabolic abnormalities in obese non-diabetic subjects A. Cignarelli1, S. Perrini1, A. Ciavarella1, S. Kounaki1, V. Quaranta2, M. Barbaro1, A. Di Trani1, V. Falcone2, A. Natalicchio1, L. Laviola1, O. Resta2, F. Giorgino1; 1Endocrinology & Metabolic Diseases, 2Respiratory Diseases, University of Bari, Italy. Background and aims: Obstructive sleep apnea (OSA) is a common under- diagnosed condition in the obese population, and has been associated with worse glycemic control in individuals with type 2 diabetes. The relationship between OSA and glycometabolic parameters was investigated in obese non- diabetic individuals. Materials and methods: Ninety-one obese subjects (57% male, mean age 45.3 ± 12 yrs, mean BMI 42.1 ± 9 kg/m2) underwent polysomnography and a 2-h oral glucose tolerance test (OGTT). Results: OSA was identified in 64% of subjects (73% in male, 41% in female, p=0.032 χ2). Obese subjects with OSA showed higher A1c (5.8% vs 5.5%, p=0.009), plasma glucose at 120 min during OGTT (PPG120) (133 mg/dl vs 102 mg/dl, p=0.001), triglyceride (140 mg/dl vs 117 mg/dl, p=0.045) and uric acid (5.8 mg/dl vs 4.9 mg/dl, p=0.035) levels than obese subjects without OSA. A1c levels and PPG12O were found to be significantly correlated with raised apnea-hypopnea index (AHI) (p=0.007 and p=0.004, respectively), oxygen desaturation index (p=0.002 and p=0.04, respectively), and percent of sleep time with oxyhaemoglobin saturation at 140g/week, n=448). In this study, diabetes was diag- nosed by prescription of antidiabetic agents, fasting plasma glucose (FPG) ≥ 6.8mmol/l, HbA1c (NGSP) ≥ 6.5% and/or a past history of diabetes in those not yet treated with antidiabetic agents. Fatty Liver Ultrasonography scores (FLUS) were also assigned as follows: 2 points, subjects with moderate or severe fatty liver; 1 point, those with mild fatty liver; 0 points, those with normal liver. Regression coefficient (B) was calculated using logistic regres- sion analysis. Results: The mean observation period was 2.23 years, with 27 LAI subjects and 15 HAI subjects with newly developed diabetes from Periods I to II. The remaining 1488 LAI subjects and 433 HAI subjects were nondiabetic. In lo- gistic regression analysis, alanine aminotransferase (ALT) was independently and significantly (B=0.025, p=0.017) associated with incident diabetes as well as age (B=0.065, p=0.008), BMI (B=0.264, p=0.050), waist circumfer- ence (WC) (B= - 0.120, p=0.036), fasting plasma glucose (FPG) (B=0.116, p S 132 1 C PS 005 Descriptive epidemiology of diabetes 301 One normal range for HbA1c in every age? The effect of aging on HbA1c in the people without diabetes J. Roth1, N. Müller1, C. Kloos1, U.A. Müller1, G. Wolf2; 1Department of Internal Medicine III, Department Endocrinology and Metabolic Disorders, 2Department of Internal Medicine III, Jena, Germany. Background and aims: Goals for HbA1c in National Guidelines and the cut- off value for diagnosis of diabetes are not adjusted for age yet. The Framing- ham Offspring Study and the National Health and Nutrition Examination Survey 2001-2004 showed a 0.6% increase of HbA1c for non diabetic people over the age of 70 compared to people below the age of 40 years. We evaluated this question in a German population. Materials and methods: We analyzed the data from 10163 visits of 3861 patients without diabetes (age 44.7y; 18-93y; 72.6% women; BMI 28,0±6,5 kg/m²) which had parallel HbA1c and blood glucose measurements and in- formation about drug treatment. The HbA1c values were divided in 3 age groups (=40 < 70y [n=5454]; >70y [n=1126]). Patients with gestational diabetes mellitus, documented use of systemic glucocorticoids or HbA1c >= 6.5% were excluded. A glucose tolerance test was not performed. Data were drawn from the electronic patients record (EMIL®) of our uni- versity outpatient department for endocrinology and metabolic disorders between 10/1992 and 01/2014. HbA1c was DCCT adjusted (mean normal range of healthy people 5.05%). Results: The mean HbA1c (95. percentile) for the groups of age was: =40 < 70 years 5.3±0.38% (5.9%), >=70 years 5.4±0.37% (6.0%). The differences between 70y (p< 0.001); and 40-70y vs. >70y (p< 0.001) were also significant. Conclusion: HbA1c increases significantly with the age of people without diabetes. The use of different cut-off values for the diagnosis of diabetes ac- cording to age should be considered. Also HbA1c goals for old people with diabetes should consider the age dependend increase of HbA1c. 302 The association of smoking cessation with HbA1c control of diabetes mellitus: a THIN database study L. Nichols1, R. Ryan1, A. Farley2, A. Roalfe1, M. Mohammed3, L. Szatkowski4, T. Coleman4, R. Morris5, A. Farmer6, P. Aveyard6, D. Lycett7; 1Primary Care Clinical Sciences, University of Birmingham, 2Public Health, Epidemiology and Biostatistics, University of Birmingham, 3University of Bradford, , 4University of Nottingham, 5University College London, 6University of Oxford, 7Coventry University, UK. Background and aims: Smoking increases the risk of developing type 2 diabetes mellitus. However several large observational studies have found during the first 3-5 years following smoking cessation the risk of developing diabetes is significantly greater than in continuing smokers; an association largely explained by weight gain. This increase in risk is temporary and after 10-12 years the risk is equivalent to never-smokers. There is also prelimi- nary evidence from small cohort studies, in those who already have diabetes that control deteriorates during the first year after smoking cessation before it improves. Our objective was to examine whether an association between smoking cessation and diabetes control exists in a large representative sample of the UK population. If deterioration in diabetes control around the time of smoking cessation does occur, this may have implications for enhanced diabetic care at that time. Materials and methods: A retrospective cohort study (01/01/2005 - 31/12/2010) was assembled using The Health Improvement Network (THIN) database. Inclusion criteria were: patients aged over 18, registered with their practice for at least one year on 01/01/2005, diagnosed with type 2 diabe- tes mellitus and whose last recorded smoking status prior to 01/01/2005 was current smoker. An adjusted multilevel regression model was developed to investigate the association between change in HbA1c and stopping smoking. Results: There were 10,692 adults with type 2 diabetes who were current smokers as at 01/01/2005. Of these, 3,131 (29%) quit smoking and remained abstinent for one year or longer. After adjustment for potential confounders, patients who quit smoking had an average increase in HbA1c after quitting of 2.3mmol/l (95% CI 1.91 to 2.77, p18 years old, aged (median IQ range) 36(26 to 61) years, 55% male with admission plasma glucose (PG mmol/L) 19.3(11.8 to 27.4). Fasting PG on OGTT in those with possible symptoms or complications of diabetes, aged 70(59 to 79) years, 63% male with admission PG 6.4(5.6 to 7.4), was 5.2(4.8 to 5.7). This was similar to fasting PG on OGTT of 5.2(4.8 to 5.9) in 108 white Caucasian, primary care patients, aged 54(46 to 61) years, who were at increased diabetes risk, p=0.65. As expected, 2hPG at 9.0(7.3 to 11.4) was higher in acutely ill patients than GP patients, 5.5(4.4 to 7.5), p S 133 1 C 304 Achievement of individualised HbA1c treatment targets in patients with type 2 diabetes and comorbid hypertension in a real world setting: results of the DIALOGUE registry D. Tschöpe1, R.E. Schmieder2, C. Koch3, P. Bramlage4, T. Ouarrak5, A.K. Gitt6; 1Herz- und Diabeteszentrum Nordrhein-Westfalen, Bad Oeynhausen, 2Med. Klinik 4, Schwerpunkt Nephrologie / Hypertensiologie, Universitätsklinikum Erlangen, 3Novartis Pharma GmbH, Nürnberg, 4Institut für Pharmakologie und präventive Medizin, Mahlow, 5Stiftung „Institut für Herzinfarktforschung“, Ludwigshafen, 6Medizinische Klinik B, Herzzentrum Ludwigshafen, Germany. Background and aims: The recent EASD consensus statement has re-focused on individualized HbA1c Treatment targets in the care of patients with type-2 diabetes mellitus. There are, however, neither clinical practice data on actu- ally pursued treatment targets nor on patient characteristics associated with either strict or loose treatment targets. Materials and methods: DIALOGUE is a prospective, observational, multi- center registry focusing on treatment targets and their achievement in clini- cal practice. Physicians were asked at baseline on HbA1c treatment targets pursued and compared to actual HbA1c values achieved at 6 months. Results: A total of 8,636 patients were included. For 3,371 of these (39.0%) an HbA1c target of ≤6.5% (strict group), for 3,647 (42.2%) a target of >6.5 to ≤7.0% (medium) and for 1,618 (18.7%) a target of >7.0 to ≤7.5% (loose) was pursued. Patients in the strict target group were younger, had lower fast- ing and postprandial blood glucose values, shorter diabetes duration and less co-morbid disease at baseline than the other groups (table). At the 6 months follow-up the mean (±SD) HbA1c achieved in the strict group was 6.8±0.9% with 25% of patients exceeding 7.1%. Corresponding values for the medium and loose target group were 7.2±0.9 (25% >7.6%) and 7.7±1.2 (25% >8.3%), respectively (p S 134 1 C Materials and methods: We assessed sex, age, BMI, duration of diabetes, education, income, ethnicity, physical activity, glucose-lowering treatment, previous cardiovascular disease, risk factors and statin treatment. We report (1) baseline characteristics; (2) progress of glycemic control from disease on- set to end of follow up (up to 10 years); (3) linear mixed models to determine the effect of the above mentioned covariates on HbA1c (mmol/mol) and time duration until HbA1c started to increase (‚turn point‘); (4) logistic regression was used to study the probability of achieving glycemic control ( S 135 1 C ducted in a locally representative sample leaving only a study from Hungary as being regionally representative and a study from Turkey being nationally representative. It is estimated that in 2013 of a predicted 10.7 million live births to women aged 20 - 49 years 15.2 % were affected by HIP. Thus, within the Region, 1.7 Million live births were affected by HIP. While the highest prevalence of HIP is found in pregnant women >40 years of age, the largest numbers of cases occur in women aged 25 - 35. Within the countries esti- mated age-standardised prevalence of HIP ranged from 5.0% of all live births affected in Belgium up to 32.1% of all live births affected in Spain. Conclusion: Prevalence estimates of HIP are sensitive to the data from which they are derived. Within the WHO-EUR-Region there is a lack of nationally representative studies on the prevalence of HIP, especially in lower middle income countries and low income countries. More data are needed, in par- ticular from low income countries, to strengthen the methodology. These are the first estimates of HIP in the WHO - EUR Region and conform to the new WHO recommendations regarding diagnosis. They indicate the importance of this issue from a public health and maternal and child health perspective. Supported by: Lilly Diabetes, Merck and Co, Novo Nordisk A/S, Pfizer, Sanofi Diabetes 309 Prevalence of metabolic syndrome and cardiometabolic risk factors in patients with long-duration psychotic illnesses D. Hopkins1, P. Gardner-Sood2, J. Lally2, D. Stahl3, K. Greenwood2, Z. Atakan2, S. Smith2, R. Ohlsen2, K. Ismail4, R. Murray2, F. Gaughran5; 1Dept of Diabetic Medicine, King‘s College Hospital NHS Foundation Trust, 2Institute of Psychiatry, King‘s College London, 3Dept of Biostatistics, King‘s College London, 4Diabetes & Psychiatry Research Group, King‘s College London, 5National Psychosis Service, South London and Maudsley NHS Foundation Trust, London, UK. Background and aims: Patients with severe mental illness have significantly reduced life expectancy with increased cardiovascular events and a high prevalence of diabetes. As a preliminary step to testing interventions to try and improve physical health outcomes in patients with established psychosis we have conducted this study to determine the prevalence of the component risk factors that comprise the metabolic syndrome in a large cohort of estab- lished patients. Materials and methods: Cross-sectional study of 450 patients with estab- lished (multi-episode) psychosis recruited at five centres in the United King- dom. Anthropometric measures (height, weight and waist circumference), and blood pressure were recorded for all subjects. In addition blood samples for HbA1c, fasting glucose and lipid profile were obtained from 296 subjects. Results: Mean age of the participants was 43.6 years (SD=10.1) and mean du- ration of psychotic illness was 15.7 years (SD=10.3). 57% of participants were male. 55% percent (n=238) were Caucasian and 33% (n=146) were of Black African or Black Caribbean ethnicity. 50% of participants were obese (body mass index, BMI>30) 44% were prescribed dibenzodiazepine antipsychotic medication (clozapine and olanzapine). There was no association between dibenzodiazepine use and BMI. Of the 296 participants with complete data 57 % fulfilled the International Diabetes Federation (IDF) criteria for diag- nosis of the metabolic syndrome. Prevalence of central obesity was very high with 82%, of participants exceeding IDF cut-off values for waist circumfer- ence. In addition, 54% had blood pressure above 130/85, 57% triglyceride above 1.7 mmol/l and 53% had HDL cholesterol below the IDF cut-off values of under 1.03 or 1.29 mmol/l for men and women respectively. 31% had fast- ing dysglycaemia (fasting glucose of 5.6 mmol/l or greater) including 26% fulfilling criteria for a diagnosis of diabetes. Conclusion: These data demonstrate a high prevalence of metabolic syn- drome and dysglycaemia in a relatively young population with longstanding psychosis and the presence of multiple adverse cardiometabolic risk factors that could be ameliorated to improve cardiovascular outcomes. The preva- lence of central obesity is considerably greater than that of obesity diagnosed by BMI indicating a need for waist measurement to predict risk in this popu- lation. Supported by: National Institute of Health Research, U.K. PS 006 Epidemiology of diabetes: comorbities and complications 310 Prevalence of undetected thyroid disorders in subjects with diabetes in South India: a cohort analysis G. Krishnan, A. Shankar, S. Jothydev, J. Kesavadev; Jothydev‘s Diabetes and Research Center, Trivandrum, India. Background and aims: Thyroid function test is not a part of routine diabetes practice in India, due to the extra cost involved and the lack of time for the physician to convince the patient on the requirement of an extra expensive investigation. Many a time, symptoms of hypothyroidism in diabetes patients are mistaken for hypoglycaemia and associated metabolic syndrome and hence goes grossly undetected. Materials and methods: We screened T2DM patients, presenting at our comprehensive diabetes speciality clinic for the first time, to evaluate the prevalence of thyroid disorders. Screening was done during the past 5 years, using the ‚sensitive TSH test‘ in Cobas E411 Elecsys. Patients with undetect- ed hypothyroidism reported mood changes, anxiety, anger, memory loss etc. which were corrected with thyroid supplementation. Results: During selection of 7402 consecutive T2DM subjects aged ≥ 21yrs (62.48% male), 5.83% self reported hypothyroidism; 29.67% showed abnor- mal TSH values, comprising of 10.42% with TSH values suggestive of sub- clinical hypothyroidism((3.04-10 μIU/mL); 17.09% overt hypothyroidism (>10 μIU/mL) and 2.16% with low TSH values suggestive of hyperthyroid- ism((60yrs age group; whereas distribution of other sub- groups were not statistically significantly different. Among females, 9.09% had mildly increased TSH levels (3.04-10μIU/mL), 22.14% had high TSH levels (>10μIU/mL) and 1.45% had low TSH levels (0.05). The level of diabetes control was possibly confounded by treatment received and HbA1c did not significantly differ among hyperthyroid, euthy- roid & sub-clinically hypothyroid patients (mean HbA1c=8.9%). Conclusion: Considering the negligible extra cost of 1.26Euros for 90 tablets (presuming an average 50mcg/day dosage) of thyroid supplementation, dia- betes patients will tremendously benefit from early detection of thyroid dis- orders. The very high prevalence of hypothyroidism in diabetes, significantly higher than in the general population, makes it imperative to include TSH test in routine diabetes evaluation in developing countries like India. 311 Prediabetes is associated with early changes in microcirculation V.M. Shyshko1, T.V. Mokhort1, N.L. Tsapaeva2, E.E. Konstantinova3; 1Endocrinology, 2Inner Disease #3, Belarussian State Medical University, 3Institute of Heat and Mass Exchange, National Academy of Science, Minsk, Belarus. Background and aims: Microangiopathy in patients with type 2 diabetes (T2D) results from previous microcirculation abnormalities (e.g. increased permeability, disturbance of intracapillary pressure and blood flow). Hyper- glycemia as well as hesitance of glucose level in patients with prediabetes (im- paired glucose tolerance and impaired fasting glucose) have negative impact on microvessel status. The aim of the study was to investigate microcircula- tion in patients with prediabetes. Materials and methods: We included 131 patients with average age 49,03 ± 8,76 years. Patients were divided into 2 groups: group 1 - 37 patients with pre- diabetes, group 2 - 35 patients with type 2 diabetes (with duration of disease no longer as 5 years and treated with oral blood glucose lowering drug) and group 3 - 59 almost healthy person. Microcirculation was measured by com- puter based conjunctival biomicroscopy (Malaja et al.), results were evaluated by the set of criteria for quantitative evaluation of conjunctival microcircula- tion: FC (number of active capillary tubes), AVA (arteriovenous anastomo- sis), Mean (vascular tortuosity), Sl (sludge), Mtr (microthrombosis). Severity of each criteria was scored and more sever changes had higher degree. Results: Microcirculation abnormalities were revealed in patients with predi- abetes: we registered statistically significant decrease of active capillary tubes (FC) (3,0[2,0;3,0]vs2,0[2,0;3,0] in control group) (Р1-3 S 136 1 C number of AVA (2,0 [2,0; 4,0]vs2,0 [2,0; 2,0] in control group) (Р1-3 S 137 1 C hazard ratios: 1.8 (95%CI 1.7;2.0) and 1.4 (95%CI 1.3 - 1.4). Compared to the period of 2000-2004, patients initiating RRT in 2005-2009 had a lower mor- tality, age and gender adjusted hazard ratios: 0.8 (95%CI 0.7;0.8). Conclusion: The incidence of RRT for DM is stable over the last decade re- flecting a decrease for T1DM and an increase for T2DM. Taken together with a steady increase in prevalence of DM in the GP this may suggest that physi- cians may be more successful in the prevention of diabetes related ESRD 315 Cardiovascular events and all-cause mortality: associations with renal function in patients with type 2 diabetes L.A. García Rodríguez1, L. Cea Soriano1, S. Johansson2, B. Stefansson2; 1Spanish Centre for Pharmacoepidemiologic Research (CEIFE), Madrid, Spain, 2AstraZeneca R&D, Mölnda l, Sweden. Background and aims: Diabetes and chronic kidney disease are independ- ent predictors of mortality and cardiovascular (CV) events. The aims of this study were to quantify the risk of death and CV events associated with esti- mated glomerular filtration rate (eGFR) and to identify other potential risk factors in patients with type 2 diabetes. Materials and methods: In this retrospective study, data were collected from The Health Improvement Network, a UK primary care database. From a co- hort of patients aged 20-90 years with type 2 diabetes, 57 946 individuals were identified who had a valid serum creatinine measurement recorded in 2000-2005. Patients were followed up from their start date (date of first ever valid recorded creatinine measurement) until they met one of the following endpoints in three separate analyses: myocardial infarction (MI), ischaemic stroke or transient ischaemic attack (IS/TIA), or death; patients were cen- sored when they reached 90 years of age or the end of the study (31 December 2010). Individuals with a record of haemodialysis or renal transplant before their start date were excluded or censored during follow-up. Incidence rates for death, MI and IS/TIA were calculated overall and by eGFR subgroup at baseline (15-29, 30-44, 45-59 and ≥ 60 mL/min). Cox regression models ad- justed for potential confounding factors were used to estimate hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) for death, MI and IS/TIA according to sex, age, eGFR category and duration of diabetes. Results: Median follow-up times were 6.80, 6.64 and 6.56 years for death, MI, and IS/TIA, respectively. Overall incidence rates for death, MI and IS/TIA were 43.65, 9.26 and 10.39 cases per 1000 person-years, respectively. Inci- dence rates were highest for patients with the lowest eGFRs (15-29 mL/min): 210.01, 31.65 and 32.48 cases per 1000 person-years for death, MI and IS/ TIA, respectively. For patients with an eGFR ≥ 60 mL/min, the corresponding incidence rates were 31.99, 7.44 and 8.65 cases per 1000 person-years. A low eGFR (15-29 mL/min) was associated with an increased risk of death (HR: 4.00; 95% CI: 3.69-4.33), MI (HR: 3.29; 95% CI: 2.68-4.04) and IS/TIA (HR: 2.47; 95% CI: 2.01-3.03) relative to eGFR ≥ 60 mL/min (Table). The risk of death, MI and IS/TIA significantly increased with age: HRs for patients aged 75 years or older relative to patients aged 20-49 years were 12.14 (95% CI: 10.79-13.67), 3.33 (95% CI: 1.79-3.98) and 6.64 (95% CI: 5.44-8.11), respec- tively. Longer duration of diabetes was associated with an increased risk of death or experiencing a CV event: HRs for patients who had had diabetes for more than 15 years relative to those who had had diabetes for less than a year were 1.74 (95% CI: 1.59-1.90) for death, 1.87 (95% CI: 1.54-2.26) for MI and 1.64 (95% CI: 1.36-1.98) for IS/TIA. Conclusion: In patients with type 2 diabetes, the risks of death, MI and IS/ TIA increased with age, duration of diabetes and decreasing eGFRs. Supported by: AstraZeneca 316 Factors associated with cognitive impairment in patients with newly diagnosed type 2 diabetes: a cross-sectional study C.D. Moulton1, S.A. Amiel2, R. Stewart1, K. Ismail1; 1Institute of Psychiatry, 2Department of Diabetes and Nutritional Sciences, King‘s College London, UK. Background and aims: Type 2 diabetes (T2DM) is a major risk factor for cognitive impairment (CI) and the pandemic of T2DM is expected to in- crease the global dementia burden. The factors associated with CI in T2DM are poorly understood. A limitation of previous cohorts examining CI in T2DM is the variable and long duration of T2DM among participants, which increases the risk of effects being obscured by differential survival and dia- betes complications. By analysing a population with newly diagnosed T2DM, we aimed to identify modifiable factors for cognitive decline at an earlier stage in the natural history of T2DM. Our primary hypothesis was that mark- ers of glucose dysregulation would be associated with cognitive impairment and secondary hypotheses were that i) depression, ii) black or South Asian ethnicity and iii) body mass index (BMI) would be independently associated with CI in T2DM. Materials and methods: We performed a cross-sectional analysis of baseline measures from the South London Diabetes (SOUL-D) cohort, a population- based multi-ethnic study of individuals diagnosed with T2DM in the last 6 months. We assessed CI using the 13-item Telephone Interview for Cognitive Status (TICS-M), which has a range of scores from 0-39. We defined CI as the lowest 10% of scores and the remainder as controls. We performed univariate and multivariate analyses of the association between CI and 1) markers of glucose dysregulation (HbA1c, fasting glucose, micro/macrovascular com- plications); 2) sociodemographic factors (age, gender, ethnicity); 3) vascular risk factors (BMI, hypertension, hyperlipidaemia, smoking) and 4) depres- sion (PHQ-9 score >12). We used Student’s t-test for continuous data and χ2 tests for categorical data and logistic regression to control for key confound- ing variables, including premorbid intelligence estimated by National Adult Reading Test (NART) score. Results: In the preliminary analyses, of 1791 patients recruited, 1680 (93.8%) had a complete TICS-M assessment. Average diabetes duration was 4.6 (SD 2.2) months. Cognitive impairment was defined as TICS-M score S 138 1 C out diabetes. The aim was to investigate which factors that contributes to the risk of AF in type 2 diabetes using national register data from Sweden. Materials and methods: Observational cohort study of 64,864 patients with type 2 diabetes, aged 30-79 years, without AF at baseline. A history of car- diovascular disease (CVD) was present in 16.9% and 3.6% had a history of congestive heart failure (CHF). Patients were, followed-up for development of AF (n=2,719) during 4.2 years from 2003 to 2009. A subgroup of 52,939 patients without history of CVD or CHF at baseline was also analysed. Results: At multivariate logistic regression, odds ratios (OR) of AF were 4.49 for a baseline history of CHF and 7.98 for in-study developed CHF. Similarly for a history of CVD, the OR was 1.28 and for in-study developed myocardial infarction an OR of 1.37. Cardiovascular risk factors associated with risk for AF were hypertension (OR, 1.42), cumulative microalbumi- nuria (OR, 1.24), obesity (OR, 1.31) and ORs for each decade of increas- ing age was 1.93, and 1.43 for male gender, with all predictors p S 139 1 C PS 007 Pharmaco-epidemiology 320 Temporal trends in use of evidence-based treatments and risk factor control: a decade of nationwide monitoring B. Eliasson1, A. Rawshani1, A.-M. Svensson1, N. Ekström1, B. Zethelius1, S. Gudbjörnsdottir2; 1Clinical and Molecular Medicine, Inst of Medicine, 2Clinical Medicine, Inst of Medicine, Uppsala, Sweden. Background and aims: Glucose-lowering therapy with metformin is a cor- nerstone of diabetes treatment. Guidelines recommend instituting metform- in at the time type 2 diabetes is diagnosed. Likewise, the primary preventive effects of statins in individuals with type 2 diabetes has been recognized for almost a decade. Little is known about the pace of adoption of these evidence- based medications. We examined this in a nationwide prospective study spanning over a decade. Materials and methods: We used the Swedish National Diabetes Register to include 131935 newly diagnosed (within one year) cases of type 2 diabetes during 2002 to 2011. The study population was divided in ten cohorts cor- responding to the year of diagnosis. To study trends in use of oral glucose- lowering drugs we included individuals who were treated with diet and lifestyle modifications at baseline. Patients were followed until initiation of oral glucose-lowering therapy or appropriate censoring. HbA1c was followed annually for each cohort. Trends in statin therapy was studied in therapy na- ive patients who did not experience any cardiovascular event throughout the study. Lipid markers were followed annually for each cohort. Time to event (instituting therapy) was analysed by means of Kaplan-Meier method. Results: We included 58407 patients treated with diet and lifestyle modifica- tions. We observed a striking pattern (figure I) towards earlier initiation of glucose-lowering drugs since 2002. Median time to drug in 2002 was 1730 (95% CI 1636, 1801) days compared to 471 (95% CI 420, 518) days in 2011. However, HbA1c levels displayed a different trend; Hba1c (at the same dura- tion of diabetes) declined steadily from 2002 to 2007, whereafter levels in- creased. Considering statin therapy, we note the same pattern as with glu- cose-lowering drugs. Median time to therapy decreased from 2450 (95% CI 2360, 2530) days in 2002 to 650 (95% CI 580, 730) days in 2011. Interestingly, the curve for 2011 leveled off and crossed the curve for 2010. LDL-cholesterol displayed a pattern similar to HbA1c - an initial decline from 2002 to 2007 and thereafter a steady increase. Post hoc analysis of trends in body mass index (BMI) and physical activity sought to find an explanation for these trends; we noted that BMI at baseline and each point thereafter has increased annually since 2002. Conclusion: We document a striking improvement in time to instituting oral glucose-lowering drugs. We also note that use of statins in primary preven- tion has accelerated in parallell. However, neither HbA1c nor lipid levels re- flected these trends. 321 Prescribing patterns of antidiabetic drugs within the first year following diagnosis of type 2 diabetes: results from the DD2 study A. Mor1, K. Berencsi1, E. Svensson1, J. Rungby2,3, J.S. Nielsen4, S. Friborg5, I. Brandslund6, J.S. Christiansen7, A. Vaag8, H. Beck-Nielsen4, H.T. Sørensen1, R.W. Thomsen1; 1Clinical Epidemiology, Aarhus University Hospital, 2Pharmacology, Aarhus University Hospital, 3Endocrinology, Gentofte University Hospital, Copenhagen, 4Diabetes Research Centre, Endocrinology, 5Endocrinology, Odense University Hospital, 6Biochemistry, Lillebaelt Hospital, Vejle, 7Internal Medicine and Endocrinology, Aarhus University Hospital, 8Endocrinology, Rigshospital, Copenhagen, Denmark. Background and aims: Real-world data are sparse on prescribing patterns of antidiabetic drugs in cases of newly diagnosed type 2 diabetes mellitus (T2DM) and on patient characteristics that may predict type of early phar- macotherapy. Materials and methods: We studied 1,317 newly diagnosed T2DM patients enrolled in the Danish Centre for Strategic Research in Type 2 Diabetes (DD2) cohort from outpatient hospital clinics and general practice, with at least one year of follow-up. We described the number and type of antidiabetic drugs prescribed in the first year after T2DM diagnosis. Using Poisson regres- sion, we calculated risk ratios (RRs) of treatment with different antidiabetic drugs associated with baseline patient characteristics. Results: Of the 1,317 newly diagnosed T2DM patients, 198 (15%) did not use any antidiabetic drugs within the first year post-diagnosis, 867 (66%) used only one drug, 197 (15%) used two drugs, and 55 (4%) used three or more drugs. Figure 1 demonstrates the ranking of individual types of antidiabetic drugs used by our 1,317 T2DM patients. The likelihood of receiving combi- nation therapy with two or more drugs in the first year post-diagnosis was substantially higher in T2DM patients aged =3) were at higher likelihood of receiving combination therapy (30%; RR = 1.54, 95% CI: 1.05-2.26) than T2DM patients with no comorbidity (19%). Weight gain > 30 kg since 20 years of age and lack of regular physical exercise also increased the likelihood of receiving combination therapy during the first year post-diagnosis (RR = 1.42, 95% CI: 1.14-1.78 and RR = 1.42, 95% CI: 1.11-1.82, respectively). High- er likelihood of receiving combination therapy also was observed in T2DM patients with fasting blood glucose > 7 mmol/L at diagnosis (RR = 3.37, 95% CI: 2.43-4.68), HbA1c >=7.5 (RR = 3.85, 95% CI: 3.00-4.94), and C-peptide < 300 pmol/L (RR = 1.78, 95% CI: 1.15-2.76). Diabetologia (2014) 57:[Suppl1]S1–S564 S 140 1 C Conclusion: Within the first year after T2DM diagnosis, 85% of the patients are treated with antidiabetic drugs, most commonly metformin (79%), an incretin (14%), or insulin (8%). Poor glycemic control and low C-peptide at T2DM diagnosis are predictors of combination therapy at one year. Other important predictors are young age, comorbidity, obesity, and lack of exer- cise. Supported by: DAS, DHMA, DDA, Novo Nordisk A/S 322 Factors associated with adherence to oral antihyperglycaemic monotherapy in patients with type 2 diabetes in the UK K. Tunceli, C. Zhao, K. Iglay, K.G. Brodovicz, C.M. Alexander, L. Radican; Merck & Co., Inc., Whitehouse Station, USA. Background and aims: Suboptimal medication adherence is a critical factor contributing to poor glycemic control and an increased risk of complications in patients with type 2 diabetes (T2DM). The current retrospective database analysis aims to estimate adherence to oral antihyperglycemic monotherapy (OAM) for T2DM patients in the UK and describe factors associated with adherence in these patients. Materials and methods: Patients with a physician prescription for metform- in, sulfonylurea (SU), thiazolidinedione (TZD) or dipeptidyl peptidase-4 in- hibitor (DPP-4i) as monotherapy between January 1, 2009 to June 30, 2012 (index period) were selected from the IMS Disease Analyser-Mediplus UK database. Patients were required to have a diagnosis of T2DM, be ≥18 years old at the time of diagnosis, and have medical records available in the da- tabase for 1 year before and after the index date. Patients were excluded if they had at least one diagnosis of type 1 diabetes mellitus, their dosing regi- men (once daily or twice daily) could not be determined at the index date, they switched between once- and twice-daily dosing regimens during the 12-month post-index period, or they violated the monotherapy requirement during the 12-month post-index period. Adherence was assessed using the proportion of days covered (PDC) defined as total days medications sup- plied divided by days in follow-up period. Logistic regression was used to assess factors (age, gender, previous treatment status [new to or previously on OAM], dosing regimen [once- or twice-daily], and number of concomitant medications) associated with whether or not a patient is adherent (yes: PDC ≥80%, no: PDC S 141 1 C in the use of dipetidyl peptidase-4 (DPP-4) inhibitors. Interestingly, within each incidence year strata, 52.1, 47.0, 48.3, 54.2, and 55.7% of patients did not have an HbA1c measurement recorded in the lab section of the EHR within the 1st year of diagnosis. Among patients with available HbA1c values, the percentages of patients with inadequate glycemic control (HbA1c>8%) at 1 year were 25.1%, 24.4%, 22.2%, 25.9%, and 28%, respectively. Conclusion: A high prevalence of HTN was observed among this population of patients with new-onset T2D. Nearly 1/3 of the new-onset T2D patients were receiving 2 or more classes of anti-diabetic agents at 1 year after di- agnosis. The use of thiazolidinediones in patients with new-onset T2D has decreased, whereas the use of DPP-4 inhibitors has increased. Of note, the use of sulfonylureas has not decreased, despite the availability of several new agents. Approximately 1/2 of new-onset T2D patients did not have an HbA1c value recorded in the lab section of the EHR within the 1st year of diagnosis. Supported by: Novo Nordisk 324 Burden of baseline comorbidities in patients with type 2 diabetes initiating various classes of antidiabetic agents - an analysis of a large US claims database T.M. Rehman1, S. Colilla2, T. Kou3, J. Wood3; 1Global Pharmacovigilance and Epidemiology, Bristol Myers Squibb, Wallingford, 2GHEOR CORDS, Bristol Myers Squibb, 3Global Pharmacovigilance and Epidemiology, Bristol Myers Squibb, Pennington, USA. Background and aims: Channeling bias is a concern when designing obser- vational comparative studies including users of different drug classes for the management of Type 2 Diabetes Mellitus (T2DM). Characterizing the differ- ences in patient baseline characteristics across comparator drug classes will aid in the evaluation of channeling bias. This study described patient demo- graphic and clinical characteristics among US patients with T2DM initiating a new class of antidiabetic drug in 2012. Materials and methods: US adult patients with T2DM who initiated a new class of antidiabetic drug during the first six months of 2012 (N=183,692) were identified in the Truven MarketScan Commercial Claims and Encoun- ters database. The date of first new class of oral or injectable antidiabetic drug during the study period was defined as the index date. Patient’s baseline char- acteristics were identified in the 12 months prior to the index date. Comor- bidities were based on ICD-9-CM diagnostic codes and NDC codes used for prescriptions. Results: We identified 183,692 patients with majority (40%) initiating bi- guanides. Similar mean age was seen in all of the drug classes except GLP- 1 RA users who were younger. Contrary to other classes, GLP-1 RA users were predominantly female. Over 50% of new users had hyperlipidemia and hypertension at baseline with the prevalence being highest in DPP-4i and GLP-1 RA users. A diagnosis of obesity was most notable in GLP-1 RA users. History of heart failure was highest in insulin users. Although rare, history of acute and chronic pancreatitis was lowest in GLP-1 RA and DPP-4i users. Compared to biguanide users, more concomitant use of lipid lowering medi- cations was observed in users of all other classes. Conclusion: Differences were observed in the burden of comorbidities across different antidiabetic drug classes initiated which suggests a degree of channeling bias. Careful consideration and accounting of these factors dur- ing analysis is needed when conducting comparative observational studies among patients treated for Type 2 diabetes. Supported by: Bristol-Myers Squibb 325 Hospitalisation frequency for hypoglycaemia and emergency calls in type 2 diabetes mellitus patients exposed to vildagliptin vs insulin-secretagogues in the French Health Insurance database B. Detournay1, S. Dejager2, J. Robert1; 1Cemka-Eval, Bourg-la Reine, 2Novartis Pharma SAS, Rueil Malmaison, France. Background and aims: To compare the frequency of severe hypoglycaemic episodes leading to hospitalization and of emergency calls for any cause in patients exposed to the DPP-4 inhibitor vildagliptin vs those exposed to in- sulin secretagogues (IS; sulphonylurea or glinide). Materials and methods: Retrospective data were extracted from the EGB (Echantillon Généraliste des Bénéficiaires) database: a data resource com- prising a sample of ~1% of all patients registered in the French National Health Insurance system (≈600,000 patients). Type 2 diabetes mellitus pa- tients exposed to regimens comprising either vildagliptin (excluding treat- ment with IS, insulin or another incretin therapy) or IS (excluding treatment with insulin and any incretin therapy) between 2009 and 2012 were selected. Hospitalizations related to hypoglycaemia during the exposure periods were identified in both cohorts. Two comparative analyses adjusting for key co- variates within the model (subjects matched for age, gender, socioeconomic status, drug exposure duration), or with multivariate logistic regression, were performed. Results: 1,440 patients were exposed to vildagliptin and 10,019 to IS. Pa- tients in the IS cohort were older than in the vildagliptin cohort (mean 67.3 (SD: 12.8) years old vs. 63.5 (SD: 11.9), p S 142 1 C Conclusion: There was a significantly lower frequency of hospitalization for severe hypoglycaemia and of all emergency calls in patients exposed to vilda- gliptin vs IS. These real-life data should be taken into consideration in the benefit/risk evaluation of the drugs. Supported by: Novartis SAS 326 Predicting glycaemic changes in the non-diabetic general population: a DIRECT study A.D.M. Koopman, S.P. Rauh, the DIRECT Consortium; Department of Epidemiology and Biostatistics, EMGO Institute for Health and Care Research, Amsterdam, Netherlands. Background and aims: Recently, glycated haemoglobin (HbA1c) levels have been introduced as a diagnostic criterion for type 2 diabetes. Our study eval- uates whether in addition to baseline HbA1c levels, non-invasive measures can be used to predict changes in HbA1c levels during 6 years follow-up. Materials and methods: Data from 2,887 initially non-diabetic subjects (based on ADA 2011 criteria) from 3 population-based cohorts (Hoorn Study, Inter99, KORA S4) were used to develop sex-specific linear regres- sion models predicting change in HbA1c levels during follow-up. By using change in HbA1c levels, lab differences were avoided. To minimize overfit- ting of the model, we performed internal validation using bootstrapping techniques. Calibration was assessed with calibration graphs. Discriminative performance was assessed with classification tables, dichotomizing HbA1c levels (≥ 5.7% vs < 5.7%). Results: At baseline, mean HbA1c level was 5.6%. During a mean follow- up of 6 years (SD: 0.7 years), median change in HbA1c levels was +0.02%. HbA1c levels increased in 51% of the subjects, 2.3% of the subjects developed HbA1c levels ≥6.5%. After backward selection, next to baseline HbA1c levels, for men: age, waist circumference, smoking, and parental history of diabe- tes were retained in the prediction model (explained variance (R2): 33%); and for women: BMI and waist circumference (R2: 23%). Calibration plots showed good agreement between predicted and observed HbA1c levels at follow-up. With respect to discrimination, our model classified 75% of the subjects correctly as having high/low HbA1c levels. Conclusion: In the non-diabetic population, non-invasive predictors can be used next to baseline HbA1c levels to predict change in HbA1c levels during 6 years follow-up. This model can be used in clinical practice to determine which patients are at high risk of glycaemic deterioration. High-risk patients can then be monitored more regularly than low-risk patients and targeted preventive interventions can be initiated. Supported by: IMI Joint Undertaking 327 Factors associated with weight gain and hypoglycaemia and the impact upon hospitalisation in type 2 diabetes patients managed with metformin plus sulfonylurea J. Gordon1, P. McEwan1,2, M. Evans3, J. Puelles4; 1Health Economics and Outcomes Research Ltd, Monmouth, 2Swansea University Centre for Health Economics, 3Cardiff and Vale University Health Board, 4Takeda Global Research and Development, London, UK. Background and aims: The importance of avoiding complications that de- crease quality of life and consume healthcare resources in people with type 2 diabetes is well understood. However, the relationship between resource utilisation and patient phenotype is less well researched. The objective of this study was to assess factors associated with weight gain and the occurrence of hypoglycaemia in type 2 diabetes (T2DM) patients managed with metformin plus sulfonylurea (M+S), and any associated impact upon hospital resource utilisation. Materials and methods: The study was a retrospective cohort study using the UK Clinical Practice Research Datalink (CPRD) and the Hospital Episode Statistics (HES) database. The cohort analysed were those with T2DM treated with M+S for at least 3 months during the period 1/1/2001 to 31/12/2011; patients with a diagnosis of malignant disease were excluded. The association between phenotypic factors at baseline (therapy escalation from metformin to M+S) and weight gain (defined as > 2kg weight change over 12 months) and primary care recorded hypoglycaemia (>=1 episode) over 12 months fol- lowing therapy escalation was assessed using logistic regression. Hospitalisa- tion associated with increasing body mass index (BMI) and hypoglycaemia was also assessed. Analysis was undertaken using R version 2.12.2 Results: A total of 11,071 patients met the study inclusion/exclusion criteria with mean age at baseline of 60.7 (SD=11.4) years, 39% female, 5.4 (SD=4.0) years duration of diabetes, HbA1c 8.7% (SD=1.6), weight 92.2 kg (SD=19.6) and BMI 32.2kg/m2 (SD=6.1). Weight gain (1=weight gain; 0 otherwise) was observed in 28.35% (n=3,139) and was significantly associated with baseline age (OR=0.99 for one-year increase in age, p S 143 1 C 328 Analysis on incidence of new-onset diabetes mellitus after renal transplantation in different eras M. Yu1, C. Lv1, M. Chen1, M. Xu2, Y. Zhang1, S. He1, M. Xue1, J. Gao3, X. Gao1, T. Zhu2; 1Endocrinology, 2Urology, 3Evidence Base Medicine Center, Zhongshan Hospital affiliated to Fudan University, Shanghai, China Background and aims: To retrospectively evaluate the long-term fluctuation of fasting plasma glucose (FPG) after renal transplantation, in order to ex- plore the incidence of new-onset diabetes mellitus after transplantation (NO- DAT) in different eras. Materials and methods: We retrospectively evaluated 709 patients receiving kidney transplantation at our center between 1 January, 1993 and 31 Decem- ber, 2008. After excluding patients with uncompleted data, graft failure or death within 1 year after transplantation, multi-organ transplant recipients, transplant more than once or previously known diabetes, 428 patients were analyzed. The incidence of new-onset diabetes mellitus after renal transplan- tation was analyzed according to FPG in different eras. Immunosuppressive treatment after transplantation might be cyclosporine-A,(CSA)+ mycophe- nolate mofetil(MMF)(or Azathioprine (AZA)) + glucocorticoid or tacroli- mus (FK506)+MMF(or AZA)+ glucocorticoid. We use AZA for anti-prolif- erative drug before 1997 and MMF after that. We started using FK506 in 1999 and in some patients using CD25 monoclonal antibody for immunosuppres- sion induction since 2001. Patients were divided into three groups 1993-1996, 1997-2000, 2001-2008 according to different immunosuppressive treatment, and the incidence of NODAT was compared between different eras. Results: Of the 428 Patients, 87 developed NODAT (20.3%) during a mean follow-up of 6 years. According to the different immunosuppressive treat- ment in the different eras, patients were divided into three groups 1993- 1996, 1997-2000, 2001-2008, of which the incidence of NODAT was 18.42%, 17.46% and 21.11% respectively. There was no difference in the incidence of NODAT among three transplantation eras. Conclusion: The incidence of NODAT was 20.3% in patients surviving for more than 1 year during a mean follow-up of 6 years at Zhongshan Hospital. There was no difference in the incidence of NODAT among three transplan- tation eras. PS 008 Type 1 diabetes: genes and biomarkers 329 Primary autoantigen specific genetic traits in the pathogenesis of type 1 diabetes J. Lempainen1,2, A.-P. Laine1, A. Hammais1, J. Toppari2, R. Veijola3, O. Simell2, M. Knip4,5, J. Ilonen1,6; 1Immunogenetics Laboratory, University of Turku, 2Department of Pediatrics, University of Turku and Turku University Hospital, 3Department of Pediatrics, University of Oulu and Oulu University Hospital, 4Children’s Hospital, University of Helsinki and Helsinki University Central Hospital, 5Department of Pediatrics, Tampere University Hospital, 6Department of Clinical Microbiology, University of Eastern Finland, Kuopio, Finland. Background and aims: Natural history studies on type 1 diabetes (T1D) as- sociated autoimmunity provide an opportunity to identify detailed mecha- nisms of autoantigen specific genetic traits of diseases pathogenesis. We hy- pothetized that if the effect of a specific T1D-risk associated genetic marker is related to the triggering autoantigen of beta-cell destruction, the effect of the single variant should be dependent on the first appearing autoantigen- specific autoantibody. Materials and methods: The study subjects were participants in the prospec- tive DIPP study and carried T1D-risk associated HLA class II genotypes. From the DIPP study cohort, we were able to identify 170 subjects with IAA as the first biochemical autoantibody and selected 325 autoantibody nega- tive controls matched for gender, date of birth and study center. Similarly we identified 151 subjects with GADA as the first biochemical autoantibody and 285 autoantibody negative controls for them. Forty-three single nucleotide polymorphisms (SNPs) associated with T1D risk were genotyped using the Sequenom platform. Differences in effect of various SNPs on the develop- ment of autoantibodies and further progresssion to clinical T1D were tested among subjects with IAA or GADA as the first biochemically-defined au- toantibody. Cox regression analysis was performed to test the effect of the gene markers on the T1D pathogenesis. Results: The effect of INS rs689, IKZF4 rs1701704 and ERBB3 rs2292239 polymorphisms differed significantly between subjects with IAA or GADA as the first biochemical autoantibody. INS SNP strongly affected the devel- opment of clinical T1D and, in particular, the appearance of beta-cell hu- moral autoimmunity among the group with IAA as the first autoantibody (p=0.0054 and 0.0024, respectively) whereas no effect of the INS SNP on the T1D pathogenesis in the GADA group could be observed. In contrast, IKZF4 and ERBB3 SNPs were associated with the development of T1D in the GADA group (p=0.0022 and 0.0064, respectively) and, the two polymorphisms af- fected the progression rate of beta-cell destruction after the appearance of au- toimmunity in the GADA group (p=0.0025 and 0.0024, respectively) whereas no effect of these SNPs on the appearance of T1D or the progression rate of beta-cell destruction could be observed in the IAA group. Conclusion: The first appearing autoantibody specific approach revealed two genetic pathways dependent on the primary autoantigen. The effect of INS gene SNP on the pathogenesis of T1D was restricted to subjects with IAA as the first autoantibody whereas the SNPs in the IKZF4-ERBB3 locus increased the T1D risk and, more specifically, the progression rate of beta-cell destruc- tion among subjects with GADA as the first autoantibody. 330 Concentrations of soluble receptor for AGEs decline at seroconversion in children with preclinical type 1 diabetes but not in autoantibody positive non-progressors K.M. Salonen1, S.J. Ryhänen1, J.M. Forbes2, T. Härkönen1, J. Ilonen3, O. Simell4, R. Veijola5, P.-H. Groop6, M. Knip1,7; 1Children´s Hospital, University of Helsinki, Finland, 2Mater Research UQ at TRI, University of Queensland, Brisbane, Australia, 3Immunogenetics Laboratory, 4Department of Paediatrics, University of Turku, 5Department of Paediatrics, University of Oulu, , 6Folkhälsan Institute of Genetics, Folkhälsan Research Center, Helsinki, 7Department of Pediatrics, Tampere University Hospital, Finland. Background and aims: Our previous studies suggest that concentrations of soluble receptor for advanced glycation end products (sRAGE) might play a Diabetologia (2014) 57:[Suppl1]S1–S564 S 144 1 C role in the process leading to type 1 diabetes. We set out to define the changes in sRAGE in prediabetic and autoantibody positive children. Materials and methods: We analyzed serum concentrations of sRAGE from samples of 168 children who progressed to type 1 diabetes and 43 children who became positive for at least 2 diabetes associated autoantibodies during prospective observation, but have not progressed to diabetes. We analyzed the sRAGE concentration before seroconversion, in the first autoantibody positive sample, and annually after that until the diagnosis of type 1 diabetes or end of follow up. Results: The children who progressed to clinical type 1 diabetes were young- er at seroconversion and at end of follow up than the non-progressors [2.04 (± 1.5) years vs. 3.04 (± 1.5) years, p S 145 1 C the correlation between the percentage of apoptosis and expression of mi- R146a was r = 0.786 (p S 146 1 C 335 Gadolinium chelate (Gd-DO3A-GAD), new contrast agent for non- invasive quantification of beta cell mass K. Louchami1,2, S. Liang2, A. Rzajeva3, M. Aswendt4, A. Sener1, U. Himmelreich2; 1Laboratory of EXperimental Hormonology, Université Libre de Bruxelles, 2Biomedical MRI Unit/MoSAIC, Dept. Imaging & Pathology, KU Leuven, 3Laboratory of Hormonology, Université Libre de Bruxelles, Belgium, 4Max Planck Institute for Neurological Research, Cologne, Germany. Background and aims: The non-invasive imaging and quantification of pan- creatic islet is considered a high-priority field of diabetes investigation. This work deals with the use of Gd-DO3A-GAD as new contrast agent in the per- spective of pancreatic islets non-invasive imaging. The paramagnetic agent consists of Gd-DO3A as the Gd(III) chelate backbone. In its inactive state, long hydrocarbon side chains are anchored to one N atom of the Gd(III) chelate with glutamate moieties as head groups, which can enter the Gd(III) coordination sphere, thus limiting the water access to the paramagnetic core. Upon decarboxylation of the glutamate moieties due to the GAD (Glutamate decarboxylase) activity, there is an increase in the hydration sphere of the Gd(III) ion, leading to an increased T1 relaxivity. The present work, aims to perform the Glutamic acid decarboxylase characterization, the effect of Gd-DO3A on the insulin secretion evoked by D-glucose and the potential toxicity. Materials and methods: The wistar rat pancreatic islets were obtained using the collagenase digestion method. Glutamate decarboxylase activity (14CO2 production) was performed using the 14C-L-glutamic acid as substrate. The insulin secretion was measured by RIA (Radio Immuno-Assay) method. MRI experiments were performed using 9.4T Biospec small animal MRI system (Bruker, Ettlingen, Germany). Each phantom contains six tubes filled with 2% agarose in H2O. Standard Bruker T1_map_RAREVTR sequence is used. (TE = 12.7ms, FOV= 5cm*5cm, Matrix size = 128*128). Results: The Glutamate decarboxylase activity (14CO2 production) was measured using the 14C-L-glutamic acid as substrate in the pancreas ho- mogenates from normal vs diabetes (Sreptozotocine) wistar rats demonstrate that the enzyme activity was 4 times fold higher in the normal animals. The time related L-glutamate decarboxylase activity in rat islets and brain ho- mogenates after 120 min incubation reached respectively 45pmol/mg wet weight and 14pmol/1000islets. The present results demonstrate that the Gd- DO3A (50mM) fails to modulate the insulin secretion evoked by D-glucose at low (2.8 mM) or high (16.7 mM) of the hexose. Last, the Pancreatic is- lets (500), pancreatic islet cells (10.6) and INS1E (10.6) cells were incubated during 3hours in the presence of Gd-DO3A (50mM), phantom contains six tubes filled with 2% agarose in H2O were prepared, and the MRI experiments performed using 9.4T Biospec small animal MRI system (Bruker, Ettlingen, Germany). The results show clearly that Gd-DO3A (inactive state) was acti- vated probably due to GAD which catalyzes the conversion of glutamate into the major inhibitory neuron transmitter GABA. Conclusion: L-glutamate decarboxylase enzyme, which catalyzes the conver- sion of glutamate seems also catalyze the Gd-DO3A. The enzyme activity is highest in the control than diabetic animals. Last, the Gd-DO3A-GAD fails to modulate the insulin secretion and didn’t show any toxicity. The present re- sults support the idea that Gd-DO3A-GAD could be used as a contrast agent for non-invasive quantification of beta cell mass. 336 How to survive type 1 diabetes - the prolong study V. Lyssenko1,2, C. Møller1, P. Storm2, W. Poon2, P. Vikman2, L. Groop2, P. Rossing1, M. Eliasson3, B. Eliasson4, K. Brismar5, P.M. Nilsson2; 1Steno Diabetes Center A/S, Gentofte, Denmark, 2Clinical Sciences, Lund University, Malmö, 3Department of Public Health and Clinical Medicine, Umeå University, 4University of Gothenburg, 5Karolinska Institutet, Stockholm, Sweden. Background and aims: To successfully prevent and treat diabetes complica- tions, identification of protective mechanisms is required for a better mecha- nistic understanding of how the disease develops. PROLONG (PROtective Genes in Diabetic Complications and LONGevity) is a cross-sectional study of selected type 1 diabetic patients in Sweden and Denmark. The aim is to identify genetic markers, biomarkers and lifestyle factors, associated with protection from micro- or macrovascular diabetic complications in patients with long-standing type 1 diabetes. Materials and methods: Patients with long diabetes duration (more than 30 years) with no major diabetic complications (i.e. nephropathy, proliferative retinopathy or laser treatment, neuropathy, myocardial infarction, stroke). As a comparator group we ascertain patients with early development of dia- betic complications within 5-20 years of diabetes duration from the same geographical regions. DNA exome sequencing performed using Infinium HumanOmniExpressExome v1.1 DNA Analysis Kit. RNA sequencing per- formed using Illumina TruSeq Stranded mRNA Kit. Results: Presently, 246 patients with type 1 diabetes are enrolled in the PROLONG study in Sweden (M/F %, 44/56, mean±SD, duration 38±12 yrs, HbA1c IFCC 64±14 mmol/mol). Based on hospital records at a diabe- tes centre in Denmark, we identified ~500 patients (M/F% 49/51, duration 40±9 yrs, HbA1c IFCC 54±10 mmol/mol) without major diabetic complica- tions and 225 (M/F%, 50/50, duration 15±4 yrs, HbA1c 71.4±16.8 mmol/ mol) who developed either macro-, or microvascular complications. Cur- rently, the first DNA exome (n=96) and RNA sequencing analyses (n=96) have been completed on the Swedish PROLONG part. In line with previous observations in the Golden Years Project (UK), we observed that patients with long-standing diabetes without complications have higher HDL choles- terol (1.87±0.56 vs 1.40±0.45 mmol/l, p=1.8×10-3) and lower triglycerides 0.78±0.39 vs 1.12±0.46 mmol/l, p=2.9×10-3), and lower heart rate (74.4±10.5 vs. 83.2±14.7, p=0.001). Notably, we saw no differences in the levels of C- peptide (p=0.66), systolic (130±17 vs 130±22, p=0.86) or diastolic (89±9 vs. 92±11, p=0.18) blood pressure between individuals with or without com- plications. Furthermore, our preliminary analyses of the whole blood RNA expression pattern using Gorilla pathway analyses illustrated that sterol me- tabolism (DHCR24, SQLE, CYP51A1, HMGCS1, MSMO1 genes, etc) (cho- lesterol and sterol biosynthetic processes, p=9.9×10-7 and p=2.2×10-6; cho- lesterol and sterol metabolic processes, p=2.4×10-5 and p=4.6×10-5) was the top pathway associated with freedom from diabetic complications. Conclusion: Our preliminary cross-sectional data suggest that patients with longstanding diabetes but free of major diabetic complications are charac- terised by a favourable lipid profile. Analyses are ongoing to verify, as well as dissect, the specific genomic markers that significantly differ in the sterol metabolism linked to this unique phenotype. Supported by: Vetenskapsrådet, Novonordisk, Crafoord, HLF, Påhlsson 337 Interleukin-15 and interleukin-6 concentrations in autoimmune diabetes K. Siewko, R. Maciulewski, A. Zielinska, A. Poplawska-Kita, D. Lipinska, M. Gorska, M. Szelachowska; Department of Endocrinology, Diabetology and Internal Medicine, Medical University, Bialystok, Poland. Background and aims: Interleukin-15 and -6 (IL-15, IL-6) play a role in an inflammation, autoimmune, infectious and cancerous processes. Their in- creased concentrations were observed in psoriasis, asthma, multiple sclero- sis and type 2 diabetes. However, the role of IL-15 and IL-6 in autoimmune diabetes pathogenesis is still unknown. The aim of our study was to evaluate of relationship between IL-15, IL-6 and CRP concentrations in persons with newly diagnosis of autoimmune diabetes in compariosn to the first degree relatives of autoimmne diabetes and healthy controls. Materials and methods: The group studied consisted of 54 persons with newly diagnosis of autoimmune diabetes (28 with Latent Autoimmune Dia- betes in Adults (LADA) and 26 with type 1 diabetes (DM1)) and 70 healthy first degree relatives of patients with autoimmune diabetes and 60 healthy controls. GADA, IAA, IA-2A concentrations were measured by radioim- munoassays method. IL-15, IL-6 concentrations by ELISA method and CRP concentration by immunoturbidimetric method. Results: We found significantly higher concentrations of IL-15 and CRP in the whole diabetes group in comparison to the group of relatives (p S 147 1 C compared to the relatives without Ab, control group and group of patients with LADA and DM1 (p S 148 1 C data shows that gene expression profiles could be the most important lead for detecting the primary cause of diabetes. Materials and methods: In order to perform a generic analysis on gene ex- pression profiles, we have used a total of 32 promoters of genes associated with classical T1D and T2D phenotypes. In order to measure the structural features of promoter sequences, we have used our original method of analysis based on DNA patterns (BMC Genomics 2012, 13:512). Thus, these DNA patterns can identify the transcription factors shared by genes associated with one or more phenotypes. Among the analyzed promoters of genes as- sociated with classical T1D, we mention the HLA-DQA1, HLA-DQB1, HLA-DRB1, HLA-DPB1, PTPN22, TLR7, CTLA4, GSDMB, STAT4, IL7R, C1QTNF6, CD55, CTSH, ERBB3 and INS gene promoters and for T2D the CAMK1D, DUSP9, HHEX, IRS1, MADD, NOTCH2, TP53INP1, VPS13C, WFS1, ZFAND6, HMGA2, PPARG, CDKN2AIP, PROX1 and TCF7L2 gene promoters (Figure A-C). In order to show the general relationship between the three phenotypes, a second distribution was made by using the center of weight of the patterns (Figure D-F). Results: The promoters of genes associated with the two main phenotypes of diabetes contain different DNA patterns. T1D promoters exhibit image- based patterns which show that they are a part of a special class of promoters called „AT-based“ (Figure A). The promoters of genes associated with T2D exhibit patterns that show they are a part of a special class of promoters called „CG-based“ (Figure B). This separation of classes shows that genes associ- ated with these two phenotypes rarely share transcription factors, therefore these genes can not be coexpressed. The third type of pattern is presented by a number of genes such as CD55, C1QTNF6, INS, ERBB3, HMGA2, CTSH, SLC30A8, CDKN2AIP, PROX1, PPARG, TCF7L2 which suggest a new phe- notype, an Intermediary Diabetes Mellitus (IDM), (Figure C). The shape of these patterns indicate that genes associated with IDM can use transcription factors from both phenotypes, further indicating that IDM may contain the driver genes for triggering T1D and T2D. Conclusion: Our current data showed that some promoters of genes associ- ated with the two main phenotypes of diabetes seem to be associated to a third intermediary phenotype (IDM). These new data highlights the hetero- geneity of different clinical phenotypes of diabetes. Supported by: CNCS-UEFISCDI Nr. PN-II-ID-PCE-2011-3-0429 341 A common variant downstream of PCSK2 is associated with reduced tolbutamide stimulated insulin release: a DIRECT study A. Jonsson1, K. Banasik1,2, A.P. Gjesing1, A. Mahajan2, N. Robertson2,3, L.M. ‚t Hart4, E. Pearson5, M. McCarthy2,3, O. Pedersen1, T. Hansen1; 1The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Denmark, 2Wellcome Trust Centre for Human Genetics, University of Oxford, 3Oxford Centre for Diabetes, Endocrinology and Metabolism, 4Department of Molecular Cell Biology, Leiden University Medical Center, Netherlands, 5Division of Cardiovascular and Diabetes Medicine, Medical Research Institute, University of Dundee, UK. Background and aims: We have previously shown that there is a high herit- ability of insulin secretion in response to a tolbutamide injection. The aim of this study was to perform a genome-wide association study to identify gene variants associating with tolbutamide stimulated insulin release. Materials and methods: A total of 700 non-diabetic individuals, from two different cohorts, one family-based and one population-based, underwent a tolbutamide-modified frequently sampled IVGTT during which glucose was injected at 0 min and tolbutamide at 20 min. Measurements of plasma glucose, serum insulin and serum C-peptide were taken at 12 time points from fasting to 180 min and insulin secretion rate (ISR) and acute insulin response (AIR) were calculated. Genotyping was performed using the Illu- mina HumanCoreExome BeadChip. Association analyses adjusted for age, sex and BMI or age, sex and insulin sensitivity (Si) and separated by cohort were performed with EMMAX using the statistical package EPACTS, in or- der to handle relatedness in the samples. The results were then meta-analyzed using the GWAMA software. Results: The strongest signal for reduced tolbutamide stimulated insulin re- lease was seen downstream of the PCSK2 gene (AIR BMI adjusted: β=-0.216, P=1.6×10-6, AIR Si adjusted: β=-0.134, P=4.2×10-4), which has previously been associated with reduced glucose stimulated insulin secretion in humans. Conclusion: Here we show that the PCSK2 locus, not only associates with reduced glucose stimulated insulin secretion as described previously, but also shows a strong signal for reduced tolbutamide stimulated insulin release. The signal is however not genome-wide significant and further replication is needed to confirm the association. Supported by: EU IMI-DIRECT project 342 Exome variation affects metformin treatment response N. van Leeuwen1, G. Nijpels2, J.M. Dekker2, L.M. ‚t Hart1; 1Molecular Cell Biology, Leiden University Medical Center, 2EMGO Institute for Health and Care Research, VU University Medical Center, Amsterdam, Netherlands. Background and aims: Metformin is the first line treatment for type 2 diabe- tes. However the glycaemic response to metformin is highly variable between individuals. A recent study showed that the heritability of the absolute re- duction in HbA1c was 34% and chromosome-wise heritability estimates sug- gested that individual genetic variants scattered across the genome account for the effect. The aim of our study was the identification of genetic factors that influence treatment response to metformin. This will help to optimize and personalize treatment in the future and it will give further insight into the working mechanism. In this study we focused on genetic variation in the exome, the part of the genome that is translated into protein because genetic variation in this region has a large a priori chance of being causal. Materials and methods: Genetic variation was measured in 800 patients on metformin therapy from the Diabetes Care System (n>8000), a large longi- tudinal cohort of Dutch type 2 diabetic patients, using the Illumina Exome chip. Twenty-five percent of the patients on metformin therapy did not reach the treatment target of an HbA1c ≤53 mmol/l (7%) within one year after initiation of therapy and we assessed if this treatment failure was associated with genetic variation using logistic regression. In addition we assessed the association with the absolute change in HbA1c after one year of treatment using linear regression. In both analyses baseline HbA1c, age, BMI, eGFR, and sex, metformin dose and metformin dual/mono therapy were included as covariates. Results: Several genetic variants affected the response to metformin treat- ment in type 2 diabetes patients. The ten most significant gene variants gave a 1.7 to 8.5 fold reduced chance of reaching the treatment goal, an HbA1c level below 53 mmol/mol and the reduction in HbA1c ranged from 2 to 8 Diabetologia (2014) 57:[Suppl1]S1–S564 S 149 1 C mmol/mol per risk allele (p 6.10-6 to 3.10-4). Most of these variants were low frequency variants and several were predicted to be damaging by SIFT/ Polyphen2. However, due to the low frequency and power, replication in ad- ditional cohorts is necessary to reach genome wide significance. This replica- tion is currently performed. Conclusion: In this study we have identified a number of gene variants with relatively low frequency but a large effect on the metformin treatment re- sponse suggesting clinical usefulness in personalizing type 2 diabetes treat- ment. Furthermore, it identifies novel pathways modulating metformin re- sponse. However, replication in additional patients and cohorts is needed before definitive conclusions can be drawn. Supported by: ZonMW, Priority Medicines Elderly program (project nr. 113102006) 343 Insulin signalling genes exert a combined effect on all-cause mortality C. Menzaghi1, A. Fontana2, M. Copetti2, S. Rizza3, B. Spoto4, G. Tripepi4, A. Marucci1, A. Testa4, F. Mallamaci4, S. De Cosmo5, S. Bacci5, M. Federici3, C. Zoccali4, V. Trischitta1; 1Research Unit of Diabetes and Endocrine Diseases, IRCCS, San Giovanni Rotondo, 2Unit of Biostatistics, IRCCS, San Giovanni Rotondo, 3University of Rome Tor Vergata, 4Research Unit of Clinical Epidemiology and Physiopathology of Renal Disease and Hypertension, CNR-IBIM, Reggio Calabria, 5Unit of Endocrinology, IRCCS, San Giovanni Rotondo, Italy. Background and aims: Type 2 diabetes (T2D) and cardiovascular (CV) dis- ease are major factors increasing all-cause mortality. Both clinical entities rec- ognize a common soil represented by insulin resistance (IR), which by itself also predicts all-cause mortality. IR is, at least partly, genetically determined. Thus, it is concievable that genetic factors, which modulate IR, play also a role in modulating T2D, CV desease and all-cause mortality. In fact, we have previously reported the combined effect of single nucleotide polymorphisms (SNPs) perturbing insulin signaling (ENPP1 K121Q, rs1044498; IRS1 G972R, rs1801278; TRIB3 Q84R, rs2295490) on IR and, as a likely consequence, T2D and major CV events. Based on these encouraging results, we investigated whether a combined effect of these 3 SNPs affects also all-cause mortality. Materials and methods: We first studied a sample comprising 742 patients (i.e. discovery sample; 238 deaths/3,520 person-years; py). Replication was assessed in a second sample of 725 diabetic patients (i.e. replication sample; 129 deaths/5,495 py). Results: In the discovery sample, weighted genetic risk score (GRS), based on each SNP’s effect size, was associated with all-cause mortality (HR=1.12, 95% CI=1.03-1.23). After stratification according to low or high genetic load (GL) (i.e. 0-1 or > 2 risk alleles), patients with high GL (n=123) were at increased risk of all-cause mortality (HR=1.36, 95% CI=1.00-1.86), as compared to those with low GL (n=619). In the replication sample, HR (95% CI) for all- cause mortality was 1.06 (0.94-1.19) for GRS and 1.58 (1.06-2.35) for GL. In a pooled analysis (1,467 individuals; 367 deaths) both GRS and GL were associated with all-cause mortality HRs (95% CI)=1.11 (1.01-1.22) and 1.41 (1.10-1.80), respectively. Conclusion: Our finding indicates that functional non-synonymous variants affecting insulin signaling exert a joint effect on all-cause mortality and is consistent with a pathogenic role of IR on life expectancy. 344 Dissecting the genetic architecture of loci with established effects on multiple cardiometabolic phenotypes and type 2 diabetes L. Marullo1, T.O. Kilpeläinen2, B.K. Cornes3, J. Dupuis4, C. Scapoli1, R.J.F. Loos5, J.B. Meigs3, A.P. Morris6, I. Prokopenko7, on behalf of the XC-Pleiotropy Group; 1Department of Life Sciences and Biotechnology, Genetic Section, University of Ferrara, Italy, 2The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Denmark, 3General Medicine Division, Massachusetts General Hospital, Boston, 4Department of Biostatistics, Boston University School of Public Health, 5The Icahn School of Medicine at Mount Sinai, New York, USA, 6Department of Biostatistics, University of Liverpool, 7Department of Genomics of Common Disease, Imperial College London, UK. Background and aims: Genome-wide association studies (GWAS) have identified hundreds of loci associated with Type 2 Diabetes (T2D) or other cardiometabolic phenotypes, many of which overlap or map to the same genomic interval. Variants associated with multiple phenotypes, such as T2D, fasting insulin, triglycerides, HDL-cholesterol and body fat percentage influencing variants at IRS1, can provide insight into biology of correlated cardiometabolic phenotypes. However, the genetic architecture of these loci is frequently complex and needs further investigation. Materials and methods: To disentangle association patterns of 630 associ- ated SNPs (Dec 2012) from GWAS meta-analyses in Europeans for 19 quan- titative phenotypes, T2D and hypertension, we defined sets of adjacent vari- ants located less than 500kb apart and harboring 446 associated SNPs within 151 genomic regions (range=2-8 SNPs/region). We undertook approximate conditional analyses (ApCA) implemented in the GCTA tool to examine whether associations with multiple phenotypes within each region could be explained by LD. Results: Across the 151 regions, we observed 14 (10%) loci in which the same SNP was associated with multiple phenotypes. Associations in 11 of these 14 loci were with epidemiologically highly correlated traits. Through ApCA, we identified 41 (27%) regions with multiple associated variants that underlie the same association signals, thus suggesting multi-phenotype ef- fects. For 19 (13%) regions, the association with one phenotype partially ex- plained the effect on another. Within 45 (30%) regions, multiple signals were explained by multiple non-related variants, whereas the remaining 32 (21%) regions showed complex architecture. Of the 44 regions associated with T2D, 15 contained the same association signal for other cardiometabolic pheno- types. Within 12 regions, including ANKRD55, SPRY2, DUSP8, PEPD and HNF4A, association with other phenotypes were not related to T2D variants. For 13 regions we observed complex architecture, while for the remaining 4 regions, the association with T2D partially explained the effect on another cardiometabolic phenotype. Conclusion: Overall, a substantial number (87 or 58%) of cardiometabolic loci, of which 28 T2D loci, show potential pleiotropic effects on multiple phenotypes, which might contribute to their shared biology. Within other regions, distinct genetic effects or more complex architecture could underlie independent regulatory mechanisms. Supported by: EFSD 345 Genetic risk factors for diabetic complications in patients with type 2 diabetes from Ukraine S. Alkayyali1, W. Poon1, L. Cherviakova2, N. Khalimon2, P. Nilsson4, L. Groop1,5, T. Svietleisha6, T. Buldenko6, H. Vasylkova6, V. Lyssenko1,7; 1Diabetes and Endocrinology, Lund University, Malmö, Sweden, 2Chernihiv Regional Hospital, Ukraine, 4Clinical Science, Lund University, Malmö, Sweden, 5Institute for Molecular Medicine Finland FIMM, Helsinki, Finland, 6Public Health Authority of Chernihiv Regional Public Administration, Ukraine, 7Steno Diabetes Center, Gentofte, Denmark. Background and aims: Chronic hyperglycaemia is associated with increased risk of progression to macro (cardiovascular diseases) and microvascular complications (neuropathy, retinopathy and nephropathy). Recently, ge- nome-wide association studies have identified a number of genetic loci for association with type 2 diabetes (T2D) and cardiometabolic traits. The ef- fects of some of these genetic variants and the risk of diabetes progression to diabetes complications has been investigated in several populations, but has never been studied in Ukrainians, a population at high cardiovascular risk. Materials and methods: We studied the association of a panel of 145 SNPs in loci previously reported to be associated with T2D/glycaemic traits (n=75), dyslipidemia (n=11), obesity (n=8), hypertension (n=13), cardiovascular diseases (CVD) (n=21) and microvascular complications (n=17) in ap- proximately 3,500 subjects with T2D from the DOLCE study (M/F% 32/68, mean±SD, age-at-onset 53.5±10.6 years, BMI 31.4±5.6 kg/m2, diabetes dura- tion 7.1±7.4 years) (Diagnostic optimization and treatment of diabetes and its complications in the Chernihiv region). Effects of genetic loci were studied using logistic regression adjusted for sex and age-at-onset for macrovascular complications, and sex and diabetes duration for microvascular complica- tions. The analyses were performed using R software, and genotyping was performed using Mass ARRAY iPLEX (Sequenom, San Diego, CA). Results: We have replicated previous associations of GIPR rs10423928 (OR=1.29, P=0.005), WDR12 rs6725887 (OR=1.42, P=0.005) and MIA3 rs17465637 (OR=1.24, P=0.04) with CVD. VEGF rs2010963 (OR=1.28, P=0.02) previously reported to be associated with diabetic retinopathy (DR), was in our study associated with diabetic nephropathy (DN). The risk C-allele in TMEM26 rs1530440 (hypertension locus) was associated with increased Diabetologia (2014) 57:[Suppl1]S1–S564 S 150 1 C CVD risk (OR=1.18, P=0.04) and DN/DR (OR=1.4, P=0.01), while other two hypertension loci ATXN2 rs653178 (OR=0.87, P=0.03) and CYP17A1 rs11191548 (OR=0.75, P=0.03) with decreased CVD risk . Also, the ERBB4 rs7588550 (DN locus) was associated with increased risk for CVD (OR=1.79, P=0.01), while CVD loci PHACTR1 rs12526453 (OR=1.8, P=0.016) and PITX2 rs6843082 (OR=1.25, P=0.04) with increased risk for DR and DN. Conclusion: We have replicated the previous association in GIPR, WDR12 and MIA3 for risk of CVD. Furthermore, we have demonstrated an associa- tion between risk alleles for hypertension and T2D/glycaemic traits loci and micro-/macrovascular complications. None of the studied BMI or dyslipi- demia loci were significantly associated with complications in our study. Supported by: Vetenskapsrådet, Novonordisk, Crafoord, HLF, Påhlsson. 346 Chemotactic cytokine receptor 5 (CCR5) gene promoter polymorphism (rs1799987) increases the risk of diabetic nephropathy in Asian type 2 diabetic patients Z. Zhang; Shandong Provincial Qianfoshan Hospital, Jinan, China. Background and aims: To evaluate the association of CCR5/CCL5 variants with the risk of DN Materials and methods: We conducted a systematic search of electronic da- tabases (Pubmed , Embase and China National Knowledge Infrastructure) on three genetic variants (CCL5-403 G/A, CCL5-28 C/G, CCR5 59029G/A) and then 11 case-control studies involving 2512 DN cases and 2358 non DN con- trol subjects were identified. The pooled odds ratio (OR) and 95% confidence interval (CI) were used to describe the strength of association with DN, the subgroup analysis was used to explore the heterogeneity bias among studies. Publication bias was tested by the Begg’s and Egger’s test. Results: In the overall analysis, we found that CCR5 59029A-positive geno- type (G/A or A/A) was an independent risk factor of DN (OR 1.69, 95% CI 1.13-2.55). Subgroup analysis demonstrated that the risk of CCR5 59029A- positive genotype was more significant among Asian patients with type 2 diabetic nephropathy (OR 2.08, 95% CI 1.68-2.57), but was nonsignificant in Caucasians or type 1 diabetes nephropathy (OR 1.23, 95% CI 0.52-2.92 for Caucasians; OR 0.82, 95% CI 0.56-1.20 for type 1 diabetes nephropathy). In addition, CCR5 59029A-positive genotype was associated with increased risk of albuminuria (OR 1.68, 95% CI 1.15-2.44 for microalbuminuria; OR 2.53, 95% CI 1.05-6.08 for macroalbuminuria). The CCL5 -403 G/A and CCL5-28 C/G gene polymorphism was not significantly associated with the risk of DN (OR 1.00, 95% CI 0.82-1.21 for CCL5-403 G/A; OR 1.02, 95% CI 0.79-1.33 for CCL5-28 C/G). Conclusion: Our studies indicated that the CCR5 59029 A gene variant is a significant susceptibility factor for DN. Subtype of DM and ethnicity might contribute for the inconsistency present in studies. The CCL5 gene polymorphism(CCL5-403 G/A, CCL5-28 C/G ) might not be a risk factor for DN. Supported by: CRD42014007495 347 Indications for potential parent of origin effects within the FTO gene using long range phasing algorithms X. Liu1,2, M. Scholz3,4, A. Tönjes5, M. Stumvoll1,5, P.F. Stadler2,6, Y. Böttcher1; 1IFB AdiposityDiseases, 2Bioinformatics Group, Department of Computer Science, 3Institute for Medical Informatics, Statistics and Epidemiology, 4LIFE Research Centre, 5Department of Medicine, University of Leipzig, Germany, 6Institute for Theoretical Chemistry, University of Vienna, Austria. Background and aims: Genome-Wide Association Studies (GWAS) were successfully applied to discover associations with obesity. However, the GWAS design is based on unrelated individuals and clear inheritance infor- mation of phases is limited. Taking into account parent of origin may provide further insights into the genetic mechanisms contributing to obesity. We hy- pothesized there may be variants within the robustly replicated fat mass and obesity associated (FTO) gene that may confer different effect sizes for obesity depending on the transmission from father or mother. Materials and methods: Genome-wide genotypes and pedigree information from the Sorbs population (N=525) were used. Phasing was done by applying long-range phasing and haplotype library imputation algorithm. Phased gen- otypes among 525 individuals were generated by AlphaImpute. Subsequently, 22 SNPs within FTO introns 1 to 3 were selected and parent of origin specific association analyses were performed using PLINK: (i) standard association test, (ii) considering paternal and (iii) maternal alleles. Results: We identified several SNPs conferring different P values depending on parental origin. Among them, rs1861868, rs1121980 and rs9939973 (all intron 1) show significantly different effect estimates beta (Student´s t-Test; P S 151 1 C PS 010 Genomics and epigenetics of type 2 diabetes 348 PARP-1 inhibitor, Olaparib, increases GLP-1 secretion and promotes both insulin secretion and TCF7L2 gene expression Q. Xia1, S. Lu1, B.T. Johnston1, S. McCormack2, S.F.A. Grant1,3; 1Division of Human Gentics, 2Division of Endocrinology, The Children‘s Hospital of Philadelphia, , 3Department of Pediatrics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, USA. Background and aims: TCF7L2 is widely considered one of the most strongly type 2 diabetes (T2D) associated loci reported to date, but the patho- physiologic mechanism by which variation within this gene modulates risk is poorly understood. We previously reported the specific protein factors that bind across the presumed causal variant at this locus, rs7903146, using oligo pulldown followed by mass spectrophotometry, with the most abundant of these being poly [ADP-ribose] polymerase type 1 (PARP-1) (Xia et al., 2014). PARP-1 has a potential role in the pathophysiology of diabetes, including evidence that PARP-1 knockout mice are protected from streptozotocin- induced diabetes. Proteomic studies have suggested that PARP-1 is also a component of the TCF7L2/beta-catenin complex. One important potential regulatory target of TCF7L2 is the pro-glucagon gene, which encodes a pre- cursor of glucagon as well as other factors important for glucose homeosta- sis, including glucagon-like peptide-1 (GLP-1). We investigated if an existing PARP-1 inhibitor developed for oncologic indications, Olaparib, could have a GLP-1 agonist effect and thereby suggest regulation of PARP-1 mediated GLP-1 activity as one candidate mechanism for TCF7L2-related modulation of T2D risk. Materials and methods: We used the human L-cell line, NCI-H716, for GLP- 1 secretion assessment. Cells were pre-treated with 10uM Olaparib and after 48 hours the cells were transferred into PBS and incubated with or without 16mM glucose for 30 minutes, after which the suspensions were collected for GLP-1 ELISA. Next, we tested the effect of Olaparib on L-cell mediated augmentation of insulin secretion. To do this, we collected the medium from the treated NCI-H716 cells and used it to culture the human beta cell line, EndoC-BH1, for 30 minutes, after which we measured insulin by ELISA. Fi- nally, we used real-time PCR to compare relative expression of TCF7L2 in NCI-H716 L-cells in the presence and absence of Olaparib. Results: Treating NCI-H716 cells with glucose increased the release of GLP- 1 into the medium as expected but, interestingly, Olaparib treatment also enhanced GLP-1 secretion significantly compared with the untreated group (51.1% average increase; P< 0.05). Furthermore, GLP-1 levels were further increased in the presence of the combination of glucose and Olaparib, com- pared with glucose or Olaparib alone (26.0% and 29.4% average increase, re- spectively; P S 152 1 C cleotides DNA sequences. Notably, 14 of these variants were found in risk loci for T1D, T2D, glucose traits and obesity (WFS1, CLEC16A, ZMIZ1, FUT2, CAPN10, TAGAP, KCNJ11, SH2B3, CTSH, HNF1B, PRKD2, SLC30A8 and KCNQ1), 2 197 were overlapped with DIAGRAM. The top nine loci (p < 1.0e-30) were found located in TCF7L2, followed by six loci (p < 1.0e-13) in CD- KAL1. However, most (81%) loci did not show associations with diabetes. 485 putative sites of RNA editing were cis eQTL of 137 genes and 12 putative sites of RNA editing were trans eQTL of 51 genes (false-discovery rate of 0.50). Conclusion: Our results suggest that RNA-DNA mismatches should be con- sidered as an important factor in genetic studies of diabetes. The consequenc- es of editing in risk alleles need to be further examined. Supported by: Swedish Research Council and Regions Skåne ALF grants 351 Global Long Interspersed Nucleotide Element 1 (LINE-1) DNA methylation in a longitudinal cohort of type 2 diabetes mellitus patients N.S. Malipatil1, K. Siddals1, B. Fu2, M. Gibson3, R.P. Narayanan4, A.H. Heald5, R. Donn6; 1Vascular Research Group, 2Centre for Biostatistics, Institute of Population Health, University of Manchester, 3Vascular Research Group, Salford Royal NHS Foundation Trust, Manchester, 4Department of Obesity and Endocrinology, University of Liverpool, 5Department of Medicine, Leighton Hospital, Crewe, 6Centre for Musculoskeletal Research, University of Manchester, UK. Background and aims: Epigenetic mechanisms impact gene expression and could predispose individuals to a particular metabolic phenotype. Increased LINE-1 methylation has previously been associated with cardiovascular risk biomarkers in healthy individuals. Here we have investigated, for the first time, the relationship between global LINE-1 DNA methylation and cardio- metabolic parameters in Type 2 Diabetes Mellitus (T2DM) patients. This has been conducted using a well characterised longitudinal cohort from Salford, UK. Materials and methods: Global LINE-1 DNA methylation was quantified by pyrosequencing in blood-derived DNA samples from 445 Caucasian T2DM patients using PyroMark Q96 CpG LINE-1 (Qiagen). The cohort consisted of Males 267 and Females 178; mean age 59.4yrs (males 58.7 years, females 60.4 years), with 7 years of longitudinal data. Global LINE-1 DNA methylation was analysed in relation to baseline anthropometric and biochemical meas- urements and their average changes per year over follow up, using multiple linear regression models adjusted for age, body mass index (BMI) and strati- fied by gender groups. Results: Methylation at 4 CpG sites was quantified. The mean across the 4 sites was 75.1% (95% Confidence Interval (CI) 74.83% - 75.33%). There were no statistically significant methylation differences by gender (males 75.2%, females 74.90% p=0.270) or by age (p=0.679). Linear regression showed no significant associations with LINE-1 DNA methylation in the total cohort with; BMI [normalised beta coefficient 2.76 (95% CI -13.91 - 19.44, p=0.745)], total cholesterol [-1.59 (95% CI -4.02 - 0.83, p=0.198)], triglycerides [-2.77 (95% CI -6.76 - 1.22, p=0.173)], HDL-cholesterol [0.64 (95% CI -0.55 - 1.83), p=0.293] or LDL-cholesterol [-8.9 (95% CI -3.09 - 1.30, p=0.425)], or with systolic Blood Pressure (BP) [-22.51 (95% CI -66.75 - 21.28, p=-0.313)] or diastolic BP [-17.81 (95% CI -44.00 - 21.28, p=0.313)]. Furthermore, there was no association with glycated haemoglobin (Hba1C) [-1.35 (95% CI -5.69 - 2.98, p=0.540)] at baseline or with average change per year in Hba1C over 7 years follow up [-0.004 (95% CI -0.014 - 0.006, p=0.443)]; nor with levels of high sensitivity C-reactive protein (CRP) [13.57 (95% CI -21.36 - 48.52, p=0.445)]. However, stratification by gender revealed significant association of global LINE1 DNA methylation in women with LDL levels [-4.52 (95% CI -8.23 - 0.80, p=0.017)], with HDL: LDL cholesterol ratio [2.78 (95% CI 1.02 - 4.53, p=0.002)] and a trend also towards hypomethylation with total cholesterol [-3.76 (95% CI -7.87 - 0.34, p=0.072)]. Conclusion: Our findings indicate a gender related difference in global LINE1 DNA methylation with specific blood metabolic parameters related to cardiovascular risk in type 2 diabetes mellitus. Supported by: Hodginson Endowment Fund 352 Altered DNA methylation and differential expression of genes influencing metabolism and inflammation in adipose tissue from subjects with type 2 diabetes E. Nilsson1,2, P. Jansson3, A. Perfilyev1, M. Pedersen4, M.K. Svensson5, P. Poulsen6, R. Ribel-Madsen2, N.L. Pedersen7, P. Almgren8, J. Fadista8, T. Rönn1, B. Klarlund-Pedersen4, C. Scheele4, A. Vaag2, C. Ling1; 1Epigenetics and Diabetes, Clinical Sciences, Malmö, Sweden, 2Diabetes and Metabolism, Department of Endocrinology, Copenhagen, Denmark, 3Wallenberg laboratory, Gothenburg, Sweden, 4Infectious Diseases, The Centre of Inflammation and Metabolism and the Centre for Physical Activity Research, Copenhagen, Denmark, 5Sahlgrenska University Hospital, Institute of Medicine, Gothenburg, Sweden, 6Novo Nordisk A/S, Global Development, Bagsværd, Denmark, 7Karolinska institutet, Department of Medical Epidemiology and Biostatistics, Stockholm, 8Diabetes and Endocrinology, Clinical Sciences, Malmö, Sweden. Background and aims: Genetics, epigenetics and environment may together affect the susceptibility for type 2 diabetes (T2D). Our aim was to dissect molecular mechanisms underlying T2D using genome-wide expression and DNA methylation data in adipose tissue from monozygotic twin pairs dis- cordant for T2D and independent case-control cohorts. Materials and methods: Genome-wide DNA methylation and mRNA ex- pression were analysed using the Infinium HumanMethylation450 BeadChip from Illumina and GeneChip® Human Gene 1.0 ST arrays from Affymetrix, respectively. Results: In adipose tissue from discordant twin pairs, we found that de- creased expression of genes involved in oxidative phosphorylation, carbohy- drate-, amino acid- and lipid metabolism, and increased expression of genes involved in inflammation and glycan degradation accompany T2D. The most differentially expressed genes included ELOVL6, GYS2, FADS1, SPP1 (OPN), CCL18 and IL1RN. We replicated these results in adipose tissue from an in- dependent case-control cohort. Several candidate genes for obesity and T2D (e.g. IRS1 and VEGFA) were differentially expressed in discordant twins. We found a heritable contribution to the genome-wide DNA methylation variability in twins. Differences in DNA methylation between monozygotic twin pairs discordant for T2D were subsequently modest. However, 15,627 sites, representing 7,046 genes including candidate genes for T2D and obe- sity showed differential DNA methylation in adipose tissue from unrelated subjects with T2D compared with controls. 1,410 of these sites did also show differential DNA methylation in the twins discordant for T2D. For the differ- entially methylated sites, the heritability estimate was 0.28. We also identified copy number variants in monozygotic twin pairs discordant for T2D. Conclusion: Taken together, subjects with T2D exhibit multiple transcrip- tional and epigenetic changes in adipose tissue relevant to the development of the disease. Supported by: The Swedish Research Council, ALF, EFSD/Lilly, Söderberg, Påhlsson, Novo Nordisk 353 Diet-induced obesity alters methylation of HOX transcription factors in mouse visceral adipose tissue L. Parrillo1, V. Costa2, R. Esposito2, G.A. Raciti1, G. Porciello1, C. Nigro1, P. Mirra1, A. Cozzolino1, C. Miele1, A. Ciccodicola2, F. Beguinot1; 1DISMET & IEOS-CNR, Federico II University of Naples, 2IGB-CNR, Italy. Background and aims: Exposure to high-fat regimens induces epigenetic modifications which might shape the obesity phenotype. Indeed, recent work has revealed altered DNA methylation at several obesity-related genes in response to high-fat diets. However, the detailed methylome profiles deter- mined by high-fat exposure remain to be elucidated. In the present study, we have implemented an unbiased genome-wide analysis of DNA methylation changes secondary to diet-induced obesity. Materials and methods: Male C57BL/6J mice were fed either high-fat (HFD) or regular chow diets (ND) for 5 months. Methylated DNA immunoprecipi- tation combined with next generation sequencing (MeDIP-seq) were per- formed on DNA extracted from the visceral epididymal fat (VF) of the ND- and the HFD-fed mice. RT-qPCR and MeDIP-qPCR were then performed on the VF of these mice. Results: Mice fed HFD gained weight (difference with the ND-fed mice sig- nificant at p S 153 1 C ND mice). MeDIP-seq analysis identified differentially methylated regions (DMRs) covering with sufficient depth almost the entire genome and showed that, in the HFD-fed mice, the number of hypermethylated DMRs was higher than that of hypomethylated DMRs. Most DMRs were located within gene bodies. Based on subsequent gene ontology analysis of the MeDIP-seq data, differentially methylated genes (DMGs) were identified in the HFD- fed mice. Interestingly, a distinct set of genes belonging to the Homeobox (HOX) gene family was found to be epigenetically modulated by chronic exposure to HFD (Hoxa1, Hoxa3, Hoxa5, Hoxb3, Hoxb6, Hoxd1, Hoxa13). Hoxa5 was selected to validate the significance of this finding by measure- ment of its promoter methylation status and its gene expression. Results of the MeDIP-qPCR analysis (p S 154 1 C blood samples at fasting for metabolic (glucose and insulin) and genetic analyses were collected. The oral glucose tolerance test (OGGT) have been performed and type 2 diabetes was diagnosed based on WHO criteria. Body composition: percentage of total body fat, visceral (VAT) and subcutaneous adipose tissue (SAT), VAT/SAT ratio were also analyzed by bio-impedance method. Genetic risk score (GRS) for each participant was calculated using previously identified 70 loci by weighting the number of risk alleles cor- responding to a particular locus by its effect size (β-estimate) on diabetes risk and summing these values. The studied genotypes distributions were in Hardy-Weinberg equilibrium (p>0.05). Results: In the logistic regression analysis the risk of type 2 diabetes develop- ment was significantly associated with age [OR=1.2 (1.1-1.4), p=2.8x10-6], HOMA-IR [OR=1.9 (1.4-2.6), p=1.9x10-5], VAT/SAT ratio [OR=11.4 (2.0- 91.7), p=0.01] and GRS [OR=1.13 (1.05-1.2), p=2x10-5] but not with BMI and gender. When particular SNPs were introduced into the regression model instead of GRS the strongest association with type 2 diabetes risk was found for rs11558471 SLC30A8 (p=0.03), rs226000 PRRC2A (p=0.018), rs11708067ADCY5 (p=0.008), rs2844479 AIF1 (p=0.02), rs10938397 GN- PDA2 (p=0.001), rs7647305 SFRS10 (p=0.01), rs174550 FAD1 (p=0.016), rs10946398 CDKAL1 (p S 155 1 C Results: Transfection of WT and mutant (K392R) STAT3 into HEK293 cells resulted in equal levels of protein expression but differing levels of transcrip- tional activation. Notably, the K392R mutant dramatically increased STAT3 reporter activity relative to the WT (by 37.1±4.7 fold; p S 156 1 C in the controls (p=0.0041). In post hoc analysis maximum IMT for GCK- MODY was statistically different from HNF1a MODY (p=0.0371) Mean IMT were as follow - 0.62±0.14 mm in GCK-MODY, 0.68±0.13 in HNF1A-MODY and 0.63±0.13 in controls (p=0.0077). Post hoc analysis showed a difference between GCK-MODY and NF1A-MODY (p=0.0108). The mean FMD% was 11.0±4.5% in GCK-MODY, 10.1±4.9 in HNF1A-MODY and 13.9 ±4.8 in the controls (p=0.0001). The differences between CK-MODY and HNF1A- MODY vs. controls were significant (p=0.0045, p=0.0001, espectively). ‚I was similar in all three groups - 24.1±4.5, 24.1±3.7 nd 24.2±3.6 in GCK-MODY, HNF1A-MODY and controls, respectively (p=0.7876). Patients were diag- nosed with diabetes at similar age (GCK-MODY: 25.1±13.5 yrs vs. HNF1A- MODY: 26.6±11.6, =0.4669). Glycemic control was similar in diabetic groups as in the GCK-MODY roup the mean HbA1c was 6.4%±0.7, while in HNF1A- MODY patients it reached 6.7±1.4 (p=0.6936). Conclusion: Both examined MODY grups showed evidence of early athero- sclerosis or endothelial dysfunction. Mild hyperglycemia in the GCK-MODY seems to have a fragile impact on the ccurrence of intermediate atheroscle- rotic phenotypes. Supported by: ODW-5224/B/P01/2011/40 362 Clinical features of MODY-HNF1a in children and adolescents with obesity E.A. Sechko1, L.I. Zilberman2, O.N. Ivanova2, T.L. Kuraeva2, V.A. Peterkova2; 1I.M. Sechenov First Moscow State Medical University, 2Endocrinology Research Centre, Moscow, Russian Federation. Background and aims: to study clinical features of MODY-HNF1a in chil- dren and adolescents with obesity. Materials and methods: 18 patients who had heterozygous mutation in HN- F1a gene were divided into 2 groups: group 1 - the patients with SDS BMI ≥ 2 (n=6; 33.3%); group 2 - the patients with SDS BMI< 2 (n=12; 66.6%). We measured fasting and stimulated glucose, C-peptide, insulin levels (OGTT). The data is presented as medians (25; 75 percentile), Mann-Whitney U-test was used to compare medians. Results: Age at diagnosis of diabetes: in group 1 was 10.4 years (9.5;14.4), in group 2 was 12.8 years (10.9;15.4), p0.05). 100% patients in group 1 and 91,6% patients in group 2 had family history of diabetes; one of the parents had insulin dependent diabetes mellitus (66.6% and 25%, respectively group 1 and group 2), non insulin dependent diabetes mellitus (16.65% and 58.3%, respectively group 1 and group 2), gestational diabetes mellitus (16.65% and 8.3%, respectively group 1 and group 2). Disease duration was 3 years (1.1; 4.5), when diagnosis MODY- HNF1a was confirmed. We indentified the novel mutations: D45fs, V119G, R229X, S249Stop. The most common Pro291 fsin- sC mutation was indentified in 5 cases (27.7%). Clinical data is presented in Table 1. Patients in group 1 had significantly higher HbA1c and fasting serum C-peptide. Two patients in group 1 were insulin resistant. Patients in groups 1 and 2 were treated with sulfonylurea (50% of all patients) and metformin (50% and 8%, respectively). 42% patients in group 2 were treated with diet only. Conclusion: MODY-HNF1a may be combined with obesity and in some cases with insulin resistance in children and adolescents. Obesity is a risk factor for early manifestation and more severe diabetes. Patients in group 1 had higher C-peptide secretion. DNA testing of patient with mild diabetes, obesity and family history of diabetes could lead to diagnosis of new cases of MODY-HNF1a. 363 Transcription activity testing can aid bioinformatics in ascribing pathogenicity to HNF1A variants identified by genomic sequencing in population-based cohorts I. Aukrust1,2, L.A. Najmi1,3, J. Molnes1,4, J. Flannick4, N.P. Burtt4, A. Molven1,5, L. Groop6, D. Altshuler4, S. Johansson1, L. Bjørkhaug1, P.R. Njølstad1,7; 1KG Jebsen Center for Diabetes Research, Dept of Clinical Science, University of Bergen, 2Dept of Biomedicine, University of Bergen, 3Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, Bergen, Norway, 4Program in Medical and Population Genetics, Broad Institute of Harvard and Massachusetts Institute of Technology (MIT), Cambridge, USA, 5Gade Laboratory for Pathology, Dept of Clinical Medicine, University of Bergen, Norway, 6Dept of Clinical Sciences, Diabetes and Endocrinology, Clinical Research Centre, Lund University, Malmö, Sweden, 7Dept of Pediatrics, Haukeland University Hospital, Bergen, Norway. Background and aims: In a population sample of randomly selected indi- viduals from 3 different cohorts, as much as 2% carry rare non-synonymous variants in 1 of 7 maturity-onset diabetes of the young (MODY) genes. Inter- estingly, the majority of these individuals remain euglycemic through middle age. To evaluate their risk for T2D later in life, detailed analyses are required to determine the true pathogenic nature of these MODY gene variants, in ad- dition to family pedigrees/phenotype analyses combined with bioinformatics in silico tools. Our aim is to test whether functional characterization of 25 rare variants (MAF S 157 1 C Materials and methods: HEK 293 cells stably expressing CEL-MUT and CEL-WT proteins were used to produce conditioned medium for use in endocytosis studies in acinar cells (rat AR42J and mouse 266-6) and β-cell models (rat INS-1E and mouse MIN6). We employed immunofluorescence confocal microscopy and assays of cell viability and glucose-stimulated insu- lin secretion. Results: In stably transfected HEK293 cells, CEL-MUT was distributed in a cytoplasmic punctate pattern, whereas the CEL-WT predominantly local- ized typically for secreted proteins in the Golgi compartments and the endo- plasmic reticulum. When untransfected acinar cells and β-cells were grown in conditioned medium from CEL-expressing cells, an accumulation of the CEL-MUT protein was observed in punctate structures. Long-term incuba- tion of acinar and β-cells in conditioned medium showed a tendency, how- ever not significant, of lower viability in the presence of CEL-MUT compared with CEL-WT. Glucose-stimulated insulin secretion in MIN6 cells was not reduced in the presence of the mutant protein. Conclusion: Our results suggest that CEL-MUT can be endocytosed both by acinar and β-cells. The endocytic process might represent a mechanism of protecting cells against the exposure to CEL aggregates by cell-mediated uptake and degradation of the potentially toxic protein. Furthermore, endo- cytosis could be directly relevant for how the CEL-MUT causes disease af- fecting both the exocrine and endocrine pancreas. PS 012 Monogenic diabetes across the world 365 Wolfram syndrome in the Japanese population: molecular analysis of the WFS1 gene and characterisation of clinical features Y. Tanizawa, K. Matsunaga, K. Tanabe, on behalf of the Japan study group of Wolfram syndrome; Division of Endocrinology, Metabolism, Hematological Science and Therapeutics, Yamaguchi University, Graduate School of Medicine, Ube, Japan. Background and aims: Wolfram syndrome (WFS) is a recessive neurologic and endocrinologic degenerative disorder, and is also known as DIDMOAD (Diabetes Insipidus, early-onset Diabetes Mellitus, progressive Optic Atrophy and Deafness) syndrome. Most affected individuals carry recessive mutations in the Wolfram syndrome 1 gene (WFS1). However, the phenotypic pleio- morphism, rarity and molecular complexity of this disease complicate our efforts to understand WFS. To address this limitation, we aimed to describe complications and to elucidate the contributions of the WFS1 mutations to clinical manifestations in patients with WFS. Materials and methods: The minimal ascertainment criterion for diagnos- ing WFS was having both early onset diabetes mellitus (DM) and bilateral optic atrophy (OA). Genetic analysis for the WFS1 was performed by direct sequencing. Results: Sixty-seven patients were identified nationally for a prevalence of one per 710,000, with 33 patients (49%) having all 4 components of DID- MOAD. In 40 subjects who agreed to participate in this investigation from 30 unrelated families, the earliest manifestation was DM at a median age of 8.7 years, followed by OA at a median age of 15.8 years. However, either OA or DI was the first diagnosed feature in 6 subjects. In 10, features other than DM predated OA. Twenty-seven patients (67.5%) had a broad spectrum of recessive mutations in the WFS1. Two patients had mutations in only one allele. Eleven patients (27.5%) had intact WFS1 alleles. Ages at onset of both DM and OA in patients with recessive WFS1 mutations were indistinguish- able from those in patients without WFS1 mutations. In the patients with predicted complete loss-of-function mutations, ages at the onsets of both DM and OA were significantly earlier than those in patients with predicted partial-loss-of function mutations. Conclusion: This study emphasizes the clinical and genetic heterogeneity in patients with WFS. Genotype-phenotype correlations may exist in patients with the WFS1 mutations, as demonstrated by the disease onset. Supported by: Grants from MHLW and MEXT of Japan 366 Czech prediction programme for type 1 diabetes: additional source of MODY patients? P. Dusatkova, L. Petruzelkova, S. Pruhova, J. Vcelakova, K. Stechova, O. Cinek, Z. Sumnik, J. Lebl, S. Kolouskova; Department of Pediatrics, 2nd Faculty of Medicine, Prague, Czech Republic. Background and aims: Since 2001, we prospectively follow up first-degree relatives of patients with Type 1 Diabetes (T1D) in the Czech Prediction Pro- gramme. Monogenic forms of diabetes, in particular Maturity-Onset Diabe- tes of the Young (MODY), are estimated to be responsible for 1-2% of all patients with diabetes. However, the MODY occurence could be largely un- derestimated because it bears several clinical features common to other types of diabetes. We aimed to search for families with MODY among participants of Prediction Programme for T1D having positive family history of diabetes. Materials and methods: First-degree relatives of patients with T1D are in- vestigated for HLA-DQ genotype and annualy screened for pancreatic au- toantibodies (IA-2Ab, GADAb, IAA). Among 557 families included in the Prediction Programme for T1D, a positive family history of diabetes (at least 2 affected family members) was reported in 53 families (9.5%). In these fami- lies, one proband with diabetes was chosen for mutation screening of most prevalent MODY genes (GCK, HNF1A, HNF4A and INS) by direct Sanger sequencing. Results: Of the 53 families with positive family history of diabetes in the Prediction Programme for T1D, 24 (45%) were genetically diagnosed with MODY. This makes up 4% of all the families from the Prediction Programme Diabetologia (2014) 57:[Suppl1]S1–S564 S 158 1 C for T1D. The most frequent mutations in our study were detected in the GCK (58%), followed by HNF1A (38%) and INS (4%) genes. The family-specific mutations were subsequently observed in 25/27 (93%) participants (first-de- gree relatives of patients with T1D) studied in the Prediction Programme for T1D: they were asymptomatic with exception of two children with HNF1A- MODY who presented osmotic symptoms (polyuria, polydipsia) during the follow up requiring insulin treatment which could be switched to administra- tion of suplhonylurea derivates. Comparing MODY to non-MODY families with positive family history of diabetes, mothers of subjects in the Prediction Programme had more often diabetes than fathers and siblings (p S 159 1 C sodes). The aim of our study was to search for m.3243A>G mutation among patients sent for MIDD or MODY testing, who lack mutations in the most common MODY genes (i.e. GCK, HNF1A or HNF4A). Materials and methods: Unrelated probands from 257 families fulfilling at least one of the following criteria, i.e. matrilinear inheritance, diabetes plus hearing impairment, diabetes development after 20th year of life, or progres- sive hearing loss, were tested for m.3243A>G mutation by the RFLP and/or Real-Time PCR. The heteroplasmy was evaluated for peripheral blood and/ or buccal mucosa. DNA testing was also extended to the family members of probands carrying the mutation. Results: The m.3243A>G mutation was found in 18 patients from 8 fami- lies (3%). Probands´ phenotypes varied from diabetes as the sole symptom to a complex picture of the MELAS syndrome (in one proband). Diabetes or impaired glucose tolerance developed all of the probands (diabetes onset ranged from 21 to 52 years), but only 3 of 10 relatives with the mutation. Five probands (62.5%), and 4 (40%) of the relatives with the mutation had hearing impairement. The heteroplasmy was higher in buccal swab samples compared to the peripheral blood (27.9±18.6% versus 12.7±17.3%). In one case, the heteroplasmy was detected in the buccal DNA only, while the blood DNA samples were repeatedly negative. Conclusion: Among 257 probands with the clinical suspicion on MIDD or MODY, 8 (3%) had the m.3243A>G mutation of the mitochondrial DNA. At the time of testing, only 5 (62.5%) of the probands were diagnosed with typi- cal combination of symptoms, i.e. diabetes and hearing loss. Therefore, DNA testing for MIDD seems to be reasonable also in diabetes patients without hearing impairement, particularly using DNA from the buccal mucosa. Supported by: APVV 0187-12, “Transendogen” (ITMS 26240220051) 370 Using targeted sequencing to investigate the prevalence of monogenic diabetes in the Norwegian childhood diabetes registry B.B. Johansson1,2, H.U. Irgens1,3, J. Molnes1,2, O. Søvik1,3, S. Levy4, D.E. Undlien5, T. Skrivarhaug6, G. Joner6,7, A. Molven1,8, S. Johansson1,2, P.R. Njølstad1,3; 1Department of Clinical Science, University of Bergen, 2Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, 3Department of Paediatrics, Haukeland University Hospital, Norway, 4Hudson Alpha Institute for Biotechnology, Huntsville, USA, 5University of Oslo, 6Department of Paediatrics, Oslo University Hospital, 7Institute of Health and Society, University of Oslo, 8Gade Institute, University of Bergen, Norway. Background and aims: Type 1 diabetes is a heterogenous group of diabe- tes. Since about 10-15 % lack detectable levels of diabetes associated auto- antibodies at diagnosis, we hypothesized that many of these may have a mo- nogenic form of diabetes. We therefore set out to systematically screen all auto-antibody negative patients in the Norwegian Childhood Diabetes Reg- istry for 13 genes known to cause monogenic diabetes. The aim of this study was to estimate the prevalence of common forms of monogenic diabetes in childhood diabetes and assess the usefulness of large scale gene sequencing in routine clinical care of young onset diabetes. Materials and methods: The Norwegian Childhood Diabetes Registry includes 95% of all children with newly diagnosed diabetes from 2002. By 22.01.2014, the registry included 3578 children with clinical data regarding onset of dia- betes, family history and treatment, as well as serum and DNA samples. In this study, we selected 478 cases that were negative for GAD and IA-2 autoan- tibodies, as well as and 478 GAD and IA-2 antibody-positive individuals as controls. The samples were screened by TruSeq Custom Amplicon targeted sequencing method, focusing on 13 monogenic diabetes genes (HNF1A, HN- F4A, GCK, NEUROD1, ABCC8, KCNJ11, HNF1B, INS, PDX1, KLF11, PAX4, BLK, CEL). The amplicons targeted the protein coding regions with flanking intronic sequences, UTRs and some selected regions; HNF4A (c.-192), INS intron (c.188-31) and GCK (c.-71). To test the sensitivity of the assay, we test- ed 22 controls with known mutations in the selected target genes. We identi- fied all these mutations. Sequencing was performed on a Miseq. All newly identified mutations were confirmed by Sanger sequencing. We used stand- ard bioinformatics tools to assess the pathogenicity of the variants identified. Results: As a first test, we screened 74 antibody negative cases that were C- peptide positive (C-peptide>200 pmol/l at diagnosis) to increase the likeli- hood of finding possible MODY. In the screened material we found 39 ex- onic and splice variants whereof twelve of the patients (16%) had a likely pathogenic mutation in HNF1A or GCK. There were no false positive vari- ants reported by the assay. Mean exonic coverage was estimated to 505X and 94% of target exons were covered at 20X. We are currently in the process of analyzing the remaining samples. Conclusion: In this preliminary study where we use targeted sequencing to investigate the prevalence of monogenic diabetes in the Norwegian Child- hood Diabetes Registry we observe that 16% of the GAD- and IA-2-antibody negative patients harbor a likely pathogenic mutation. This panel-sequencing assay provides a specific and sensitive method for simultaneous analysis of 13 genes associated with monogenic diabetes. Supported by: Norwegian Diabetes Association 371 Clinical features of MODY2 in children and adolescents in Russian Federation L. Zilberman1, E. Sechko1, V. Peterkova1, T. Kuraeva1, O. Ivanova2; 1Children, 2Laboratory Immunology, FGBU Endocrinology Research Centre, Moscow, Russian Federation. Background and aims: Maturity-onset diabetes of the young, type 2 (MODY2) is is a monogenic disease caused by mutations in the glucokinase gene (GCK). Aim: To determine the prevalence of GCK gene mutations and to delineate the clinical phenotype of identified GCK mutation carriers in monogenic diabetes patients in Russia. Materials and methods: We included 96 patients with suspected MODY. All patients were screened for HNF1A and GCK mutations by direct sequencing. Results: Mutations in the GCK were found in 46 patients (48 %), and HNF1A mutation - in 18 patients (19%), 32 patients (33%) were negative for MODY2 and MODY3. The age at diagnosis was 10.4 yrs (7;14). 11% of patients were investigated because of clinical manifestations, 66% were diagnosed occa- sionally, 23% were screened because of family history. Duration of the disease was 1.3 (0.6;2.3). BMI was in normal range in 95.6% of the patients (17.0 (15.8;18.5)).BMI was increased to 28.8 (SDS BMI +3.2) in only one 8-year- old girl.HbA1c at the time of diagnosis was 6.6% (6.4;6.8).autoantibodies (ICA, GAD, IAA) were negative in all patients. GTT: Glu (mmol/L)0`-6,6 (5,9; 7), 1-h -11,15 (9,2; 12,78), 2-h -9,15 (8,475; 10,35).Seven patients (15%) aged 11.8-13.2 yrs with normal BMI were found to be insulin-resistant (Insu- lin at baseline was 14.2 U/l(11.0;15.2), HOMA>3.5). Three patients requested insulin therapy (6.5%), and three other patients received sulfonylureas. 87% (40 patients) were able to maintain normal blood glucose levels with a diet and physical exercise only. Conclusion: MODY3 is seemed to be less prevalent than MODY2 in Russian population. Most of the patients have asymptomatic disease and have been revealed occasionally in our study by impaired fasting glucose and slightly elevated HbA1c. Treatment for the majority of MODY2 children can be re- stricted by diet and physical exercise. 372 The length of the deletion in the region 17q contributes to the individual variability of the phenotype of patients with renal cysts and diabetes syndrome (RCAD, HNF1B-MODY) S. Pruhova1, P. Dusatkova1, M. Malina1, Z. Slamova2, K. Blahova1, R. Kotalova1, J. Dusek1, J. Lebl1, T. Seeman1, O. Cinek1; 1Department of Pediatrics, 2Department of Biology and Medical Genetics, 2nd Faculty of Medicine, Charles University in Prague and University Hospital Motol, Czech Republic. Background and aims: The renal cysts and diabetes (RCAD) syndrome caused by defects in a gene for hepatocyte nuclear factor 1 beta (HNF1B) is characterized by broad spectrum of clinical features. While heterozygous point mutations detectable by Sanger sequencing are relatively rare, we fo- cused on gross deletions of the HNF1B gene that are determined by MLPA (Multiplex Ligation Probe-dependent Amplification). Rather importantly, the deletions most often extend beyond the single HNF1B gene, thus more deleted genes may participate in the clinical picture. Objective and hypoth- eses. We compared the clinical phenotype of patients carrying the gross dele- tions whose extent was precisely determined by array comparative genomic hybridisation (aCGH), and point mutations in the HNF1B gene. Materials and methods: In thirteen patients (6 males, median age 15.5 years) carrying the HNF1B gene deletion was performed aCGH on CytoChip Oligo 8x60K. The clinical data were compared with 5 patients (1 male, median age 15.5 years) having point mutations in HNF1B. Diabetologia (2014) 57:[Suppl1]S1–S564 S 160 1 C Results: The average length of heterozygous deletion was 1.69 Mb. The long- est deletion reached 2.5 Mb affecting 47 genes and the shortest deletion found in three patients was 1.4 Mb long and deleted 16 genes. All patients lost also LHX1 gene encoding transcription factor important for the development of the renal and urogenital system. Compared to other deletion carriers, pa- tients having longest deletions (2.5 and 2.1 Mb) manifested renal dysfunction at older age (10 and 30 years) with milder changes of the kidney structure (isolated cysts and functional changes only) and both presented diabetes as a first clinical feature of RCAD. Patients with shorter deletion manifested re- nal changes (polycystosis) prenatally and are mostly without diabetes so far. Comparing deletion and point mutation carriers, prenatal ultrasound kidney changes were found in 10/13 and 4/5 patients, respectively. Diabetes mani- fested at the median age of 17 years in 5/13 and 2/5 patients. Hypomagnesae- mia was present in 11/13 and 2/5 patients. Conclusion: Regardless to the length of the deletion in the region 17q, the dominating clinical phenotype of the patients is similar compared to the pa- tients having point mutation in the HNF1B gene. However, the length of the deletion can contribute to the individual variability in the age of manifesta- tion and other variability of the phenotype. Supported by: Czech Ministry of Health NT 11402 and NT 1145 PS 013 Islet gene expression in type 2 diabetes 373 Beta cell function, more than beta cell area, affects plasma glucose levels and influences the progression of type 2 diabetes in humans M. Suleiman, F. Olimpico, M. Bugliani, F. Syed, M. Occhipinti, F. Scatena, F. Filipponi, D. Focosi, U. Boggi, P. Marchetti, L. Marselli; Department of Clinical and Experimental Medicine, Islet Cell Laboratory, University of Pisa, Italy. Background and aims: Type 2 diabetes (T2D) is characterized by reduced beta cell functional mass. However, the direct relationships between beta cell amount or function with in vivo blood glucose levels and disease progression are unclear. Materials and methods: Pancreases were obtained from 388 non-diabetic (ND) donors [age: 60±17 years, BMI: 25.2±3.8 kg/m2, mean glycemia (MG) in intensive care unit: 154±45 mg/dl] and 68 T2D subjects [(age: 71±8 years, BMI: 26.9±4.0 kg/m2, MG: 218±72 mg/dl, diabetes duration (DD): 9±7 years)]. Pancreatic insulin area (PIA) was assessed by immunocytochemis- try; islets (HI) were isolated and ex-vivo insulin release (IR) was evaluated in response to 3.3 and 16.7 mM glucose (G). Results: PIA was lower in T2D than ND subjects (0.51±0.20 vs 0.84±0.42%, p S 161 1 C Results: 461 out of 10,294 mSigDB gene sets were significantly and consist- ently altered across datasets (P value of meta-ES S 162 1 C 377 Dephosphorylation and nuclear translocation of the CREB-regulated transcription co-activator 1 (CRTC1) elevates miR-212 and miR-132 expression in insulin secreting cells H.A. Malm1, C. Berggreen2, M. Orho-Melander1, L. Eliasson1, O. Göransson2, I.G. Mollet1; 1Dept Clinical Sciences in Malmö, Lund University Diabetes Centre, Malmö, 2Dept of Experimental Medical Sciences, Lund University, Lund University Diabetes Center, Sweden. Background and aims: We have previously demonstrated miR-212 and miR- 132 to be upregulated in diabetic GK rat islets. Others have demonstrated cAMP-responsive element binding protein (CREB) to bind CRE-sites on the miR-212/132 promotor. Here we aim to investigate mechanisms underlying cAMP-mediated transcriptional regulation of the miR-212/132 cluster in in- sulin secreting cells. Materials and methods: INS-1 832/13 cells were cultured for 0.5, 2, 6 and 24h in various glucose concentrations and in the absence or presence of for- skolin/IBMX or GLP-1 (2h). In addition, cells were transfected with siRNA against all three isoforms of CRTC or salt inducible kinases (SIKs) or with mature miR-212 and miR-132. Relative expression of miRNA and mRNA was determined by qPCR; Insulin secretion at 2.8 and 16.7 mM glucose was analyzed using RIA; protein expression and phosphorylation was evaluated with western blot analysis (WB) and CRTC1 translocation with confocal mi- croscopy. Results: Expression of both miRNAs was increased after 2h in forskolin/ IBMX (3-4-fold, p S 163 1 C 380 Intracellular cholesterol transporters, islet dysfunction and insulin secretion in type 2 diabetes mellitus J.B. Pinto; Biological and Biomedical Sciences, Glasgow Caledonian University, UK. Background and aims: Elevated intracellular cholesterol levels prevent glu- cose stimulated insulin secretion from pancreatic beta cells, by impairing iCa2+ elevation and its signaling, and hindering PiP2 hydrolysis, affecting ac- tin dynamics and plasma membrane potential. Conversely, cholesterol deple- tion restores exocytosis of insulin granules in sterol-loaded cells. Mechanisms that improve the efficiency of intracellular cholesterol transport may there- fore prove valuable therapeutically. Cholesterol-binding members of the ster- oidogenic acute regulatory protein domain (StarD) family of lipid trafficking proteins (STARTs) are known to modulate lipid homeostasis in a number of differing cell types, including macrophages, keratinocytes and hepatocytes; however, despite their potential therapeutic utility, little is known about the expression or function of StarD proteins, or their ligands, in beta-cells, or their relationship with insulin secretion. Our aims were to (i) investigate the expression of StarD cholesterol trafficking proteins, in BRIN-BD11 insu- linoma cells, following manipulation of cellular cholesterol levels, and (ii) to explore the role of endosomal cholesterol transfer protein, StarD3, in choles- terol homeostasis and insulin secretion. Materials and methods: Methodologies included radioisotopic labelling of cellular lipid pools, colorimetric assays for lipid mass, Q-PCR, immunoblot- ting and ELISA. Results: Treatment with methyl beta-cyclodextrin (MCD; 1-10mmol l-1; 1h) efficiently removed cholesterol from BRIN-BD11 cells, dramatically increas- ing endogenous cholesterol biosynthesis, while cholesterol lipid complex (CLC 1:200) induced significant increases in cholesterol mass (p S 164 1 C PS 014 Beta cell proliferation and differentiation 382 The pocket protein family involved in beta cell replication during pregnancy Y. Xu1, L. Gao1, H. Zhang1, M. Woo2, X. Wu1; 1First Affiliated Hospital with Nanjing Medical University, China, 2Institute of Medical Science, University of Toronto, Canada. Background and aims: The pocket protein family consists of retinoblastoma protein (Rb), p107 and p130, which can repress E2F-mediated gene tran- scription, recruit chromatin remodeling enzymes, and undergo phosphoryla- tion by cyclin-dependent kinases, comprising the central regulatory check- point that controls the mammalian G1/S-phase cell cycle transition. In spite of the sequence similarity within the DNA binding domain, they play distinct roles in the regulation of cell cycle progression. Islet beta cell replication is the main source for beta cell mass expansion during pregnancy. Activation of prolactin receptor signaling plays a key role in this process. Whether the pocket protein family is critical for beta cell replication during pregnancy is unclear. The present study is to investigate the change of Rb, p107 and p130 during pregnancy and its related mechanism in prolactin-mediated beta-cell replication. Materials and methods: Pancreatic islets of C57BL/6 mice were isolated and purified by collagenase digestion at non-pregnant (NP), pregnant day 10.5 (P10.5), P14.5, P18.5 and after parturition day 4 (AP4), AP8. The gene and protein expression of Rb, p107, p130 and E2F1~5 were detected by real- time quantitative PCR and Western blotting methods. Rb and phosphor-Rb expression of islets were detected by immunofluorescence. INS-1 cells were transfected with siRNA mediated by liposome to knockdown Rb expression, and detected the change of cell cycle with prolactin stimulation by flow cy- tometry. INS-1 cells were cultured with inhibitors of AKT, Stat5, ERK, JNK and Pim respectively with prolactin stimulation to observe the change of Rb/ E2F1 expression. Results: During pregnancy, Rb, p107 and E2F1~3 gradually increased and re- turned back to NP level after parturition, with peak of Rb, phosphor-Rb and E2F1 at P14.5, p107 and E2F2 at AP4, E2F3 at P18.5 (p S 165 1 C 10ng/ml Wnt3a: 114.8±1.03% (p S 166 1 C 387 Beta cell mass expansion is impaired in aged rats exposed to 90% pancreatectomy and gastrin treatment N. Tellez1, Y. Martí1, M. Vilaseca1, A. Pla1, E. Montanya1,2; 1Diabetes, Nutrition and Endocrine diseases, CIBERDEM-IDIBELL- University of Barcelona, 2Hospital Universitari de Bellvitge, Barcelona, Spain. Background and aims: Aging is a non-modifiable risk factor for the develop- ment of diabetes. The contribution of impaired beta cell regeneration to this increased risk remains unknown. In young rats, 90%-pancreatectomy (Px) results in β-cell mass regeneration which is further enhanced with gastrin treatment. The aim of our study was to investigate the β-cell regeneration potential of aged rats using the 90%-Px model and gastrin treatment. Materials and methods: 1 and 12 month-old Wistar rats underwent 90%-Px and were treated from the day of surgery with [15leu] gastrin-17 (150μg/kg • 12h, Px+G, n=21) or with vehicle (Px+V; n=21). A group of sham-operated rats treated with vehicle was included for each age group (S+V; n=18). Pan- creatic remnants were harvested on days 3 and 14 after surgery for morpho- metric, immunohistochemical and gene expression analysis. Results: Young Px rats showed increased β-cell mass that was further in- creased with gastrin treatment. Gene expression and nuclear immunolo- calization of nkx6.1 in ductal cells, and the percentage of extra-islet β-cells (indirect markers of β-cell neogenesis) were also increased in gastrin- treated young Px rats. β-cell apoptosis was similar among groups, and β-cell replica- tion and size were similarly increased in gastrin and vehicle-treated young Px rats. In aged rats, β-cell mass was not increased in any of the Px groups, de- spite the increased β-cell replication and individual β-cell size. Gene expres- sion and nuclear immunolocalization of nkx6.1 in ductal cells, and the per- centage of extra-islet β-cells were not increased in aged gastrin- or vehicle- treated Px rats. The dedifferentiation-related transcription factors Neurog3 and Sox9 were significantly upregulated in islet β-cells from aged Px rats. Conclusion: The potential for compensatory β-cell hyperplasia and hyper- trophy is retained in aged rats. In contrast, impaired β-cell neogenesis along with beta cell dedifferentiation may contribute to the limited beta cell regen- eration in aged rats. Supported by: ACD, SED, CIBERDEM which is a project of ISCIII 388 Cell surface marker expression and colony formation in islet-depleted pancreatic digests P. Augstein1,2, E. Bandala-Sanchez2, G. Naselli2, T. Loudovaris3, L. Lee2, A. Neale2, P. Campbell3, W.J. Hawthorne4, P. Heinke1, I. Banakh2, L.C. Harrison2; 1Institute of Diabetes, Karlsburg, Germany, 2The Walter and Eliza Hall Institute for Medical Research, Parkville, 3St. Vincent’s Institute of Medical Research, Fitzroy, 4Westmead Hospital, University of Sydney, Australia. Background and aims: Stem cells (SCs) are characterized by self-renewal ca- pacity and pluripotency. Hematopoietic SCs (HSC) are identified by a com- bination of markers and functional assays. In the adult human pancreas, the existence of SCs is debated. To identify putative stem cells, we screened islet- depleted pancreas digests, a by-product of the islet isolation procedure, for expression of surface markers and epithelial colony-forming potential (CFP). Materials and methods: Islet-depleted pancreas digests from 18 cadaveric organ donors were stained with antibodies to hematopoietic, mesenchymal (MSC) and epithelial stem cell (EpSC) markers. Cells in the ‘side population’ (SP) were identified by Hoechst 33342 dye efflux in a BD-LSR-II-W flow cy- tometer. EpCam, CD133, Hpd1 and Hoechst 33342 alone or in combination were used to separate populations in a FACSAria-C sorter. CFP was evaluated in a 2D assay. Data are presented as mean ± SEM. Results: Very few cells expressed HSC markers (CD45 2.7 ± 0.8 %, CD117 1.2 ± 0.6 %, CD31 10.8 ± 0.8 %, CD56 2.8 ± 1.0 %). The MSC marker CD29 was detected on a majority of cells (97.1 ± 1.0 %), whereas the other MSC mark- ers CD44 (46.4 ± 7.8 %), CD34 (34.1 ± 10.8 %), CD90 (17.2 ± 4.5 %), CD105 (36.1 ± 11.1) and CD73 (27.6 ± 5.3 %) were expressed on fewer cells. Certain EpSC markers, EpCam (75.3 ± 3.9 %) and CD49f (60.9 ± 7.1 %), were ex- pressed by a majority of cells whereas others, CD133 (44.6 ± 7.5 %) and CD26 (20.5 ± 3.6 %), were expressed by fewer cells. The CFP of pancreas digest cells was however significantly enriched by selection for CD133highSP cells, such that CD133highSP > SP > CD133high > CD133low. The majority of CD133high (61.4 ± 6.3 %) and SP (58.32 ± 6.97 %) cells expressed the ductal marker CA19-9. In the human pancreas, ductal cells express CD133 and comprise a 2-fold increased proportion (3.1 vs. 7.5 %) of the SP. Conclusion: We identified two markers for the isolation of colony-forming cells. Our findings suggest that ductal cells that efflux Hoechst 33342 and express CD133high are a reservoir of stem cells in the adult human pancreas. Supported by: JDRF, EFSD, NHMRC 389 Highly-efficient MSCs fusion with beta cells results in beta cell-like phenotype Z. Azizi1,2, F. Paroni1, C. Lange2, A.R. Zander2, C. Westenfelder3, K. Maedler1; 1Center for Biomoleculare interactions, Unversity of Bremen, 2Center of Bone Marrow Transplantation, Universitätsklinikum Hamburg-Eppendorf, Germany, 3Departments of Medicine and Physiology, University of Utah and George E Wahlen Veterans Affairs Medical Centers, Salt Lake City, USA. Background and aims: Mesenchymal stromal cells (MSC) have anti-inflam- matory, anti-apoptotic and immunosuppressive effects and they are a potent source for cell therapy. Cell fusion enables rapid generation of functional new reprogrammed cells. In this study, we aimed to establish a fusion protocol of bone marrow derived human MSCs with the rat β-cell line (INS1-E) as well as isolated human pancreatic islets in order to generate functional insulin producing β-MSCs as a cell-based treatment for diabetes. Materials and methods: Human MSCs were isolated from bone marrow and characterized by colony forming units in culture and flow cytometry for expression of CD73, CD105 and CD90 as well as lack of CD45, CD34 and MHCII. Multipotency was tested by differentiation into the adipogenic and osteogenic lineages. To induce cell fusion, eGFP transfected MSCs with puromycin as selection marker were cultured with either mCherry-INS1E or human dispersed isolated islets and treated with phytohemagglutinin and Polyethylene glycol (PEG). Results: Of the whole cell population, 44 ± 10% polykaryons were identi- fied based on polyploidy flow cytometry at 2 days after fusion. 29± 6% of all MSCs were identified as β-MSC heterokaryons based on double positivity for mCherry and eGFP. Six days after fusion and four days after puromycin selection, the β-cell transcription factors Nkx6.1, MafA and PDX1 and im- portant genes for regulating β-cell function insulin, GLUT2 and GCK and insulin content were increased in the β-MSCs from fused human dispersed islets/MSCs as well as INS-1E/MSCs compared to non-fused controls after puromycin selection. Insulin positive β-MSCs also expressed nuclear PDX1. Conclusion: Our results show an efficient established protocol for fusion of human MSCs and β-cells, which resulted in a β-cell like phenotype; this could be a novel tool for cell-based therapies of diabetes. 390 A novel and reliable protocol for human embryonic stem cell differentiation into the definitive endoderm based on dispersed single cells U. Diekmann, S. Lenzen, O. Naujok; Institute of Clinical Biochemistry, Hannover Medical School, Germany. Background and aims: The differentiation of embryonic stem (ES) cells into organs derived from the endoderm germ layer, such as liver or pancreas re- quire as initial step the formation of the definitive endoderm (DE). Human ES cell differentiation is usually initiated from colonies, which is the routinely used cell culture procedure. However, differences in colony size, colony num- ber and the potentially altered sensitivity of these cells to media supplements render colonies into a suboptimal starting material. The aim of this study was the development of a reliable and highly efficient protocol for the differentia- tion of human ES cells into the DE lineage from dispersed single cells and their further differentiation into the pancreatic lineage. Materials and methods: Three different human ES cell lines were passaged as dispersed single cells and defined cell numbers were subjected to differen- tiation. Different protocols for the DE induction were tested by combining various concentrations of Wnt3a, CHIR-99021 and Activin A for four days. The expression of the marker genes T, GSC, MIXL1, SOX17 and FOXA2 was analyzed by qPCR, flow cytometry, and immunofluorescent staining (IF). Furthermore, differentiation of these DE cells into the pancreatic lineage with FGF10, retinoic acid, dorsomorphine plus SB-431542 was performed. PDX1- and NGN3-positive cells were quantified by IF. Diabetologia (2014) 57:[Suppl1]S1–S564 S 167 1 C Results: Differentiation with CHIR-99021 and Activin A for the first 24 h and a subsequent treatment with Activin A alone resulted in the highest num- bers of DE committed cells from all three tested cell lines. The expression of the marker genes SOX17 and FOXA2 were significantly increased under these conditions compared to random differentiated cells and to a classical DE differentiation protocol used as positive control. The quantification by IF and flow cytometry of DE committed cells revealed efficiencies ranging from ~70% (Hues4 and HES3) up to >80% (Hues8), a more than 2-fold increase compared to the reference protocol (~33-40%). The ES cells proliferated un- der this condition resulting in a more than 5-fold increased cell number after four days of differentiation. In addition, the dispersed single cells differenti- ated into DE were able to further differentiate into PDX1-positive progeni- tors (~40% after 10 days) and subsequently into NGN3-positive endocrine precursor cells (~10% after 14 days). High expression levels for the marker genes FOXA2, HNF6, MNX1 (HB9), NKX2.2, NKX6.1 and NGN3 in a man- ner similar to the in vivo pancreatic development were detected proofing the maturation of the pancreatic precursor cells. Conclusion: This novel protocol was able to differentiate three human ES cell lines in a highly efficient manner into the DE. Therefore only low initial cell numbers and reduced concentrations of expensive growth factors were required without growth limitations. In addition, these cells were able to dif- ferentiate further into the pancreatic lineage and endocrine progenitor cells. Supported by: REBIRTH Cluster of Excellence (DFG)) 391 Isolation and characterisation of embryo stem cell-like cells in adult human pancreas S. Lee1,2, S. Jeong2, H. Park2, S.-C. Kim1; 1Surgery, 2Asan Institute for Life Science, Asan Medical Center, Seoul, Republic of Korea. Background and aims: Within recent years, stem/progenitor cells research has become a very important part of regeneration or differentiation into in- sulin-producing cells. However, the stem/progenitor cells are extremely rare attendance and expressing location and moment of these stem cells are not clearly demonstrated in adult human pancreatic tissue. Therefore we have identified undifferentiated embryo stem cell-like cells expression in adult hu- man pancreas and isolated for characterization. Materials and methods: Enriched human exocrine cells are obtained after COBE purification of islet isolation. For islet-depleted pancreatic exocrine cells culture, endocrine cells were sorted out with microbead conjugated PSA- NCAM antibody using magnetic-activated cell sorting (MACS) and purified CA19-9 positive pancreatic ductal cells or non-purified cells were cultured for 6 days. We observed morphology changes and RNA expression pattern of embryo stem cell markers. To identification of stem cells present location, we SSEA-4 positive cell selected from enriched exocrine cell fraction. Results: Non-purified crud duct cells attached easily and epithelial-like cells extended grow up quickly from primary attached cells but purified ductal cells were showed insufficient growth. Expression of classic stem cell mark- ers: Oct4, c-Myc, Klf4, Nanog, Sox2 and SSEA-4 mRNAs was found in crude duct and purified duct cell fraction. However, these stem cell markers was not detected or weakly expressed in PSA-NCAM negative and CA19-9 negative cell fraction. Classic stem cell mRNA markers were only expressed in SSEA-4 positive cells and SSEA-4 positive cells were detected in pancreatic duct cell by immunocytochemistry. Conclusion: We characterized and isolated of SSEA-4 positive embryo stem cell-like cells in adult human pancreatic duct and hypothesize that these cells differentiate to insulin- producing cells in adult human pancreas. 392 Expression of insulin genes in foetal liver S. Takasawa1, S. Murakami-Kawaguchi2, K. Nata3, A. Itaya-Hironaka1, S. Sakuramoto-Tsuchida1, A. Yamauchi1, H. Ota1, M. Takeda1, C. Tsuchida1, H. Tsujinaka1, T. Fujimura1, H. Okamoto4; 1Biochemistry, Nara Medical University, Kashihara, 2Internal Medicine, Murakami Shinmachi Hospital, Aomori, 3Medical Biochemistry, Iwate Medical University School of Pharmacy, Yahaba, 4Tohoku University, Sendai, Japan. Background and aims: Pancreatic beta cells are the major source of insulin in adult mammals. The cells have a limited capacity for regeneration and the destruction causes type 1 diabetes. Recently, a possible permanent cure for diabetes has been explored by using several different approaches, one of them being the transdifferentiation of nonpancreatic cells to insulin-producing cells. Hepatocytes seem to be an ideal target for the production of insulin, because both the liver and ventral pancreas appear to arise from the same cell population located within the embryonic endoderm during embryogenesis. Despite the similarity in development of these two cell types, the expression of the insulin gene in fetal liver has remained elusive because of its pancreatic beta cell-specific expression in adult animals. In the present study, we dem- onstrate insulin mRNA and immunoreactive cells for insulin in fetal liver and the promoter activity for insulin in fetal hepatocytes. Materials and methods: Timed pregnant ICR mice were obtained from Ja- pan SLC. On embryonic days 13.5 (E13.5), E16.5 and E18.5, livers were dis- sected. The livers of newborn mice were dissected at Day0. Livers from adults were dissected from 14-week-old female mice. Gene expression was analyzed by RT-PCR. Immunohistochemical analysis was carried out using antibod- ies against insulin, proinsulin, processed insulin, glucagon, somatostatin, and pancreatic polypeptide. Promoter activity of mouse Ins1 (-703~+14) and Ins2 (-830~+14) was measured using primary cultured fetal hepatocytes by lucif- erase assay. Electrophoretic mobility shift assay was carried out using whole- cell extracts of fetal liver. Human fetal (22-40 weeks) and adult liver RNAs, purchased from Clontech, were also analyzed by RT-PCR. Results: The expression of insulin and transcription factors for insulin is in- vestigated in mouse fetal liver. We detected mRNAs for Ins1 and Ins2 and proinsulin- and mature insulin-positive cells in mouse fetal liver by RT-PCR and immunohistochemistry. Glucagon, somatostatin and pancreatic poly- peptide were not expressed throughout development. Mouse Ins2 and Ins1 promoters were transiently activated in mouse fetal hepatocytes of E13.5 and E16.5, respectively. Pancreatic and duodenal homeobox 1 (Pdx1) mRNA was not expressed during development of the liver. In contrast, mRNAs and pro- teins of neurogenic differentiation (NeuroD)/beta cell E-box transactivator 2 (Beta2) and v-maf musculoaponeurotic fibrosarcoma oncogene homolog (MafA) were almost simultaneously expressed with insulin genes in the liver. Ins2 and Ins1 promoters were activated in hepatoma cells by the transfection of the expression vector for NeuroD/Beta2 alone and by the combination of NeuroD/Beta2 and MafA, respectively. In addition, we also analyzed human fetal and adult liver RNAs and found that mRNAs for insulin, NeuroD/Beta2, and MafA were expressed in human fetal liver. Conclusion: These results indicate that the expression of NeuroD/Beta2 and MafA is linked temporally with the transcription of insulin gene in fetal liver and suggest the potential usage of fetal hepatocytes to make insulin-produc- ing cells by introducing/inducing transcription factors. Diabetologia (2014) 57:[Suppl1]S1–S564 S 168 1 C PS 015 Metabolic coupling in insulin secretion 393 MICU2: a modulator of stimulus-secretion-coupling in INS-1 832 cells A. Bagge1, L. Andersson1, P. Spégel1, B. Valtat1, V.K. Mootha2, H. Mulder1; 1Unit of Molecular Metabolism, Lund University Diabetes Center, Malmö, Sweden, 2Broad Institute, Cambridge, USA. Background and aims: In β-cells, glucose-stimulated insulin secretion (GSIS) is tightly coupled to the metabolism of glucose, which ultimately trig- gers a rise in cytosolic calcium (Ca2+) levels activating the exocytotic machin- ery. Ca2+ is also involved in activation of the more sustained second-phase of insulin secretion. However, the mechanism is less clear. The regulated mitochondrial Ca2+m-influx during metabolic activation is critical for GSIS but the function of Ca2+m is still not fully understood. Mcu and Micu1 are subunits of the Mitochondrial Calcium Uniporter complex (MCU-complex) and have been shown to be important for ATP synthesis and GSIS. Micu2 was recently identified in mouse liver as a part of the MCU-complex functioning as a gate-keeper for Ca2+m-uptake, similar to Micu1. Our preliminary data show that an expression quantitative trait locus (eQTL) associated with the MICU2 gene is correlated with β-cell function (HOMA-β) and fasting glu- cose in humans. The aim of this study was to investigate the role of Micu2 in GSIS and mitochondrial function in the clonal beta-cell line INS-1 832/13. Materials and methods: Micu2 was silenced for 72 h by siRNA in INS-1 832/13 cells, and insulin secretion was determined by radioimmunoassay, ATP content by luciferase-dependent luminescence, and lactate release as wells as NADH/NAD+-ratio by colorimetric assays. Glucose utilization was measured as 3H2O after stimulation with 5- 3H-labeled glucose, seahorse XF24 was used to measure respiration and metabolite profiling was done by GC/MS. Results: KD of Micu2 in INS-1 832/13 cells resulted in a 33% (N=3, p S 169 1 C low-affinity cAMP-binding domains. The latter domain is not required for translocation, but it extends the range of effective cAMP concentrations and increases the magnitude of the translocation response. Glucose-stimulated Epac2 translocation in β-cells involves additional signals than cAMP eleva- tion alone. 396 Does increased mitochondrial pool improves beta cell function? Trying to create super beta cells with PGC-1a mediated mitochondrial biogenesis A. Besseiche1, G. Guillemain1, E. Ramond1, J.-P. Riveline2, P. Ferré1, J.-F. Gautier2, B. Bréant3, B. Blondeau4; 1Inserm-CRC, 2Hopital Lariboisière, 3Inserm, 4UPMC-CRC, Paris, France. Background and aims: Understanding diabetes and try to cure it requires defining the regulators of pancreatic beta-cell function. Because beta cells rely on their cellular metabolism, mitochondrial activity and energy produc- tion to secrete insulin properly, we tested the idea that augmenting mito- chondrias in beta cells might improve their insulin secretion capacity, thus generating super beta cells that would be adapted to situations of increased insulin demand. To this aim, we overexpressed in beta cells the transcrip- tional coregulator PGC-1α, a potent inducer of mitochondrial biogenesis, and studied both in vivo and in vitro the consequences on insulin secretion. Materials and methods: To carry the study, we used two models of PGC-1α overexpression in beta cells: first Min6 cells overexpressing PGC-1α with an adenovirus and second, transgenic mice with a specific PGC-1α overexpres- sion in mice using the Tetracycline inducible system. We measured insulin secretion in response to glucose, glucose tolerance in mice, mitochondrial function (cellular respiration, ATP turnover and reactive oxygen species - ROS - production) and assessed the beta-cell energy status by measuring the phosphorylation of the AMP-Kinase. Results: In Min6 cells, PGC-1α overexpression increases mitochondrial biogenesis and β-cell respiration but paradoxically, does not increase ATP turnover. PGC-1α overexpression increases ROS production and leads to the activation of AMPK, suggesting both an oxidative and energetic stress in these cells. Instead of being improved, insulin secretion in response to glucose was blunted when PGC-1α is overexpressed. Interestingly, rescu- ing insulin secretion was achieved using a powerful antioxidant (Tempol) to prevent ROS production or using compound C-induced, an inhibitor of AMP-kinase activation. Similar results were obtained on isolated islets from PGC-1α overexpressing mice: increased ROS production, activated AMPK and blunted insulin secretion. Conclusion: In beta cells, increasing mitochondria with the use of PGC- 1α overexpression induces both an oxidative stress and an energetic stress. Therefore, strategies to increase insulin function through increased mito- chondrial function may not be efficient. Specific mechanisms underlying these defects still need to be identified. In contrast with other metabolic tis- sues, increased mitochondrial pool using PGC-1α is not beneficial for insulin production by beta cells. Supported by: SFD and INSERM 397 Mitochondrial coupling efficiency is swiftly increased by hyperoxia reciprocally to effects by hypoxia I. Hals1, S. Bruerberg1, V. Grill1,2; 1Inst of Cancer Researsch and Molecular Medicine, NTNU, Faculty of Medicine, 2Department of Endocrinology, St Olav‘s Hospital, Trondheim University Hospital, Trondheim, Norway. Background and aims: Previous exposure to hyperoxia was recently shown to negatively affect beta cell function and viability. These findings are impor- tant in the context of possible damage to beta cells during hyperoxic treat- ment in humans, such as in preterm babies, and therefore prompt further studies. The early time course of effects of hyperoxia has not been clarified and may provide information on mechanisms. Accordingly, here we tested for acute effects of hyperoxia on mitochondrial function Materials and methods: INS-1 derived 832/13 cells were cultured in RPMI medium, usually together with 11 mM glucose. For some experiments INS- 1 cells overexpressing uncoupling protein 2 (UCP-2) were used. Oxygen consumption was measured in intact cells by high-resolution respirometry (OROBOROS) before and after sequential administration of oligomycin (which blocks the production of ATP by ATP synthase), FCCP (which induc- es respiration that maximises oxidative capacity), rotenone and lastly antimy- cin. Elevated oxygen concentration (400 µM) was achieved in one chamber, the other containing normal levels of oxygen (200 µM). Results: The presence of hyperoxia decreased the inhibitory effect by oli- gomycin when the inhibitor was added 15-20 min after the introduction of hyperoxia. The ratio of oligomycin uninhibited respiration in relation to FCCP-induced respiration was increased by 59.4% (from 0.32 ± 0.02 (n = 6) at normoxia to 0.51 ± 0.02 (n = 20) at hyerpoxia, p < 0.0003). Similar effects of hyperoxia were seen in INS-1 cells overexpressing UCP-2. The glucose con- centration ( 5.5 or 11 mM) during respirometry did not modify the uncou- pling effect of hyperoxia. Opposite effects (decreased ratio oligomycin/FCCP respiration) was seen in cells pre-exposed for 8 h to hypoxia (0.3% oxygen) (0.29 ± 0.01 vs. 0.26 ± 0.02, n = 5, p< 0.05). Conclusion: 1) Beta cell mitochondria swiftly adapt to hyperoxia by increas- ing un-coupling 2) This effect is not mediated by UCP2, 3) The effect is re- ciprocal to that exerted by previous hypoxia. Altogether our findings demon- strate swift and relevant adaptability of beta cell mitochondria to differences in the oxygen environment, a feature that should, in the short term at least, promote their survival. Supported by: Liaison Committee between RHA and NTNU, the Norwegian Diabetes Association 398 Inhibition of the malate-aspartate shuttle abolishes glucagon secretion without affecting insulin secretion from mouse pancreatic islets L.E. Andersson1, J. Stamenkovic1, A. Bagge1, V.V. Sharoyko1, C.B. Wollheim2,3, H. Mulder1, P. Spégel1; 1Lund University Diabetes Center, Unit of Molecular Metabolism Department of Clinical SciencesLund University, 2Department of Clinical Sciences, Lund University Diabetes Center, Lund University, Malmö, Sweden, 3Department of Cell Physiology and Metabolism, University Medical Centre, Geneva, Switzerland. Background and aims: Altered secretion of both insulin and glucagon has been implicated in the pathogenesis of Type 2 diabetes (T2D). The main fo- cus of research has been on understanding the impaired insulin secretion from the pancreatic ß-cells, leaving the mechanisms controlling glucagon secretion from the pancreatic α-cells largely unresolved. Materials and methods: The α-cell line, αTC1-6, and ß-cell line, INS-1 832/13 were stimulated with glucose, in presence or absence of phenylsuc- cinate, and pyruvate after which hormone secretion, metabolite levels, glu- cose utilization, respiration and lactate release were assayed. Key experiments were replicated in isolated mouse islets. Results: INS-1 832/13 and αTC1-6 cells respectively secreted insulin and glucagon dose dependently in response to glucose. Glycolytic metabolism was similar in the two cell lines; tricarboxylic acid (TCA)-cycle metabolism, respiration and ATP-production were less glucose-responsive in αTC1-6 cells. Hence, a tight coupling of glycolytic and mitochondrial metabolism was observed only in INS-1 832/13 cells. Inhibition of the malate-aspartate shuttle, using phenyl succinate, impacted glucose-provoked ATP production and glucagon secretion from αTC1-6, but not INS-1 832/13 cells. Blocking the malate-aspartate shuttle increased levels of glycerol-3-phosphate only in the INS-1 832/13 cells. Accordingly, expression of components of the glyc- erolphosphate shuttle relative to expression of the malate-aspartate shuttle was found to be lower in αTC1-6 cells. These results were confirmed in pri- mary mouse islets, where phenyl succinate abrogated secretion of glucagon but not insulin. Conclusion: Our data suggest that the glycerolphosphate shuttle augments the malate-aspartate shuttle in the ß-cell but not so in the α-cell. A suppressed activity of the glycerolphosphate shuttle in the α-cell prevents the α-cell from maintaining a high rate of glycolysis and glucagon secretion at high glucose levels. Importantly, pyruvate- and lactate-provoked glucagon secretion re- mains unaffected since they are independent of mitochondrial shuttle ac- tivity. Consequently, secretion of glucagon becomes more sensitive to fuels available during physical exercise. Supported by: Novo Nordisk Foundation Diabetologia (2014) 57:[Suppl1]S1–S564 S 170 1 C 399 Shared and unique motifs of nutrient sensing in GLP-1 (GLUTag) and insulin (INS-1 832/13) secreting cells P. Spégel1, L. Shcherbina2, L. Andersson1, C. Wollheim3, N. Wierup2, H. Mulder1; 1Unit for Molecular Metabolism, 2Unit of Neuroendocrine Cell Biology, 3Lund University Diabetes Centre, Malmö, Sweden. Background and aims: Glucagon-like peptide-1 (GLP-1), secreted from intestinal L-cells, promotes beta-cell proliferation and enhances glucose- stimulated insulin secretion. Despite this important role for GLP-1, much less is known about stimulus-secretion coupling in the L-cell as opposed to the beta-cell. GLP-1 secretion from L-cells has been suggested to be mainly regulated by nutrient up-take rather than metabolism of such fuels. However, it has also been shown that a, as of yet, unknown metabolic signal may po- tentiate secretion of the hormone. We hypothesized that levels of a potential coupling factor of GLP-1 secretion should rise in nutrient stimulated L-cells in parallel to increased hormone secretion. Materials and methods: GLUTag and INS-1 832/13 cells were stimulated with a range of different secretagogues in presence and absence of various pharmacological inhibitors. Nutrient-provoked hormone secretion was as- sessed, in addition to profiling of alterations in metabolite levels by gas chro- matography/mass spectrometry. Results: Both glucose and glutamine stimulated TCA-cycle metabolism (3- 9-fold, p S 171 1 C PS 016 Modulation of islet function through cell surface receptors 401 Differential action of GLP-1 and GIP on human pancreatic islet function and viability Y. De Marinis, P. Storm, J. Fadista, L. Groop; Dept of Clincal Sciences Malmö, Sweden. Background and aims: Gastric inhibitory peptide (GIP) and glucagon-like peptide 1 (GLP-1) have been studied intensively in various animal models. Physiological studies of these incretins in human pancreatic islets are still lacking. Here, we have assessed the direct effects of GIP and GLP-1 on islets from non-diabetic and diabetic adult human donors. Materials and methods: We investigated how the incretins affect glucose in- duced human islet cell activity by MTS assay. Insulin section from islets of 61 human donors was measured at 1 and 16.7 mM glucose. To varify various gene expression, RNA-sequencing was performed on human islets. Results: When challenged by high glucose (20 mM, 72 hr), the number of metabolically active human islet cells decreased (60% decrease vs. 5.5 mM glucose, p S 172 1 C Conclusion: Active DPP-4 is expressed in islets in a species-specific manner. Local DPP-4 activity is directly increased by high glucose per se rather than via insulin or GLP-1 indirectly. The underlying regulatory mechanism of of DPP-4 activity and its clinical implications, as well as the importance of the species difference, remain to be further investigated. Supported by: National Key Clinical Specialties of China. 404 Defective compensatory insulin secretion causes HFD-induced glucose intolerance in mice with point mutation in free fatty acid receptor 1 G. Kaiser1,2, S. Sabrautzki3,4, F. Gerst1,2, G. Przemeck4,3, H.-U. Häring1,2, M. Hrabĕ de Angelis3,4, S. Ullrich1,2; 1Institute for Diabetes Research and Metabolic Diseases (IDM) of the Helmholtz Center Munich, German Center for Diabetes Research (DZD e.V.), Tübingen, 2Internal Medicine IV, Endocrinology, Diabetology, Angiology, Nephrology und Clinical Chemistry, University Hospital Tübingen, 3German Center for Diabetes Research (DZD e.V.), Germany, 4Institute of Experimental Genetics, Helmholtz Zentrum München, German Research Center for Environmental Health, Munich, Germany. Background and aims: The fatty acid receptor 1 (FFAR1/GPR40) mediates the fatty acid-dependent augmentation of glucose-induced insulin secretion (GIIS) in pancreatic beta-cells. This feature led to the generation of specific FFAR1-agonists for the treatment of hyperglycemia. Intriguingly, Ffar1- /- mice were protected against liver steatosis, suggesting that antagonists of FFAR1 could have beneficial effects on metabolism. To further understand the impact of FFAR1 on the regulation of glucose homeostasis and insulin disposal, mice carrying point mutations in Ffar1 were generated. Materials and methods: 16,800 archived sperm samples were screened for N-ethyl-N-nitrosourea-induced point mutations in Ffar1. Mice were gener- ated on C3HeB/FeJ background using three different sperm samples, each carrying one missense mutation in Ffar1. Heterozygous and wild-type lit- termates were fed chow (CD) or high fat diet (HFD) for 8 weeks. Thereafter, glucose tolerance and insulin sensitivity as well as total pancreatic insulin content were assessed. Insulin secretion was measured in isolated islets of homozygous mice. Results: Three mice strains, two with single point mutations in the extracel- lular (T146S and R258W) and one in the intracellular domain (L106P) of the receptor showed normal development, behavior and growth. HFD-induced insulin resistance and male mice deviated from normal glucose tolerance, although the pancreatic insulin content was almost doubled in all the mice regardless of genotype. Most interestingly, heterozygous mice carrying the T146S-mutation in FFAR1 developed significant glucose intolerance de- spite the adaptive increase in insulin production. Isolated islets from T146S homozygous animals showed impaired insulin secretion in response to the FFAR1 agonist TUG-469. Conclusion: Proper activation of FFAR1 is necessary for insulin hyperse- cretion in the insulin resistant state, in order to protect from HFD-induced glucose intolerance, but FFAR1 receptor signaling may not contribute to the adaptive increase in pancreatic insulin synthesis. Supported by: BMBF to DZD e.V. 405 Stimulation of free fatty acid receptor 1 reduces thioredoxin interacting protein and exhibits anti-apoptotic properties in insulin secreting cells M. Panse1, G. Kaiser1,2, F. Gerst1,2, H.-U. Haering1,2, S. Ullrich1; 1Internal Medicine IV, University Hospital Tübingen, 2Institute for Diabetes Research and Metabolic Diseases of the Hemholtz Center Munich at the Eberhard-Karls-University of Tübingen (IDM), Germany. Background and aims: Multiple observations suggest that glucose-de- pendent stimulation of thioredoxin interacting protein (TxNIP) expression contributes to glucotoxicity. TxNIP inhibits thioredoxin thereby increasing the mitochondrial oxidative stress. Glucose upregulates TxNIP through a ChREBP-dependent transcriptional activation. ChREBP is negatively regu- lated by FOXO1. Palmitic acid activates FOXO1 through PKCδ and is known to downregulate TxNIP, inspite of inducing apoptosis under glucolipotoxic conditions. Since palmitic acid also acts as a ligand for Free fatty acid receptor 1 (FFAR1), TxNIP downregulation could occur via metabolism or receptor activation. Recently we observed that FFAR1 agonists exhibit anti-apoptotic properties under glucolipotoxic conditions. Present study investigates the underlying mechanism of the putative protective effects of FFAR1 agonists. Materials and methods: Human islets received from the European Centers of Islet Transplantation, mouse islets isolated from C57BL/6 and Ffar1-/- mice and INS-1E cells were treated either with palmitic acid (50-600 µmol/l) or the FFAR1 agonist and antagonist (TUG-469 and TUG 761 respectively, 0.1-10 µmol/l). Changes in expression were analysed by qRT-PCR and western blot. Protein expression was downregulated by siRNA techniques. Results: In human and mouse islets, as well as INS-1E cells, palmitic acid (600 µmol/l) reduced TxNIP mRNA by more than 50% under 11 mmol/l glu- cose. This effect was also observed in Ffar1-/- mouse islets. Downregulation of PKCδ reversed the effect of palmitic acid on TxNIP, while, INS-1E cells overexpressing PKCδ and displaying high levels of nuclear FOXO1, had ex- tremely low levels of TxNIP. In INS-1E cells, FFAR1 agonist (3-10 µmol/l) re- duced TxNIP by more than 60% whereas the antagonist (10 µmol/l) doubled it. In contrast to palmitic acid, the FFAR1 agonist (10 µmol/l) downregulated TxNIP in a FFAR1 dependent manner. However, downregulation of either PKCδ or FOXO1 could not reverse the effects of the ligands. Conclusion: The observations suggest that inhibition of TxNIP could be me- diated by two distinct pathways: one pathway involves PKCδ dependent nu- clear accumulation of FOXO1, stimulated by high concentrations of palmitic acid, while the other involves FFAR1 and does not seem to involve PKCδ and FOXO1, however, it may contribute to the anti-apoptotic property of FFAR1. Supported by: GRK1302 406 Expression and function of an omega-3 fatty acid receptor GPR120 in Islets of Langerhans M. Dogan, S.J. Persaud, P.M. Jones; Diabetes Research Group, Division of Diabetes and Nutritional Sciences, King‘s College London, UK. Background and aims: GPR120 and GPR40 are both activated by medium and long chain fatty acids, although they share only 10% amino acid homol- ogy. Several previous studies have demonstrated important effects on islet Diabetologia (2014) 57:[Suppl1]S1–S564 S 173 1 C function of GPR40 activation, but much less are known about the expres- sion and function of GPR120 in islets, although it has been implicated in other metabolic processes. The current study investigated the expression of GPR120 in mouse and human islets, and measured the effects of a GPR120 agonist on islet function. Materials and methods: Expression of GPR40 and GPR120 mRNAs was as- sessed by quantitative RT-PCR. Fluorescence immunohistochemistry was performed to localise GPR120 protein expression in mouse and human pan- creas. Insulin, glucagon and somatostatin secretion were quantified by radio- immunoassay. Caspase 3/7 activities were measured to determine apoptotic responses to a mixed cytokine challenge (50 U/ml IL-1β, 1000 U/ml TNF-α, 1000 U/ml INF-γ). Results: Both human and mouse islets expressed GPR120 mRNA, which was 6 times higher than GPR40 mRNA expression in mouse islets (data are expressed relative to β-actin mRNA levels; GPR120: 0.4±0.08; GPR40: 0.07±0.01; n=6; *p S 174 1 C tive correlation with TNF-α, associated with adipose tissue mass, insulin re- sistance, and inflammation, could reflect a beneficial anti-inflammatory role of sAxl. Gas6/Axl has been reported to repress TNF-α expression in various cell types including β-cells. Further studies are needed to understand the bio- logical importance of the correlations. Supported by: BerGenBio AS, Novo Nordisk Foundation, Danish Council f. Strategic Research 409 Persephin: a new player in beta cell proliferation A.J. Ineson1, A. Clarke1, W. Vernon2, D.M. Smith3, N.G. Morgan1, M. Sorhede-Winzell3, H.J. Welters1; 1Institute of Biomedical and Clinical Science, University of Exeter Medical School, 2Astra Zeneca, Macclesfield, UK, 3Astra Zeneca, Molndal, Sweden. Background and aims: Defects in beta cell proliferation and survival are known to contribute to the development of Type 2 diabetes. Previously glial cell line derived neurotrophic factor (GDNF) has been shown to improve beta cell proliferation and survival. GDNF is a member of the GDNF family ligands (GFLs) which also includes artemin, neurturin and persephin. These molecules are all potent neurotrophic factors which preferentially signal via GDNF family receptor α 1-4 (GFRα1-4) respectively. Pancreatic beta cells share a range of properties with neurons, such as expression of glutamate receptors and maintenance of membrane electrical properties which are criti- cal for the development of action potentials involved in the release of insu- lin. Expression of GDNF family receptors extends these similarities. Indeed, GFLs are essential for the appropriate innervation of the developing pancreas, these nerves are associated with Schwann cells which are known to release GFLs. Currently the effects of GFLs on pancreatic beta cells are unknown and we aim to elucidate their role using the murine beta cell line, MIN6c4, as a model. Materials and methods: Quantitative RT-PCR and western blotting were used to measure mRNA and protein expression respectively. Click- iTTM EdU kit was used to assess cell proliferation. Trypan blue was used to assess cell viability. Results: Expression of GFRα1-4 mRNA was detected in MIN6c4 cells, mouse islets and human islets (n=3, except human islets where n=2). GFRα4 dis- played at least 10 fold lower expression than all other receptors (Table 1). However protein expression of these receptors did not correlate with the mRNA levels. In MIN6c4 cell lysates both GFRα1 and GFRα4 protein was detected, but the truncated form (isoform 2) of GFRα2 was more abundantly expressed than the complete receptor and GFRα3 was undetected. Following 16hr serum starvation, MIN6c4 cells were incubated for 6 hrs with 100ng/ ml of each GFL. Both GDNF and persephin significantly increased prolifera- tion, whereas neurturin and artemin had no effect (fold increase compared to untreated (mean ± sem); GDNF: 1.8 ± 0.21, p S 175 1 C PS 017 Transgenic animal models of type 1 and type 2 diabetes 411 Expression of the chemokine receptors CXCR6 and CX3CR1 on CD8+ T cells in diabetic RIP-LCMV mice P. Müller1, S. Lasch1, M. Bayer1, J.M. Pfeilschifter1, A. Ludwig2, E. Hintermann1, P. Gutwein1, U. Christen1; 1Goethe University Hospital Frankfurt, 2RWTH Aachen University, Germany. Background and aims: In Type 1 Diabetes (T1D) self-destructing lympho- cytes are attracted to the islets of Langerhans by proinflammatory chemokines. The IFNγ inducible transmembrane chemokines CXCL16 and CX3CL1 with their receptors CXCR6 and CX3CR1 and their sheddase ADAM 10 could be novel targets for T1D therapy. Here we investigated the expression of, and ADAM10 during the pathogenesis of T1D in a mouse model. Materials and methods: We used the well-established RIP-LCMV model of T1D. As a target autoantigen in the β-cells, RIP-LCMV mice express the glycoprotein (GP) or the nucleoprotein (NP) of the lymphocytic choriomen- ingitis virus (LCMV) under control of the rat insulin promoter (RIP). RIP- LCMV-GP mice develop T1D within two to three weeks after infection with LCMV, RIP-LCMV-NP mice within one to six months after infection. In order to evaluate the influence of CXCL16/CXCR6 and CX3CL1/CX3CR1, RIP-LCMV mice were crossed to either CXCR6 or CX3CR1-deficient mice. Results: Analysis of pancreas lysates from RIP-LCMV-GP mice collected at different times after infection revealed an upregulation of CXCL16 and CX3CL1 in a time dependent manner after infection with LCMV. There is a strong increase of CXCL16 and CX3CL1 production during the acute phase of infection until day 14. Thereafter in the chronic phase of the autoimmune destruction of the beta-cells the chemokine production is decreased com- pared to the peak at day 14, but remains elevated throughout the observation time. Histologic studies of pancreata from RIP-LCMV-GP mice collected at the same times after LCMV infection revealed an enhanced CX3CL1/ CX3CR1 expression. In particular, CX3CR1 was predominantly found in in- filtrates of CD4 and CD8 T cells. Stimulation of lymphocytes from different organs (spleen, pancreatic lymph nodes, and pancreas) with immunodomi- nant LCMV-peptides followed by intracellular staining of IFNγ to determine autoantigen-specific T-cells revealed a time dependent expression of CX3CR1 on autoantigen-specific CD8 T-cells. Importantly, CX3CR1 was expressed in a significantly higher frequency in pancreatic lymph nodes and spleen on specific CD8 T-cells, but was equally expressed in the pancreas. These data indicate a selective disappearance of specific CX3CR1+ T cells after encoun- ter of their target autoantigens in the pancreas. The peak of CX3CR1+ anti- gen-specific T-cells from pancreatic lymph nodes to pancreas is delayed for two days indicating a migration of autoaggressive T-cells into the pancreas. Preliminary, incidence studies of RIP-LCMV-GP/NP x CXCR6GFP/GFP mice showed a significant reduction of T1D development in both models com- pared with RIP-LCMV. Conclusion: In summary our data demonstrate that both CXCL16 and CX- 3CL1 are expressed after LCMV infection and remain elevated in the pancreas during the autoimmune destruction of beta-cells. However further analyses suggest that both the CX3CL1/CX3CR1 and the CXCL16/CXCR6 axis might be important for islet infiltration by antigen-specific T cells and subsequently the development of T1D. Supported by: TRIP / Goethe University Hospital 412 Commensal bacteria-LPS translocation does not promote the breakdown of T cell tolerance and autoimmune diabetes G. Espinosa Carrasco, M. Villard, P. Plence, J. Hernandez; Inserm U844, Université de Montpellier 1, France. Background and aims: Type 1 diabetes results as a failure of the mechanisms that maintain immune tolerance. Genetic and environmental factors con- tribute to render potentially autoreactive T cells into pathogenic effectors. Both, CD8+ and CD4+ T cells override tolerance and cooperate to progres- sively destroy the beta cells of the pancreas. Interestingly, it has recently been shown that commensal bacteria can shape and induce anit-self T cell re- sponses. These observations suggest that bacterial products may promote the breakdown of T cell tolerance. Here we have addressed whether the systemic translocation of LPS from commensal bacteria promotes the breakdown of peripheral T cell tolerance to pancreatic antigens. Materials and methods: To address this question, we have utilized transgenic mice that express a well-characterized model antigen, influenza HA, in the beta cells of the pancreas. These mice are profoundly tolerant of the HA an- tigen in both the CD8+ and the CD4+ T cell compartments. Mild irradiation promotes the breakdown of T cell tolerance and the onset of autoimmune diabetes. One of the effects of irradiation is the systemic translocation of LPS from commensal bacteria. To assess whether bacterial LPS is responsible for the breakdown of tolerance observed we have treated mice with a cocktail of antibiotics in the drinking water in order to prevent translocation. Results: Our results demonstrated that antibiotic treatment can efficiently prevent the systemic translocation of LPS induced by irradiation. Surpris- ingly, the absence of LPS translocation did not prevent the onset of autoim- mune diabetes in transgenic mice. Beta cell-specifc CD8+ T cells proliferated extensively in response to self-antigen cross-presentation in the draining lymph nodes of the pancreas, differentiated into effector cells and infiltrated the islets of Langerhans inducing disease in both antibiotic treated mice and controls. Analyses of the CD11c+ antigen presenting cells demonstrated that irradiation induces their activation as measured by the enhanced expression of CD80, CD86, CD70, CD40 and MHC II. However, antibiotic treatment did not completely reverse this activation in all the dendritic cell subsets ana- lyzed. Conclusion: Our results indicate that commensal bacteria LPS translocation is not sufficient to promote the breakdown of CD8+ T cell peripheral toler- ance and the onset of autoimmune diabetes. Supported by: LNCC 413 Predictive value of decision-tree analysis of blood gene expression profiles for islet infiltration in the LEW.1AR1-iddm rat model of type 1 diabetes T. Schoeppe1, H. Weiss1, R. Schmidt2, M. Tiedge1; 1IBIO, 2IBIMA, University of Rostock, Germany. Background and aims: The LEW.1AR1-iddm rat is an animal model of spon- taneous autoimmune diabetes. Islet infiltration occurs within a narrow time range between 40 and 60 days after birth resulting in progressive beta cell destruction and overt diabetes. It was the aim of this study to generate a pre- diction model (decision-tree) for islet infiltration that based on gene expres- sion profiles of peripheral blood mononuclear cells (PBMC) of prediabetic normoglycaemic LEW.1AR1-iddm rats. Materials and methods: Normoglycaemic LEW.1AR1-iddm and LEW.1AR1 control rats were killed at the age of 40 - 60 days. Serial pancreatic sections were stained with Haematoxylin-Eosin (HE) to document the state of islet infiltration. RNA was isolated from purified PBMCs for PCR-Array analy- sis and RT-PCR. Gene expression data of proinflammatory cytokines, anti- inflammatory cytokines and T-cell markers were analyzed by WEKA data mining tool using C4.5 algorithm to generate a decision-tree to predict islet infiltration. Results: At the stage of organ preparation all rats were normoglyceamic (blood glucose level 5 - 7.5 mmol/l) and were discriminated by the status of infiltration (uninfiltrated, infiltrated). The gene expression profile of PBMCs followed a two-peak model with high expression of chemokines at day 40, 60 and downregulation at day 50. The cytokine- and chemokine genes Ccl2, Cxcl11, Ccl21, Il-18, TNF (2.1-2,7fold), Cxcl1, Cxcl2, Cxcl3 (3.6-5.4fold) and Ccl7, Ccl24 (6.9-8.8fold) were upregulated at day 40 in rats with islet infiltra- tion and facilitate recruitment of immune cells to the pancreas. A downregu- lation of cytokine- and chemokine genes was observed for C4b, Ccl1, Ccl17, Ccl24, Ccr2, IL-1a, IL-1b, IL-6, IL-6r, IL-9, Tlr4, Tlr7 and TNF in combination with upregulated genes Cxcl2, Cxcl3 and IL-22 at day 50 in rats with pro- gressive islet infiltration. Upregulation was also detected for C4b, Ccl1, Ccl2, Ccl21, Ccl22, IL-18 (2.3-5.3fold) and was accompanied by strong expression of Cxcl1 (33.8fold), Cxcl3, Cxcl2 (10.2-11.7fold), Ccl7 (49.7fold) and Ccl24 (129.2fold) in infiltrated animals and highlight a significant inflammation state in the blood. Decision tree analysis showed specific combinations of genes with predictive character at different time points of islet infiltration. At the early stage of islet infiltration (d40) IL-1β proved to be the most predic- tive marker of islet infiltration. At d45 IL-1β and TNF-α were characteristic for islet infiltration while with progression of islet infiltration and beta cell destruction the regulatory genes CTLA4 and CD25 (d50), NRP1 and IL-1β (d55) and IL-10 (d60) gained decisive power. Diabetologia (2014) 57:[Suppl1]S1–S564 S 176 1 C Conclusion: The process of islet infiltration is mirrored by a two peak model of proinflammatory and antiinflammatory chemokines/cytokines expression in PBMCs. A proinflammatory gene expression profile has predictive power for the early period of islet infiltration while anti-inflammatory/regulatory genes dominate the progressive period of islet infiltration. In the late period of islet infiltration before onset of manifest hyperglycaemia proinflammtory genes showed a second peak. Thus, in risk patients the gene expression pro- files in PBMCs could be an attractive strategy to monitor the state of islet autoimmunity in combination with the titer of autoantibodies. Supported by: EFSD/GSK 414 A LADA subform in the LEW.1AR1-iddm rat, a model for human type 1 diabetes A. Jörns, S. Lenzen; Institute of Clinical Biochemistry, Hannover Medical School, Hannover, Germany. Background and aims: The LEW.1AR1-idddm rat as an animal model of hu- man type 1 diabetes was characterized during diabetes development by an immune cell infiltrate of CD4 and CD8 T cells as well as CD68 macrophages with proinflammatory cytokine expression in pancreatic islets leading to ap- optotic beta cell loss. In this rat model a late onset and slowly developing subform of diabetes could be distinguished from the early and quickly pro- gressing one. Materials and methods: Twice before and immediately after diabetes mani- festation pancreatic tissue was analyzed, obtained from the rats by sequential biopsies and of the residual organ. The analyses were performed by immuno- histochemistry and in situ RT-PCR for the immune cell composition and the cytokine pattern of the infiltrate including proliferation and apoptosis rate of the pancreatic beta cells in comparison with the normoglycaemic controls. Results: The pancreatic islets from normoglycaemic LEW.1AR1-idddm rats showed no signs of immune cell infiltration and cytokine expression. Diabe- tes manifestation took place between day 55 and day 65 for the early develop- ing subform. In this fulminant subform the immune cell infiltrate was equally composed of CD8 T cells and CD68 macrophages started 3 to 5 days before diabetes manifestation. Immediately after infiltrating the islets the immune cells expressed the proinflammatory cytokines, IL-1β and TNF-α. Ultrastruc- turally confirmed apoptotic beta cell death showed an apoptosis rate between 1.7-3.7 % in the TUNEL assay whereas the proliferation rate of beta cells ranged from 0.7-2.5 %, as identified by double immunofluorescence staining of Ki67 and insulin. Diabetes manifestation occurring after day 80 showed a prolonged infiltration process of more than 10 days before disease onset. The immune cell infiltrate composition comprised 2/3 CD68 macrophages and only 1/3 CD8 T cells. These infiltrating islets strongly expressed the chemokine Ccl2 especially attracting macrophages into the islets in contrast to islets from the fulminant subform. The proinflammatory cytokine expres- sion pattern showed a strong expression of IL-1β before diabetes manifesta- tion and later on after onset concomitant with the expression of TNF-α. The apoptosis rate was lower than in the fulminant subform with values between 0.8-2.0 % whereas the proliferation rate of beta cells ranged from 0.9-3.2 %. Conclusion: Heterogeneities in the LEW.1AR1-idddm rat colony were ob- served during diabetes development. Animals with a later onset of diabetes showed a milder diabetes form with a high number of surviving beta cells. This subform of the LEW.1AR1-idddm rats starting later than day 80 with an infiltration dominated by macrophages and a slower beta cell loss resembles very well the LADA (Latent Autoimmune Diabetes in Adults) subform in hu- man adults developing autoimmune diabetes at an age >35 years. The lower frequency of T cells and the lower expression level of TNF-α in the immune cell infiltrate opens the perspective to develop specific immunomodulatory treatments for the LADA type of autoimmune diabetes. Supported by: DFG (JO 395/2-1) and EU (NAIMIT) 415 Aberrant hypothalamic expression of the master clock gene Per1 at diabetes onset in non-obese diabetic mice V. Burkart1, A. Reinbeck1,2, M. Ingenwerth3, A. Stahr3, H.-J. Partke1, C. von Gall3, M. Roden1,4; 1Institute for Clinical Diabetology, German Diabetes Center, Düsseldorf, 2Kidney Diseases Research, Bayer Pharma AG, Wuppertal, 3Institute for Anatomy II, Heinrich-Heine-University Düsseldorf, 4Department of Endocrinology and Diabetology, University Clinics Düsseldorf, Germany. Background and aims: Fundamental metabolic and immunological pro- cesses follow a circadian periodicity, which is under control of an endoge- nous rhythm generator in the hypothalamic suprachiasmatic nucleus (SCN). Previous studies demonstrated disturbed rhythmicity of energy metabolism and physical activity in newly diabetic non-obese diabetic (NOD) mice. We therefore hypothesized that the pathogenesis of insulin-deficient diabetes in NOD mice is associated with disturbed expression of the period circadian clock protein homolog 1 (Per1), a master regulator of circadian rhythmicity in the hypothalamus. Materials and methods: Female C57BL6 mice and prediabetic (normoglyce- mic) NOD mice (70-80 days old) as well as newly diabetic (< 5 d) NOD mice were kept at a 12-h light / 12-h dark cycle. The physical activity (registered as “counts”) and the respiratory exchange rate (RER) of the animals were moni- tored in an automated modular system for comprehensive metabolic pheno- typing. Immunohistochemistry of the SCN was performed to quantify the expression of the master clock gene Per1 and of the transcription factor cFos, which reflects general cerebral activity. Results: C57BL6 mice as well as non-diabetic and newly diabetic NOD mice showed comparable low physical activity during light phases whereas dur- ing dark phases diabetic NOD mice showed a two-fold lower activity than C57BL6 and non-diabetic NOD mice (p < 0.05). The altered pattern of physi- cal activity of diabetic NOD mice was associated with a constantly decreased RER in the light (0.82 ± 0.01) and dark phases (0.83 ± 0.01) when compared to non-diabetic NOD mice (RER light phase: 0.88 ± 0.01 (p < 0.001); RER dark phase: 0.95 ± 0.01 (p < 0.001)). In non-diabetic NOD mice, hypotha- lamic expression of Per1 was high (Per1 positive cell nuclei in SCN: 126 ± 19) at Zeitgeber (ZT) 14 (2 h after light-on). However, at ZT02 (2 h after light-off) the Per1 expression in these animals did not show the expected reduction that is observed in C57BL6 mice (26 ± 11) but remained elevated (108 ± 14) (p < 0.01). In NOD mice, the high levels and the lack of circadian rhythmicity of Per1 expression persisted after the onset of diabetes (ZT14: 121 ± 23, ZT02: 97 ± 19). Higher Per1 expression in NOD mice was associated with a loss of the periodicity of their hypothalamic cFos expression after diabetes onset. Conclusion: In conclusion, the finding of an aberrant hypothalamic expres- sion of the master clock gene Per1 in prediabetic NOD mice points to a role of disturbed circadian rhythmicity during the development of insulin-deficient diabetes in this model of type 1 diabetes. Supported by: German Center for Diabetes Research 416 Targeted deletion of SIRT6 in pancreatic beta cells leads to glucose intolerance M. Song1, S. Ka1, E. Bae2, B. Park1; 1Biochemistry, Chonbuk National University, Jeonju, 2Pharmacy, Woosuk University, Wanju, Republic of Korea. Background and aims: Sirtuin 6 (SIRT6) has emerged as a key regulator of glucose and fatty acid metabolism in liver and adipose tissues. However, the role of SIRT6 in pancreatic beta-cells has not reported. In this study, we ex- amined the physiological role of SIRT6 in pancreatic beta-cell function and systemic glucose homeostasis. Materials and methods: A Cre-LoxP system was used to generate beta-cell- specific SIRT6 knockout (BS6KO) mice. Metabolic and histological analyses were carried out in BS6KO mice as well as littermate controls after normal chow diet (NCD) or high fat diet (HFD) for 16 weeks. Additionally, isolated islets from BS6KO mice were used to investigate the effect of SIRT6 deletion on insulin secretion and beta-cell viability. Results: While their fasting glucose and body weight were comparable with wild-type mice on NCD, BS6KO mice fed the HFD induced more weight gain and exhibited hyperglycemia in both nonfasting and 16-h fasting conditions. Oral glucose tolerance test revealed that BS6KO mice had severe glucose in- tolerance, which was paralleled by a reduction in insulin-immunoreactive Diabetologia (2014) 57:[Suppl1]S1–S564 S 177 1 C pancreatic beta-cell mass. Mimicking this, islets from BS6KO mice exhibited profound defects in the glucose-stimulated insulin secretion. On the con- trary, normal islets transduced with SIRT6 adenovirus were resistant against cytokine- and lipo-toxicity. Conclusion: In addition to being essential for maintaining glucose homeo- stasis, SIRT6 is involved in beta-cell function by protecting beta-cells against exogenous insults and thereby maintaining insulin secreting capacity. 417 Glycogen over-accumulation in beta cells affects insulin secretion and glucose homeostasis J. Mir1,2, R. Gasa2, J. Duran1, R. Gomis2, J.J. Guinovart1; 1Institute for Research in Biomedicine, 2Institut d‘Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain. Background and aims: The presence of detectable amounts of glycogen in β-cells of diabetic patients has led to the notion that excessive glycogen ac- cumulation is involved in glucotoxicity and β-cell dysfunction. However, the causal contribution of β-cell glycogen to the pathophysiology of diabetes re- mains elusive. The regulatory subunit Protein targeting to glycogen (PTG) induces the dephosphorylation of glycogen synthase, resulting in a hyper- activation of this enzyme and enhanced glycogen buildup. In this study, to directly investigate the effect of increased glycogen storage in β-cells on glu- cose homeostasis and β-cell function, we have generated a transgenic mouse model over-expressing PTG in these cells. Materials and methods: Mice over-expressing PTG in β-cells (βPTG) were generated by crossing conditional PTG transgenic mice with RIPCre mice. RIPCre littermates were used as controls (Ctrl) in all experiments. Glucose homeostasis was measured by intraperitoneal glucose tolerance test (IP- GTT) and insulin sensitivity test (ITT). Pancreatic insulin positive area was measured by morphometry. Glycogen was detected by PAS staining and by enzymatic methods after precipitation in 66% ethanol and digestion with α-amyloglucosidase. Insulin secretion was measured by static incubation of isolated islets and quantified by ELISA. Results: βPTG mice were born at expected ratios and showed no differences in growth and body weight. PAS staining revealed heterogeneous glycogen accumulation in β-cells. βPTG mice exhibited a mild impairment in glucose tolerance (AUC βPTG vs Ctrl: 665 ± 65 vs 429 ± 56, p S 178 1 C tivated MIN6 cells after Tbc1d1 knockdown (n=5) was significantly increased compared to respective controls (0,9 ± 0,038 fmol x min-1 x islet-1 vs. 1,28 ± 0,069 fmol x min-1 x islet-1, p S 179 1 C tive metabolism affecting cytosolic ATP/ADP levels, such as amino acids or methyl succinate, had no effect on the second phase of secretion. At the same time, changes in [Ca2+]i show a strong temporal correlation with exocytosis. The effects of [Ca2+]i on insulin exocytosis were potentiated by agents produc- ing an elevation of intracellular cAMP (like GLP-1). Conclusion: These findings illustrate the power of on-line high-resolution monitoring of multiple parameters in living systems. Our data suggest a hier- archy of intracellular signalling events in the control of insulin secretion. We propose that insulin release kinetics is principally determined by variations of signalling events (including [Ca2+]i) rather than energy metabolism. The lat- ter is likely to play the triggering role but the temporal fine-tuning is exerted via more distal events. Supported by: Wellcome Trust Programme Grant (PR), OxBRC Fellowship (AIT) 422 Identification of Rab27a-GAP-interacting proteins and its functional analysis in pancreatic beta cells M. Yamaoka1, T. Ando1, M. Okamoto1, T. Terabayashi1, K. Matsunaga2, R. Ishizaki2, T. Izumi2, I. Niki1, T. Ishizaki1, T. Kimura1; 1Pharmacology, Oita University, Faculty of Medicine, Yufu, 2Laboratory of Molecular Endocrinology and Metabolism, Department of Molecular Medicine, Institute for Molecular and Cellular Regulation, Gunma University, Maebashi, Japan. Background and aims: Membrane trafficking is crucial for the regulation of the size of the readily releasable pool of insulin granules and for the recovery of insulin secretory membranes. Rab27a, which is a member of the Rab fam- ily, have been considered to control of pre-exocytosis in pancreatic-beta cells through its GTP-dependent effectors. In contrast, the GDP-bound form has been regarded as an inactive form, because the functional binding partner of this form has not been identified. We previously identified coronin 3 as a novel GDP-dependent effector of Rab27a in pancreatic beta-cells. Moreo- ver, we found that the insulin secretagogue glucose caused a shift from the GTP- to GDP-bound Rab27a via Rab27a-GAP, resulting in the regulation of endocytosis of insulin secretory membranes. Taken together, GTP- and GDP-bound Rab27a regulates insulin secretion at the pre-exocytotic and the endocytotic stages, respectively. In the present study, we investigated the un- derlying mechanisms by which Rab27a-GAP controls endocytosis of insulin secretory membranes in pancreatic beta-cells. Materials and methods: Affinity column chromatography was performed to identify Rab27a-GAP interacting proteins. Extracts from the insulin-secret- ing beta-cell line, MIN6, were loaded onto glutathione-sepharose 4B beads coated with GST-Rab27a-GAP. The proteins bound to the columns were eluted and were analyzed by peptide mass fingerprinting. For binding assay, immunoprecipitation analysis was performed using COS7-cells expressing Rab27a-GAP. For the observation of the intracellular localization of Rab27a- GAP and Arf6-GEF, MIN6-cells were analyzed with a TIRF microscopy sys- tem or a confocal laser microscopy system. Results: We identified Arf6-GEF as a Rab27a-GAP interacting protein. Arf6-GEF is a guanine nucleotide exchange factor that converts Arf6 from the GDP- to the GTP-bound form, resulting in the regulation of endocyto- sis through clathrin-coated vesicle formation. We found that Rab27a-GAP directly interacted with Arf6-GEF. The C-terminal region of Rab27a-GAP bound the PH domain of Arf6-GEF. Glucose stimulation induced the redis- tribution of both Rab27a-GAP and Arf6-GEF to the vicinity of the plasma membrane. Silencing of Arf6-GEF inhibited the glucose-induced redistribu- tion of Rab27a-GAP. In contrast, the redistribution of Rab27a-GAP did not occur in Arf6-GEF-silencing cells. Conclusion: These results indicate that Arf6-GEF is required for the glucose- induced redistribution of Rab27a-GAP. We previously reported that GDP- bound Rab27a regulates the retrograde transport of the internalized secre- tory membrane, the stage after the GTP-bound Arf6-dependent formation of clathrin-coated vesicles. We therefore propose a model where Rab27a-GAP is involved in both Rab27a and Arf6 signaling and plays a pivotal role in mem- brane trafficking for insulin secretion. Supported by: Gunma Univ. (11028) 423 Temperature-dependent effects of high glucose and potassium depolarisation on insulin granule mobility and exocytosis in insulin-secreting MIN6 cells K. Schumacher1, M. Matz2, K. Baumann2, I. Rustenbeck1; 1Institute of Pharmacology and Toxicology, 2Institute of Medicinal and Pharmaceutical Chemistry, University of Braunschweig, Germany. Background and aims: It is widely assumed that the first phase of glucose- induced insulin secretion can be mimicked by K+ depolarization. We found, however, that the insulinotropic efficacy was differently affected by lowering the temperature. We thus compared parameters of insulin granule mobility and secretion during glucose stimulation and K+ depolarization at different temperatures. Materials and methods: Insulin granules were labeled by transient transfec- tion with insulin-EGFP and the granules present in the submembrane space were imaged by TIRF microscopy. Single MIN6 cells were continuously peri- fused with oxygenated HEPES-buffered Krebs-Ringer medium containing the respective stimuli at 37 °C, 32 °C, and 22 °C. The image files (1 sequence = 200 images) were evaluated by an in-house written program. The insulin secretion of MIN6 pseudo-islets was measured by perifusion with ELISA of the fractionated efflux. Results: At 37 °C stimulation of the perifused MIN6 pseudo-islets with 30 mM glucose and, after a wash-out period of 10 minutes, 40 mM K+ increased insulin secretion. K+ was about twofold more effective than glucose. At 32 °C, K+ was even more effective than at 37 °C, and was markedly but only tran- siently effective at 22 °C. Relative to K+, glucose lost efficacy at 32 °C and was virtually ineffective at 22 °C. When the stimuli were applied in the reversed sequence K+ was of unchanged efficacy, but glucose was almost ineffective at all temperatures. The same experimental protocol was used to measure the mobility of insulin granules in the submembrane space of perifused single MIN6 cells. Under control conditions (3 mM glucose, 37 °C), the mean num- ber of submembrane granules was 357 ± 37 per cell footprint at the beginning of image acquisition. 33.5 ± 1.4 % of these remained visible throughout the sequence (long-term residents). Newly arrived granules made up 8.9 ± 0.6 % per image, equivalent to about 32 granules. The number of departing granules closely mirrored that of arriving granules (mobility in the z-dimension). The short-term residents (present for ≤ 1 s) amounted to 82.2 ± 0.9 % of the total number of granules (7547 ± 928) in a sequence. At all temperatures stimula- tion with glucose and K+ led to an increase in the number of arriving and departing granules, independent of the sequence of exposure. Consequently, the total number of granules per sequence was increased during phases of stimulation. The caging diameter (mobility in the x/y-dimension) in contrast, decreased during glucose- and K+-stimulation at 37 °C, but not at 32°C and 22°C. The number of the long-term resident granules was similar for all three temperatures at basal glucose and when 40 mM K+ was applied first. How- ever, when 30 mM glucose was applied first the number remained stable at 37°C in contrast to a steady decline at 32°C and 22°C. Conclusion: At all temperatures, glucose and K+ depolarization enhance the granule turnover in the submembrane space. This characteristic fits to K+- induced secretion, whereas the temperature-dependency of the caging diam- eter and of the long-term resident granules reflects the secretion pattern of the glucose stimulus when it is applied first. Additionally, glucose-induced secretion is associated with limited lateral mobility and/or longer presence of the granules at the plasma membrane. Supported by: DFG, DDG, DDS 424 CaMKII-mediated metabolis memory of pancreatic beta cells controls insulin secretion and is inhibited by palmitate G.J. Santos1,2, S.M. Ferreira1, F. Ortis1, L.F. Rezende1, A.C. Dias1, E.M. Carneiro1, K.H. Kaestner2, A.C. Boschero1, T.R. Nardeli1; 1Department of Anatomy, Cell Biology and Physiology and Biophysics, State University of Campinas, Brazil, 2Department of Surgery and Institute for Diabetes, Obesity and Metabolism, University of Pennsilvanya, Philadelphia, USA. Background and aims: Ca2+/calmodulin-dependent protein kinase II (CaM- KII) functions both in regulation of insulin secretion and neurotransmitter release through common downstream mediators. Memory is the ability to acquire, to store and to evocate any kind of information. In CNS, the process behind this phenomenon in the Long-Term Potentiation (LTP) and is known Diabetologia (2014) 57:[Suppl1]S1–S564 S 180 1 C that it requires Ca2+ to occur. In additional, CaMKII is necessary to store information during LTP. In pancreatic ß-cells, CaMKII plays pivotal role during GSIS process. Palmitate is a Saturated-fatty-acid-induced important for beta cell metabolism, especially during DM2 onset. It is well known that palmitate inhibits glucose-induced CaMKII phosphorylation and potentiates GSIS. Therefore, we investigates that pancreatic ß-cells acquire and store the information as a form of “metabolic memory”, just as neurons store cognitive information, and, that palmitate inhibits this process. Materials and methods: To test this hypothesis, we developed a novel par- adigm of pulsed exposure of mice and human ß-cells to intervals of high glucose, followed by a 24-hour consolidation period to eliminate any acute metabolic effects. After this period, we analyzed insulin secretion (by RIA), protein expression (by Western blot), response to a glucose-ramp and the glucose-induced Ca2+ influx. Results: Strikingly, ß-cells exposed to this high-glucose pulse paradigm ex- hibited significantly stronger insulin secretion. This metabolic memory was entirely dependent on CaMKII and was inhibited by Palmitate. We also ob- served, in pulse group, an increase in Ca2+ influx induced by glucose. In ad- ditional, metabolic memory was reflected on the protein level by increased expression of proteins involved in GSIS and Ca2+-dependent vesicle secre- tion, such as GCK, Cav1.2, SNAP25, pCaMKII and pSynapsin. Finally, we observed in human islet elevated levels of the key ß cell transcription factor MAFA. Conclusion: In summary, like neurons, human and mouse ß-cells are able to acquire and retrieve information in a process dependent on CaMKII that is inhibited by Palmitate. Supported by: NIH, FAPESP and CNPq 425 Islet microcapillary endothelial cells secrete an attenuating factor of insulin secretion S. Sasson1, B. Daniel1, A. Livne1, K. Nurit2, G. Cohen1; 1Pharmacology, 2Medicine, Hadassah- Hebrew University School of Medicine, Jerusalem, Israel. Background and aims: Paracrine interactions and chemotactic signals ex- changed within islet of Langerhans’ endothelium and the endocrine cells have been proven critical for the development and organization of functional islet units. Moreover, autocrine and paracrine signals within the microen- vironment of the islets can also modulate β-cell glucose-stimulated insulin secretion (GSIS). For instance, it has been shown that several extra-cellular matrix components, derived from islets’ microcapillary endothelium (IME), regulate β-cell spreading and amplify insulin secretion. We aimed at investi- gating whether IME secrete other soluble factors that affect β-cell function and viability. Materials and methods: Conventional conditioned-medium (CM) from IME (the MS-1 cell line and primary cultures of endothelia cells from rat islets) were prepared to study contact-independent IME-β-cell interactions. Standard GSIS assay in the absence or presence of insulin secretagogues was employed. In addition, cell viability assays (cell count, MTT assay and an- nexin/PI flow cytometry) and size-based fractionation of the CM were per- formed. Results: Our data show that IME-derived CM contained soluble factor(s) that markedly attenuated GSIS from both INS-1E β-cell line and freshly iso- lated rat islets. Fractionation of the CM indicated that the active factor was a high molecular weight protein. This factor reduced GSIS without changing the rate of insulin biosynthesis, cell viability and cellular ATP content. Using the insulin secretagogues IBMX, KCl, tolbutamide and L-arginine, we found that only the second phase of insulin secretion was markedly reduced by the CM. This was accompanied with a significant decrease in cellular cAMP level following glucose stimulation. Furthermore, mass spectrometric analysis of several high molecular weight proteins in the active CM fraction are cur- rently being analysed to identify the active attenuating factors of GSIS. Conclusion: These findings support the hypothesis that a paracrine interac- tion between IME and β-cells modulates glucose-dependent insulin release and may contributes to maintaining glucose homeostasis. Supported by: David R. Bloom Center for Pharmacy and the Dr. Adolf and Klara Brettler Centre 426 Glucose-stimulated insulin secretion is reduced after treatment with rosuvastatin calcium V.A. Salunkhe, O. Elvstam, J.L.S. Esguerra, L. Eliasson, J. Vikman, A. Wendt; Diabetes Centre, Islet Cell Exocytosis, Clinical Research Centre, Lund University, Malmö, Sweden. Background and aims: Statins are a family of drugs widely used to reduce the risk of cardiovascular disease by lowering the cholesterol levels. Rosuvastatin calcium (RS) is one of these highly prescribed statins and there are studies suggesting increased risk of diabetes for subjects on RS therapy. Whether RS influences insulin secretion from the beta-cell is yet to be established, but our group has previously shown that insulin secretion is highly sensitive to changes in plasma membrane cholesterol. Here, we aim to investigate the ef- fects of RS on insulin secretion, exocytosis and Ca2+ influx through voltage- dependent Ca2+ channels in pancreatic beta-cells. Materials and methods: The study was performed in INS-1 832/13 cells. These cells were treated with different concentrations of RS (20 nM, 200 nM, 2 μM and 20 μM) for various time spans (24/48 hour (h) and 1/2 month). Insulin secretion at 2.8 and 16.7 mM glucose (G) was measured by radio- immunoassay. Exocytosis was detected as changes in membrane capacitance evoked by a train of 10 depolarizations from -70 mV to 0 mV using whole- cell patch-clamp configuration. Results: First we studied influence on glucose-stimulated insulin secretion (GSIS) in INS-1 832/13 cells treated with above mentioned concentrations of RS for 48 h. In cells treated with 200 nM, 2 μM and 20 μM RS insulin secre- tion at 16.7 mM G was significantly reduced (26%, 35% and 29% reduction; p S 181 1 C to overexpress the core protein Sdc4 and to clarify the contribution of Sdc4 to beta-cell function. Materials and methods: Subcloning of MIN6 cells, mouse pancreatic beta- cell line, was performed by the limiting dilution method. The cells were then screened and selected by an index of GSIS, KCl-induced insulin secretion, HS expression and mRNA expression of Sdc4. A stable transformant expressing the mouse Sdc4 was established by transfecting pCMV-mSDC4 and selected with hygromycin B. The stable transformant expressing the highest level of Sdc4 mRNA (by quantitative RT-PCR analysis) was used for further experi- mentation. GSIS activity was measured with a basal (2.8 mM) or stimulatory (11.2 or 25 mM) glucose concentrations. Results: MIN6 cells were subcloned, and 30 clones were obtained. Two sublines were selected for this study, designated T3 and T16. T3 cells ex- hibit GSIS in a concentration-dependent manner, whereas T16 cells re- spond poorly to glucose. In the presence of KCl, T16 cells secreted al- most the same amount of insulin as T3. Next, the expression of HS in these subclones was analyzed by western blot analysis. Immunoblot with anti-HS antibody showed a single band of 35 kDa, consistent with the re- ported size of the Sdc4, was detected in T3 cells. On the other hand, HS was not detected in T16 cells. In addition, the mRNA of Sdc4 was de- tected in T3 cells by RT-PCR analysis, while not detected in T16 cells. In Sdc4 overexpressed T16 (T16/mSdc4) cells, insulin secretory response exhibited 3.1-fold (p S 182 1 C in diabetes and whether such activation triggers β-cell death, are not known and is investigated in this study. Materials and methods: Isolated human islets and the rat β-cell line INS1 were exposed to a diabetic milieu (IL-1beta/IFNgamma or increased glucose concentration (22.2 mM)). Bcl-xL, MOB1 (LATS2-associated protein), Ring- E3 lygase Praja2 (MOB1 negative regulator) and β-cell apoptosis (Caspase 3 & PARP cleavage) were analyzed by Western blotting. Praja2 and Bcl-xL ex- pression were also analyzed by real-time PCR. LATS2 was overexpressed by plasmid transfection, LATS2 inactivation was performed by overexpressing dominant-negative LATS2 (dn-LATS2) or specific siRNA to LATS2. Bcl-xL turnover was performed in HEK 293 cells overexpressing Bcl-xL and LATS2. Results: LATS2 was activated in human islets and INS1 cells exposed to IL- 1beta/IFNgamma or increased glucose concentrations. This correlated with MOB1 up-regulation, decreased Praja2 expression and increased β-cell apop- tosis. Overexpression of LATS2 itself increased MOB1 levels and β-cell apop- tosis in INS1 cells and human islets indicating that LATS2 alone is sufficient to promote β-cell death. MOB1 interacts with and activates LATS2 kinase activity. Interestingly, MOB1 co-precipitation with LATS2 was increased in the presence the diabetic milieu compared to untreated INS1 cells suggesting that MOB1 is essential for the higher LATS2 activity under diabetic condi- tions. LATS2 overexpression in INS1 cells as well as in Bcl-xL -overexpressing HEK293 cells decreased mitochondrial anti-apoptotic protein Bcl-xL without changes in Bcl-xL mRNA levels; this suggested that the decrease in Bcl-xL expression occurred at a post transcriptional level and mediated the pro-ap- optotic function of LATS2. Reciprocally, inhibition of endogenous LATS2 ac- tivity by siRNA knockdown or overexpression of dominant negative LATS2 protects β-cells from gluco- and cytokine-induced apoptosis demonstrated by decreased caspase-3- and PARP-cleavage as well as increased Bcl-xL levels. LATS2 knockdown down-regulated MOB1 and restored Praja2 levels under diabetic conditions in INS1 cells indicating a major role of MOB1 and Praja2 in the mechanism of stress-induced β-cell apoptosis. Our results indicate that LATS2, MOB1 and Praja2 regulate each other and may be a component of the previously uncharacterized stress-sensitive apoptotic pathway. Under normal conditions, Praja2 promoted cell survival by degrading MOB1 and then in- hibiting LATS2 action, but prolonged diabetogenic stress decreased Praja2 expression, which allowed higher MOB1 levels and pro-apoptotic LATS2 signaling. Conclusion: Our results show that LATS2 plays an important role in β-cell apoptosis and its inhibition may provide a new strategy to restore β-cell sur- vival in diabetes. Supported by: ERC 430 Identification of the ubiquitin ligase SCF(FBW7) as a novel regulator of the NF-κB pathway in pancreatic beta cells K. Meyerovich1, M. Fukaya1, L.F. Terra2, D.L. Eizirik1, A.K. Cardozo1; 1Université Libre de Bruxelles, Belgium, 2Universidade de São Paulo, Brazil. Background and aims: Activation of the transcription factor NF-κB by pro- inflammatory cytokines contributes to β-cell demise in animal models of type 1 diabetes (T1D). Thus, understanding on how NF-κB is activated, is crucial to prevent β-cell death. SCF(FBW7) is an ubiquitin ligase with three isoforms, localised in the nucleolus, nucleus, and cytosol. It has been previ- ously reported that FBW7 positively contributes to NF-κB activity in cancer- ous cells. The aim of this study was to evaluate whether FBW7 modulates NF-κB activation and survival of pancreatic β-cells. Materials and methods: Specific siRNAs were used to downregulate FBW7 and βTrCP expression. Overexpression of FBW7 was carried out with plas- mid encoding FBW7. Cell viability was assessed by the DNA-binding dyes Propidium Iodide and Hoechst 33342. NF-κB activity was measured using a luciferase reporter assay. mRNA expression of NF-κB-dependent genes, FBW7 and βTrCP was quantified by qPCR, while the proteins regulating NF- κB activation were analysed by Western blot. Results: IL-1β+INF-γ decreased FBW7 mRNA expression in primary rat β-cells by 45%. Knockdown of FBW7 by siRNA (FBW7) significantly in- creased IL-1β+IFN-γ-mediated NF-κB activation in INS-1E cells, primary beta cells as well as in human β-cell line as measured by a luciferase reporter promoter. This was accompanied by increased expression of the NF-κB tar- get genes Rantes and CCL19. There was also a two-fold induction of nitric oxide production (p S 183 1 C 432 IL-13 improves beta cell survival and protects against IL-1beta-induced beta cell death S. Rutti1, G. Ruozi2, L. Braga2, V. Jimenez3, A. Casellas3, C. Mallol3, E. Casana3, F. Bosch3, M. Giacca2, P. Halban1; 1Genetic Medicine and Development, University of Geneva, Switzerland, 2Molecular Medicine Laboratory, International Centre for Genetic Engineering and Biothechnology, Trieste, Italy, 3School of Veterinary Medicine, Universitat autonoma de Barcelona, Spain. Background and aims: IL-13 is a cytokine classically produced by anti-in- flammatory T-helper-2 lymphocytes. Interestingly, IL-13 is decreased in the circulation of type 2 diabetic patients and has recently been shown to impact positively on liver and skeletal muscle. Although IL-13 can impact positively on transformed beta-cell lines, its impact and mode of action on primary hu- man or rodent beta-cell function and survival remain to be explored. Materials and methods: Human and rat dispersed islet cells and sorted beta- cells were maintained in culture for 48 h in the presence of IL-13 (1 to 50 ng/ml) alone or in combination with IL-1β (20 ng/ml). Insulin secretion in response to glucose (1h; 16.7 mmol/l) was measured by radioimmunoassay, proliferation by incorporation of BrdU over 48 h, cell death by TUNEL and protein expression/phosphorylation by western blot. Results are presented as mean±SE, n=3-5 independent experiments with statistical significance of dif- ferences assessed by Student’s t test or ANOVA with Bonferroni post hoc test. Results: IL-13 did not affect human or rat beta-cell basal or glucose-stim- ulated insulin secretion, or rat beta-cell proliferation. IL-13 (10 ng/ml) de- creased basal cell death in human (IL-13: 0.6±0.2% TUNEL+ cells vs. con- trol: 1.1±0.3%; p S 184 1 C in gene expression profiles elicited uniquely by acute exposure to low but not high concentrations of IL-1beta and that could drive improved beta cell function. Supported by: FNSNF 435 The hepatokine fetuin-A triggers inflammation without inducing cell death but improves cAMP-dependent insulin secretion of pancreatic islets F. Gerst1,2, G. Kaiser1,2, N. Stefan1,2, H.-U. Häring1,2, S. Ullrich1,2; 1Internal Medicine IV, University Hospital Tübingen, 2Institute for Diabetes Research and Metabolic Deseases of the Helmholtz Center Munich at the Eberhard-Karl-University of Tübingen, Germany. Background and aims: Nonalcoholic fatty liver displays an altered secretion of hepatokines. Previously, we found that increased plasma levels of the hepa- tokine fetuin-A strongly predict the incidence of type 2 diabetes in humans. As mechanisms of action, fetuin-A directly inhibits insulin receptor signaling and augments inflammation by interacting with fatty acids to activate TLR4. In addition, we found a negative relationship of fetuin-A levels with adjusted insulin secretion only in subjects with impaired glucose tolerance. The pre- sent study aims to examine the impact of fetuin-A on fatty acid-induced ef- fects in pancreatic beta-cells. Materials and methods: Human islets received from the European Centers of Islet Transplantation (ECIT) and mouse islets were cultured in medium supplemented with human fetuin-A or serum albumin as control (0.6 mg/ ml each) for 48h. Due to the low concentration of albumin palmitic acid was used at a concentration of 60 µmol/l, which does not induce metabolic stress. Changes in gene expression were analyzed by microarray and qRT-PCR. Ap- optosis was estimated by TUNEL staining. Insulin secretion was evaluated after static incubation by radioimmunoassay. Results: In agreement with previous observations in other cell types, fetuin- A increased 2-fold the mRNA levels of IL1ß, IL33, IL24 and IL17RB in hu- man islets. The stimulation of cytokine production depended on the activa- tion of TLR2/4, as fetuin-A elevated the mRNA amount of IL1ß and MCP-1 in WT, but not in TLR2/4 KO islets. On the contrary, palmitic acid neither stimulated nor augmented the fetuin-A-mediated cytokines production. Sur- prisingly, in spite of increasing the cytokine levels, fetuin-A failed to induce apoptosis and, moreover, significantly inhibited palmitic acid-induced cell death. In addition, chronic exposure to fetuin-A did not affect glucose and palmitic acid stimulated insulin secretion but it selectively and specifically improved forskolin and exendin-4 stimulation of secretion. Conclusion: These results suggest that fetuin-A stimulates cytokine produc- tion in a TLR-dependent manner. However, it exerts rather beneficial effects on beta cell function, as it does not induce apoptosis and improves cAMP- amplified insulin secretion. Supported by: DZD e.V. PS 020 Clinical immunology of type 1 diabetes 436 The promoters of genes may be the closest link between type 1 diabetes and other autoimmune diseases C. Guja1, P. Gagniuc2, C. Ionescu-Tirgoviste1; 1Diabetes, Nutrition and Metabolic Diseases, „Carol Davila“ University of Medicine and Pharmacy Bucharest, 2Institute of Genetics, University of Bucharest, Romania. Background and aims: Type 1 diabetes (T1D) is a T-cell mediated autoim- mune disease that selectively targets pancreatic beta cells. T1D is thought to arise from the progressive loss of beta cells occurring over months upto years. Despite extensive genetic studies in T1D, rarely analyzes were performed with respect to promoters of genes associated with autoimmunity and T1D. Materials and methods: For our study we chose a set of representative genes for several diabetes phenotypes. We analyzed the promoters of genes asso- ciated with T1D, including HLA-DQA1, HLA-DQB1, HLA-DRB1, HLA- DPB1, PTPN22, TLR7, CTLA4, GSDMB, STAT4, IL7R, C1QTNF6, CD55, CTSH, ERBB3 and INS and with T2D, including CAMK1D, DUSP9, HHEX, IRS1, MADD, NOTCH2, TP53INP1, VPS13C, WFS1, ZFAND6, HMGA2, PPARG, CDKN2AIP, PROX1 and TCF7L2. In order to detect a possible re- lationship between T1D and other autoimmune diseases, we analyzed the promoters of genes associated with more than only one autoimmune disease, including RGS1, IL10, AFF3, IL21, INS, KIF3A, GPR183, GSDMB, ORM- DL3, PTPN2, FUT2, TLR7, PTPN22, CTLA4, CD55 and STAT4. The DNA pattern” methodology that we have used, determines the coexpression rela- tionships between genes. Thus, in the resulting distribution, the overlapping positions (or close positions) of promoters of these genes indicate that those promoters use transcription factors in common. Results: Our distribution shows that there is a common denominator be- tween promoters of genes involved in autoimmunity. Both the promoters of T1D genes and those associated with other autoimmune diseases occupy the same overlapping positions in our distribution. The figure (Figure A) below shows the general distribution of different phenotypes, namely T1D (in blue), T2D (in red), genes associated with both phenotypes (included in ,,Intermediary Diabetes Mellitus” - IDM phneotype) in green and promot- ers of genes associated with autoimmunity (in brown). In a general distribu- tion of the average positions of gene promoters for each phenotype (Figure B), the close functional relationship between T1D and autoimmune diseases gene promoters is becoming increasingly clear. The clustering of these genes, very close to each other, suggests that transcription factors involved in their expression may be part of the triggering mechanism of anti beta cell auto- immunity. The close distribution of the promoters of genes associated with the analyzed phenotypes represents a functional correlation, in the sense that these genes are more likely to be co- expressed. Conclusion: Promoters of genes associated with T1D and with other autoim- mune diseases share one common factor in causing T1D, namely the tran- scription factors involved in long term gene expression. The wide distribution between promoters of genes associated with T1D and T2D strongly suggests that these two phenotypes are not using transcription factors in common. Clinical Trial Registration Number: CNCS-UEFISCDI Nr. PN-II-ID- PCE-2011-3-0429 Diabetologia (2014) 57:[Suppl1]S1–S564 S 185 1 C 437 A novel autoantibody detected in patients with fulminant type 1 diabetes F. Haseda1, A. Imagawa2, H. Nishikawa3, S. Mitsui1, C. Tsutsumi1, R. Fujisawa1, H. Sano1, Y. Murase-Mishiba1, J. Terasaki1, S. Sakaguchi3, T. Hanafusa1; 1Department of Internal Medicine (I), Osaka Medical College, Takatsuki, 2Department of Metabolic Medicine, Graduate School of Medicine, 3Experimental Immunology, Immunology Frontier Research Center, Osaka University, Suita, Japan. Background and aims: Fulminant type 1 diabetes (FT1D) is a subtype of type 1 diabetes characterized by complete insulin deficiency resulting from the destruction of pancreatic beta cells even at the disease onset. Massive cellular infiltration of T-cells and macrophages has been detected in islets and exocrine pancreas, suggesting that immune disorder might contribute to the development of FT1D. However, islet-related autoantibodies, such as GAD Ab or IA-2 Ab, were usually negative, indicating that we have no useful biomarker to diagnose FT1D. Therefore, we performed serological antibody analysis cyclopedically to discover a novel diagnostic marker of FT1D. Materials and methods: First, we analyzed a total of 6 serum samples from 3 patients with FT1D (1 sample in acute and 1 in chronic phases from each patient) on over 9000 human protein arrays (Invitrogen ProtoArrayR Human Protein Microarray v5.0) by fluorescence. Second, titres of the antibody were measured in sera from 20 patients with FT1D (both in acute and chronic phases), 32 patients with type 1A diabetes (T1AD), 30 patients with type 2 diabetes (T2D), 22 patients with autoimmune thyroid disease (AITD) and 30 healthy control subjects (HC) using ELISA assay. Duration of FT1D in acute and chronic phase was 8.1 ± 6.3 (mean ± SD, day) and 24.2 ± 9.1, respectively. Results: By serological antibody analysis of over 9000 antibodies, we detected 8 antibodies which showed high signals from all 3 patients with FT1D in acute phase (acute/chronic phase ratio>1.4). We focused on one novel anti- body that has not yet been reported in any conditions and measured its titre by ELISA assay. Titres of the antibody were 0.1071 ± 0.0374 (mean ± SD, arbitrary unit) in FT1D patients (acute phase), 0.0823 ± 0.0182 in FT1D pa- tients (chronic phase), 0.0771 ± 0.0171 in T1AD patients, 0.0710 ± 0.0171 in T2D patients, 0.0701 ± 0.0116 in AITD patients and 0.0641 ± 0.0078 in HC. Significantly high titre of the antibody was detected in sera from FT1D pa- tients in acute phase (versus T1AD patients; P=0.0002, versus T2D patients; P S 186 1 C Conclusion: 1) There are close associations between different immune me- diators 2) These associations are not specific for autoimmune diabetes 3) BMI is a major confounder 4) The associations of beta cell decline to individual immune mediators need confirmation in further studies. Clinical Trial Registration Number: NCT00131755 Supported by: Research Council of Norway, Norwegian Diabetes Assoc. and St.Olavs Hospital 440 Relationship between autoantibodies combination, metabolic syndrome and its components in autoimmune diabetes in adults T. Bulum, K. Blaslov, J. Knežević-Ćuća, L. Duvnjak; Vuk Vrhovac Clinic for Diabetes, Endocrinology and Metabolic Diseases, University Hospital Merkur, Zagreb, Croatia. Background and aims: Although clinical presentation of latent autoimmune diabetes in adults (LADA) at diagnosis frequently features clinical traits of type 2 diabetes mellitus (T2DM), it is the autoimmune process that drives LADA patients through slower beta-cells destruction and insulin depend- ency within few years. Majority of LADA patients has pronounced positivity for glutamic acid decarboxylase enzyme antibodies (GAD-Ab) frequently ac- companied by moderate levels of circulating autoantibodies to islet cell cy- toplasmic antigens (ICA-Ab). Positivity for tyrosine phosphatase-like trans- membrane glycoprotein (IA2-Ab) has also been reported but to a lesser ex- tent. While simultaneous positivity for ICA-Ab, and GAD-Ab with eventual presence and of IA2-Ab indicate patients with insulin deficiency and typical type 1 LADA phenotype, GAD antibody positivity, even in higher titers, rep- resent a marker for slower progressing LADA. The role of IA2Ab positivity in LADA phenotypisation is poorly understood and controversial. The aim of our study was to establish a possible association between autoimmune profile of patients regarding double or triple antibody positivity and LADA pheno- type in the context of metabolic syndrome (MS) prevalence, individual com- ponents of metabolic syndrome and chronic diabetic complications. Materials and methods: This cross-sectional study population comprised 69 islet cell antibody positive patients coming for their comprehensive an- nual review. Rate of positivity was calculated by determining end-point titres of samples that were converted to the units of Juvenile Diabetes Foundation (JDF-U) by comparison with a standard curve of end-point titer of standard sera. The threshold of detection was >5 JDF units. GAD and IA2-Abs were detected with an enzyme-linked immunosorbent assay (Euroimmun AG, Luebeck, Germany) and the results were expressed in arbitrary units. The cut-off limit was 10 U/ml for GAD Abs and 15 U/ml for IA2-Abs. MS was defined according to the International Diabetes Federation definition, arte- rial hypertension (AH) was considered as blood pressure > 130/85 mmHg or the use of antihypertensive drugs. Patients were divided into three groups: Gad Ab only positive (n=28), GAD Ab+ICA Ab positive (n=26) and GAD Ab+ICA Ab+IA2 Ab positive (n=15). Results: Twenty five (36.2%) patients were male, mean age approximately 51 years with disease duration 8 years. The lowest value of waist circumference (80 vs 82 vs 89.5 cm), MS (1 vs 11 vs 15) and AH (1 vs 11 vs 16) prevalence was found in the group positive for all three antibodies compared the oth- ers two groups. In the multinominal multivariate logistic regression model higher waist circumference (OR 0.931 (0.869-0.988)), MS prevalence (OR 0.062 ( 0.007-0.537)) and AH prevalence (OR 0.065 (0.007-0.577)) were neg- atively associated with triple Ab positivity compared to single Ab positivity, and compared to double Ab positive group with ORs as follows: 0.940 (0.877- 0.990), 0.097 (0.011-0.855), and 0.099 (0.011-0.879). Conclusion: Our results highlight the importance of the inverse association of simultaneous Abs positivity for ICA, GAD and IA2 with the presence of MS and its components in LADA patients. This inverse relationship might implicate the presence of LADA patients’ phenotype closer to type 1 diabetes. 441 Autoimmune gastrointestinal markers in patients with type 1 diabetes O. Derevyanko, T. Nikonova, N. Dalantaeva, O. Ivanova, O. Smirnova; Endocrinology Research Centre, Moscow, Russian Federation. Background and aims: In patients with type 1 diabetes the prevalence of other autoimmune diseases is 3 to 5 fold higher than in non-diabetic pa- tients. Gastrointestinal autoimmune diseases are generally asymptomatic up to an advanced stage when they become major causes of mortality if not di- agnosed and treated on time. The presence of immunological markers may be indicative of gastrointestinal autoimmune diseases even in a clinically asymptomatic case. The aim of our study was to investigate the frequency of specific autoantibodies occurrence in patients with T1 DM, to clarify the role of polyautoimmunity in the appearance of autoantibodies and the associa- tion of autoantibody titer with clinical features of accompaning autoimmune diseases. Materials and methods: 84 patients with T1DM and 21 healthy subjects (control group) were included in the study. T1DM patients were divided into two groups. The first group included 58 patients with T1DM only. The second group consisted of 26 patients with T1DM and another autoimmune disease (thyroid, celiac and Addison‘s disease, rheumatoid arthritis, vitiligo, myas- thenia gravis). Anti-gastric parietal cells antibodies (AGPA) - markers of au- toimmune gastritis, antinuclear antibodies (ANA), and anti-smooth muscle antibodies (ASMA) -markers of autoimmune hepatitis, antimitochondrial antibodies (AMA) - primary biliary cirrhosis marker were detected by using indirect immunofluorescence antibody test. All subjects were tested for ICA, GADA, IA-2A, C-peptide, HbA1c. Gastroscopy and ultrasound liver exami- nation were performed to every patient. Results: Significant difference was observed in frequency of ANA, AGPA, ASMA occurrence between the study groups and control group, p S 187 1 C portions of natural killer (NK) cells (P S 188 1 C PS 021 Measuring and preserving insulin secretion in type 1 diabetes 445 111In-exendin uptake in the pancreas correlates with the beta cell mass but not with the alpha cell mass M. Brom, C. Frielink, L. Joosten, M. Gotthardt; Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands. Background and aims: A method to non-invasively determine the beta cell mass in vivo would enable us to study the pathophysiology of diabetes and is still a major unmet need. Targeting of the GLP-1R with 111In-labeled exendin is an attractive approach for determination of the beta cell mass and preclini- cal studies as well as a proof-of-concept study in type 1 diabetic patients and healthy subjects showed a direct correlation of beta cell mass and radiotracer uptake. Despite these highly promising initial results, the influence of alpha cells on the uptake of the radiotracer remains a matter of debate. In this study we examined the specificity of 111In-exendin in a rat model for beta cell loss by comparing the uptake of the tracer with the beta and alpha cell mass. Materials and methods: Brown Norway rats were treated with 45 or 60 mg/ kg alloxan (n=4 per group) in order to destroy the beta cells and 4 rats were injected with vehicle as a control. One week after injection of alloxan, 15 MBq 111In-exendin (corresponding to 0.1 µg exendin) was injected and the pan- creas was dissected 1 h after injection of 111In-exendin. The radioactivity con- centration was measured in a gamma counter and the beta cell and alpha cell mass were determined by morphometric analysis after immunohistochemi- cal staining for insulin and glucagon, respectively. Results: The uptake of 111In-exendin (percentage of the injected dose) showed a strong positive linear correlation with the beta cell mass (Pearson r = 0.90). The absolute alpha cell mass was similar in healthy rats (2.2 ± 0.4 mg) and rats treated with 45 or 60 mg/kg alloxan (2.3 ± 0.3 mg and 1.8 ± 0.3 mg, respectively) and there was no significant correlation between the alpha cell mass and the uptake of 111In-exendin (Pearson r = 0.31). The total mass of the endocrine pancreas was reduced in the alloxan treated rats (8.5 ± 1.9 mg, 5.5 ± 1.7 mg, and 2.8 ±0.8 mg for healthy, 45 and 60 mg/kg alloxan, respectively) and percentage of glucagon positive cells of the total endocrine mass was in- creased after alloxan treatment (26% ± 4%, 43% ± 8%, and 69% ±21% alpha cells of the total endocrine mass for healthy, 45 and 60 mg/kg alloxan, respec- tively). The uptake of 111In-exendin showed a negative linear correlation with the alpha cell fraction (calculated by dividing the alpha cell mass by the total endocrine mass, Pearson r = -0.81). Conclusion: The uptake of 111In-exendin correlated with the beta cell mass, but not with the alpha cell mass. Together with the increased relative alpha cell mass and the negative linear correlation between the 111In-exendin up- take and alpha cell fraction, these data clearly indicate towards specificity of 111In-exendin to beta cells and that the influence of the alpha cells on 111In- exendin uptake is negligible. Supported by: NIH 1R01 AG 030328-01 and EC (FP7/2007-2013), BetaImage, grant n° 222980 446 Towards clinical PET imaging of pancreatic beta cells with [18F]exendin-4 K. Mikkola1, C.-B. Yim1, P. Lehtiniemi1, V.-V. Elomaa1, T. Ishizu1, J. Rajander2, O. Solin1,2, P. Nuutila1; 1Turku PET Centre, University of Turku and Turku University Central Hospital, 2Accelerator Laboratory, Åbo Akademi University, Turku, Finland. Background and aims: [18F]exendin-4 is a radioligand specific for the gluca- gon like peptide-1 receptor (GLP-1R) abundantly expressed in beta cells. Quantitative non-invasive PET (positron emission tomography) imaging of pancreatic beta cells would serve as a valuable diagnostic tool and provide means to monitor therapeutic interventions. Several exendin based tracers have been intensively explored for beta cell imaging. The obstacle with these has been high uptake in the kidneys. The aim of this study is to develop a novel [18F]exendin-4 tracer for clinical imaging of beta cells with PET. Materials and methods: Biodistribution and kinetics of [18F]exendin-4 was evaluated in Sprague-Dawley rats (N=3-9 per time point) weighing 250−300 g. After intravenous injection of [18F]exendin-4 (19 ± 3 MBq/kg, mass 0.3 ± 0.2 nmol/kg) rats were sacrificed at 15 min, 30 min, 1 h, 2 h, 4 h or 6 h. The organ-specific radioactivity was reported as a percentage of the injected dose per gram of tissue (%ID/g). The GLP-1R specificity was assessed using cold exendin-3 (N=1). Intrapancreatic distribution of radioactivity was assessed using autoradiography. Islet labelling was verified by immunohistochemis- try and islet-to-exocrine tissue ratios were analysed. Radioactive metabolites were determined by HPLC. For PET scans, rats were imaged up to six hours (dynamic 0-1 h, static 3.5-4 h and static 5.30-6 h) using Inveon Multimodal- ity PET/CT. Results: Radioactivity was nearly constant in the pancreas over the course of the study (0.18 ± 0.03 %ID/g at 1 h and 0.08 ± 0.03 %ID/g at 6 h p.i., Fig. 1A). Blocking studies indicated GLP-1R specific uptake in the islets. Autoradiog- raphy analysis of pancreatic sections showed that the islet-to-exocrine tissue ratio was 78 ± 29 at 1 h p.i. The amount of unchanged tracer in plasma was 28.3 ± 3.5 % at 1 h p.i. In line with other exendin based tracers, [18F]exendin-4 uptake by kidneys was high at 1 hour time point (16.7 ± 3.0 %ID/g), but thereafter its clearance was fast and retention decreased (1.5 ± 0.4 %ID/g, at 6 h p.i., Fig. 1B). After 6 h PET imaging, the highest tracer uptake was found in kidney, lung and stomach wall. Radioactivity in bone was low, indicating low defluorination of the tracer. Conclusion: In conclusion, we found a specific and sustained uptake of [18F] exendin-4 in the pancreatic islets and high renal clearance of the tracer. These indications are promising for the development of novel [18F]exendin-4 to- wards clinical imaging of beta cells. Currently we are investigating [18F]ex- endin-4 biodistribution and pancreatic uptake in larger animals. Figure 1. Biodistribution of [18F]exendin-4 radioactivity in rat at various time points after tracer injection. Supported by: Finnish Centre of Excellence and EC 7th framework programme BetaImage 447 111In-exendin imaging in patients with type 1 diabetes and healthy controls W. Woliner - van der Weg1, M. Brom1, M. Béhe2, C.J.J. Tack3, O.C. Boerman1, M.J.R. Janssen1, M. Gotthardt1; 1Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands, 2Paul Scherrer institute, Center for Radiopharmaceutical Sciences, Villingen-PSI, Switzerland, 3Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands. Background and aims: Currently the beta cell mass (BCM) can, in contrast to beta cell function, not be measured non-invasively. A non-invasive test would help to increase our knowledge about changes in beta cell mass during the development and treatment of diabetes. Preclinical studies showed that the pancreatic uptake of 111In-labeled exendin, which targets to the glucagon-like peptide-1 (GLP-1) receptor, directly correlates with BCM. As a first step into clinical use of 111In- exendin imaging for beta cell quantification, we acquired and analyzed single photon emission computed tomography (SPECT) images after 111In-exendin administration in T1D patients and healthy subjects. Materials and methods: Patients with long standing T1D (BMI below 27, age 21-60 yr, minimal 5 year duration of T1D) and healthy controls (normal glu- cose tolerance measured by OGTT, matched for age, gender and BMI) were eligible for the study. Four, 24 and 48 hours after i.v. injection of 150 MBq 111In-Exendin-4, SPECT images were acquired for quantitative assessment of the uptake into the pancreatic beta cells. The pancreas volume was measured by CT. Uptake calculation was based on the counts within two spherical vol- umes of interest (VOIs) placed in the head and corpus region, multiplied by the pancreas volume. Counts were corrected for administered activity and time after injection, leading to the pancreatic uptake in counts per MBq (c/ MBq). Results: To date, 10 T1D patients and 8 healthy volunteers were included. The T1D patients showed an average uptake of 451 c/MBq in the whole pancreas Diabetologia (2014) 57:[Suppl1]S1–S564 S 189 1 C (SD 426 c/MBq). The average uptake of the healthy controls was more than two and a half times higher; 1258 c/MBq per pancreas (SD 645 c/MBq). Both groups showed large interindividual differences: the uptake in T1D patients ranges from 81 to 1445 c/MBq per pancreas and in the healthy subjects from 264 to 1998 c/MBq per pancreas.(See figure 1) Also the concentration of 111In-Exendin in the pancreas (not corrected for pancreas size) is more than two times lower in the group with T1D. The images and subsequent quantita- tive analysis show that the 111In retention in the pancreas remains stable for at least 4 to 48 hours after injection in the pancreas. Conclusion: This clinical study indicates that 111In-Exendin imaging with SPECT could indeed be the first technology to enable non-invasive beta cell quantification. In line with literature, it visualizes considerable interindivid- ual differences in both groups and shows a, in general, much lower uptake of 111In-Exendin in T1D patients than in healthy volunteers. Therefore it could be a valuable tool for further elucidating the complex pathofysiology in dia- betes. Clinical Trial Registration Number: NCT01825148 Supported by: NIH 1R01 AG 030328-01 and EC (FP7/2007-2013), BetaImage, grant n° 222980 448 Glucagon and mixed-meal tests to estimate beta cell and incretin functions in normals: relevance for studies in type 1 diabetes C. Guglielmi1, L. Valente1, S. Fallucca1, S. Angeletti1, S. Briganti1, M. De Pascalis1, J.M. Lachin2, P. Pozzilli1; 1University Campus Bio Medico, Roma, Italy, 2The George Washington University, Rockville, USA. Background and aims: Measurement of C-peptide after the administration of different stimuli is the only instrument able to directly measure in vivo beta cell function and it is used as primary enpoint in several immunointervention trials. The two tests currently used are Glucagon Stimulation Test (GST) and Mixed Meal Tolerance Test (MMTT). Materials and methods: We carried out a study in 10 healthy subjects (25- 40 yrs) to evaluate beta cell function, Gastric Inhibitory Peptide (GIP) and Glucagon Like Peptide 1 (GLP1) which may affect C-peptide response to the two tests. GST and MMTT were carried out one week apart according to standard tests. Results: At the end of GST (20 min) stimulated C-peptide showed a mean increase from baseline of 147.1% while at the end of MMTT (120 min), the mean increase of C-peptide was equal to 99.8% (Δincrease= 47.2%) while Max C-peptide reached during MMTT was greater than that gained dur- ing GST (C-pept maxMMTT=2.3nmol/L vs C-pep maxGST=1.9nmol/L). A positive and linear correlation was found between the incremental AUC (iAUC) C-peptide in GST and iAUC C-peptide in MMTT (r=0.61, p=0.05). These data show that the two tests can be considered equivalent in assessing beta cell function although there are differences: in MMTT, but not in GST, the incretin response affects C-peptide levels. A positive and linear correla- tion between GIP and C-peptide levels was found during MMTT (r=0.92, p=0.008) and a positive and exponential correlation was found between GIP and insulin during MMTT (R2=0.82). This was not the case with GST. No correlation was found between GLP-1 and C-peptide levels during MMTT and between GIP or GLP1 levels and C-peptide during GST. Conclusion: Although the two stimulation tests may produce a similar re- sponse in C-peptide, they diverge in that the beta cell response to GST test is independent of the incretin axis whereas this is not the case with MMTT. This result indicates the two tests differ in their mechanisms of action of stimulat- ing C-peptide secretion. 449 Functioning beta cells in type 1 diabetes may not be as low as it is presumed G. Tamer1, B. Dogan2, I. Ocakoglu3, O. Kostek4, I. Kartal1, G. Sagun4, M. Adas5, I. Tamer6, H.H. Mutlu2, A. Orhun3; 1Division of Endocrinology and Metabolism, Department of Internal Medicine, 2Department of Family Medicine, 3Department of Biochemistry, 4Department of Internal Medicine, Medeniyet University, Goztepe Training and Research Hospital, 5Division of Endocrinology and Metabolism, Department of Internal Medicine, OkmeydanıTraining and Research Hospital, 6Department of Family Medicine, Dr Lutfi Kirdar Kartal Training and Research Hospital, Istanbul, Turkey. Background and aims: Enhancing endogenous insulin production in type 1 diabetic patients(T1DP) can improve glycemic control and decrease compli- cations and rates of mortality. However, it can be succesfull even if sufficient β-cell function is present. We aimed to evaluate the extent of β-cell function by determining fasting levels of C-peptide and those after meal stimulus. Materials and methods: Sixtythree T1DP were enrolled. Ethics comittee of our hospital approved the study protocol, which was in accordance with the Helsinki Declaration. Fasting C-peptide levels of all participants and stimu- lated (at 90 th minute post mixed meal) C-peptide levels of 54 were measured by using an electrochemiluminescence assay. Two categorizations were done using fasting (the first categorization ) and at 90th minute post mixed meal test (the second categorization) of C-peptide levels. For the first categoriza- tion; the groups were classified as follows: patients with undetectable ≤0.1 ng/ mL (group 1); with minimal 0.1-0.8 ng/ml (group 2); and with sustained ≥0.8 ng/mL(group 3) C-peptide levels. For the second categorization, groups were as follows: patients with undetectable ≤0.1 ng/mL (group 1); with minimal 0.1-0.8 ng/ml (group 2); and with sustained ≥0.8 ng/mL (group 3) C-peptide levels which increased at the 90th minute after the meal ≥150% of fasting C- peptide level. Results: For the first category; 25.4%, 49.2%, 25.4% of T1DP were in group 1, group 2 and group 3, respectively. For the second category; 22.2% of T1DP were in group 1, 24.1% of them in group 2 and 53.7 % of them in group 3. For the first categorization 74% and for the second categorization 77.8% of T1DP had detectable C-peptide levels. More than 50% of T1DP had a re- sponse to meal stimulus with C-peptide levels ≥0.8 ng/mL which increased up to ≥150% of fasting C-peptide levels like non-diabetics. Conclusion: The finding that presence of insulin secretion which increases after meals, suggests the possibility of undergoing β-cell regeneration within many T1DP. 450 Residual beta cell function in long-standing childhood onset type 1 diabetes P.J. Bingley1, R.J. Aitken1, I.V. Wilson1, A.E. Long1, A.J.K. Williams1, T. McDonald2, S. Wong3, A.T. Hattersley2, K.M. Gillespie1; 1School of Clinical Sciences, University of Bristol, 2University of Exeter Medical School, 3Institute of Molecular and Experimental Medicine, Cardiff University School of Medicine, UK. Background and aims: Some people with type 1 diabetes have functioning beta cells years after diagnosis, but the reason is unclear. We aimed to de- termine the frequency of residual C-peptide secretion in people with long- standing type 1 diabetes recruited to the Bart’s-Oxford study at diagnosis and relate this to current and baseline islet autoantibodies. Materials and methods: We measured two-hour post-meal urine C- peptide:creatinine ratio (UCPCR) and autoantibodies to glutamate decar- boxylase (GADA) and islet antigen-2 (IA2A) in samples collected from 140 patients (median age at diagnosis 11.9 years, 50% male) a median of 22 years Diabetologia (2014) 57:[Suppl1]S1–S564 S 190 1 C (range 12.2-28 years) after diagnosis. Baseline status for GADA, IA2A and zinc transporter 8 autoantibodies was determined in samples collected within 24 months of diagnosis (median 33 days). UCPCR thresholds equivalent to mixed meal-stimulated serum C-peptide ≥0.2 nmol/L and ≥0.03nmol/L were used to define ‘preserved’ and ‘minimal’ endogenous insulin secretion. HLA class II genotype was established by PCR-SSP. Associations were examined by chi-squared and non-parametric testing. Results: Of the 140 patients, 23 (16.4%) still had detectable endogenous in- sulin secretion; seven (5.0%) ‘preserved’ and 16 (11.4%) ‘minimal’. 62% of participants had at least one antibody in current samples; 29% were GADA+ and 50% IA2A+. Persistent C-peptide secretion was inversely related to age at diagnosis (p=0.0009), and no one diagnosed before age 10 had ‚preserved‘ secretion. UCPCR was independent of diabetes duration and baseline or current autoantibody status. Of 7 individuals with ‘preserved’ endogenous insulin secretion, 6 had GADA at diagnosis with at least one diabetes risk- associated HLA class II haplotype. Conclusion: A subset of patients with proven autoimmune-mediated type 1 diabetes has endogenous insulin secretion many years after diagnosis, but this is rarely found in individuals diagnosed in early childhood. Ongoing islet autoimmunity is also common. We hypothesise that, while aggressive early onset autoimmunity results in complete beta cell destruction, the less aggressive autoimmune process associated with later onset type 1 diabetes allows residual beta cells to provide a focus for persistent function as a result of regulated autoimmunity and/or beta cell renewal. Supported by: JDRF and Diabetes UK 451 High GADA titer significantly increases the risk of insulin requirement in LADA: a 7-year follow-up (NIRAD Study 7) S. Zampetti1, G. Campagna1, C. Tiberti1, M. Songini3, M. Arpi4, G. De Simone5, E. Cossu6, L. Cocco7, G. Leto2, E. Bosi8, F. Giorgino9, M. Spoletini2, R. Buzzetti2, NIRAD Study Group; 1Sapienza University, Rome, 3San Michele Hospital, Cagliari, 4Catania University, 5ASL Napoli, 6University of Cagliari, 7Cardarelli Hospital, Campobasso, 8San Raffaele Institute, Milan, 9University of Bari, Italy. Background and aims: The aim of the study was to determine whether glu- tamic acid decarboxylase antibody (GADA) titer and other clinical param- eters could define the risk of progression to insulin therapy in LADA patients during a 7-years follow-up. Materials and methods: N= 222 LADA and n=430 type 2 diabetes subjects were followed for 7-years from the time of GADA screening to evaluate their progression towards insulin therapy. Kaplan-Meier curves and multivariate logistic regression analysis were performed to identify markers able to influ- ence this progression. Results: During the follow-up, drop out was 4% in both groups. N=119 (56.1%) out of n=212 LADA and n=86 (20.9%) out of n=412 type 2 diabe- tes required insulin. Kaplan-Meier plots showed that 74/104 (71.1%) of high GADA titer required insulin compared to 45/108 (41.6%) of low GADA titer and to 86/412 (20.9%) of type 2 diabetes (p S 191 1 C a higher age (p S 192 1 C oxidative stress. Peroxiredoxin 4 (Prdx4), an ER-specific antioxidative per- oxidase can utilize luminal H2O2 as driving force for reoxidizing PDI family members, thus efficiently contributing to disulfide bond formation. Therefore the aim of this study was to examine the functional significance of Prdx4 on β-cell function with emphasis on insulin content and secretion during stimu- lation with nutrient secretagogues. Materials and methods: The ER-specific Prdx4 was specifically overex- pressed in insulin-producing INS-1E cells. The Prdx4 overexpression was verified by Western blot, while its antioxidative effect was assessed by DCF in the presence of 5 mM dithiothreitol (DTT). Insulin content and secre- tion were determined by RIA and proinsulin transcription by qRT-PCR after stimulation with glucose (3, 10, 30 mM) and a combination of leucine (10 mM) plus glutamine (2 mM). Results: Immunoblotting revealed that INS-1E cells stably transduced with a lentiviral construct carrying Prdx4 exhibited a significant increase in the Prdx4 expression compared to control cells. Exposure of control cells to DTT resulted in a significant generation of reactive oxygen species (ROS), whereas Prdx4 overexpression completely prevented the DTT-mediated ROS gen- eration. Overexpression of Prdx4 led to an improved insulin secretion after stimulation with 10 and 30 mM glucose compared to control cells, while the basal insulin secretion at 3 mM glucose was not affected. An augmented insu- lin secretion could also be observed in Prdx4 overexpressing cells incubated with leucine plus glutamine. In addition, Prdx4 overexpressing cells exhibited an enhanced proinsulin mRNA transcription and insulin content when com- pared to control cells. Conclusion: These data strongly suggest that enhancing ER-specific perox- iredoxin 4 expression in glucose-responsive insulin-secreting INS-1E cells significantly metabolized the DTT- mediated H2O2 generation within the ER and improved the glucose-induced insulin-secretion, which was accom- panied by the enhancing proinsulin gene transcription and insulin content. This beta cell beneficial effect was also observed upon stimulation with leu- cine, another nutrient insulin secretagogue, indicating that the effect is not restricted to glucose. Thus, Prdx4 improves the ER folding capacity and could contribute to the preservation of beta cell function under conditions of high insulin requirement. 456 Targeted overexpression of glycosylation-negative catalase mutations in the endoplasmic reticulum of insulin-producing cells S. Lortz, I. Mehmeti, S. Lenzen; Institute of Clinical Biochemistry, Hannover Medical School, Germany. Background and aims: In reaction to peripheral insulin resistance in type 2 diabetes mellitus, β-cells show initially an increased insulin secretion. Es- calated insulin biosynthesis and secretion require an enhanced capacity for protein folding and disulfide bond formation, resulting in an increased pro- duction of H2O2 in the endoplasmic reticulum (ER). A direct quantification is not possible due to the oxidative environment in the ER and catalase over- expression for compensating an elevated H2O2 production failed due to N- glycosylation of the catalase protein. Therefore, a glycosylation-negative ER- targeted catalase variant, designed by mutagenesis, should be overexpressed to reduce the H2O2 concentration in the ER of insulin-secreting cells and to proof its potential impact on insulin secretion and ER stress. Materials and methods: The two potential N-glycosylation sites of the hu- man catalase protein were eliminated by PCR-mediated mutagenesis. The generated catalase variants were peroxisomally overexpressed to confirm that catalase enzyme activity was unaffected by the mutations, before the ER- specific expression was carried out in RINm5F and INS-1E cells. Catalase functionality in the ER and prevention of N-glycosylation was examined by enzyme activity measurement, incubation with H2O2 and Western blot analy- sis. Glucose-induced insulin secretion was quantified by RIA and typical ER stress markers were quantified by qRT-PCR analysis. Results: Despite of the mutation of the N-glycosylation motives at aspara- gine-244 and -439, catalase activity was not affected by both single mutants (N244: 938±114, N439: 972±101, untransfected: 16±4 U/µg protein). After overexpression of both variants in the ER only the ER-catalase-N244 showed enzyme activity and lack of glycosylation, whereas ER-catalase-N439 was still glycosylated and only a slightly elevated catalase activity could be detected (N244: 935±51, N439: 74±10 U/µg protein). Quantification of catalase en- zyme activity of the double mutant N244/439 after peroxisomal and ER-tar- geted overexpression revealed lower catalase activity in both compartments compared with N244 (peroxisome: 278 ± 17, ER: 168 ± 12 U/µg protein). Measurement of glucose-induced insulin secretion and of ER stress-induced genes showed no significant changes compared with untransfected control cells. Conclusion: With the overexpression of ER-catalase-N244 a highly effective H2O2 inactivation within the ER could be achieved for the first time. Double mutation of both N-glycosylation motives N-244 and -439 was associated with partial loss of function, possibly due to the destruction of the three- dimensional enzyme structure. Since catalase has a high H2O2 inactivation capacity and is not involved in the protein folding process, the enzyme is an ideal tool for the investigation of insulin biosynthesis-associated oxidative ER stress, postulated for type 2 diabetes, or oxidative stress induced by misfolded protein aggregation in the ER. However, increased H2O2 inactivation through overexpression of ER-Catalase-N244 had no influence on insulin biosynthe- sis and the ER-stress response. 457 ATF6β regulates the Wfs1 gene and has a cell survival role in the ER stress response in pancreatic beta cells A. Volchuk, T. Odisho, L. Zhang; Cellular & Molecular Biology, Toronto General Research Institute, Canada. Background and aims: Endoplasmic reticulum (ER) stress has been impli- cated in the development of pancreatic β-cell dysfunction and death resulting in type 2 diabetes. The Unfolded Protein Response (UPR) is the cellular sys- tem that responds to ER stress and consists of three main ER stress sensors, PERK, IRE1 and ATF6. Activating transcription factor 6 (ATF6) is an essen- tial component of the UPR in cells undergoing ER stress that consists of two distinct genes, ATF6α and ATF6β. The ATF6β isoform has been less studied and its role in the UPR in β-cells is unclear. Materials and methods: Rodent islets and insulinoma cell lines were used to examine ATF6β function. Microarray and qPCR validation was employed to identify ATF6β target genes and siRNA knock-down and adenoviral overex- pression was used to assess the role of ATF6β in ER stress-induced apoptosis. Results: ATF6β mRNA and protein were detected in pancreatic β-cell lines and rodent and human islets and the protein was proteolyzed to the nuclear active form (ATF6βp60) in response to pharmacological ER stress. Knock- down of ATF6β using siRNA in INS-1 832/13 insulinoma cells did not af- fect mRNA induction of several major ER stress response genes in response to tunicamycin-induced ER stress, suggesting ATF6β is not essential for the basic UPR. Expressing active ATF6βp60 and ATF6αp50 followed by micro- array analysis revealed that ATF6β and ATF6α regulate similar UPR genes, including chaperones and ERAD components, although some genes such as Wfs1 are ATF6β-specific. Interestingly, knockdown of ATF6β increased the susceptibility of INS-1 832/13 β-cells to apoptosis under both control and ER stress conditions, while overexpression of active ATF6βp60 reduced ap- optosis. Conclusion: ATF6β is not essential for induction of major ER stress response genes, but is required to maintain cell survival in β-cells undergoing chronic ER stress. The ATF6β pro-survival role may relate to induction of genes that promote cell survival such as Wfs1. Supported by: CIHR 458 The novel translation initiation factor, eIF2A, is up-regulated by endoplasmic reticulum stress and protects beta cells from apoptosis E. Panzhinskiy1, F. Taghizadeh1, K.-Y. Chu1, Y. Yang1, E. Jan2, J.D. Johnson1; 1Cellular and Physiological Sciences, 2Biochemistry, University of British Columbia, Vancouver, Canada. Background and aims: Endoplasmic reticulum (ER) stress is an important mediator of pancreatic beta cell loss in diabetes. Recent reports have associat- ed inhibition of protein synthesis by the ER stress-induced unfolded protein response (UPR) with apoptosis. ER-stress associated translation reprogram- ming increase translation of stress-response mRNAs, but the signaling path- ways that modulate this process in beta cells are not well studied. Recently, the novel translation initiation factors eIF2A and eIF2D have been implicated in the translation of specific mRNAs under stress conditions in other systems. We investigated the role of eIF2A and eIF2D in the regulation of UPR gene expression and ER-stress-induced apoptosis in beta cells. Materials and methods: Thapsigargin (Tg, 1 μmol/l), palmitate (PA, 0.5 mmol/l in BSA) were used to induce ER stress in primary mouse islets and mouse insulinoma cells (MIN6). Real-time RT-PCR, Western Blot and im- Diabetologia (2014) 57:[Suppl1]S1–S564 S 193 1 C munocytochemistry were used to assess gene expression. Apoptosis was as- sessed by flow cytometry of propidium iodide (PI) stained cells. Differences between means were considered statistically significant when p was < 0.05. Results: We found that eIF2A and eIF2D were expressed in mouse islets at both mRNA and protein levels. We compared eIF2A protein expression levels between different mouse tissues (heart, brain, liver, muscle, spleen, pancreatic islets) and found that they were highest in the pancreatic islets. eIF2D protein levels did not significantly differ between islets and other tissues. Fluorescent microscopy revealed that eIF2A and eIF2D were mainly localized in the cy- toplasm of primary beta cells. Using MIN6 cells, we found gradual induction of eIF2A protein expression over the time course of Tg-induced ER stress, with the maximal fold increase of 1.78 ± 0.18 (n = 4) after 24 h, which coin- cided with activation of pro-apoptotic markers C/EBP homologous protein (CHOP, 137 ± 24 fold, n = 4) and cleaved caspase-3 (234 ± 11 fold, n = 4). We also observed a 1.66 ± 0.17 fold increase (n = 4) in eIF2D protein levels after 6 h of Tg-treatment that coincided with maximum Activating Transcription Factor-4 protein levels (8.7 ± 1.3 fold increase, n = 4). However, the effect was transient and after 24 h of Tg treatment, eIF2D protein levels were not different from the basal level. Treatment of MIN6 cells with PA for 24 h also significantly increased protein levels of eIF2A (2.1 ± 0.5 fold, n = 3) with no significant differences in eIF2D levels observed. Isolated mouse islets treated with either Tg or PA for 24 h showed elevated levels of eIF2A protein (by 1.35 ± 0.5 and 1.43 ± 0.2 fold respectively, n = 3) when compared to basal conditions, but not eIF2D. Importantly, we demonstrated that overexpression of recombinant eIF2A (10-fold) in MIN6 cells reduced thapsigargin-induced apoptosis by 51 ± 15% (n = 3). We found that decreased death of beta cells overexpressing eIF2A after 24 h of thapsigargin treatment was accompanied by decreased expression of UPR pro-apoptotic marker CHOP by 45 ± 13% (n = 3). Conclusion: We conclude that eIF2A and eIF2D are expressed and differ- entially regulated over the course of UPR in beta cells. Also we identified a novel protective role for eIF2A in the context of ER-stressed beta cells via the inhibition of CHOP. Thus, eIF2A may potentially serve as a new therapeutic target in diabetes. Supported by: CDA 459 The balance between adaptive and apoptotic unfolded protein responses regulates beta cell death through JNK, XBP1 and CHOP J.Y. Chan, J. Luzuriaga, D.R. Laybutt; Garvan Institute of Medical Research, Sydney, Australia. Background and aims: The loss of β-cell mass due to increased apoptosis is critical in both type 1 and type 2 diabetes. Pro-inflammatory cytokines and saturated fatty acids are potential mediators of β-cell apoptosis, although the mechanisms are poorly understood. Endoplasmic reticulum (ER) stress is present in β-cells in both type 1 and type 2 diabetes. ER stress is resolved through adaptive actions of the unfolded protein response (UPR). However, β-cells are prone to failure of the adaptive UPR with consequent activation of pro-apoptotic UPR, but the mechanisms are unclear. Here, we investigated the role of key ER stress-responsive transcription factors, XBP1 and CHOP, in the UPR induced by cytokines (IL-1β, TNF-α and IFN-γ) or the saturated fatty acid, palmitate. We also examined the influence of Jun N-terminal ki- nase (JNK) activity in the UPR and the significance of these responses in β-cell survival. Materials and methods: MIN6 β-cells and mouse islets were exposed for 24- 48 h to the combination of pro-inflammatory cytokines, IL-1β (100 U/ml), TNF-α (100 U/ml), IFN-γ (250 U/ml) or the saturated fatty acid, palmitate (0.4 mM coupled to 0.92% BSA). siRNA was used to silence XBP1 and CHOP in MIN6 cells. IRE1/XBP1 inhibitor (4μ8c - 30 μM) was used in mouse is- lets. SP600125 (20 μM) was used to inhibit JNK activity. β-cell death was measured using a cell death detection ELISA. Gene and protein expression changes were assessed by real-time PCR and Western blot. Results: In MIN6 cells, exposure to cytokines or palmitate resulted in in- creased cell death and ER stress, which featured marked increases in pro- apoptotic UPR markers, CHOP, Atf3 and Trib3. siRNA-mediated inhibition of XBP1 reduced the expression of adaptive UPR markers, including ER chaperones and foldases (BiP, Grp94, Fkbp11, Erp72, Edem1) and potenti- ated cytokine- and palmitate-induced apoptosis (p S 194 1 C of the Bcl-2 gene family members, caspase-9, HSTF1, P38, pP38, STAT-1 and pSTAT-1were assessed by Western blotting. Caspase 3 and 8 activities were studied by fluorimetric assays. Results: Our data showed that upon cytokines and rhPRL treatment, the pro- portion of fragmented nuclei was increased in HSPB1 silenced cells (p S 195 1 C PS 023 Mechanisms of lipotoxicity 463 Inhibition of de novo lipogenesis by lysosomal acid lipase in beta cells helps counter-regulate insulin secretion G.L. Pearson1, P.J. Meikle2, N. Mellett2, H. Du3, T.J. Biden1; 1Garvan Institute of Medical Research, Darlinghurst, 2Baker IDI, Melbourne, Australia, 3School of Medicine, Indiana University, Indianapolis, USA. Background and aims: Endogenous lipolysis is hypothesised to contribute to the amplification phase of glucose-stimulated insulin secretion (GSIS) in pancreatic β-cells, mediated acutely by neutral lipases. Recent work shows that lysosomal acid lipase (LAL), an acidic lipase responsible for the break- down of lipid delivered to the lysosome, is a negative regulator of GSIS. How- ever, the exact mechanism of action of LAL and its role in whole body glucose homeostasis are unclear. Our current aims were to: (1) investigate glucose homeostasis in global LAL knock-out (KO) mice and (2) further characterise its mechanism of action. Materials and methods: Wild-type (WT) and LAL KO mice were subjected to intraperitoneal glucose-tolerance tests (i.p. GTT). GSIS was measured by insulin RIA in islets isolated from WT and LAL KO mice. After 24 h LAL in- hibition with Lalistat (5μmol/l) in MIN6 cells, gene expression was analysed by RT-PCR and lipid mass was assessed using mass spectrometry. Results: LAL KO mice had significantly lower glucose excursions during an i.p.GTT compared to WT mice (AUC mean±SEM: WT 1514.3±104.8 n=10; LAL KO 1087.0±95.2 n=7; P S 196 1 C newly synthesized Cer for SM biosynthesis. Moreover, these results support a role of Cer transport between ER and Golgi apparatus in the regulation of beta cell death induced by gluco-lipotoxicity. 466 Palmitate and glucose induce autophagy in INS(832/13) cells A.S. Medina, H. Bennet, M. Fex; Diabetes and Celiac Disease, Lund University, Malmö, Sweden. Background and aims: Autophagy is an important biological process by which proteins and organelles are sequestered in autophagosomal membrane vesicles and delivered to the lysosome for degradation. Eukaryotic cells can degrade long-lived proteins and intracellular organelles through macro- autophagy, for instance during periods of an increased energy demand. In addition, autophagy may function as a first line cellular defense to protect cells against various stressors and pathogens. The ATG genes, are crucial in regulating the formation of autophagy membranes, and dysfunctional regula- tion/expression of these may be involved in disease processes. The role of au- tophagy and/or regulation of ATG genes in type 2 diabetes are incompletely understood, although some evidence points to the involvement of autophagy in this disease. Therefore, the aim of this study was to investigate the exact cellular stressors that induce autophagy and ATG gene expression in beta cells in-vitro. Materials and methods: The INS(832/13) beta cell line was cultured in com- bination with 30mM glucose, 0.5mM palmitate, or a combination of both, cytokines, LPS and tamoxifen (as a positive control) and untreated cells as a negative control. Incubations were performed for 6, 12, 24, and 48 hours to investigate the degree of autophagy visualized by fluorescent LC3B stain- ing and confocal microscopy. Quantitative (Q)-PCR for ATG genes was per- formed in cells treated under each condition and insulin secretion analysis were performed following standard protocols. Results: LC3B expression was used as the classical readout for autophagy. A preliminary time-course experiment evaluating INS(832/13) cells after 6, 12, 24, and 48 hours exposure to 30mM glucose and 0.5mM palmitate and a combination of both (glucolipotoxicity) indicated that autophagy response was maximal between 12 and 24 hours, and more prominent in cells treated with palmitate alone as compared to glucose treatment alone. The combina- tion of both glucose and palmitate did not induce more autophagy than pal- mitate treatment, which was evaluated by counting the LCB3 positive dots in each sample. The LC3B dot-like structures appear in the cytoplasm and peri- nuclear region and were more evident in all treated groups when compared to non-treated cells. An increase in expression was observed for ATG5, and ATG12 after 24 h incubation with palmitate. Insulin secretion was measured in cells treated in these conditions and preliminary data suggest a decreased in response to treatment with both palmitate and glucose. This indicates that the autophagic response induced by glucolipotoxicity inhibits insulin release, perhaps by mechanism relating to changes in autophagic response. Conclusion: Our findings in vitro in INS(832/13) beta cells indicate that increased palmitate levels represent a powerful activator for the autophagy process which is coupled to an increase of specific ATG genes, necessary for the formation of the autophagosome. These changes are accompanied by a reduction in glucose stimulated insulin release. Supported by: The Childhood Diabetes Foundation, Swedish Research Council 467 L-type voltage-gated calcium channels mediate sensitivity to glucolipotoxicity in beta cells via activation of nuclear receptors Nr4a and inhibition of autophagy P.J. Buda1, E. Zhang1, J. Esguerra2, E. Renstrom1; 1Islets pathophysiology, 2Islet cells exocytosis, Lund University, Malmö, Sweden. Background and aims: Glucolipotoxicity is considered a major factor driving the gradual demise of pancreatic beta-cells mass in type 2-diabetes. Coun- teracting this, autophagy allows cells to survive under adverse conditions. Recent reports underscore the protective role of autophagy in animal models exposed to free-fatty acids and high concentrations of glucose. However, little is known about the involvement of autophagy in diabetes progression and its regulation in pancreatic B-cells is poorly understood. Under high glucose and palmitate conditions voltage-gated calcium channels (VGCC) become activated and in turn stimulate extracellular calcium-dependent gene expres- sion. In this study we investigated which VGCC-activated genes are essen- tial for autophagy regulation and whether deliberate tuning of expression of these genes under glucolipotoxic conditions could improve beta-cell viability through the activation of autophagy. Materials and methods: Genes activated by VGCC were revealed by use of affymetrix microarray performed on INS-1 cells after stimulation with 70mM potassium chloride. Amongst them autophagy related genes were identified based on the literature studies. Expression of these genes was measured by quantitative PCR and western blot. Their relation to the type of VGCC was obtained owing to specific VGCC inhibitors. Autophagy was assessed by use of western blot where the ratio of LC3B II over I form was calculated, and as well as by use of confocal microscopy on INS-1 cells wherein the number of LC3B-GFP puncta was counted. Apoptosis was measured by use of both flow cytometry (the number of AnnexinV and ViaProbe positive cells) and ELISA assay measuring enrichment of nucleosomes in cytosol. Results: Affymetrix microarray revealed that activation of VGCC with 70mM potassium chloride upregulated autophagy related nuclear receptors Nr4a: NUR77 and NOR-1. Quantitative PCR showed 10-times increase of NUR77 and NOR-1 genes under high potassium conditions. At the same time acti- vation of VGCC inhibited autophagy, which was indicated by 20% decrease in LC3B II/I ratio and 50% decrease in the number of LC3B-GFP puncta in beta-cells. This coincided with 50% reduction of AMPK phosporylation - the main autophagy activator. Inhibition of VGCC-activated NUR77 and NOR- 1 expression by isradipine or downregulation of both genes with siRNA re- stored AMPK activity and ablated inhibitory effect of high potassium on au- tophagy. 25 mM glucose with 0,5 mM palmitate inhibited autophagy which was indicated by 70% decrease in LC3B-GFP puncta and increased 10 times the number of AnnexinV and ViaProbe positive cells as compared to low glu- cose conditions. Isradipine or autophagy inducing rapamycin reduced apop- tosis by 20% under glucolipotoxic conditions. Finally, cells depleted of both NUR77 and NOR-1 gained 25% improvement in cell viability at the presence of high glucose and palmitate. Conclusion: Our study shows that inhibition of autophagy by L-type VGCC makes beta-cells more vulnerable to apoptosis under glucolipotoxic condi- tions. Beneficial for cell viability activation of autophagy can be obtained by inhibition of L-type calcium currents as it prevents induction of anti-au- tophagic - NUR77 and NOR-1. Supported by: SRC 468 Lipotoxicity alters the genome-wide epigenetic pattern in human pancreatic islets C.A. Ling1, E. Hall1, P. Volkov1, T. Dayeh1, K. Bacos1, T. Rönn1, M. Dekker Nitert2; 1Clinical Sciences, Lund University, Malmö, Sweden, 2The University of Queensland, Brisbane, Australia. Background and aims: Circulating levels of free fatty acids are often in- creased in subjects with type 2 diabetes (T2D). Long-term exposure to lipids has harmful effects on islet function and insulin secretion. Epigenetic modi- fications such as DNA methylation may contribute to T2D. However, there is limited information on whether fatty acids alter the epigenetic pattern in human pancreatic islets. Our aim was therefore to analyse the genome-wide DNA methylation pattern in human pancreatic islets exposed to palmitate for 48 hours and relate methylation to gene expression and insulin secretion in the islets. Materials and methods: mRNA expression and DNA methylation were ana- lysed genome-wide in human islets using microarrays. Results: Palmitate treatment for 48 hours decreased glucose-stimulated insu- lin secretion but did not affect apoptosis in the human islets. We found 1860 genes with differential expression in palmitate-treated human islets. These include candidate genes for T2D such as GLIS3, HNF1B and SLC30A8. Ad- ditionally, palmitate altered the expression of genes in glycolysis/gluconeo- genesis, pyruvate metabolism, fatty acid metabolism, glutathione metabolism in human islets. The global DNA methylation level and DNA methylation levels of CpG island shelves and shores, 5’UTR, 3’UTR and gene body regions were altered in human islets exposed to palmitate. Moreover, 290 genes with differential expression had a corresponding change in DNA methylation e.g. several candidate genes for T2D. Importantly, 67 of these genes were also associated with BMI and 37 were differentially expressed in islets from T2D patients. Diabetologia (2014) 57:[Suppl1]S1–S564 S 197 1 C Conclusion: We demonstrate that lipotoxicity gives rise to epigenetic modi- fications as well as transcriptional changes in human pancreatic islets. These changes may contribute to impaired insulin secretion and T2D. Supported by: Swedish Research council, ALF, Påhlsson, Novo Nordisk founda- tion 469 EPA and DHA protect pancreatic islets against palmitate toxicity C.F. Lucena, L.P. Roma, K. Veras, M. Graciano, A.R. Carpinelli; Physiology and Biophysics, University of São Paulo, Brazil. Background and aims: In previous studies, we have shown that fish oil supplementation improve the antioxidant defense in pancreatic islets from healthy rats. To test whether these effects were due to ω-3 fatty acids present in high concentrations in the supplemented diet, we decided to test whether EPA and DHA have the same effects in vitro and protect pancreatic beta cells from palmitate toxicity. Materials and methods: Pancreatic islets were obtained by collagenase diges- tion and cultured with RPMI 1640 medium containing 10mM glucose, 10% FBS, penicillin and streptomycin. After overnight culture, islets were divided in four groups: control (vehicle ethanol), ω-3 fatty acids (50μM EPA + 50μM DHA), 100 µM Palmitate (P), Pω3 (50 μM EPA + 50 μM DHA + 100µM palmitate). Islet cell death, glucose stimulated insulin secretion (GSIS) and superoxide (ROS) content (cytosolic and mitochondrial) were analyzed after 48 h culture. Some of the experiments were performed using dispersed cells. Briefly, after 48 h culture, islets were dispersed into small cell clusters using trypsin and gentle pipetting in a Ca2+-free medium. Cells were then analyzed by flow cytometry. Results: 48h culture in the presence of P increased beta cell ROS production and apoptosis and impaired GSIS. Under these conditions, EPA+DHA trig- gered a parallel ~60-65 % reduction in ROS production and beta-cell apopto- sis induced by P, and induced significant protection against the impairment in GSIS (~1,28ng insulin in P vs ~1,60ng insulin in Pω3, per islets/hour). The reduction in insulin secretion was not due to decreased insulin content. When dispersed cells were analyzed by flow cytometry, there was no differ- ence in mitochondrial ROS production (mitosox probe). On the other hand, ROS production was increased with P and reduced to control levels when P cells were cultured in the presence of EPA and DHA, confirming the results obtained with whole islets using confocal microscopy. Apoptosis was also in- creased in P vs control and the addition of ω-3 is not significantly different from control. Conclusion: Our results show that EPA+DHA protect pancreatic islets from alterations induced by palmitate, i.e. increased superoxide production, de- creased viability and beta cell function. The effects may result from decreased cytosolic superoxide production, which in turn can lead to increased viability and beta cell function. As we previously reported, in vivo ω-3 supplementa- tion decrease protein levels of NAD(P)H oxidase subunits and consequently superoxide production in pancreatic islets. Considering that in vitro effects do not involve mitochondrial superoxide production, we can speculate that this protection might involve downregulation of NAD(P)H oxidase. How- ever, further experiments are needed to elucidate the mechanisms involved in the protection of ω3 fatty acids against palmitate toxicity. Supported by: Fapesp, CNPq, Capes 470 Different regulation of beta cell proliferation induced by short-term and long-term high-fat diet loading A. Nakamura1, N. Kitao1, H. Nomoto1, H. Miyoshi1, Y. Terauchi2, T. Atsumi1; 1Division of Immunology and Metabolism, Hokkaido University, Sapporo, 2Yokohama City University, Japan. Background and aims: Evidence has mounted that insufficient adaptation of beta cell mass is symptomatic of type 2 diabetes. Because the adaptation mechanism of beta cell mass in response to insulin resistance is not fully un- derstood, elucidating this mechanism is absolutely required for developing strategies of diabetes treatment. Previously, we demonstrated that insulin receptor substrate-2 (Irs2) is critically required for beta cell proliferation to occur in response to high-fat (HF) diet-induced insulin resistance. Recently, it was reported that beta cell proliferation began within the first 7 days of HF diet loading. However, it is unclear whether Irs2 is required for beta cell proliferation induced by short-term HF diet loading. Here, we investigated the effect of short-term HF diet loading on the regulation of beta cell pro- liferation. Materials and methods: Eight-week-old C57bl/6J mice were given free ac- cess to either standard chow (SC) or a HF diet. After 7 days on the above diets, we investigated body weight, blood glucose, visceral fat weight, liver weight and pancreatic weight in these mice. Also, insulin tolerance test and immunohistochemical analysis to assess beta cell proliferation and mass were performed. Furthermore, we evaluated the changes in expression levels of genes involved in beta cell proliferation and function in islets isolated from these mice. Results: Body weight and fed blood glucose levels were significantly higher in the mice on the HF diet than those in the mice fed SC. Although visceral fat weight were significantly higher in the mice on the HF diet than those in the mice fed SC, there were no differences in liver weight or pancreatic weight between the two groups. The glucose-lowering effect of insulin in the mice on the HF diet was equivalent to that in the mice fed SC on day 6. Immunohisto- chemical analysis revealed that there was a significant increase in stimulated the BrdU incorporation rate in the mice on the HF diet in comparison with the mice fed SC on day 7, although there was no difference in the area of the beta cells relative to that of the whole pancreatic tissues between the two groups. Real-time quantitative PCR showed that Ki67 and Cyclin A2 mRNA were significantly increased in the mice on the HF diet in comparison with the mice fed SC. However, no increase in the expression levels of Irs2 or genes involved in beta cell function, such as pancreatic and duodenal homeobox-1, glucokinase, insulin-1, or insulin-2 were noted in the mice on the HF diet as compared with that in the mice fed SC. Conclusion: Beta cell proliferation was induced by HF diet loading only for 7 days without up-regulation of Irs2. Our results suggest different regulation of beta cell proliferation induced by short-term and long-term high-fat diet loading. 471 Metabolic and pancreatic effects of transplantation of mesenchymal stem cells in a model of insulin resistance and type 2 diabetes A.M.O. Leal1, P.G. Bueno2, G.D. Pisani2, J.U. Yaochite3, K.F. Malmegrin3, L. Avó1; 1Medicine, 2Physiology, Federal University of São Carlos, 3University of São Paulo, Ribeirão Preto, Brazil. Background and aims: Type 2 Diabetes Mellitus (DM2) is associated with insulin resistance and dysfunction of pancreatic β cells. The regenerative cel- lular therapy, in particular with multi/pluripotent cells has been investigated as a potential therapeutic strategy for DM2. Among them, mesenchymal stem cells (MSCs) due to its immuno regulatory role are important thera- peutic candidates. The purpose of this study was to investigate the effects of multiple infusions of MSCs on glucose homeostasis and morphometry of the pancreatic islets in high fat diet-induced diabetes in Swiss mice. Materials and methods: Swiss mice were fed a standard diet or a high fat diet for eight weeks. The animals were then divided into 3 groups: non-diabetic group (fed a standard diet), untreated diabetic group and MSCs transplanted group. The transplanted mice received 4 intraperitoneal infusions of MSCs cells (5-8 x 106 MSCs resuspended in buffer). Diabetic untreated animals re- ceived only buffer injection and non-diabetic group did not receive injections. Fasting plasma glucose (FPG) was determined weekly and glucose (GTT) and insulin (ITT) tolerance tests were performed at 1, 2, 3, and 4 months after the infusions of MSCs. Four months after infusion of the MSCs, the animals were decapitated and pancreas and serum were collected for analysis. Results: The MSCs transplanted animals were classified as responder (FPG < 180mg/dL) or non-responder (FPG > 180mg/dL). According to this cri- terion, 72.2 % and 27.8 % of MSCs transplanted animals were classified as responders and non-responders, respectively. Fasting glycemia decreased significantly (p S 198 1 C than in non-diabetic mice, and significantly lower in responders mice than in untreated diabetic and non-responders animals. Islet cell proliferation was significantly (p S 199 1 C rate, plasma insulin levels, and fraction of endocrine cells and proliferating β-cells in graft immunohistochemistry. Co-transplantation with BM-spheres increased the vessel density derived from the donors and recipients. Conclusion: Co-transplantation of islets and angiogenic myeloid cells de- rived from the BM-spheroid improved the outcome of marginal mass islet transplantation by facilitation of revascularization. Supported by: NRF 2011-0023257 474 First evaluation of a closed, continuous media renewal system for human islets of Langerhans M. Ståhle1, A. Friberg1, S. Ingvast1, O. Korsgren1, A. Lams2, B. Delorme2; 1Clinical Immunology, Immunology, Genetics and Pathology, Uppsala, Sweden, 2Macopharma, Lille, France. Background and aims: Many clinical islet centres have adopted a short stor- age period for islets prior to transplantation. A storage period provides time for quality assessment of islets, recipient matching, transport of islets and patients to transplant centre and initiation of immunosuppressive protocols. There are, however, risks for loss of islet tissue and contamination during the storage period. To reduce contamination risks and improve handling and storage conditions, the present study presents the first evaluation data from of a new closed PReservation ISlet system (called PRISM) specifically developed for islet storage with regulated continuous media renewal. Materials and methods: At the end of the isolation process, purified islets were split and stored either in single transfer packs for platelets (Fenwal, Swe- den) as control or in the new V1.4 PRISM automate (Macopharma, France) for 4-5 days. Islets were both kept in CMRL-1066 supplemented with 10 mM HEPES, 10 mM nicotinamid, 2 mM L-glutamine, 50 µg/mL gentamicin, 5 mM sodium pyruvate, 20 µg/mL ciprofloxacin and 10% blood group-com- patible human serum. After a first over night storage at 37°C, the temperature was lowered to 25°C for the rest of the study period. Culture media exchanges were either performed manually for islets in the platelet bags or automatically using the PRISM automate. Islet quality assessments were performed at day 1 and last day (4 or 5) and included glucose-stimulated insulin release, intra- cellular insulin content, ADP/ATP ratio, cytokine expression and recovery. Results: No differences were found between islets kept in the PRISM auto- mate compared to islets kept in single transfer packs for platelets regarding stimulated insulin release, intracellular insulin content, ADP/ATP ratio or expression of MCP-1, tissue factor, IL-6 or IL-8. Conclusion: The closed, automatic media renewal PRISM technology seems to preserve functional integrity of clinical grade human islets kept in stor- age as well as standard clinical practice. It presents an attractive method for standardization and automation of islet storage. 475 Feasibility of islet magnetic resonance imaging using ferumoxytol in intraportal islet transplantation J. Kim, S.-M. Jin, M.-K. Lee, M.-S. Lee, K. Hur, J. Jun; Samsung Medical Center, Seoul, Republic of Korea. Background and aims: There is a clinical need for an alternative labeling agent for magnetic resonance imaging (MRI) in cell therapy including is- let transplantation, because superparamagnetic iron oxide (SPIO) has been withdrawn from the market. We aimed to evaluate the feasibility of islet mag- netic resonance imaging using ferumoxytol, which is a clinically available ul- trasmall superparamagnetic iron oxide (uSPIO) as an iron supplement drug. Materials and methods: We compared islet function and viability of control islets and islets labeled with ferumoxytol. In vitro efficacy of ferumoxytol la- beling was assessed with prussian blue stain, electron microscopy and ex vivo MRI of labeled islets. In vivo efficacy of labeling was assessed in both renal subcapsular and intraportal islet transplantation models. Results: Labeling with ferumoxytol up to 800μg/mL did not compromise the viability and glucose-stimulated insulin secretion of labeled islets. Prussian blue stain of labeled islets showed internalized ferumoxytol particles. Ex vivo magnetic resonance imaging of islets labeled with ferumoxytol (up to 800μg/ mL) for 48hr revealed visible hypointense spots representing labeled islets. In syngeneic renal subcapsular islet transplantation model, islet MRI at 14 days-post-transplantation (DPTs) showed visible hypointense spots repre- senting islet graft. In islet MRI at 7 and 14 days after syngeneic intraportal islet transplantation, there was a difference in the total area of hypointense spots between recipients with normoglycemia and hyperglycemia at 28 DPTs. Conclusion: Islet MRI using ferumoxytol was feasible in terms of in vitro and in vivo efficacy and safety. Labeling islet with ferumoxytol could be a useful option for estimation of islet mass in clinical islet transplantation. 476 Outcomes for adults with type 1 diabetes referred with severe hypoglycaemia and/or referred for islet transplantation to a specialist hypoglycaemia service M.L. Byrne1, P. Srinivasan2, N.D. Heaton2, D. Hopkins3, S.A. Amiel1, P. Choudhary1; 1Diabetes Research Group, 2Institute of Liver Studies, 3Department of Diabetes, King‘s College Hospital, London, UK. Background and aims: Severe hypoglycaemia (SH) and fear of hypoglycae- mia remain major barriers to achieving good diabetes control. In adults with type 1 diabetes (T1D), structured education in flexible insulin therapy (e.g. DAFNE, Dose Adjustment For Normal Eating) reduces SH by 65%, and tech- nology (insulin pump therapy and sensors) 4-fold. For persistent SH, despite optimised medical therapy, islet cell transplantation is indicated, reducing SH from 20.0 [7.0-50.0] to 0.3 [0.0-1.6] episodes per patient/year. We examined outcomes for patients referred to a diabetes centre with an islet transplant unit for islet transplantation and/or problematic hypoglycaemia. Materials and methods: Retrospective case note audit of all people with T1D referred to islet transplant unit for islet transplantation and/or with prob- lematic hypoglycaemia between 2009 and 2012 [n=82]. 45 met criteria for islet cell transplantation having >1 SH in previous year. Optimal outcome was defined as ≤1 severe hypoglycaemic episode over the past 12 months as documented at most recent visit. Results: The cohort was 55.6% male, mean (±SD) age 44.8 (±11.7) years, BMI 25.2 (±3.6) kg/m2 and duration of diabetes 28.1 (±13.4) years. HbA1c at index visit was 8.5% (±1.8), median [IQR] frequency of severe hypogly- caemia was 6.0 [2.0-21.5] per patient/year and 84.4% had impaired aware- ness of hypoglycaemia. 84.4% were referred from secondary diabetes ser- vices, 13.3% had completed DAFNE and 31.1% were using insulin pumps. Nine patients (20.0%) had initial assessment only (3 died, 2 did not attend follow up and 4 were referred back to local team with a new management plan). Follow up of the remaining 36 had a median duration of 28.5 [17.8- 42.5] months, during which SH fell from 6.0 [2.0-24.0] to 0.0 [0.0-3.0] events per patient/year; p S 200 1 C 477 Meal tests to assess beta cell function and mass in islet-transplanted patients J.-F.R. Brun1, L. Bories2, T. Berney3, J. Mercier1, P.-Y. Benhamou4, A. Wojtusciszyn2; 1INSERM U1046, Equipe d‘Explorations Métaboliques, 2INSERM U1001, Department of Endocrinology, Diabetes, and Nutrition, Montpellier, France, 3Department of Surgery, Cell Isolation and Transplant Center, Geneva, Switzerland, 4Inserm, U1055, Grenoble, France. Background and aims: Current model-derived calculations of insulin secre- tion based on C-peptide kinetics during a meal test allow to measure the two phases of insulin secretion. The second phase expressed as a β-cell glucose sensitivity (BCS) has been shown to be a strong predictor of diabetes pro- gression, and recent studies also suggest that it is closely related to the size of pancreatic beta-cell mass. Since beta-cell mass is a critical factor for the outcome of pancreatic islet transplantation, we measured these parameters in transplanted patients in order to see whether they were related to glucose tolerance. Materials and methods: We performed 13 standardized breakfast tests (76 g of carbohydrates) in 7 patients treated with islet transplantation (5 F /2 M, age 32-60 yr; weight: 51,3-72,8kg, time after transplantation 3-44 months) and compared them to 103 non-diabetic controls (C). Calculation of insulin sen- sitivity (Caumo‘s oral minimal model) and insulin secretion rate (ISR) (Van Cauter‘s model), with calculation of beta-cell sensitivity to glucose (BCS) and insulin secretion parameters given by the classical models of Breda and of Mari were done with the results of the tests. Results: According to fasting glycemia and glycemia at 2 hours, breakfast tests showed 6 normoglycemic profiles (NP) and 7 diabetic profiles (DP). NP patients had significant lower HbA1c (5,9% VS 7,1%; p < 0,001) and were insulin-independent compared to DP who all resumed insulin. No differ- ences were observed between groups regarding time after transplantation. Comparison of NP and DP with C showed that total ISR was higher in C than in the two groups of transplanted patients : 110,98±4,11 pMol/m² in C vs 61,17±9,75 (p DP 181,80±52,25 (C vs NP p DP 12,89±3,72 (C vs NP p S 201 1 C depletion present at baseline (SPK vs. C: 0.8±1.3 vs. 11.4±4.2 ENF/mm skin surface; p0.05 vs. baseline) with total ENF absence in 11 biopsies. Similarly, all AFT and QST results were clearly abnomal in SPK recipients in comparison with C and no amelioration occurred after long-term normoglycemia. Although some improvement was seen at follow-up in several electrophysiological pa- rameters, statistical significance for the SPK group as a whole was achieved in median motor NCV only (median; interquartile range, pre- vs. post-SPK: 47.5; 43.0 to 50.0 vs. 51.4; 50.9 to 55.3 m/s; p=0.004). Conclusion: Lower limb epidermal nerve fibre depletion and neurological function tests - except for median motor nerve conduction velocity - were not significantly improved following establishment of long-term normogly- cemia in pancreas/kidney transplant recipients. These results confirm the poor reversibility of advanced structural and functional changes in diabetic peripheral neuropathy. Supported by: NT13014-4 grant of the IGA MZ CR agency 480 Effect of simultaneous kidney-pancreas transplantation in patients with type 1 diabetes to stabilise / progression of complications M. Shamkhalova1, A. Glazunova1, M. Biragova1, S. Gracheva1, M. Shestakova1, Y. Moysyuk2, A. Pinchuk3, I. Dmitriev3, S. Arzumanov4; 1Diabetic Nephropathy, National Research Centre for Endocrinology, 2Transptantation of liver and kidney, Academician V.I. Shumakov Federal Research Center of Transplantology and artificial organs, 3Transptantation of the pancreas and kidney, Scientific-Research Institute of Emergency Care. N.a. N.V. Sklifosofskiy, 4Kidney transplantation and vascular surgery in urology and hemodialysis, Federal Research Centre of Urology, the Ministry of Health of Russia, Moscow, Russian Federation. Background and aims: To evaluate the effect of successfully simultaneous kidney- pancreas transplantation (SPK) in patients with type 1 diabetes mel- litus (T1DM) to stabilize/progression of complications. Materials and methods: The study included 8 patients on standard immuno- suppressive triple therapy. The average duration of T1DM 25 years [20,5;25], the duration of diabetic nephropathy (DN) - 7,96 лет[6,8;9,0]. All patients (3 men and 5 women) remained in the study for at least months [12;25,5] after the transplantation. Results: The mean level of HbA1c in the group before the study was 8,65 % [8,4;9,1], then decreased to individualising glycemic targets - 5.75% [5,55;6.0] after SPK. According to a continuous glucose monitoring system using «iPRO2» marked euglycemia during the day (glycemia 3,9-8,9 mmol/l - 89 %, to lower than 3.9 mmol / l - 11% , higher than 8,9 mmol/l - 0 % of the time of day). The examination determined normoalbuminuria, GFR 80,125 [71;90,5]. All patients had normal levels of hemoglobin 120,125 [112,5;130,0], parathormone 64,94[61,37;67,5], phosphorus 1,2 [1,05;1,4], blood pressure 110[110,0;111,5]. The progression of initial proliferative diabetic retinopathy (DR) in the post-transplantation period was observed in 37.5 % of patients, followed by performing a vitrectomy and additional sessions laser panretinal photocoagulation. At 87.5% (7 people) identified nonstenotic atherosclero- sis of the lower extremities, 1 patient - significant stenosis of the popliteal, posterior tibial artery to the right, requiring holding endovascular balloon angioplasty and stenting. In 75 % (6 people) developed ulcerative defects in the lower limbs and 4 people observed the progression of the chronic stage of osteoarthropathy. Conclusion: In addition to the recovery of renal function and euglycemia in patients with T1DM undergoing SPK, noted the progression of DR and diabetic foot syndrome, which bear witness to the genesis of multivariate dia- betic complications requiring verification and timely therapy. PS 025 Hypoglycaemia rates and their economic burden 481 Self-reported hypoglycaemia: a global study of 24 countries with 27,585 insulin-treated patients with diabetes: the HAT study K. Khunti1, S. Alsifri2, R. Aronson3, M. Cigrovski Berković4, C. Enters-Weijnen5, T. Forsén6, G. Galstyan7, P. Geelhoed-Duijvestijn8, M. Goldfracht9, R. Kapur10, N. Lalic11, B. Ludvik12, E. Moberg13, U. Pedersen-Bjergaard14, A. Ramachandran15; 1University of Leicester, UK, 2Alhada Armed Forces Hospital, Taif, Saudi Arabia, 3LMC Diabetes & Endocrinology, Toronto, Canada, 4University Hospital Centre Sestre milosrdnice, Zagreb, Croatia, 5Julius Clinical and Julius Center, UMC Utrecht, Netherlands, 6University of Helsinki, Finland, 7Endocrine and Metabolic Consultants, Moscow, Russian Federation, 8Medical Centre Haaglanden, The Hague, Netherlands, 9Clalit Health Services, Haifa, Israel, 10Novo Nordisk A/S, Søborg, Denmark, 11Faculty of Medicine University of Belgrade, Serbia, 12Medical University of Vienna, Austria, 13Department of Medicine, Karolinska Institutet, Stockholm, Sweden, 14Nordsjællands University Hospital Hillerød, Denmark, 15India Diabetes Research Foundation and Dr. A. Ramachandran‘s Diabetes Hospitals, Chennai, India. Background and aims: Hypoglycaemia is an important concern for patients with diabetes and physicians when setting glycaemic targets. The Hypogly- caemia Assessment Tool (HAT) study, the largest and most comprehensive of its kind, assessed self-reported hypoglycaemia and associated predictive factors in a global population of patients with insulin-treated diabetes. Materials and methods: HAT was a non-interventional, multicentre, 6-month retrospective and 1-month prospective study of hypoglycaemic events in 24 countries using self-assessment questionnaires and patient dia- ries (for 28 days) in people aged ≥18 years with type 1 (T1D) or type 2 (T2D) diabetes using insulin for ≥12 months attending routine clinics. Associations between predictive factors and hypoglycaemia were examined using negative binomial regression models adjusted for period and country. Results: 27,585 patients completed the study (Table 1). 83.4% of patients with T1D and 50.8% of patients with T2D experienced ≥1 hypoglycaemic event in the 4 weeks before baseline (51.5 and 16.5 events per patient year). Higher (p S 202 1 C Conclusion: In this large, multinational population of patients with T1D or insulin-treated T2D, rates of overall, nocturnal and severe hypoglycaemia were higher than previously reported. An increased incidence of overall hy- poglycaemia in the prospective study indicated significant under-reporting of hypoglycaemia. Clinical Trial Registration Number: NCT01696266 Supported by: Novo Nordisk A/S 482 Hypoglycaemia study questionnaire in patients with diabetes R. Tahhan; Endocrinology, Al-Zahra Hospital, Dubai, United Arab Emirates. Background and aims: To identify the frequency of self-reporting Moder- ate and severe hypoglycemia hypo and its relation to demographic, clinical variables, fear of hypoglycemia FoH and mood among adults with diabetes in outpatients clinics in United Arab Emirates. Materials and methods: The study adapted hypoglycemia patient question- naire by ADA/Endocrine society 2013 workshop on hypo. After obtaining approval of hospital ethical committee, patient verbal consent, adults‘ pa- tients with diabetes attending the clinics were encouraged to participate. FoH was assessed by Hypoglycemia fear survey HFS-II. WHO-5 survey is conducted for mood assessment when total HFS-II score ≥9. Hypoglycemia incidence was grouped according to its impact on patient‘s function into two groups; mild / no hypo and moderate/severe hypo group. Analysis was done using SPSS 21 for windows. Comparison between quantitative variables used student T-test or ANOVA when applicable and Chi-square test for qualitative variables. Results: Total study sample is 104 patients with diabetes, mainly DM2 92.3% 96/104.59.6 %( 42) females and 40.2% (62) males. Mean age, BMI, diabe- tes duration and HbA1c; 45.88±9.6, 30.88±6.04, 7.68±5.89, and 7.60±1.33 respectively. 56.7% reported mild/no hypo 59/104 versus 43.3 % (45) mod/ severe hypo. Diabetes duration increased the incidence of moderate/severe hypo; 6.41±5.71 (59) vs 9.36±5.75 (45) for no/mild hypo and mod/severe hypo respectively p= 0.002 but not for the other variables; age, HbA1c, BMI, RBG. Total HFS-II and its sub-scale worry and behavior scores were signifi- cantly higher among those subjects reporting mod/severe hypo; 26.47±22.7, 16.86±15.1, 9.35±9.44 respectively vs 9.83±11.57, 6.19±7.66, 3.64±6.77 for with mild/no hypo group P S 203 1 C were documented in 15% of hypoglycemic events reported by women and in 11% of episodes reported by men (p S 204 1 C did not, incurred a total healthcare cost of $36,680 (95% CI: $34,245-$39,288) vs. $16,850 (95% CI: $16,549-$17,156), diabetes medical costs ($10,350 vs. $2,403; p24 h) (Table). Although the number of events were lower in patients with T2D compared with T1D, the proportion resulting in hospital treatment for ≤24 h were similar across treatment regimens, whereas a higher proportion (47.6%) in the T2BOT group required hospital treatment for >24 h vs T1BB (5.0%) and T2MDI (5.3%). This resource utilisation and associated financial burden are significant: the UK tariffs for “Admitted Patient Care & Outpatient Procedures - Diabetes with Hypoglycaemic Disorders“ are £1,269 for ≤69 years and £2,187 for ≥70 years, in addition to £235 for an ambu- lance transfer. This yields an average cost per event across treatment regi- mens of £305 ((11.9%+6.7%)*£1269+29.3%*£235) for patients ≤69 and £476 ((11.9%+6.7%)*£2187+29.3%*£235) for patients ≥70 years Conclusion: This analysis suggests that severe hypoglycaemic events often result in emergency/ ambulance calls and treatment in hospital, and was ob- served with all insulin regimens, incurring a substantial economic burden. Thus, preventative measures to reduce severe hypoglycaemia are likely to re- duce this burden. Supported by: Novo Nordisk 488 Derivation and validation of a risk prediction tool for hypoglycaemia in hospitalised medical patients B.R. Shah1,2, S. Walji3, A. Kiss2, J.E. James1, J.M. Lowe1,2; 1Department of Medicine, University of Toronto, 2Sunnybrook Research Institute, 3Department of Family and Community Medicine, University of Toronto, Canada. Background and aims: There is no objective way of identifying, on admission to hospital, those patients with diabetes who are at risk for hypoglycaemia. At present, most admitted patients with diabetes have their blood glucose moni- tored frequently but many never have glucose readings below an acceptable range. A method of identifying those at greatest risk of hypoglycaemia would lead to efficiencies in care delivery and improvements in quality of care and patient safety. The aim of this study was to develop and validate a tool using data available at the time of admission to predict the risk of hypoglycaemia during hospitalisation. Materials and methods: A derivation cohort was identified of 300 random- ly selected patients with diabetes admitted to a medical inpatient unit at a tertiary care teaching hospital in Toronto, Canada between November 2009 and October 2010. All point-of-care glucose tests were collected, and hypo- glycaemia was defined as any result during the hospitalisation ≤3.9 mmol/L. Pre-specified candidate variables were abstracted from hospital records for each patient. Multivariable logistic regression was used to identify the inde- pendent predictors of hypoglycaemia using a backwards stepwise elimination method. The regression coefficients from the model were converted into an integer points score, and receiver operative characteristic curves were used to determine the threshold score that maximized sensitivity and specificity. The model was tested in a validation cohort of 300 randomly selected patients with the same inclusion criteria from a different Toronto tertiary care teach- ing hospital. Discrimination of the model was assessed using the c-statistic, and calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test. Results: In the derivation cohort, 105 (35%) patients had hypoglycaemia during their hospitalisation. The predictor variables remaining after back- wards elimination were sulfonylurea use, insulin use, beta-blocker use, al- pha-blocker use, an emergency department visit for any reason in the prior 6 months, fever and serum creatinine. The integer points scores derived from the model are shown in the Table. A summary score of ≥15 had a sensitivity of 79% and a specificity of 55% for predicting hypoglycaemia. In the validation cohort, this model had a c-statistic of 0.703 and the Hosmer-Lemeshow test had a p value of 0.50. Diabetologia (2014) 57:[Suppl1]S1–S564 S 205 1 C Conclusion: We have derived and validated an easy-to-use index that could be applied at the time of admission, which is moderately discriminative for predicting hypoglycaemia in hospitalised medical patients with diabetes. Supported by: Sunnybrook AFP Innovation Fund 489 The ‚Local Impact of Hypoglycaemia Tool (LIHT)‘ for estimating the economic impact of hypoglycaemic episodes in national, local and user- defined populations W.A. Parekh1, D. Ashley1, B. Chubb1, M. Evans2, H. Gillies3; 1Health Economics & Outcomes Research, EU Market Access, Novo Nordisk Ltd, Crawley, 2Consultant Diabetologist, University Hospital, Llandough, Cardiff, 3Abacus International, Bicester, UK. Background and aims: Insulin is the most efficacious glucose lowering ther- apy for patients with diabetes, however, a common short-term adverse effect of insulin therapy is hypoglycaemia. Hypoglycaemia has a major impact on patients’ lives affecting employment, driving, relationships, travel, and leisure activities. Severe and non-severe hypoglycaemic episodes also represent a substantial and often under-estimated cost burden to the National Health Service (NHS) and society through increased treatment costs and reduced productivity. The purpose of the Local Impact of Hypoglycaemia Tool (LIHT) is to help clinicians and budget holders estimate the cost of hypoglycaemia in the UK and specifically in their locality at the Primary Care Organisation (PCO), General Practice (GP) or user defined population level. Materials and methods: The LIHT is a versatile model which allows the incorporation of real-life local data. The user selects a region/population of interest and based on the epidemiology of diabetes in the UK the model es- timates the number of insulin-treated adults with type 1 and type 2 diabe- tes in that region. Using hypoglycaemia rates from the UK Hypoglycaemia Study Group (UKHSG) observational study, and the cost of a hypoglycae- mic episode, the annual cost of severe and non-severe hypoglycaemic epi- sodes in that population is estimated. The cost of a hypoglycaemic episode is dependent on the utilisation and unit costs of healthcare resources (blood glucose testing, glucose preparations, health care professional consultations, ambulance and hospital costs, derived from MIMS, Personal Social Services Research Unit (Health & Social Care) and NHS Tariff information), and is es- timated to range from £1.67 to £2,195. In addition to the direct costs incurred by the NHS, indirect costs associated with lost productivity are estimated. Results: The model highlights the cost burden of hypoglycaemia in the UK for insulin-treated adults with diabetes. On a national level, with a UK popu- lation of almost 67 million people, the total cost of managing hypoglycaemic episodes in insulin-treated adults with diabetes is an estimated £363.6m per year (£235.6m for severe and £128m for non-severe episodes, respectively). In addition, hypoglycaemic episodes are associated with indirect and hidden costs, such as lost productivity and higher rates of falls and fractures. On a lo- cal health economy level the model can provide estimations for each particu- lar PCO within the UK. Using a hypothetical population of 100,000, the total cost of managing hypoglycaemic episodes is estimated to be £543,493 per year (£352,163 for severe and £191,330 for non-severe episodes, respectively). Conclusion: This model highlights the substantial cost burden to the NHS of hypoglycaemia in insulin-treated adults and may aid clinicians and NHS budget holders with choices regarding insulin treatments. The model offers an opportunity to explore how reducing hypoglycaemia can improve the quality of diabetes care for patients and result in potential budget savings. PS 026 Insulin secretion in animal models 490 Influence of fasting duration on glucose metabolism in C57BL/6J mice during intravenous glucose tolerance test A.-H. Thorén Fischer, M. Abels, N. Wierup; Clinical Sciences in Malmö, Lund University, Neuroendocrine Cell Biology, Malmö, Sweden. Background and aims: The intravenous glucose tolerance test (IvGTT) is commonly used to study insulin release and beta cell function in both hu- mans and animal models. IvGTT is often preceded by a period of fasting and different laboratories use different protocols for this. Because the IvGTT is frequently used and a lot of conclusions are drawn from it, it is important to standardize the performance and use conditions that best reflect the situation aimed to study, e.g fasting or starvation. In this study we wanted to investigate how different fasting time affects insulin secretion and glucose elimination during IvGTT in mice. The time durations were chosen based upon com- monly used fasting times to find a fasting duration that gave the strongest insulin response and glucose elimination. Materials and methods: 20g C57BL/6J BomTac female mice (n=5 in each group) were fasted for different times (12 hour, 4 hour and 1 hour). The mice were anaesthetized with Hypnorm/Dormicum 30 minutes prior to the ex- periment. A basal blood sample was taken from all the mice. Thereafter, glu- cose (1 g/kg) was injected in a tail vein, and six following blood samples were taken at 1, 5, 10, 20, 50 and 75 minutes. All blood samples were taken from the retro orbital plexus. Glucose (colorimetric assay) and insulin (ELISA) were analyzed in plasma. Results: The mice fasted for 12 hours had higher basal glucose levels, com- pared with mice fasted for 1 and 4 hours (p S 206 1 C by dual emission x-ray absorptiometry using a PIXImus imager. IVGTTs were performed on 5 hour fasted mice. After injecting glucose in a tail vein, samples were taken from a retrobulbar intraorbital capillary plexus at 1, 5, 10, 20, 30 and 50 minutes for determination of insulin and glucose concentra- tion. Glucose was dosed per total body mass (0.35g/kg), per lean body mass (0.35 g/lean kg) or equal for all animals (10 mg). Results: Each regimen detected a similar decrease in insulin sensitivity in DIO mice. The different glucose dosing regimens gave, however, diverging results in regards to glucose elimination and the acute insulin response. Thus, the fixed dose regimen was the only that revealed impairment of glucose elimination in HFD (glucose elimination rate 1.4±0.2 vs. 2.4±0.2 %/min, P S 207 1 C their roles in IPC viability, insulin-like peptide (dILP) mRNA expression and dILP2 secretion. None of the kinases had an effect on the viability of IPCs. One of the kinases, Tousled-like kinase, showed significantly reduced dILP3 expression. Further, knockdown of several identified genes like ADCK, PKC98E, Rio2, aPKC, CDK12 and Rio1 showed accumulation of dILP2 in IPCs, implying disturbed dILP2 secretion. One of the genes, aPKC, has pre- viously been linked to insulin secretion in mammals, suggesting conserved function. The mechanisms of action of other genes which are not previously studied with respect to insulin signaling are being investigated. Conclusion: In conclusion, we have identified twelve protein kinases, which regulate insulin production in Drosophila melanogaster. One of the genes, Tlk, regulates dILP3 transcription in IPCs. Further, ADCK, PKC98E, Rio2, aPKC, CDK12 and Rio1 play a role in secretion of insulin like peptide 2 from IPCs. Supported by: ILS Doctoral Program, Sigrid Juselius Foundation. 495 Antidiabetic effects of tigerinin-1R in mice chronically maintained on high fat diet Y.H.A. Abdel-Wahab, O.O. Ojo, D.K. Srinisavan, B.O. Owolabi, P.R. Flatt; School of Biomedical Sciences, University of Ulster, Coleraine, UK. Background and aims: We previously reported in vitro and acute in vivo insulinotropic effects of tigerinin-1R (RVCSAIPLPICH.NH2), an amphib- ian host defence peptide, isolated from the skin secretion of Hoplobatrachus rugulosus. The present study investigated plasma degradation and chronic metabolic effects of tigerinin-1R in Swiss NIH mice fed a high fat diet to induce obesity-diabetes. Materials and methods: Proteolytic degradation of synthetic tigerinin-1R was investigated by reversed-phase HPLC and MALDI-TOF mass spectrom- etry using GLP-1(7-36) as positive control. Changes in glycaemic responses and metabolic parameters were measured in mice with high fat diet-induced obesity-diabetes treated with twice-daily intraperitoneal injection of tigerin- in-1R (75nmol/kgbw) for 15 days. Glucose concentrations were measured by glucose oxidase method using an Analox GM9 glucose analyser and insulin concentrations were measured by radioimmunoassay. Indirect calorimetry and body composition were measured by CLAMS and DEXA whole body scanning. Insulin secretory responses of islets isolated from treated and un- treated mice were examined. Results: Unlike GLP-1(7-36), tigerinin-1R was resistant to in vitro degrada- tion by plasma enzymes. Twice-daily injection of tigerinin-1R (75nmol/kg bw) for 15 days had no significant effect on food intake or body weight. Non- fasting glucose levels were significantly lowered (16%, P S 208 1 C release at 5.6mM glucose (p S 209 1 C PS 027 Determinants of insulin secretion in humans 500 Across glucose tolerance spectrum, men display greater decreases in insulin secretion than women S.S. Shankar1, R.H. Raymond2, D. Chen3, C. Cobelli4, C. Dalla Man4, D.S. Lee3, R.P. Robertson5, D. Stefanovski6, M.T. Vassileva7, D.A. Fryburg8, for the Beta Cell Team of FNIH Biomarkers Consortium; 1Eli Lilly Research Laboratories, Eli Lilly and Co., Indianapolis, 2RSquared Solutions, Skillman, 3Pfizer Pharmatherapeutics R & D, Cambridge, USA, 4University of Padova, Italy, 5Pacific Northwest Diabetes Research Institute, Seattle, 6Cedars-Sinai Medical Center, Los Angeles, 7Foundation for the NIH Biomarkers Consortium, Bethesda, 8ROI BioPharma Consulting, East Lyme, USA. Background and aims: Little is known about gender differences in insulin secretion (IS). As part of a project examining the utility of beta cell func- tion tests, we studied effect of gender on IS response to the arginine (ARG) stimulation test (AST) and the mixed meal tolerance test (MMTT) in over- night fasted, obese men (M) and women (W) with normal glucose tolerance (NGT), prediabetes (PDM) and type 2 diabetes (T2DM). Materials and methods: For AST, acute IS (AIRarg) was measured over 5 min at baseline glucose after an i.v. ARG bolus (5 gm over 30 sec). Immedi- ately following these samples, a 60 min infusion (900 mg/min) of glucose was initiated; ARG was administered again after 50 min of the glucose infusion; IS samples were then repeated (AIRmax). All arg results are baseline-adjusted. For MMTT, IS (Φtot and DItot) were measured in response to a standardized 450 kCal meal and estimated using minimal model. Results: The table below summarizes the results. Within each gender sig- nificant declines in insulin secretion were detected for all 4 parameters (NGT and PDM were largely similar v. T2DM). To compare the changes in insulin secretion of women to men across the glucose tolerance groups, an ANCOVA model was developed. Of covariates tested (BMI and age), only age was in- cluded in the model as a significant covariate. ANCOVA showed that from NGT to T2DM, as seen in both the AST and MMTT, the decline in insulin secretion is greater in men than in women. Conclusion: We conclude that 1) in a cross-sectional analysis of men and women with NGT, PDM, and T2DM, insulin secretion declines in both gen- ders; 2) From NGT to T2DM, men show a greater decline than women in 2 different tests of insulin secretion. These results may have implications for gender balance in study designs measuring insulin secretion. Clinical Trial Registration Number: NCT01663207 Supported by: FNIH Biomarkers Consortium Sponsors 501 Circulating non-esterified fatty acids (NEFA) do not associate with beta cell dysfunction: evidence from the RISC study E. Rebelos1, A. Natali1, A. Mari1, B. Balkau2, A. Golay3, P. Piatti4, N. Lalic5, M. Laakso6, E. Ferrannini1; 1Department of Clinical and Experimental Medicine, University of Pisa, Italy, 2INSERM U 258, Villejuif, France, 3Department of Clinical and Experimental Medicine, University Hospitals of Geneva and University of Geneva, Switzerland, 4Internal Medicine Department, Cardio-Diabetes Trials Unit, Scientific Institute San Raffaele, Milan, Italy, 5Faculty of Medicine, Clinic for Endocrinology, Diabetes and Metabolic Disease, University of Belgrade, Serbia, 6Department of Medicine, University of Eastern Finland, Kuopio University Hospital, Finland. Background and aims: It is commonly held that chronically elevated NEFA levels adversely affects insulin secretion and insulin action (lipotoxicity). Al- though the association between NEFA and insulin resistance is well estab- lished, the effect of raised NEFA on ß-cell function has only been explored in small experimental protocols using acute elevations in NEFA. Our aim was to analyse the relationship between endogenous NEFA levels and insulin secre- tion and ß-cell function both cross-sectionally and longitudinally. Materials and methods: In 701 women and 566 men from the Relationship Between Insulin Sensitivity and Cardiovascular Disease (RISC) cohort (mean age 44 years, BMI range 17-44 kg/m2, 14% with impaired fasting glycaemia (IFG) and 9% with impaired glucose tolerance (IGT)) followed up for 3 years, we measured insulin sensitivity (by a euglycaemic insulin clamp) and ß-cell function (by modelling of the C-peptide response to oral glucose and as the acute insulin response (AIR) to intravenous glucose). Results: At baseline, both fasting and insulin-suppressed NEFA were signifi- cantly (p S 210 1 C of beta-cell function [HOMA-beta, fasting immunoreactive insulin (µIU/ ml) x 20 / (fasting plasma glucose (mmol/l)- 3.5)], C-peptide index [fasting C-peptide (nmol/l)/ fasting plasma glucose (mmol/l)], insulinogenic index [∆ serum insulin 0-30 min (pmol/l) / ∆ plasma glucose 0-30 min (mmol/l) in 75 g-OGTT] and ΔC-peptide by glucagon test [ΔC-peptide, increment of serum C-peptide (nmol/l) at 6 min after intravenous injection of 1-mg gluca- gon]. Pancreatic samples were fixed in formaldehyde, embedded in paraffin for subsequent analysis and cut from these paraffin blocks into 5-μm thick sections. We used immunohistochemistry to determine pancreatic beta-cell area. The relative beta-cell area represented the proportion of insulin-positive cell area to whole pancreatic section (%). Results: The mean of relative beta-cell area was 1.072±0.424, 0.998±0.419 and 0.762±0.441%, in NGT, IGT and DM, respectively. ΔC-peptide (r=0.64, p=0.002), HOMA-beta (r=0.50, p=0.003) and C-peptide index (r=0.36, p=0.042) correlated significantly and positively with the relative beta-cell area. Insulinogenic index tended to correlate positively but not significantly with the relative beta-cell area (r=0.33, p=0.078). Among them, ΔC-peptide showed the closest association with the relative beta-cell area. Age tended to correlate negatively but not significantly with the relative beta-cell area (r=-0.33, p=0.066). Neither BMI nor HbA1c showed a significant correlation with the relative beta-cell area (r=0.26, p=0.147; r=-0.16, p=0.378, respec- tively). Glucose levels at 0, 30, 60, 120 min in 75 g-oral glucose tolerance test (OGTT) tended to correlate negatively with the relative beta-cell area (r=- 0.26, p=0.145, r=-0.30, p=0.105, r=-0.31, p=0.087, r=-0.34, p=0.064, respec- tively), while the area under the curve of plasma glucose level from 0 to 120 min in 75 g-OGTT correlated significantly and negatively with the relative beta-cell area (r=-0.36, p=0.045). Conclusion: ΔC-peptide by glucagon test, HOMA-beta, and CPI correlated closely with beta-cell area, and ΔC-peptide was the most valuable index for the prediction of beta-cell mass. 503 Normal meal tolerance test is more valuable than glucagon stimulation test in type 2 diabetes Y. Fujioka, T. Ohkura, K. Shoji, K. Sumi, H. Shiochi, N. Yamamoto, K. Matsuzawa, A. Murawaki, S. Izawa, H. Ohkura, M. Kato, S. Taniguchi, K. Yamamoto; Tottori University Faculty of Medicine, Yonago, Japan. Background and aims: The glucagon stimulation test (GST) is a standard method of evaluating endogenous insulin secretion. However, the GST has adverse effects and takes effort. In contrast, the meal tolerance test (MTT) has no adverse effect and cheaper than GST. Moreover, to perform the MTT by normal meal is important to evaluate insulin secretion in real life. However, the MTT is affected by the glucotoxicity than GST in the hyperglycemic state. The aim of this study is to evaluate the properties of the glucagon stimulation test and normal meal tolerance test in type 2 diabetes patients. Materials and methods: We enrolled 142 patients of type 2 diabetes (mean: age 61.0, M/F: 80/62, BMI 24.7, HbA1c 9.4) and performed the GST and the MTT. We performed the MTT by using a normal diabetes food of 30kcal/ day per ideal body weight, which contains 60% of carbohydrates, protein 20%, lipid 20%. In the MTT, patients continued oral hypoglycemic agents and in- sulin treatment as usual. In the MTT, we measured plasma glucose (PG) and serum C-peptide immunoreactivity (CPR), and plasma insulin level before and 120minutes after meal load. We calculated increment of CPR (ΔCPR) by subtracting fasting CPR (FCPR) from 6 minutes after glucagon injection and 120 minutes after meal load. Results: The mean fasting PG (FPG) was 154 mg/dl, the mean FCPR was 2.2 ng/ml, the mean GSTΔCPR was 2.0 ng/ml, and the mean MTTΔCPR was 3.1 ng/ml. 104 patients represent higher ΔCPR in the MTT than the GST, and the mean MTTΔCPR was significantly higher than the mean GSTΔCPR (P S 211 1 C 505 Decrease in plasma mannose level after glucose load is associated with glucose tolerance: plasma mannose level as a putative indicator of glycogenolysis S. Hirano1, K. Yoshimura1, H. Takata1, S. Ohmi1, T. Taguchi2, S. Yamada2, I. Miwa2, Y. Terada1, S. Fujimoto1; 1Department of Endocrinology, Metabolism and Nephrology, Kochi University, Nankoku, 2Department of Pathobiochemistry, Meijo University, Nagoya, Japan. Background and aims: Mannose is a monosaccharide constituent of gly- coproteins and glycolipids. Experiments in rats showed previously that the plasma mannose level decreases after glucose load which causes insulin se- cretion in normal rats, but does not decrease in diabetic rats and that hepatic glycogenolysis is a source of this plasma mannose, but these results are not fully elucidated in human. Subjects with glucose intolerance have impaired secretion and sensitivity of insulin, which are the most important factors in suppression of glycogenolysis. To explore the possibility of the plasma man- nose level as an indicator of glycogenolysis in human, the decrease in plasma mannose levels after glucose load in subjects with various degrees of glucose intolerance was examined and was analyzed to clarify association with clini- cal factors. Materials and methods: 75g OGTT was performed on Japanese subjects without diabetic medication from whom informed consent was obtained. Based on OGTT data, subjects were classified as normal (NGT), impaired glucose tolerance (IGT), and diabetes (DM) according to WHO criteria. In each group, 25 subjects were consecutively recruited [total 75 subjects, age: 65±11 (mean±SD); male/female: 34/41; BMI: 24.9±3.8 (mean±SD)]. In- sulinogenic index (IGI) as an index of insulin secretion and QUICKI and Matsuda index as indices of insulin sensitivity were calculated. Mannose was assayed using established method. Briefly, after labeling with 4-aminoben- zoyl ethyl ester, the concentration of mannose was determined using HPLC. Glucose contained in samples was confirmed not to affect the assay. The study protocol was approved by the ethics committee of the institute. Results: Plasma mean levels of glucose during 120 min (Gm) were signifi- cantly different among groups [NGT: 138±20, n=25; IGT: 178±27, n=25; DM: 226±24 mg/dl, n=25, (means±SD)], while mean IRI during 120 min (Im) did not significantly differ. Basal levels of mannose were similar among groups (NGT: 40.8±10.0; IGT: 43.7±11.2; DM: 45.0±12.0 μM). After glucose load, the plasma mannose level was decreased gradually and reached plateau at 90 min in NGT, but was not decreased significantly in DM. Plasma changes of mannose during 120 min from base line (M120-M0) were significantly different among groups (NGT: -16.7±8.2; IGT: -9.0±9.8; DM: -1.4±9.3 μM, P S 212 1 C muscle) carbohydrate oxidation and its effect on insulin secretion deserve further research. Supported by: CNRU P30 DK072476 & FONDECYT 1130217 508 Role of parasympathetic activity in insulin secretion in metabolically healthy obese patients C. Cussac-Pillegand, S. Chiheb, Y. Jaber, C. Cyrille, R. Duteil, E. Cosson, P. Valensi; AP-HP, Jean Verdier Hospital, Bondy, France. Background and aims: Cardiac parasympathetic activity is often altered in obese non-diabetic patients, even in the metabolic healthy obese individuals. We here hypothesized that the persistance of a good parasympathetic activ- ity could contribute to a better insulin secretion in the presence of insulin resistance. Materials and methods: We recruited 47 obese patients (BMI 42.3±6.5 kg/ m², 35.5±12.7 years). Among them 31 with only one criterion of the meta- bolic syndrome (fasting hyperglycemia in only one patient) in addition to a large waist circumference was considered as metabolically healthy. Car- diac parasympathetic activity (HF-HR) and vascular sympathetic activity (LF-SBP) were evaluated by spectral analysis of heart rate (HR) and systolic blood pressure (SBP) variations (Finapres®), respectively. An oral glucose tolerance test was performed. Two indexes of insulin resistance (Matsuda and HOMA-IR), and composite indexes of insulin secretion (insulinogenic index=IGI=Δinsulinemia(T0-T30)/Δglycemia(T0-T30); and oral disposition index=ODI=IGI/insulinemia) were calculated. Results: Patients with a lower parasympathetic activity (HF-HR S 213 1 C PS 028 Insulin secretion in diabetes 510 Most people with long duration of type 1 diabetes are insulin microsecretors and produce their own endogenous insulin T.J. McDonald1, R. Oram1, B. Shields1, E. Pearson2, A. Hattersley1; 1University of Exeter Medical School, 2University of Dundee, Medical Research Institute, UK. Background and aims: Ultrasensitive assays that can detect C-peptide un- der5 pmol/L allow detection of very low levels of c-peptide. We aimed to use urine c-peptide creatinine ratio (UCPCR) to assess endogenous insulin in a large cross-sectional population-based study of patients with Type 1 diabetes (T1D). Materials and methods: We recruited 944 patients from primary and sec- ondary care in 2 UK centres. All diagnosed under 30 years, duration >5 years, clinical diagnosis of T1D. Median(IQR) age of diagnosis 11(6-17)y, duration 18(11-26)y, HbA1c 8.7(7.9-9.8)%, insulin dose 0.78(0.60-0.97)u/kg/24hr, and BMI 25.6(23.3-28.6)Kgm-2. All provided a home post-meal UCPCR. Results: 81% (790/944) had detectable endogenous production (median(IQR) UCPCR 0.012 (0.004-0.038)nmol/mmol). Most had very low, historically un- detectable, levels (492/944, 53%, UCPCR >0.001-0.03 nmol/mmol). 8% had C-peptide levels above the DCCT cut off of significant endogenous insulin. Absolute UCPCR levels fell with duration but the proportion with detectable UCPCR never fell below 73% (maximum duration 47 years).Age of diagnosis and duration independent predictors of C-peptide in multivariate modelling. Conclusion: The majority of patients with long duration T1D are insulin microsecretors and have detectable urine c-peptide. Some rare individuals with T1D maintain higher levels of endogenous insulin for many years after diagnosis of diabetes. The fact that some beta cells remain in most with long- standing T1D may reflect escape from immune attack, or beta cell regenera- tion. Understanding this may lead to a better understanding of pathogenesis in T1D and open new possibilities for treatment. Supported by: HICF 511 Presence and titer of GAD antibodies are determinants of beta cell function in patients with newly diagnosed type 2 diabetes mellitus: further insights in the metabolic phenotype of LADA patients M. Trombetta1, L. Boselli1, C. Brangani1, F. Zardi2, C. Negri3, V. Stoico3, E. Bonora1, R. Bonadonna4; 1Medicine, University of Verona, 2Hospital of Bolzano, 3AOUI of Verona, 4Medicine, University of Parma and AOUI of Parma, Italy. Background and aims: Latent Autoimmune Diabetes in Adults (LADA) is a metabolic disorder at the crossroad between type 1 (T1DM) and type 2 diabetes (T2DM). Aim of our study was to carefully assess beta cell function and insulin sensitivity in patients with LADA, in comparison to patients with either type 2 diabetes clinically undistinguishable from LADA or typical type 2 diabetes. Materials and methods: In 35 (M/F=19/16) patients (mean±SEM: age 57.4±1.6 years, BMI 27.5±0.9 kg/m2) with newly diagnosed LADA were com- pared to 35 patients with newly diagnosed T2DM matched for age, gender, BMI and HbA1c (LADA-like). The latter group was extracted from the data- base of the Verona Newly Diagnosed Type 2 diabetes (VNDS; N=589 GADA- negative patients) The rest of VNDS patients herein represent typical T2DM. LADA patients were further divided in two groups according to GADA levels (median 4 kU/L): low GAD-LADA (GADA below 4kU/L)and high GAD- LADA (GADA above 4kU/L). In all patients we performed on separate days: 1. prolonged (5-hours) frequently sampled OGTT to assess derivative control (DC) and proportional control (PC) of beta cell function by state of art math- ematical modeling of glucose and C-peptide curves; 2. standard euglycemic insulin clamp to assess insulin sensitivity (SI). Results: SI was not statistically different (p S 214 1 C Conclusion: In conclusion, either in remission, shortly after ketosis resolu- tion or during long term follow-up, latent HHV-8 infection is associated with better insulin secretion in response to glucose in patients of sub-Saharan Af- rican origin with KPD. Mechanisms by which latent HHV-8 infection impact on β-cell functions need to be unraveled. 513 Correlation between the indices for glycaemic variability and 24-hour urinary C-peptide values in patients with type 1 diabetes mellitus as assessed by CGM C. Seo, R. Nishimura, Y. Onda, D. Tsujino, K. Andou, K. Ustunomiya; Division of Diabetes, Metabolism and Endocrinology Department, The Jikei University School of Medicine, Tokyo, Japan. Background and aims: Loss of pancreatic cell function, abnormal secretion of the counter-regulatory hormones, inappropriate treatment, insulin resist- ance, malingering/manipulation, concomitant diseases such as infection and some other endocrine disease, and mental/somatic stress are reported to be among the factors that tend to make glycemic control difficult in patients with type 1 diabetes. First and foremost of these are, however, loss of pan- creatic β cell function and abnormal secretion of the counter-regulatory hor- mones. Therefore, stable glycemic control becomes progressively lost with the loss of pancreatic β cell function. While, generally, it often proves extremely difficult to achieve glycemic control in type 1 diabetes, not all patients with type 1 diabetes exhibit unstable diabetes and a subset of patients exhibit stable glycemic control. In the latter patients, pancreatic β cell function is shown to be preserved to a certain extent, thus contributing to their stable glycemic control. However, to date, there are very few studies in which residual pan- creatic β cell function was examined for correlation with glycemic variation evaluated by using Continuous Glucose Monitoring(CGM) that allows glu- cose levels to be measured on a continuous basis. Materials and methods: Of all patients being treated with intensive insulin therapy (excluding Continuous Subcutaneous Insulin Infusion) and moni- tored by CGM, a total of 59 patients with type 1 diabetes (males/females, 20/39) with detectable urinary C-peptide immunoreactivity (U-CPR) from 24-hour samples were included in the study. Values measured for continu- ous 24 hours and free from measurement errors were used for current analy- sis. U-CPR values were used as an index for insulin-secretory capacity and the SD of 24hour-glucose concentration was used as an index for stability of glycemic control. The subjects’ U-CPR values were examined for correla- tion with their duration of disease, BMI, and HbA1c values, as well as the SD of their glucose concentration. Statistical analyses were performed by using SPSS 17.0. Spearmann’s rank sum correlation coefficients were used to test for correlation between the variables examined. Results: All data are shown as median and interquartile range. The subjects’ median U-CPR value was 1.4 (0.9-6.9) μg/day, their median age 39 (32-58) years, duration of disease 11 (6-22) years, BMI 21.2 (19.4-23.2) kg/m2, HbA1c value 8.0 (7.1-9.4)% ,their total insulin dose 36 (28-48) units, and their basal insulin ratio(basal insulin dose/total insulin dose) 44.4 (37.5-54.3)%. A sig- nificant negative correlation was noted between their U-CPR values and the SD of their glucose concentration (r = -0.305; P = 0.019) , as well as between their U-CPR values and their duration of disease (r = -0.353; P = 0.006),and between their U-CPR values and their basal insulin ratio (r = -0.268; P=0.040) . However, there was no significant correlation between their U-CPR values and their age, BMI, HbA1c values or total insulin units. Conclusion: Our study demonstrated that the lower the U-CPR values in patients with type 1 diabetes, the greater their glycemic variability and that the longer their duration of disease, the lower the U-CPR values, suggesting that U-CPR assessment may have a key role in predicting glycemic variations in patients with type 1 diabetes. 514 Factors affecting glycaemic variability in patients with type 2 diabetes S.-Y. Yoo1, S.-A. Lee1, H.-J. Chin2, D.-H. Lee3, G. Koh1; 1Department of Internal Medicine, Jeju National University Hospital, 2Department of Internal Medicine, Hankook General Hospital, Jeju, 3Department of Internal Medicine, Wonkwang University Hospital, Jeollabuk-do, Republic of Korea. Background and aims: It is not widely known about factors determining glu- cose fluctuations in diabetic patients. Also, the role of glycemic variability in the development of cardiovascular diseases remains controversial. We inves- tigated the relationship between indices of glycemic variability, cardiovascu- lar (CV) risk factors, and clinical and laboratory variables in type 2 diabetic patients. Materials and methods: We enrolled 236 type 2 diabetic patients which per- formed a 7-point SMBG more than once during each month for 3 consecu- tive months. From their SMBG data, glycemic variability indices (standard deviation (SD) and M-value) were calculated monthly. HbA1c was measured on the last day of the third month. BMI, waist circumference (WC), blood pressures (BP), duration of diabetes, hsCRP, fibrinogen, ALT, GGT, creati- nine, uric acid, total cholesterol, triglyeride (TG), HDL-C, LDL-C, urine albumin:creatinine ratio (UACR) and ankle-brachial pressure index (ABI) were assessed. Treatment stage of diabetes was defined based on the number of oral hypoglycemic agents and the administration of insulin. The 10-year risk for CVD was calculated using 2013 ACC/AHA Prevention Guidelines Atherosclerotic Cardiovascular Disease (ASCVD) Risk Estimator. Results: The SD was significantly correlated with duration of diabetes (r=0.291; p S 215 1 C Conclusion: In this study, duration of diabetes rather than CV risk factors was an independent variable of indices of glycemic variability. And the in- dices were positively correlated to treatment stage of diabetes more than 10- year ASCVD risk. These findings suggest that glycemic variability is largely determined by β-cell function which deteriorates with increasing duration of diabetes, but not cardiovascular complication. 515 Loss of glucose sensitivity and islet neogenesis predict the occurence of diabetes after acute beta cell reduction A.V. Sun1, T. Mezza1, G. Sorice1, C. Conte1, C. Cefalo1, R.N. Kulkarni2, A. Mari3, A. Giaccari1; 1Catholic University, Rome, Italy, 2Joslin Diabetes Center, Boston, USA, 3CNR, Padova, Italy. Background and aims: Progressive deterioration in β cell function and de- crease in β cell mass represent the main mechanisms involved in type 2 dia- betes. To investigate if the deterioration of the β cell function corresponds to a loss of β cell mass, we performed oral glucose tolerance tests (OGTT), hy- perglycemic clamps (HC) and followed by arginine stimulation in 16 patients undergoing pancreatoduodenectomy (PD), pre- and post-surgery. To further explore whether Islet features could be justified by in vivo beta cell function, we explored neogenesis from duct cells, islet size and trans-differentiation of α cells to β cells. Materials and methods: Based on post-surgery OGTT, subjects were divided into 3 groups of glucose tolerance: normal (NGT, n=5), impaired (IGT, n=4) or diabetes (DM, n=7) (8 F/8 M, 51±15 yrs.). To evaluate β cell function, β cell glucose sensitivity (GS) during HC was calculated as the ratio of insulin secretion and glucose increments. During surgery, pancreas samples were collected for IHC for glucagon, insulin and somatostatin+ cells to assess islet morphology. Ductal cells were stained by CK19. Results: Before surgery, Arginine-stimulated Insulin Secretion (AIS) was similar across groups, whereas GS was lower in IGT and DM as compared with NGT subjects (62.9±23.1 and 45.5±11.2 vs 90.6±18.7 pmol•min-1•m-2, respectively). Following 50% PD, GS decreased in all patients (p S 216 1 C 517 Improved beta cell and disposition function leads to reduction in daily insulin dosage over time in patients with type 2 diabetes treated by long- term CSII therapy S. Choi1, E. Hong1, K.-J. Kim1, H. An1, Y. Noh2; 1Endocrinology and Metabolism, 2Biochemistry, School of Medicine, Konkuk University, Seoul, Republic of Korea. Background and aims: We observed the daily insulin dosage required to maintain normoglycemia decreased over time during long-term continu- ous insulin infusion (CSII) therapy in patients with type 2 diabetes. To see if the decreased daily insulin dosage is related to changes in beta cell function and insulin resistance during CSII therapy, we examined changes in C-pepti- dogenic Index (CI), Matzda Index (MI), and disposition Index (DI; product of CI and MI) for 1 year. Materials and methods: Two hundred seventeen patients with type 2 dia- betes (age, 59.1 ± 11.1 years; male, 113; female, 104; duration, 12.6 ± 10.2 years; body mass index, 23.7 ±3.7 kg/m2; HbA1c 8.9 ± 2.1 %) were treated by CSII.Blood samplings were performed at baseline and 6 months and 1 year after CSII therapy (after overnight fasting and 2 hours after ingestion of 500 kcal-mixed meal). Results: After 1 year of CSII therapy, HbA1c decreased from 8.9 ± 2.1 to 7.3 ± 1.7 % (p < 0.001), serum C-peptide level increased from 4.8 ± 2.8 to 6.0 ± 2.9 ng/ml (p < 0.001), and CI increased from 0.021 ± 0.018 to 0.029 ± 0.017 (p < 0.001) at 2 hours after meal ingestion. DI increased from 0.060± 0.059 to 0.069 ± 0.045 (p < 0.003). Daily insulin dosage decreased from 71.7 ± 34.4 to 46.3 ± 31.1 IU/day (p S 217 1 C PS 029 Gut hormones 519 Impact of Roux-en-Y gastric bypass on distribution of enteroendocrine cells and their gene expression in obese patients with type 2 diabetes and non-diabetic controls N.A. Rhee1, C.D. Wahlgren1, J. Pedersen2, E. Langholz3, E.P. Wandall3, S.U. Friis3, P. Vilmann4, S.J. Paulsen5, V.B. Kristiansen6, J. Jelsing5, S.S. Poulsen2, J.J. Holst2, T. Vilsbøll1, F.K. Knop1; 1Department of Internal Medicine, Gentofte Hospital, Diabetes Research Division, Hellerup, 2Department of Biomedical Sciences, University of Copenhagen, 3Department of Medicine, Unit of Enteroscopy, University of Copenhagen, Hellerup, 4Gastrounit, Copenhagen University Hospital Herlev, 5Gubra Aps, Hørsholm, Denmark, 6Gastro Unit, Copenhagen University Hospital Hvidovre, Denmark. Background and aims: RYGB is a weight-reducing surgical procedure, which changes macroscopic gut anatomy and the route of ingested nutrients and secretory products through the gastrointestinal tract, resulting in weight loss and improved glucose homeostasis in the majority of subjects. The aim of this study was to describe the impact of Roux-en-Y gastric bypass (RYGB) on the distribution of small intestinal enteroendocrine cells and the expres- sion of their hormonal products in obese patients with type 2 diabetes and non-diabetic controls. Materials and methods: Twelve obese subjects with type 2 diabetes and 11 age and BMI matched controls underwent RYGB and enteroscopy at least 4 months later. Mucosa biopsies from both procedures were immunohis- tochemically stained for cholecystokinin (CCK), glucose-dependent insu- linotropic polypeptide (GIP), glucagon-like peptide-1 (GLP-1) and peptide YY (PYY) and gene expression analyses were performed looking at ghrelin (GHRL), CCK, secretin, GIP, GLP-1, PYY, neurotensin and farnesoid X re- ceptor (FXR). Results: Mean time (±SEM) of enteroscopy after RYGB was 10.4±0.8 months. Immunohistochemically there was an increase in the density of cells stain- ing positive for CCK, GIP, GLP-1 and PYY (controls only) after vs. before RYGB. The gene expression of GHRL, secretin and GIP mRNA was down- regulated after RYGB. CCK, PYY, neurotensin and FXR mRNA gene expres- sion was unaltered after RYGB whereas GLP-1 mRNA was up-regulated in both groups. Conclusion: Numerous alterations in the distribution of enteroendocrine cells and their expression of hormonal products is seen after RYGB. These changes may play an important role in the metabolic improvements seen after RYGB. Supported by: Novo Nordisk Foundation 520 Insulin-like peptide 5 is an orexigenic gastro-intestinal hormone secreted from colonic L-cells F. Reimann1, H. Heffron2, P. Richards1, J. Powell2, M. Binda3, J. Doran2, Y. Toyoda2, J.D. Wade4, S. Aparicio5, M. Carlton2, A.P. Coll1, S. O‘Rahilly1, F.M. Gribble1, J. Grosse2; 1University of Cambridge, 2Takeda, Cambridge, UK, 3Frimorfo Ltd, Fribourg, Switzerland, 4University of Melbourne, Australia, 5BCCAB, Vancouver, Canada. Background and aims: Intestinal hormones secreted from enteroendocrine L-cells orchestrate the fate of ingested nutrients. Glucagon-like peptide-1 (GLP-1) acts as an incretin, thus boosting postprandial insulin secretion, and both GLP-1 and peptideYY (PYY) inhibit further food intake. Using a transgenic mouse model fluorescently labelling proglucagon expressing cells, we identified insulin-like peptide-5 (Insl5) as another hormone expressed in colonic L-cells, and characterised its role in murine physiology. Materials and methods: Expression of Insl5 was investigated by RT-PCR and immunohisto-chemistry. Plasma levels of Insl5 after fasting and re-feeding were assessed by ELISA (Kamiya). Wild-type and newly generated Insl5-re- ceptor (relaxin/insulin-like family peptide receptor-4) knock-out (Rxfp4-/-) mice, accustomed to fasting and/or i.p. injections, were injected with Isnl5 (either from Phoenix or chemically synthesised in Melbourne) or polyclonal Insl5 antibodies (Phoenix) and assessed for food intake. All animal studies were in accordance with the UK Home Office legislation and approved by the appropriate ethical committee. Results: RT-PCR detected specific Insl5 expression in colonic, but not duo- denal/jejunal L-cells and immunofluorescence microscopy confirmed co-ex- pression with proglucagon in an overlapping set of vesicles. In spite of a previ- ous report of reduced insulin responses in Insl5-/- mice, we failed to detect significant effects of Insl5 on insulin secretion from isolated islets, and Rxfp4- /- mice did not exhibit impaired glucose tolerance. However, we observed that over-night fasted mice had elevated plasma Insl5 levels (4.2 ± 0.5 pg/ml) which fell after 10 hrs of re-feeding (2.7 ± 0.2 pg/ml, n = 10). This effect was pronounced in mice restricted to 60% caloric intake for 2 weeks ( 32.7 ± 6.4 pg/ml fasting; 2. 5 ± 0.2 pg/ml 10 h refed, n=10) but blunted in mice receiving a 45% high-fat diet, consistent with the finding that colonic Insl5 mRNA ex- pression was elevated in the calorie restricted group. Over-night fasted mice injected with Insl5 (0, 8, 40, 200 or 1000 ng/25g body weight) showed a dose dependent increase in food intake (225 ± 25 mg chow/20 min at the highest dose compared to 13 ± 12.5 mg chow/20 min in the placebo group). This effect was absent in Rxfp4-/- mice. Tail vein injection with polyclonal Insl5 antibodies blunted re-feeding responses in fasted mice (5.8 ± 0.2 g cumula- tive food intake/24 h at 100 μg Insl5-IgG/mouse compared to 7.1 ± 0.1 g after injection of an unspecific control IgG, n= 5 respectively) and this effect was also absent in Rxfp4-/- mice. Conclusion: These results are consistent with Insl5 acting as an orexigenic peptide secreted from the distal intestine, joining the only other established orexigenic gut-peptide, ghrelin, which is secreted from the stomach. Co- expression in a subset of anorexigenic peptide (GLP-1 and PYY) secreting L-cells is surprising and requires further investigations. Supported by: Wellcome Trust, MRC and Takeda Cambridge 521 Mechanism and physiological roles of glucose sensing in enteroendocrine cells E. Lee, T. Miki; Department of Medical Physiology, Chiba University, Graduate School of Medicine, Japan. Background and aims: Glucagon-like peptide 1 (GLP-1) is secreted from enteroendocrine L-cells in response to oral glucose ingestion and plays an important role in dampening the rise in postprandial blood glucose levels. However, the mechanism of glucose sensing in the L-cells and its physio- logical consequence is not fully understood. In this study, the mechanism of glucose sensing in the L-cells and the physiological importance of luminal glucose sensing in the maintenance of glucose homeostasis was examined in mice in vivo. Materials and methods: C57BL/6 mice were administrated with carbohy- drates (glucose or maltose) together with or without an antagonist of SGLT1 (phloridzin) or GLUT2 (phloretin) and the portal blood was collected and subjected to GLP-1 assay. The activation of enteroendocrine cells and neu- rons in the brainstem was detected by immunostaining of phospho-CaMK2. In addition, glucose-induced suppression of gluconeogenesis was evaluated by both oral (p.o.) and intraperitoneal (i.p.) administration. Results: Significant GLP-1 secretion was induced by intraduodenal glu- cose administration at 5 min after infusion. The secretion was significantly blocked by co-administration of phloridzin but not by phloretin. I.p. glucose administration failed to trigger GLP-1 secretion. While plasma GLP-1 con- centration was not increased at 30 min after oral glucose administration, it was significantly increased by oral maltose. Interestingly, GLP-1 secretion at 30 min after maltose administration was not suppressed by phloridzin, sug- gesting that glucose-sensing mechanism in L-cells differs between early and late phases after stimulation. Immunostaining of duodenum revealed that en- teroendocrine cells and small numbers of nonepitherial cells in the villi were detected positive for phospho-CaMK2 at 5 min after intraduodenal glucose administration. Similarly, several cells in the medulla oblongata were shown to be activated by intraduodenal glucose administration. We also examined the gluconeogenesis and STAT3 phosphorylation in the liver in response to either p.o. or i.p. glucose administration. Interestingly, the suppression of glu- coneogenesis and induction of STAT3 phosphorylation by p.o. administra- tion was significantly potent than that by i.p administration. Conclusion: Glucose sensing in L-cells is dependent on SGLT1-mediated signaling in the early phase but becomes independent in the late phase after luminal glucose stimulation. Our data suggested that glucose in the gut lu- men first activates enteroendocrine cells and then the information is trans- ferred to several nuclei in the brain stem, influencing the glucose metabolism in the liver. Supported by: JSPS Diabetologia (2014) 57:[Suppl1]S1–S564 S 218 1 C 522 Free fatty acid receptor GPR120 is highly expressed in enteroendocrine K cells of upper small intestine and has a critical role in GIP secretion after fat ingestion K. Iwasaki1, N. Harada1, K. Sasaki1, S. Yamane1, K. Iida2, K. Suzuki1, A. Hamasaki1, E. Joo1, D. Nasteska1, K. Shibue1, T. Harada1, A. Hirasawa2, N. Inagaki1; 1Department of Diabetes, Endocrinology and Nutrition, Graduate School of Medicine, 2Department of Genomic Drug Discovery Science, Graduate School of Pharmaceutical Sciences, Kyoto University, Japan. Background and aims: Gastric inhibitory polypeptide (GIP) is an incretin secreted from enteroendocrine K-cells in response to fat and glucose inges- tion. Recently, free fatty acid receptor (FFAR) GPR120 was identified as a “li- pid sensor” and was involved in GLP-1 secretion. However, GPR120 expres- sion and role in K-cells remain unclear. In this study, we elucidated GPR120 gene expression in K-cells using GIP-GFP knock-in (GIP-GFP) mice in which K-cells can be visualized by GFP fluorescence. Furthermore, we clari- fied the effects of GPR120 on GIP secretion using GPR120-defficient mice (GPR120-/-) and GPR120 antagonist. Materials and methods: The number of GFP-positive cells was measured by flow cytometry analysis and immunohistochemistry. GFP-positive cells and GFP-negative cells were collected as K-cells and non-K-cells, respectively, from GI tract of GIP-GFP mice by flow cytometer. GIP content in GFP-posi- tive and GFP-negative cells were measured by ELISA. GIP mRNA and FFAR mRNA expressions (GPR40, GPR41, GPR43, GPR119, and GPR120) in GFP- positive and GFP-negative cells were assessed by sq RT-PCR. Oral glucose tolerance tests (OGTTs) and lard tolerance tests (OLTTs) were performed using GPR120-/- mice and wild-type (WT) mice to evaluate total GIP secre- tion. After oral or intravenous administration of GPR120 antagonist in C57/ BL6 mice, total GIP levels during OLTTs were measured. Results: GFP-positive cells were observed in small intestine, but not in stom- ach and colon. K-cell number and GIP content and mRNA expression in K- cells were significantly higher in upper small intestine than those in lower small intestine. GPR120 mRNA was highly expressed in K-cells (but not in GFP-negative cells) of upper small intestine, while GPR40 mRNA and GPR41 mRNA were highly expressed in K-cells (but not in GFP-negative cells) of lower small intestine. GPR120-/- mice had lower GIP secretion (75% reduc- tion) during OLTT compared to WT mice, but not during OGTT. Oral ad- ministration of GPR120 antagonist significantly attenuated lard oil-induced GIP secretion. On the other hand, intravenous administration of GPR120 antagonist did not reduce GIP levels after lard oil ingestion. Conclusion: We confirmed that GPR120 is substantially expressed in K cells of the upper small intestine. Our results using GPR120-/- mice and a GPR120 antagonist demonstrate that GPR120 plays a critical role in lipid-induced GIP secretion from K-cells population located mainly in upper small intestine. 523 CART is a regulator of GIP and GLP-1 expression and secretion in vitro L. Shcherbina, N. Wierup; Lund University Diabetes Centre, Lund University, Malmö, Sweden. Background and aims: Cocaine- and amphetamine-regulated transcript (CART) is a regulatory peptide that controls islet hormone secretion and beta-cell survival. We have recently shown that CART is expressed in hu- man enteroendocrine cells, including L- and K-cells in the human duodenum and jejunum. Furthermore, we have shown that CART plasma levels increase after a meal in humans. In the present study we aimed to examine whether endogenous L- and K-cell CART regulates incretin hormone secretion and expression. Materials and methods: CART gene expression was silenced using siRNA in GLUTag and STC-1 cells, used as L- and K-cell models respectively. Gene ex- pression was assessed with qPCR, protein content and secretion with ELISA, and cell survival using WST-1 assay. Results: CART silencing resulted in a 84.6±5.9% reduction in CART mRNA levels in GLUTag cells (p S 219 1 C 525 Effect of acute GIP- and GLP-1- infusion on circulating levels of pro- atrial natriuretic peptide in subjects with different stages of glucose tolerance N.N. Rudovich1, A. Sparwasser2, W. Doehner3, O. Pivovarova1, A. Arafat4, A. Bergmann5, M. Nauck6, A.F.H. Pfeiffer1; 1Clinical Nutrition, DIFE Potsdam, Nuthetal, 2Thermo Fisher Scientific, B.R.A.H.M.S GmbH, Biotechnology Centre Hennigsdorf, 3Interdisciplinary stroke research, 4Campus Mitte, Charité University Medicine, 5Sphingotec GmbH, Biotechnology Centre Hennigsdorf, Berlin, 6Diabeteszentrum Bad Lauterberg, Germany. Background and aims: Glucagon-like peptide-1 (GLP-1) receptor agonist acts antihypertensive via release of atrial natriuretic peptide in mice. Moreo- ver, reduction of glycemic load by alpha-glucosidase inhibitors increase circulating levels of pro-atrial natriuretic peptide (proANP) suggesting an evidence of gut-heart axis also in humans. Whether GLP-1 and glucose-de- pendent insulinotropic peptide (GIP) interact with NP system in humans is unknown. Materials and methods: Twelve patients with type 2 diabetes (T2DM) (nine men and three women; 61 ± 10 years; BMI 30.0 ± 3.7 kg/m²; HbA(1c) 7.3 ± 1.5%) were studied. In randomized order, intravenous infusions of GLP-1(7- 36)-amide (1.2 pmol • kg⁻¹ • min⁻¹), GIP (4 pmol • kg⁻¹ • min⁻¹), GLP-1 plus GIP, and placebo were administered over 360 min after an overnight fast (≥ 1 day wash-out period between experiments). Additionally, effect of 240 min GIP-infusion (2 pmol • kg⁻¹ • min⁻¹) on proANP was studied in glucose- tolerant male subjects (n=8). Plasma proANP concentrations were measured by an automated mid-region-directed proANP immunoassay. Results: GLP-1 decreased proANP levels (mean± sem, 44.1 ± 3.1 vs. 34.2 ± 3.0 pmol/l; p < 0.05). During placebo, GIP and GIP + GLP-1 infusion no significant changes of circulating proANP were observed in T2DM subjects. In glucose tolerant subjects, proANP levels are decreased at the end of the ex- periments with placebo and GIP (51.9 ± 6.5 vs. 40.0 ± 5.4 pmol/l in placebo, p < 0.05; 49.5 ± 6.3 vs. 40.8 ± 6.5 pmol/l in GIP-tests, p = 0.051). Conclusion: As proANP concentration reflects ANP secretion, our data could not confirm interactions between two major incretins and ANP in hu- mans. Clinical Trial Registration Number: NCT00774488 526 Novel GIP receptor-mediated bioassay more accurately reflects changes of GIP activity than traditional assays T. Yanagimachi, Y. Fujita, Y. Takeda, J. Honjo, H. Kitsunai, H. Sakagami, Y. Takiyama, A. Abiko, Y. Makino, M. Haneda; Division of Metabolism and Biosystemic Science, Department of Internal Medicine, Asahikawa Medical University, Japan. Background and aims: GIP and GLP-1 are incretins released from the gut rapidly inactivated by dipeptidyl peptidase-4 (DPP-4) after secretion into truncated forms that are no longer insulinotropic forms of incretins. Al- though recently developed ELISA kits may be able to measure active incretins levels, it is unclear if the available assays detect totally biological active forms. Moreover, recent reports suggest GIP has a short-form isoform GIP (1-30) secreted from gut and pancreas, which may not be detectable by the available methods. Therefore, we determined active GIP and GLP-1 levels from human blood samples using novel cell-based, receptor-mediated bioassays. Materials and methods: We utilized the cell lines stably co-transfected with human-form GIP or GLP-1 receptors and a cAMP-inducible luciferase ex- pression construct for bioassays. We performed a 75 g OGTT in six non- diabetic subjects and measured plasma total GIP (Millipore), total GLP-1 (Millipore), insulin, C-peptide and glucagon levels by ELISA, active GIP and active GLP-1 levels by bioassays. Next, we performed another OGTT in iden- tical subjects after administration of DPP-4 inhibitor (sitagliptin: 100 mg/ day) for three days. We measured plasma total/active GIP and GLP-1 levels as above. Additionally, we compared the efficacy of our GIP bioassay with that of a commercial active GIP ELISA kit (IBL, Japan) by measuring samples from identical subjects before and after sitagliptin administration. Results: To evaluate the specificity of the bioassays, we confirmed their re- sponsiveness with several synthetic incretin receptor agonists. In the GIP bio- assay, GIP (1-42) and short-form GIP (1-30) almost equivalently increased luciferase activity in a concentration dependent manner. In the GLP-1 bio- assay, exendin-4 produced almost comparable luciferase activity to GLP-1 (7-36 amide) but liraglutide did less than GLP-1 (7-36 amide). In samples collected during on OGTT, total GIP levels rapidly increased from 0 min to 15 min and gradually increased up to 120 min (68.8 ± 11.4 pmol/l). In con- trast, active GIP levels peaked at 30 min (42.3 ± 4.3 pmol/l) and dropped at 60 min. Notably, active GIP had a positive correlation with the insulin levels. Total GLP-1 levels increased to 30 min and remained at similar levels up to 120 min (25.4 ± 2.7 pmol/l). In contrast, active GLP-1 levels reached a peak at 30 min and then dropped. Administration of a DPP-4 inhibitor to identi- cal subjects induced 1.4-fold increase of active GLP-1 levels at 15 minutes after oral glucose load. Interestingly, active GIP measured by bioassay levels increased approximately by 20-fold after DPP-4 treatment. But total and ac- tive GIP levels by ELISA were increased just 1.7 and 2.1-fold, respectively. Other DPP-4 inhibitors (linagliptin: 5 mg/day and vildagliptin: 100 mg/day) showed similar effect to sitagliptin. Then we investigated the discrepancy in measured active GIP levels between bioassay and ELISA. We first confirmed that Sitagliptin per se did not increase luciferase activity. The bioassay could detect short-form GIP (1-30), but total or active ELISA kits could not at all. Conclusion: We propose that the bioassays more accurately reflect the con- tribution of the incretins in the entero-insular axis and short-form GIP (1-30) may be more important than GIP (1-42) in DPP-4 inhibitor treatment. Supported by: JDF Diabetologia (2014) 57:[Suppl1]S1–S564 S 220 1 C PS 030 Gut endocrinology in vivo 527 Effects of a non-glucose food preload on oral glucose tolerance in subjects with impaired glucose tolerance and type 2 diabetes D. Trico1,2, A. Tulipani1, S. Baldi1, C. Morgantini2, M.P. Macedo3, A. Mari4, E. Ferrannini1, A. Natali1; 1Department of Clinical and Experimental Medicine, University of Pisa, 2Scuola Superiore Sant‘Anna, Pisa, 3Centro de Estudos de Doenças Crónicas, Universidade Novade Lisboa, 4Institute of Biomedical Engineering, National Research Council, Padova, Italy. Background and aims: Protein and fat ingestion delays gastric emptying and stimulates gut hormones release. Recent studies have suggested that these ef- fects could be exploited to improve the plasma glucose excursions induced by carbohydrates in type 2 diabetic patients (T2D). The present study was designed to measure the impact of a small non-glucose food preload on the major components of glucose homeostasis in subjects with impaired glucose tolerance (IGT) and T2D and to test whether gut hormones participate in this response. Materials and methods: Ten diet-controlled T2D patients (8 males, age 55±7 years, BMI 28.7±4.7 kg/m2, HbA1c 48±2 mmol/l) and thirteen IGT subjects (7 males, age 47±5 years, BMI 26.3±1.6 kg/m2, HbA1c 38±1 mmol/l) were enrolled. On two separate days, after an overnight fast, they were randomised to a preload of either 500 ml of water (control) or Parmesan cheese (50 g) plus a boiled egg with 300 ml of water; 30 minutes later they underwent a standard 75 gr oral glucose tolerance test (OGTT). Arterialised timed blood samples were collected to measure plasma glucose, insulin, C-peptide, GLP-1 and GIP. Two stable glucose tracers were administrated ([6,6-2H2]glucose and [U-13C]glucose, i.v. and per os respectively) to measure the rate of appear- ance of ingested (RaO) and endogenous glucose and the glucose clearance. Three major components of β-cell function, namely glucose sensitivity (βGS), rate sensitivity (βRS) and potentiation (POT) were evaluated by modeling insulin secretion (estimated from C-peptide deconvolution) and plasma glu- cose. Insulin sensitivity was estimated using the OGIS method. Results: In T2D patients, the plasma glucose rise during the OGTT was significantly lower after the food preload than after water ingestion (iAUC -49.4±5.9%, p S 221 1 C across glucose tolerance (GT) groups. In the same subjects we report total GLP-1 responses to i.v. arg in 53 obese subjects with normal glucose tolerance (NGT), prediabetes (PDM), and type 2 diabetes (T2DM). Materials and methods: Following an overnight fast, samples were acquire for total GLP-1 pre- and for 10 min post an i.v. arg bolus (5 gm over 30 sec) at basal glucose. Pre-and post-arg GLP-1 samples were also acquired during the last 10 min of a 60 min glucose infusion (900 mg/min). Results: The table summarizes results during basal glucose period; previously reported and baseline adjusted AIRarg insulin responses are also included. Pre-arg fasting GLP-1 differed across GT spectrum, highest in T2DM. ARG elicited GLP-1 secretion in all 3 GTs, with greater changes in PDM and T2DM than NGT (Delta GLP-1). Pre-arg GLP-1 did not correlate with AIRarg, whereas change in GLP-1 correlated with AIRarg for NGT and T2DM but not PDM in basal glucose state (NGT/T2DM: r= 0.52/0.49 P S 222 1 C 532 Gallbladder emptying and single-dose metformin elicit robust and additive glucagon-like peptide-1 responses U. Rohde1, D.P. Sonne1, M. Hansen1, A. Brønden1, M. Christensen1, J.F. Rehfeld2, J.J. Holst3, T. Vilsbøll1, F.K. Knop1; 1Diabetes Research Division, Gentofte Hospital, Hellerup, 2Department of Clinical Biochemistry, Rigshospitalet, 3Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Denmark. Background and aims: Preclinical studies suggest that gallbladder emptying and subsequent activation of the bile acid receptor TGR5 on enteroendocrine L cells leads to glucagon-like peptide-1 (GLP-1) secretion. Drugs affecting bile acid binding (colesevelam (COL)) or reabsorption (metformin (MET)) seem to increase postprandial GLP-1 secretion in humans. We hypothesised that gallbladder emptying stimulates human GLP-1 secretion and that COL and MET, respectively, would potentiate any GLP-1 secretion induced by gallbladder emptying. Materials and methods: Ten subjects (age (mean±SD): 23.4±3.8 years; BMI: 21.9±1.8 kg/m2; HbA1c: 5.1±0.3%) were studied on 6 randomised days. In a double-blind fashion the subjects received 1) COL (3.75 g); 2) MET (1.5 g); or 3) placebo (PLA) in 50 ml water admixed 1.5 g paracetamol (for evaluation of gastric emptying) administered via nasogastric tube, with a concomitant 60-minute iv infusion of saline and cholecystokinin-8 (CCK), respectively. Blood was sampled for 4 hours for measurements of plasma GLP-1, glucose, insulin, C-peptide and glucagon. Gallbladder emptying was measured by ul- trasound. Food intake was assessed at the end of each day. Results: CCK infusion during PLA induced complete gallbladder emptying and a significant GLP-1 response (incremental AUC) compared to saline in- fusion (392±173 (mean±SEM) vs. -277±94 pM×min, p=0.02). MET without CCK elicited a significant GLP-1 response (215±87 vs. -277±94 pM×min (saline+PLA), p=0.002), which was potentiated by CCK-induced gallblad- der emptying (963±202 pM×min (CCK+MET), p=0.03). COL did not elicit significant GLP-1 responses. Plasma glucose was not affected by the interven- tions, nor was insulin, C-peptide, glucagon or food intake. Conclusion: CCK-induced gallbladder emptying and single dose metform- in, respectively, elicit robust and additive GLP-1 responses in humans. We, therefore, speculate that metformins mode of action includes stimulation of GLP-1 secretion by both bile acid-dependent and independent mechanisms. 533 Effect of the bile acid chenodeoxycholic acid and the bile acid sequestrant colesevelam on glucose metabolism in patients with type 2 diabetes and in healthy control subjects M. Hansen1,2, M. Scheltema1,3, D.P. Sonne1,2, J.F. Rehfeld4, J.J. Holst2,5, T. Vilsbøll1, F.K. Knop1,2; 1Department of Internal Medicine, Gentofte Hospital, Diabetes Research Division, Hellerup, 2University of Copenhagen, Department of Biomedical Sciences, Denmark, 3University of Amsterdam, Department of Endocrinology and Metabolism, Netherlands, 4Rigshospitalet, Department of Clinical Biochemistry, Copenhagen, 5University of Copenhagen, Novo Nordisk Foundation Center for Metabolic Research, Denmark. Background and aims: In patients with type 2 diabetes, rectal administra- tion of bile acids, and oral bile acid sequestrants lower plasma glucose (PG). We evaluated the effects of chenodeoxycholic acid (CDCA) and the bile acid sequestrant colesevelam (COL), delivered by intragastric tube, on glucagon- like peptide-1 (GLP-1) (primary endpoint), PG, insulin, C-peptide, glucagon, cholecystokinin (CCK), gastric emptying, gallbladder volume, appetite and food intake, in a placebo-controlled, double-blinded study. Materials and methods: On 4 separate days 10 patients with type 2 diabetes (age (mean±SD): 62±7 years, BMI: 28.9±2.3 kg/m2; HbA1c: 53±12 mmol/ mol) and 10 matched healthy control subjects (age: 61±10 years, BMI: 28.2±3.2 kg/m2; HbA1c: 37±4 mmol/mol) received 1) CDCA (1.25 g); 2) COL (3.75 g), 3) CDCA+COL, or 4) placebo; suspended in 100 ml of water with 1.5 g paracetamol (for evaluation of gastric emptying). During 180 min blood was drawn, gallbladder volume was evaluated by ultrasound, and ap- petite perception was evaluated by visual analogue scale. At the end of each day ad libitum food intake was evaluated. Results: In both the patients with type 2 diabetes and the healthy subjects, CDCA elicited a significant increase in GLP-1 vs. COL, CDCA+COL and pla- cebo (p S 223 1 C PS 031 Gut hormone action 535 Inhibition of GIP receptor signalling in adipose tissue reduces insulin resistance in high fat diet-induced obesity E. Joo, N. Harada, T. Fukushima, A. Hamasaki, S. Yamane, K. Iwasaki, D. Nasteska, K. Shibue, T. Harada, K. Suzuki, N. Inagaki; Department of Diabetes, Endocrinology and Nutrition, Graduate School of Medicine, Kyoto University, Japan. Background and aims: Gastric inhibitory polypeptide (GIP) is an incretin that potentiates insulin secretion. GIP receptor (GIPR) is expressed not only in β-cells, but in adipose tissue as well, and previous in vitro studies show that GIP directly induces energy accumulation into the adipose tissue. Therefore, GIP has a direct and indirect effect (via insulin) on induction of adiposity. However, the effect of direct GIP action on adiposity in vivo remains un- clear. In this study, we generated adipose tissue-specific GIPR-deficient mice (GIPRfat-/-) and clarified the direct GIP action in adipose tissue in vivo. Materials and methods: GIPRfat-/- were generated from floxed GIPR mice (Lox) and adipocyte protein 2 (aP2)-Cre transgenic mice (aP2). GIPRfat-/-, Lox, and aP2 were fed normal fat diet (NFD: 10% fat) or high fat diet (HFD: 60% fat) and their body weight gain was measured. After 15 weeks of NFD or HFD feeding, oral glucose tolerance test (OGTT), insulin tolerance test (ITT), and histological analyses (liver and adipose tissue) were performed. Fat volume (CT scan), energy expenditure, activity count, and triglyceride (TG) content in liver were measured. Results: In GIPRfat-/-, expression levels of GIPR mRNA were substantially lower in visceral and subcutaneous adipose tissue (90% reduction). Body weight gain, blood glucose levels, and insulin levels during OGTT did not differ among three types of mice under NFD condition. Under HFD condi- tion, body weight gain was slightly lower in GIPRfat-/- compared to that of control mice (Lox and aP2). Fat volume and adipocyte size did not differ be- tween HFD-fed GIPRfat-/- and control mice. OGTT data showed that insulin and glucose levels were significantly lower in HFD-fed GIPRfat-/- compared to control mice. HOMA-IR and ITT data showed that insulin resistance was reduced in HFD-fed GIPRfat-/-. Furthermore, fat content in liver was signifi- cantly lower in HFD-fed GIPRfat-/- compared to control mice. Histological analysis showed that hepatic steatosis was reduced in HFD-fed GIPRfat-/-. There was no significant difference in energy expenditure and mouse activity among three types of mice under HFD feeding. Conclusion: GIP receptor signaling in the adipose tissue plays an important role in HFD-induced insulin resistance without increase in fat mass in vivo. 536 The long-acting analogue of short-form GIP does not induce obesity in normal mice but ameliorates chronic hyperglycaemia in low-dose streptozotocin-induced diabetic mice Y. Fujita, T. Yanagimachi, K.K. Atageldieyeva, J. Honjo, Y. Takiyama, A. Abiko, Y. Makino, M. Haneda; Department of Internal Medicine, Asahikawa Medical University, Japan. Background and aims: It is still ambiguous whether GIP can be beneficial as potential therapeutics in the diabetic state. However, we reported that a short-isoform GIP, GIP (1-30), which is released chiefly from pancreatic al- pha cells, may function not only as glucose-dependent insulin release in a paracrine manner, but also as the fetal islet development. Again, it is previ- ously reported that GIP (1-30) less potentiates fat accumulation in contrast to GIP (1-42) secreted from the gut. To investigate the possibility of short-form GIP, we tested the efficacy of a long-acting GIP (1-30) analogue in vivo and investigated whether the analogue can suppress the progression to hypergly- cemia and improve glycaemia in low-dose streptozotocin (LD-STZ) mice. Materials and methods: We synthesized dipeptidyl peptidase-4 (DPP-4) resistant GIP (1-30) analogue via modification of the second position ala- nin and then attaching 40kDa polyethylene glycol (PEG) to the C-terminus. First, we administrated PEGylated (350-7000 pmol/body) or non-PEGylated GIP to the non-diabetic control mice by the subcutaneous single injection, and evaluated pharmacokinetics of the peptides from 60 min until day 7. To measure plasma GIP activity, we employed the novel receptor-mediated bioassay reflecting cAMP production in vitro. Then, we used non-diabetic and LD-STZ-induced diabetic mice, with or without administration of the PEGylated GIP analogue. We designated mice to four groups, non-STZ, STZ, non-STZ+GIP, STZ+GIP, respectively. We injected the PEGylated GIP analogue or vehicle subcutaneously every 3 days (initial dose 100 pmol/body and maintenance dose 70 pmol/body) and monitored body weight and non- fasted blood glucose until day 30. We measured HbA1c levels and conducted intraperitoneal glucose tolerance tests (IPGTT) at the end. Finally, we evalu- ated the islet morphology via double immunofluorescent staining with insu- lin and glucagon. Results: The single subcutaneous injection of PEGylated GIP analogue gradually increased GIP activity in plasma but maintained much longer (the approximate half-life was 48 hours), whilst the injection of non-PEGylated analogue rapidly increased the activity but vanished within 2 hours. The maximum-dose injection of PEGylated GIP still augmented to detect bio- logical GIP activity until day 7. Persistent PEGylated GIP treatment neither induced obesity nor affected glycemic levels in non-STZ groups. In contrast, PEGylated GIP treatment significantly decreased non-fasted blood glucose levels, reduced HbA1c levels (STZ+GIP: 4.9±0.2% vs STZ: 5.7±0.4 %, p S 224 1 C 538 Effects of exogenous glucagon-like peptide-1 on the blood pressure, heart rate, mesenteric blood flow and glycaemic responses to intraduodenal glucose in older subjects L.G. Trahair1,2, M. Horowitz1,2, T. Hausken3, C. Feinle-Bisset1,2, C.K. Rayner1,2, K.L. Jones1,2; 1Discipline of Medicine, The University of Adelaide, 2NHMRC Centre of Research Excellence in Translating Nutritional Science to Good Health, The University of Adelaide, Australia, 3Section for Gastroenterology, Department of Clinical Medicine, University of Bergen, Norway. Background and aims: A postprandial fall in blood pressure (BP) occurs frequently in healthy older subjects and type 2 patients, the magnitude of which is related to the rate of gastric emptying (GE) and rise in splanchnic blood flow. Postprandial hypotension (PPH) is now recognised as an impor- tant clinical problem in these groups. In humans, exogenous glucagon-like peptide-1 (GLP-1) has been reported to increase BP in two studies, but not another, while clinical trials of GLP-1 agonists in type 2 diabetes and obe- sity have reported a modest reduction in BP and a rise in heart rate (HR). Studies relating to the cardiovascular effects of GLP-1 and its agonists have not, however, discriminated between fasting, as opposed to postprandial, BP and HR. The aims of this study were to determine whether exogenous GLP-1 modulates the effects of an intraduodenal (ID) glucose infusion on BP, HR and splanchnic blood flow in healthy older subjects. Materials and methods: On two separate days, 10 healthy ‘older’ subjects (9M,1F; age 73 ± 2 yr) received an infusion of GLP-1 (0.9 pmol/kg/min IV), or saline for 90 min (t = -30 - 60 min) in random order. Between t = 0 - 60 min, ID glucose was infused at 3 kcal/min. BP and HR were assessed with an automated device (Dinamap, GE Medical Systems), superior mesenteric artery (SMA) flow by Doppler ultrasonography and blood glucose and serum insulin were also measured. Results are shown (mean ± SEM). Results: During the ‘fasting’ period (t = -30 - 0 min), GLP-1 had no effect on BP (P = 0.40) or HR (P = 0.59). In response to ID glucose (t = 0 - 60 min), systolic BP decreased (P < 0.001; Figure 1A), and both HR (P < 0.001; Figure 1B) and SMA flow (P < 0.05; Figure 1C) increased, on both days. GLP-1 at- tenuated the maximum fall in systolic BP (saline: -14 ± 3 vs. GLP-1: -9 ± 2 mmHg; P < 0.05), tended to increase the maximum rise in HR (saline: 13 ± 3 vs. GLP-1: 15 ± 3 BPM; P = 0.09) and increased the maximum rise in SMA flow (saline: 500 ± 52 vs. GLP-1: 837 ± 102 ml/min; P < 0.01). GLP-1 dimin- ished the peak glycaemic response to ID glucose (saline: 10.3 ± 0.5 vs. GLP-1: 9.3 ± 0.3 mmol/L; P < 0.05) and increased the AUC 0 - 60 min insulin/glucose ratio (saline: 3.5 ± 0.7 vs. GLP-1: 4.9 ± 0.9; P < 0.05). Conclusion: In healthy older subjects, acute administration of GLP-1 at- tenuates the hypotensive response to ID glucose, apparently independent of changes in HR, and potentiates the increase in SMA flow. Given that GLP-1 also slows GE, a role in the management of PPH warrants evaluation. Supported by: NHMRC #627189 539 Impaired incretin effect and gastrointestinal-mediated glucose disposal in non-diabetic patients with cirrhosis F.K. Knop1, L. Gluud2, A.E. Junker1, J.J. Holst3, T. Vilsbøll1; 1Diabetes Research Division, Gentofte, 2Department of Gastroenterology, Hvidovre, 3Department of Biomedical Science, Copenhagen, Denmark. Background and aims: Cirrhosis is often complicated by glucose intolerance, but the pathogenesis not yet completely understood. We evaluated the in- cretin effect and gastrointestinal-mediated glucose disposal (GIGD) in non- diabetic patients with biopsy-verified liver cirrhosis. Materials and methods: Patients with compensated cirrhosis (Child Pugh A or B) and matched healthy controls underwent a 4h 50g-oral glucose toler- ance test (OGTT) and an isoglycaemic intravenous glucose infusion (IIGI) on two separate days. We calculated the incretin effect [100% × (AUCC- peptide, OGTT - AUCC-peptide,IIGI/AUCC-peptide,OGTT)], the GIGD [100% × (glucoseOGTT - glucoseIIGI/glucoseOGTT)] and insulin resistance (according to the homeostatic model assessment (HOMAIR)). Characteris- tics of participants are summarised as median±interquartile range and com- pared using non-parametric analyses. Results: Ten patients (5 women) with cirrhosis (age: 54±15 years; BMI: 26±6 kg/m2; fasting plasma glucose (FPG): 5.8±0.7 mM; HbA1c: 38±6 mmol/ mol (5.6±0.4%)) and 10 matched healthy controls (age: 58±17 years; BMI: 29±1 kg/m2; FPG: 5.2±0.4 mM; HbA1c: 34±6 mmol/mol (5.5±0.2%)) were included. Patients with cirrhosis were more glucose intolerant (AUCOGTT: 609±458 (cirrhosis) vs. 180±109 mM×min (controls), P S 225 1 C 541 Preserved incretin effect despite significant hyperinsulinaemia following glucose loading in South Asians, in comparison to Caucasians R. Ahluwalia1, T.L. Vilsboll2, L. Ranganath1, J.J. Meier3, J. Vora1, F.K. Knop2; 1The Royal Liverpool & Broadgreen University Hospitals NHS Trust, Liverpool, UK, 2Diabetes Research Division, Copenhagen University Hospital Gentofte, Denmark, 3Department of Medicine, St. Josef-Hospital, Ruhr-University, Bochum, Germany. Background and aims: Circulating hyperinsulinaemia in normal glucose tolerant South Asians is reflective of underlying insulin resistance. Role of incretin hormones in this adaptive islet response in South Asians, remains to be explored. Our study aims to investigate the incretin effect in matched groups of South Asians and Caucasians. Materials and methods: Ten South Asian subjects (age: 34±4 years (mean±SEM); BMI 24.5±1 kg/m2; fasting plasma glucose (FPG): 4.8±0.2 mmol/l;) and twelve age and BMI-matched Caucasian subjects (age: 31±3 years; BMI 24.6±1 kg/m2; FPG: 4.6±0.1 mmol/l,) with normal glucose toler- ance were recruited. On two different days, subjects underwent a 4h 50g- OGTT and an isoglycaemic i.v. glucose (20%) infusion (IGII), respectively, involving sampling of insulin, C-peptide, glucagon and incretin hormones. The incretin effect was measured using the following formula based on beta cell secretory responses (AUCs): 100x(AUCOGTT-AUCIGII)/AUCOGTT. Results: Fasting insulin levels were higher in South Asians vs. Caucasians (12.0±1.6 vs. 8.0±0.8 mU/l, p=0.02). The oral glucose curves were replicated successfully during the IGII studies. The integrated insulin responses after OGTT and IIGI, respectively, were significantly higher in South Asians vs. Caucasians (OGTT: 10239±1683 vs. 5377±616 mU/l x min, p=0.008; IGII: 5610±1347 vs. 2703±321 mU/l x min, p=0.03). Incretin effects were similar in the two groups (48±6 vs. 49±2%, p=0.8). Conclusion: Our findings show preserved incretin effect despite insulin re- sistance and hyperinsulinaemia in normal glucose tolerant South Asians. 542 Difference in insulin responsiveness between meal- and glucagon- loading test in Japanese type 2 diabetes T. Kitamoto1,2, E. Nara2, Y. Matsuzawa2, J. Saito2, M. Omura2, T. Nishikawa2; 1Clinical Cell Biology and Medicine, Chiba University, 2Endocrinology and Diabetes Center, Yokohama Rosai Hospital, Japan. Background and aims: Responsiveness of c-peptide (CPR) to mixed meal tolerance test (MT) is a useful marker to evaluate the insulin secretion ability such as postprandial responsiveness of insulin production in type 2 diabetes. On the other hand, the total capacity of endogenous insulin secretion can be examined by the response of CPR after glucagon loading test (GT). Basal ability of insulin production was reported to be relatively impaired in Japa- nese type 2 diabetes, rather than Western patients, suggesting that incretin- related agents can be much more effective for glycemic control in Japanese than Western patients. Then, we tried to compare the responsiveness of CPR to MT and GT in Japanese type 2 diabetes for clarifying effectiveness of incretin-based therapy, such as DPP4 inhibitor (DPP4i) or GLP-1 analog (GLP-1A). Materials and methods: We enrolled 95 Japanese type 2 diabetes patients prescribed DPP4i or GLP-1A from 2011 to 2012. CPR response (⊿CPR) to MT and GT was analyzed. Then, we divided the patients into 2 groups; controlled patients who reached glycemic target (HbA1c8.0% in DPP4i-treated group, while in GLP-1A-treated group, that of ⊿CPR by GT was over 1.39 and under 1.26 ng/ml in cases showing HbA1c8.0%, respectively. Moreover, 86% of un- controlled patients by DDP4i-treatment showed more than 1.39 in ⊿CPR in GT, suggesting that they can be successfully treated with GLP1-A. Conclusion: The present study clearly demonstrated that the gut-mediated insulin secretion estimated by MT can predict the efficacy of DPP4i, while total capacity of cAMP-mediated insulin secretion examined by GT may pre- dict usefulness of GLP-1A. We can, therefore, simply decide whether DPP4i or GLP-1A might be effective for controlling hyperglycemia in Japanese type 2 diabetes by performing MT and GT. Diabetologia (2014) 57:[Suppl1]S1–S564 S 226 1 C PS 032 Novel approaches for identifying and targeting determinants of glucose tolerance 543 Glypican-4 is increased in human subjects with impaired glucose tolerance and decreased in patients with newly diagnosed type 2 diabetes H. Wang1,2, L. Li1,2, K. Li3, G. Yang3; 1College of Laboratory Medicine, 2Key Laboratory of Diagnostic Medicine (Ministry of Education), 3Department of Endocrinology and Department of Surgery, Chongqing Medical University, China. Background and aims: Glypican-4(GPC-4), a membrane-bound hormone owing to a glycosyl-phosphatidylinositol (GPI) anchor, is synthesized and secreted by adipose tissues.GPC-4 differentially expressed in visceral and subcutaneous adipose tissue, and the expression in human white adipose tissue (WAT) was highly correlated with BMI and the WHR. However, no report has demonstrated the relationship of circulating levels of GPC-4 with impaired glucose tolerance (IGT) or T2DM in humans. The aim of the cur- rent study is to investigate whether GPC-4 correlates with obesity and insulin resistance by cross-sectional and interventional studies on anthropometric, metabolic, and hormonal predictors of circulating GPC-4. Materials and methods: Circulating GPC-4 was measured with ELISA in subjects with NGT, IGT, and nT2DM who are involved in the study from June 1, 2013 to November 10, 2013. EHC were performed in healthy and T2DM subjects. Plasma glucose was measured by the glucose oxidase method. HbA1c was measured by HPLC. Plasma insulin was measured by radioimmunoassay. Plasma FFA was measured with a commercial kit. To- tal cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglycerides were analyzed enzymatically using an autoana- lyzer.Real-time RT-PCR and Western blotting were used to assess mRNA and protein expression of GPC-4. Results: Circulating GPC-4 levels were significantly higher in IGT subjects and lower in nT2DM subjects compared to controls. Circulating GPC-4 was positively correlated with BMI, WHR, HOMA-IS and FAT%, while it was in- versely correlated with FBG and HbA1c. Excluding diabetic subjects, increas- ing GPC-4 levels were associated with HOMA-IR and M values(r = 0.247 and r = -0.491; both P S 227 1 C are generally called α-glycosidases, with α-amylase being primary among them. BTI320 is a novel, non-systemic therapy that safely reduces postpran- dial glucose excursions with reduced side effects compared with acarbose. Acarbose is a natural microbial pseudo-tetrasaccharide that binds revers- ibly and competitively to the oligosaccharide binding site of α-glucosidases. BTI320 is composed of non-glucose-containing polysaccharides. It is es- sentially a composite of two modified galactomannans: GM-α (1⇒1) linked polymer; and GM-β (1⇒4) linked polymer. We believe that BTI320 functions by targeting several polysaccharide hydrolyzing enzymes and that the active ingredient is GM-α. Materials and methods: The present study was focused on assessing the mo- lecular mechanism of action of BTI320 in relationship to α-amylase. We used 1H and 13C nuclear magnetic resonance (NMR) spectroscopy to investigate interactions between BTI320 components, GM-α and GM-β, and the enzyme α-amylase, as well as the effects that GM-α and GM-β have on the rates of amylase-mediated starch hydrolysis towards glucose. The amylose iodine as- say was also used to assess starch hydrolysis. Results are compared with those on acarbose. Results: Chemical shift changes in NMR spectra of α-amylase demonstrate that GM-α interacts with the enzyme, possibly at or near its active site; GM-β appears to have no effect on α-amylase. GM-α and GM-β both interact with starch and apparently change the amylose structure, thus affecting how am- ylase hydrolyses the starch. Under certain conditions, the rate constants for starch (1 mg/ml) hydrolysis with α-amylase goes from 22.5 s-1 in the absence of GM-α to 2.7 s-1 in the presence of 4 mg/ml GM-α (p S 228 1 C 548 The identification of novel metabolites that track with improvements in glycaemia following a 12-week lifestyle intervention in high risk individuals A. Kennedy1, J. Cobb2, G. O’Donoghue3, G. Anderson3, N. McCaffrey1, S. Cleary1, E. Durkan1, H. Kenny1, J. Monedero1, K.-P. Adam2, M. Sinnott4, B. Carr4, T. Thybo3, J. Nolan3, D. O’Gorman1; 1Centre of Preventive Medicine and School of Health & Human Performance, Dublin City University, Ireland, 2Metabolon Inc, Durham, USA, 3Steno Diabetes Centre, Copenhagen, Denmark, 4Vhi Healthcare, Dublin, Ireland. Background and aims: A number of studies have identifed fasting metabo- lites associated with insulin resistance, dysglycemia and type 2 diabetes. It is not known if metabolites respond to lifestyle modifications that reduce the risk of type 2 diabetes. The purpose of this study was to identify metabolite biomarkers, linked to dysglycaemia, that track with improvements in plasma glucose following a lifestyle intervention in high risk individuals. Materials and methods: As part of the DEXLIFE programme, individuals with a FINDRISC score >12 or impaired fasting/2-hr glucose volunteered to participate in a 12-week diet and exercise intervention (n=104). An oral glucose tolerance test, body composition assessment(DEXA)and maximal aerobic capacity were determined before and after the intervention. Using the stable isotope dilution technique, quantitative assays were developed for a set of 23 candidate biomarker metabolites previously linked to dysglycemia. This set included: α-hydroxybutyric acid (AHB), linoleoylglycerophospho- choline (LGPC), oleic acid, α-ketoglutaric acid, 2-aminoadipic acid, glycine, aromatic amino acids, and the 3 branched-chain amino acids and several of their catabolites. Results: After the 12-week intervention fasting levels of 12 of the 23 metab- olites were significantly different (p10% (n=16) there was a significant de- crease in plasma tyrosine, α-ketoglutarate and phenyalanine (p S 229 1 C Conclusion: The lower levels of P4 and PR in P-DEX islets is likely to be a factor involved in the previously described up-regulation of miR-29. The up- regulation of ERβ induced by DEX treatment could be preserving the secre- tory responsiveness to glucose. Finally, the down-regulation of inflammatory markers could disrupt beta cell resetting after parturition, thus reflecting in an inappropriate maternal glucose homeostasis. Supported by: FAPESP, CNPq PS 033 Determinants of insulin sensitivity and glycaemic control 551 Determinants of insulin sensitivity in Middle Eastern immigrants and native Swedes L. Bennet1, L. Groop2, P.W. Franks3; 1Department of Clinical Sciences, Family Medicine, Lund University, 2Department of Clinical Sciences, Lund University Diabetes Center, 3Department of Clinical Sciences, Genetic and Molecular Epidemiology, Lund University, Malmö, Sweden. Background and aims: Little is known about the metabolic, lifestyle and her- itable risk factors that underlie the observation that type 2 diabetes (T2D) is highly prevalent amongst Middle Eastern immigrants to Northern Europe. Hence, in the current report from the MEDIM study (the impact of Migra- tion and Ethnicity on Diabetes in Malmö), we aimed to (1) investigate insulin sensitivity and secretion (2) study risk factors associated with T2D and (3) investigate the contribution of metabolic-, demographic- and lifestyle-related risk factors for impaired insulin sensitivity index (ISI) in immigrants from Iraq, with comparisons to native Swedes. Materials and methods: Population-based, cross-sectional study conducted 2010-2012 including residents 30 to 75 years of age born in Iraq or Sweden, in whom 10-h fasting blood samples and oral glucose tolerance tests were per- formed. Further, sociodemography, lifestyle behaviors and comorbidity were assessed by questionnaires. Insulin sensitivity and secretion were assessed by Matsuda indices. Outcomes were T2D and ISI. Associations were assessed by logistic- and linear regression analysis. Results: In Iraqi (n=1398) compared to Swedish (n=757) participants, T2D was twice as prevalent (11.6 vs. 5.8%, p S 230 1 C function) in T2DM. We have shown that lowering plasma FFA and plasma glucose concentrations separately in T2DM individuals improves beta cell function and ameliorate insulin resistance. In the present study, we examined the effect of lowering both the plasma FFA and glucose concentrations on beta cell function and insulin sensitivity in T2DM individuals. Materials and methods: 12 T2DM subjects (age=49 ±3, BMI=31.1± 1.8, HbA1c = 8.5 ± 0.3, diabetes duration = 8.5 ± 0.3 years) received treatment with dapagliflozin (10 mg/day) for 3 weeks and during the third week they also received acipimox (250 mg four times daily). 75 gram OGTT and eugly- cemic insulin clamp were performed at baseline and at weeks 2 and 3. Results: Compared to baseline, dapagliflozin lowered the fasting plasma glucose concentration at 2 weeks (181±9 to 146±9, p S 231 1 C Results: After 15 days from the beginning of diet and therapy for GA1 we ob- served a drop of anti GAD to 233KUI/L and a marked reduction of hypogly- cemias (mean AUC S 232 1 C 557 Effects of chronic variable stress and dietary fat on insulin sensitivity S. Müller-Lühlhoff1, T. Jelenik2, S. Hartwig1, S. Lehr1, P. Nowotny2, M. Roden2,3, H. Al-Hasani1, T.R. Castañeda1; 1Institute for Clinical Biochemistry and Pathobiochemistry, 2Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research, 3Department of Endocrinology and Diabetology, University Clinics Düsseldorf, Heinrich Heine University, Düsseldorf, Germany. Background and aims: Stress is a state of threat to homeostasis that may contribute to metabolic disorders such as visceral obesity and type 2 diabetes. However, there is a lack of studies analyzing the long-term consequences of chronic stress on insulin sensitivity and its relation to fat consumption. Our aim was to characterize how chronic variable stress (CVS) in relation to di- etary fat affects insulin sensitivity in mice. Materials and methods: Three-months old, body weight (BW)-matched male C57BL/6 mice were exposed to a random series of stressors for 15 days (CVS), the unstressed control mice were housed separately (Ctrl). Body com- position was analyzed with NMR. After CVS, blood was collected (08:00- 12:00 am) for corticosterone and insulin analysis. Subsequently, mice were consuming a low- (Chow) or a high-fat diet (HFD) for three months and insulin sensitivity was measured in vivo with hyperinsulinemic-euglycemic clamps. Statistical differences were considered significant at p S 233 1 C receptor ratio of IR protein and reduced Akt phospholylation by insulin stim- ulation (at concentration of 0.1 or 1.0nM) compared to wild type. Conclusion: In two cases of hyperinsulinaemic reactive hypoglycaemia, we found novel mutations in the IR gene considered to be linked to their hypo- glycemia. We propose a disease entity of a syndrome of hyperinsulinaemic reactive hypoglycaemia associated with mutations in the human IR gene. PS 034 Glucose uptake and glucose action 560 APPL1 affects on glucose uptake induced by uniaxial stretch in C2C12 myotubes T. Saito, Y. Shimoda, Y. Tagaya, E. Yamada, S. Okada, M. Yamada; Medicine and Molecular Science, Gunma University, Maebashi, Japan. Background and aims: Adaptor protein, phosphotyrosine interaction, PH domain and leucine zipper containing 1 (APPL1) was identified as protein kinase B or Akt (PKB/Akt) binding protein. It has been reported that APPL1 highly expresses in adipose and muscle tissues and has an important role in GLUT4 translocation in response to insulin and adiponectin via Akt or 5’ adenosine monophosphate-activated protein kinase (AMPK) dependent pathway. We found that Akt and AMPK were activated in mouse derived muscle cell line, C2C12 cell, in response to uniaxial stretch stimulation. Therefore, this study aimed to investigate the function of APPL1 in C2C12 myotubes on GLUT4 translocation mechanism stimulated by mechanical stretch. Materials and methods: C2C12 myoblasts were grown on an elastic silicone chamber and promoted differentiation. Differentiated C2C12 myotubes were stimulated by cyclic uniaxial stretch (10% of initial length, 10 cycle/min) for 5 hours. Plasmid DNA was transfected into C2C12 myoblasts by electropora- tion. Glucose uptake was measured by enzymatic assay and the localization of APPL1 was examined by immunofluorescence. Compound C, an AMPK inhibitor, or protein kinase C (PKC) zeta pseudosubstrate, a PKC inhibitor, was administrated to block the activity of AMPK or PKC zeta pathway dur- ing stretch. Results: Stretch increased glucose uptake, Akt, AMPK, and PKC zeta phos- phorylation without affecting protein expression of these proteins and GLUT4. APPL1 localized at perinuclear in basal condition and moved to plasma membrane after stretch stimulation. Stretch-induced glucose up- take, AMPK, and PKC zeta phosphorylation were statistically increased by 26%, 28%, and 30%, respectively, in APPL1 transfected cells compared to non-transfected. On the other hand, stretch-induced Akt activation was sup- pressed by 30% in APPL1-overexressed cells. Although compound C didn’t affect on glucose uptake, PKC zeta pseudosubstrate suppressed glucose up- take induced by APPL1 by 67% in stimulated condition. Conclusion: These results suggest that APPL1 is regulated by stretch stimula- tion and promote stretch induced glucose uptake in C2C12 myotubes mainly through PKC zeta dependent signaling pathway, but not AMPK dependent pathway. Supported by: JSPS KAKENHI 561 Acetic acid enhances glucose uptake and blood flow rates in the skeletal muscle in humans with impaired glucose tolerance E. Petsiou1, P. Mitrou2, E. Papakonstantinou1, E. Maratou2, V. Lambadiari1, F. Spanoudi1, S.A. Raptis1,2, G. Dimitriadis1; 12nd Department of Internal Medicine and Research Institute, Athens University Medical School, Attikon University Hospital, 2Hellenic National Center for Research, Prevention and Treatment of Diabetes Mellitus and its Complications (H.N.D.C), Athens, Greece. Background and aims: Previous studies support the antiglycaemic effect of vinegar. However, the effect of vinegar on endothelial function and muscle glucose metabolism has not been studied in humans. This cross-over study aims to investigate the effects of vinegar on plasma glucose, insulin and lipid levels, as well as blood flow and muscle glucose uptake in subjects with im- paired glucose tolerance (IGT), using the arteriovenous difference technique. Materials and methods: Eight subjects with IGT (age 46+4years, BMI 29+2), without any medication therapy, were randomised to consume vinegar (30ml vinegar containing 6% acetic acid, 20ml water) or placebo (50ml water) be- fore a mixed meal (566kcal; 75g carbohydrates, 26g protein, 6g fat). Plasma samples were taken at 15-60min intervals for 300min from the radial artery and from a forearm vein for measurements of glucose, insulin, triglycerides, non-esterified fatty acids (NEFA) and glycerol. Muscle blood flow was meas- ured with strain-gauge plethysmography. Glucose flux was calculated as the Diabetologia (2014) 57:[Suppl1]S1–S564 S 234 1 C arteriovenous difference of glucose multiplied by the blood flow rates. The cross-over study was conducted 1 week later. Results: Vinegar compared to placebo: 1) decreased arterial plasma insulin (8101+1410 vs 12008+1645mU/L*min, p=0.043) 2) increased forearm blood flow (404+97 vs 134+71 ml/min/100ml tissue*min, p=0.048) 3) increased muscle glucose uptake (933+74 vs 603+108μmol/100ml tissue, p=0.02) 4) de- creased arterial plasma triglycerides (45+11 vs 86+10 nmol/L*min, p=0.036), without changing NEFA and glycerol. Conclusion: In individuals with impaired glucose tolerance, vinegar addi- tion to a mixed meal results in an enhancement of muscle blood flow rates, an improvement of glucose uptake by the forearm muscle and a reduction of postprandial hyperinsulinaemia and hypertriglyceridaemia. From this point of view vinegar may be considered beneficial for improving insulin resistance and metabolic abnormalities in the atherogenic prediabetic state. 562 Disclosing acute caffeine action on insulin sensitivity: effects on skeletal muscle M.J. Ribeiro1, J.F. Sacramento1, S. Yubero2, B.F. Melo1, A. Obeso2, C. Gonzalez2, M.P. Guarino1,3, S.V. Conde1; 1CEDOC, Faculdade de Ciências Médicas, Universidade Nova de Lisboa, Portugal, 2Departamento de Bioquímica y Biología Molecular y Fisiología, Universidad de Valladolid, Facultad de Medicina. IBGM, CSIC. Ciber de Enfermedades Respiratorias, CIBERES, Instituto de Salud Carlos III, Valladolid, Spain, 3UIS-Unidade de Investigação em Saúde- Escola Superior de Saúde de Leiria- Instituto Politécnico de Leiria, Portugal. Background and aims: Caffeine, a non-selective adenosine antagonist, is the behaviorally active substance most widely consumed in the world. When consumed regularly, this xanthine presents beneficial effects on type 2 diabe- tes and metabolic syndrome. However, the sensitizer effect of chronic caffeine intake contrasts with acute caffeine administration that has been associated with an increase in insulin resistance (IR). The aim of this work was to in- vestigate the effect of acute caffeine administration on insulin sensitivity and the involvement of adenosine receptors. Additionally, the mechanism behind caffeine-mediated effects in skeletal muscle was assessed. Materials and methods: In vivo experiments were performed in Wistar rats of both sexes, aged 3 months (200-350g) anaesthetized with pentobarbitone (60mg/Kg). The effect of the acute administration of caffeine (0.001-5μM), DPCPX (A1 antagonist, 0.0005-5μM), SCH58261 (A2A antagonist, 0.0005- 5μM) and MRS1754 (A2B antagonist, 0.001-5μM) on insulin sensitivity was evaluated by means of an insulin tolerance test. Skeletal muscle Glut4 and AMPKα1 expression were quantified by Western-blot. The effect of A1 and A2B adenosine agonists on glucose uptake was evaluated. Sodium nitroprus- siate (SNP, 10nM), a nitric oxide (NO) donor was used to evaluate the effect of NO on adenosine antagonism induced-IR. Results: Acute caffeine decreased insulin sensitivity in a concentration de- pendent manner (Emax=55.54 ± 5.37%, IC50=11.61nM), an effect that is mediated by A1 and A2B adenosine receptors. Additionally, in skeletal mus- cle, acute caffeine administration did not modify AMPK expression, however it significantly decreased Glut4 by 23.23% and 31.81% (0.05 and 0.5μM of caffeine, respectively). We found that A1, but not A2B agonists significantly increased glucose uptake to 2.11±0.04 nmol.mg-1 tissue in skeletal muscle when compared to control (1.69±0.04nmol.mg-1tissue). SNP partially re- versed DPCPX and MRS1754 induced-IR by 77.4 and 51.1%, respectively, when compared with the application of adenosine antagonists alone (KITT DPCPX=2.12±0.44%glucose.min-1; KITT MRS1754=2.16±0.08%glucose. min-1). Conclusion: Acute caffeine administration decreases insulin sensitivity in a concentration dependent manner being this effect mediated by A1 and A2B adenosine receptors. In skeletal muscle, the effect of caffeine on insulin sen- sitivity involves a decrease in Glut4 expression and in insulin-dependent glu- cose uptake that is mediated mainly by A1 adenosine receptors. Additionally, adenosine-mediated insulin sensitivity seems to involve NO production and its sensitizer actions. Supported by: FCT: PTDC/SAU-ORG/111417/2009, BFU2012-37459 and CIBER CB06/06/005 563 Glucose uptake independent of insulin action: role of S-nitrosilated insulin chains J.M. Gaspar1,2, I.S. Lima1,2, J. Caldeira3, Y.-B. Kim2, M.P. Macedo1,4; 1Chronic Disease Center (CEDOC), New Medical School,, Lisbon, Portugal, 2Division of Endocrinology, Diabetes and Metabolism, Beth Israel Deaconess Medical Center, Boston, USA, 3REQUIMTE, Faculty of Sciences and Technology, 4Portuguese Association of Diabetes and Education Research Center (APDP-ERC), Lisbon, Portugal. Background and aims: Portal insulin is removed (50-70%) during first pass transit in the liver by a process called insulin clearance. Protein disulfide isomerase (PDI) mediates insulin clearance by reduction of insulin disulfide bonds, breaking it into A and B chains by a mechanism that requires glu- tathione (GSH). Previously, we reported that A/B chains of insulin could be S-nitrosated (A-SNO and B-SNO) in vitro, forming nitrosothiols. After a meal, the increase of GSH levels together with the activation of parasym- pathetic nervous system which leads to nitric oxide (NO) release results in an increase in peripheral insulin sensitivity. A concurrent increase in hepatic PDI activity cleaves and nitrosilates A and B chain of insulin (A-SNO e B- SNO). Our hypothesis is that S-nitrosilated modified derivatives of insulin, A-SNO and B-SNO, produced during the insulin clearance process stimulates glucose uptake in skeletal muscle by a mechanism independent of the insulin signalling pathway. Materials and methods: The presence of BSNO in the liver of Wistar rats was analyzed by immunoprecipitation of S-nitrosilation proteins following by western blot for B-chain. Physiological effects of S-nitrosilated modified derivatives of insulin were evaluated in differentiated skeletal muscle cells (L6-mycGLUT4 cells) and adipocytes (3T3-L1 cells). The cells were stimu- lated with 100nM insulin, A-chain, B-chain, A-SNO and B-SNO to evalu- ate glucose uptake using 3H-Glucose method. To analyse detailed signalling pathways activated by these insulin derivatives we performed cellular extracts and by western blotting technique we investigate the enrolled proteins. Results: We detected the existence of BSNO in liver homogenates of rats and that BSNO levels were decreased (23% comparing with sham rats) in animals that were subjected to an hepatic denervation of parasympathetic nerves - NO pathway. We observed that A-SNO and B-SNO significantly stimulates glucose uptake in L6 muscle cells (159.8 ± 19.2% and 204.8 ± 30.2% to con- trol, respectively), similar to insulin stimulus (203.1 ± 15.0). The stimulation of glucose uptake was mediated by the activation of AMPK pathway, detected by the increase in phosphorylation of Thr172 AMPK. We also observed that these insulin derivatives did not activate the canonical insulin pathway. In adipocytes we observed that A-chain and B-chain insulin derivatives increase glucose uptake (309.2 ± 83.2% and 319.0 ± 23.6% to control, respectively). In a lesser extent B-SNO also stimulates glucose uptake (193.6 ± 13.8% to control). In adipocytes these derivatives does not activate either the canonical insulin pathway or AMPK signaling pathway. Conclusion: As a conclusion insulin clearance process can lead to the for- mation of A/B chain and/or the respective S-nitrosilated derivatives. In adipocytes, mainly A-chain and B-chain stimulates glucose uptake. On the other hand when insulin derivatives are S-nitrosilated (A-SNO and B-SNO) they act on skeletal muscle promoting glucose uptake, through activation of AMPK signaling pathway. Supported by: Albert Renold (EASD), EMBO, PTDC/DTP-EPI/0207/2102 (FCT) 564 Treatment with small, subthyrotoxic doses of thyroxine can improve peripheral cellular glucose transport and insulin sensitivity in patients with type 2 diabetes E. Maratou1, V. Lambadiari2, F. Spanoudi2, E. Vassilatou2, G. Matsagouras2, E. Hatziagelaki2, P. Mitrou1, S.A. Raptis1,2, G. Dimitriadis2; 1Hellenic National Center for Research, Prevention and Treatment of Diabetes Mellitus and its Complications (HNDC), 22nd Department of Internal Medicine, Research Institute and Diabetes Center, Athens University, “Attikon” University Hospital, Athens, Greece. Background and aims: Variation of plasma thyroid hormones’ levels influ- ences insulin sensitivity and peripheral glucose disposal. Administration of high doses of thyroxine to healthy humans induces insulin resistance, where- as moderate doses increase peripheral glucose disposal. We investigated the effect of the administration of small doses of thyroxine to healthy humans Diabetologia (2014) 57:[Suppl1]S1–S564 S 235 1 C and patients with type 2 diabetes on the insulin sensitivity indices, in vitro glucose uptake and GLUT4 recruitment in the plasma membrane of periph- eral monocytes. Materials and methods: Eleven healthy, euthyroid subjects (CON) with a micronodular texture of the thyroid gland, aged 40.8±2.7 yrs, BMI 26.6±0.83 kg/m2 , T3 118.1±5 ng/dl (1.818±0.077 nmol/l), T4 7.1±0.3 μg/dl (91.377 ±3.861 nmol/l), TSH 1.235±0.109 μU/ml, and eleven treatment-naive sub- jects with type 2 diabetes (DM) [aged 42.9±3.8 yrs, BMI 27.48±1.39 kg/m2 , T3 119±5.7 ng/dl (1.832±0.087 nmol/l), T4 8,13±0.46 μg/dl (104.8±5.93 nmol/l), TSH 1.51±0.14 μU/ml], were studied before and after administra- tion of 50 μg of thyroxine once daily for 2 months. The study was approved by the hospital ethics committee, and subjects gave informed consent. A meal was given to the participants. Blood was drawn before the meal for the study of glucose uptake and GLUT4 abundance in peripheral monocytes and sev- eral blood samples were drawn for 300min from a forearm deep vein and the radial artery for measurements of glucose, insulin and HOMA-IR and Matsuda indices. The insulin’s effect on GLUT4 translocation from cytoplas- mic depots to plasma membrane was studied by stimulating monocytes with various concentrations of the hormone (0, 25, 50,100 and 200uU/ml). The increment of enrichment of plasma membrane to GLUT4 is represented as %increment from baseline (0mU/l) to maximal concentration (200mU/l). Plasma membrane GLUT4 abundance and glucose uptake (fluorescent ana- logue 6-NBDG) were studied in isolated monocytes by flow cytometry. Results: The area under the curve of the 6-NBDG uptake was significant- ly higher in both CON (pre treatment 7392.7+577.4 vs. post treatment 9383.8+460, P S 236 1 C 567 Impaired hepatic insulin extraction in type 2 diabetes F. Piccinini1, C. Dalla Man1, A. Vella2, C. Cobelli1; 1Department of Information Engineering, University of Padova, Italy, 2Mayo Clinic, Rochester, USA. Background and aims: The evaluation of hepatic insulin extraction (HE) during a meal is fundamental for understanding the regulation of carbo- hydrate metabolism in healthy (H), prediabetic (PD), and type 2 diabetic (T2DM) subjects. In this study we use a new physiologically-based math- ematical model for estimating HE across the spectrum of glucose tolerance. Materials and methods: To this purpose we studied 18 H (age=49.6±8.2 yr, BMI=27.7±3.1 kg/m2), 35 PD (both impaired glucose tolerant and im- paired fasting glucose; age=53.2±7.7 yr, BMI=30.1±5.0 kg/m2) and 22 T2DM subjects (age=53.9±6.9 yr, BMI=32.7±4.9 kg/m2). Each subject underwent a standard mixed meal (75 g glucose, 43% carbohydrate, 17% protein, 40% fat). Frequent plasma sampling was undertaken in all the subjects in order to measure plasma glucose, insulin and C-peptide concentrations. The model assumes that C-peptide kinetics is described by a two-compartment model, insulin kinetics by a three-compartment model, and insulin secretion is a sum of two components, i.e. one proportional to glucose rate of change and one proportional, with a constant delay, to glucose concentration. HE suppres- sion is assumed to be linearly dependent on plasma glucose concentrations. Basal (HEb) and total (HEtot) HE indices can be derived from the model pa- rameters, as well as a new index quantifying HE sensitivity to glucose (SG HE). Results: The model well fitted C-peptide and insulin data in all the subjects, also providing precise parameter estimates. Rank sum test showed that HEb and HEtot are significantly lower in T2DM than H subjects (HEb=49% vs. 64%, p=0.01; HEtot=43% vs. 59%, p=0.04). Moreover, SG HE is significantly lower in T2DM than PD (SG HE=0.10 vs. 0.14 l/mmol, p=0.006) and H sub- jects (SG HE=0.10 vs. 0.13 l/mmol, p=0.02). Figure 1 shows the average HE profiles reconstructed by the model in the three groups: evidently the basal state is impaired in T2DM, as it is confirmed by the statistical test, moreover, in T2DM, the profile has a longer low phase, and recovers slower than H and PD subjects. Conclusion: The new physiological model for the quantification of HE, ini- tially developed in H subjects, has been successfully used both in PD and T2DM. This model allows the estimation of a new index measuring HE sen- sitivity to glucose. This was significantly lower in T2DM compared to H and PD, likely due to an inability of plasma glucose to suppress HE. Supported by: FIRB 2008, DK82396, DK78646 568 The effect of glycaemic variability on expression of genes involved in the development of hyperglycaemia induced tissue damage K. Kuricová1, L. Pácal1, V. Dvořáková1, J. Šoupal2, M. Prázný2, K. Kaňková1; 1Department of Pathological Physiology, Masaryk University, Brno, 223rd Department of Internal Medicine, Charles University, Prague, Czech Republic. Background and aims: Large prospective studies provided evidence of rela- tionship between long-term diabetes compensation expressed as HbA1c and the risk of diabetic complications. However, HbA1c and diabetes duration explain only small portion of complications risk. It is well known that pa- tients with the same HbA1c differ in amplitudes and duration of glycaemic excursions. Therefore, short-term (i.e. during the day) glycaemic excursions, so called glycaemic variability (GV), are debated currently. Animal experi- ments and in vitro studies showed that oscillating glucose may have more deleterious effect on cells due to higher production of superoxide than hyper- glycaemia itself. Therefore, aim of our study was to evaluate the effect of GV on the expression of genes whose products are involved in the development of tissue damage in diabetes. We compared the effect of low and high GV of diabetics, GV of healthy subjects and chronic normo- and hyperglycae- mia. Specifically, we detected gene expression of glyoxalase 1 (GLO1), heme oxygenase (HMOX), receptor for advanced glycation end-products (RAGE), p53, superoxide dismutase (SOD2), DJ-1 (an enzyme with glyoxalase activ- ity), transketolase (TKT) and thiamine pyrophosphokinase (TPK1). Materials and methods: Primary human umbilical vein endothelial cells (HUVEC) were cultured 24 hours in the following settings mimicking situa- tion in humans (both diabetics and healthy) and in commonly used in vitro hyperglycaemia model: (A) high GV (standard deviation [SD] > 5), (B) low GV (SD < 3), (C) non-diabetics (SD = 1), (D) continuous normoglycaemia (5 mmol/l) and (E) continuous hyperglycaemia (25 mmol/l). Based on real 24 hours glycaemic curves from type 1 diabetics (3 patients with low GV and 3 with high GV) and healthy subjects (2 curves) we created 24 hours profiles with medium change after 2 hours. After 24 hours cells were harvested and RNA isolated and reverse transcribed using commercial kits (Roche). Gene expression was determined using quantitative PCR with predesigned probes (TaqMan™ Assay) with β-actin as a reference gene. Results: HMOX expression was significantly lower in both low and high GV compared to GV of healthy subjects (P = 0.04, Mann-Whitney test). Similar pattern was also found in four of the selected genes (p53, SOD2, GLO1, DJ- 1) although statistically not significant (all P > 0.05, Mann-Whitney test). Expression of all genes with exception of DJ-1 was significantly higher in continuous normoglycaemia when compared to low and high GV (all P < 0.04, Mann-Whitney). Finally, no significant difference was found between continuous hyperglycemia and low and high GV (all P > 0.05). Conclusion: Results of our pilot study indicate that the effect of either low or high GV on expression of selected genes involved in diabetes induced tissue damage does not significantly differ from the effect of continuous hypergly- caemia. Supported by: IGA MZ NT13198 Diabetologia (2014) 57:[Suppl1]S1–S564 S 237 1 C PS 035 Mechanisms of insulin action in non-human models 569 Beneficial roll of TAZ modulator, TM-25659, in obese and diabetic conditions J.-G. Jeong1, J. Jeon1, E. Ha1, S.-E. Choi2, S.-A. Lee1, Y. Kang2, M. Bae3, J. Ahn3, H. Jeong4, E. Hwang4, S. Han1, D. Kim1, K.-W. Lee1; 1Endocrinology and Metabolism, Ajou University School of Medicine, Suwon, 2Institute of Medical Science, Ajou University School of Medicine, Suwon, 3Korea Research Institute of Chemical Technology, University of Science & Technology, 4College of Pharmacy, Graduate School of Pharmaceutical Sciences, and Global Top5 Research Program, Ewha Womans University, Seoul, Republic of Korea. Background and aims: The transcriptional co-activator with PDZ-binding motif (TAZ) inhibits adipocyte development by intervention with per- oxisome proliferator-activated receptor (PPAR)-gamma. Recently, PPAR- gamma has been implicated in muscle lipid accumulation. Therefore, we investigated the effects of TAZ modulator on skeletal muscle function in C2 myotubes and in C57BL/6J mice. In a previous study, Jang et al identified 2-butyl-5-methyl-6-(pyridine-3-yl)-3-[2’-(1H-tetrazole-5-yl)-biphenyl-4- ylmethyl]-3H-imidazo[4,5-b]pyridine] (TM-25659) as a TAZ modulator and was shown to decrease weight gain in a high fat diet-induced obese model. Materials and methods: We measured insulin signaling and glucose uptake using immunoblotting and 2-NBDG uptake. To elucidate the preventing mechanism of TM-25659 on palmitate-induced C2 myotubes insulin signal- ing, we performed real-time PCR with pro-inflammatory cytokines (TNF- alpha, IL-1-beta, IL-6 and MCP-1) mRNA. Male C57BL/6J mice that were fed a HFD for 8 weeks were randomly assigned for an additional 6 weeks to 3 groups: normal diet (ND), HFD, HFD+TM-25659. After treatment, oral glucose tolerance test (OGTT) and intraperitoneal insulin tolerance test (IP- ITT) were performed to evaluate anti-diabetic effects. After sacrifice, tissue extracts of the soleus muscle were obtained for quantification of insulin sign- aling through Akt phosphorylation proteins by immunoblotting. Results: In our present study, we confirmed TM-25659 considerably in- hibits palmitate-induced insulin resistance on C2 myotubes. Moreover, we showed that TM-25659 significantly suppresses high fat diet (HFD) induced insulin resistance in C57BL/6J mice skeletal muscle. Consequently, the TM- 25659-treated group showed improved HFD-reduced glucose tolerance, in- creased insulin sensitivity, and increased AKT phosphorylation protein. Also, TM-25659 down-regulated expression of genes involved in pro-inflamma- tion. Conclusion: These results suggest TAZ modulator, TM-25659, play a ben- eficial role in controlling insulin resistance in obese and diabetic conditions. 570 Human skeletal muscle clock: implications in myokine secretion in physiology and pathophysiology S. Skarupelova1, L. Perrin1, E. Lefai2, H. Vidal2, J. Philippe1, C. Dibner1; 1Division of Endocrinology, Diabetes and Nutrition, Geneva University Hospital (HUG), Switzerland, 2Faculté de Médecine, INSERM, Lyon, France. Background and aims: Circadian rhythms are present in many organisms of different complexity, representing an adaptation mechanism allowing adjustment of organism’s internal processes to the external world. In mam- mals circadian system is organized in a hierarchical structure with a central pacemaker residing in hypothalamic suprachiasmatic nuclei that in its turn controls peripheral secondary clocks in organs including skeletal muscle. The aim of our study was to characterize molecular clock and its function in hu- man primary skeletal muscle of healthy, overweight and obese subjects. Materials and methods: Human primary myoblasts were derived from muscle biopsies of Musculus gluteus maximus. Bioluminescent reporters Bmal1-luciferase and Per2-luciferase were introduced by lentiviral transduc- tion, and allowed visualization of the oscillatory profiles of Bmal1 and Per2 continuously in living cells. For clock ablation myotubes were transfected with siRNA-targeting Clock. Secretory profiles of myokine IL6 (ELISA meas- urements), assessed in parallel with bioluminescence profile recording, were compared in myotubes with functional and ablated clock. Results: Our experiments revealed that fully differentiated in vitro synchro- nized myotubes exhibited circadian rhythm with a period length of about 24.6 hours. Oscillation period tended to be longer in case of overweight and obese subjects. Endogenous core clock transcripts as well as bioluminescent profiles showed sustained circadian expression pattern with pronounced am- plitude. Around 90% of Clock knockdown significantly flattened oscillatory profile of Bmal1-luciferase reporter and amplitude of endogenous core clock transcripts. Our results revealed that the primary myotubes with intact clock secret IL6 in the circadian manner. Moreover, remarkable down regulation of IL6 secretion was observed upon clock ablation, meaning that the clock machinery might regulate IL6 secretion in human skeletal muscle. Conclusion: Our data demonstrate that the human skeletal muscle exhibits pronounced circadian oscillations that impacts on the muscle IL6 secretory function. Moreover, changes in rhythmicity of circadian machinery in hu- man skeletal muscle might be associated with metabolic disorders such as obesity. Supported by: FN 3100A-146475 571 Role of the RNA binding protein Sam68 in leptin and insulin signalling in granulosa cells V. Sanchez-Margalet1, T. Vilariño-García1, A. Perez-Perez1, F. Sanchez-Jimenez1, V. Blasco2, N. Prados-Nodd2, M. Fernandez-Sanchez2; 1Medical Biochemistry and Molecular Biology and Immunology, Virgen Macarena University Hospital, 2Instituto Valenciano de Infertilidad, Seville, Spain. Background and aims: Obesity is a medical problem not only because it is associated with type 2 diabetes and cardiovascular risk, but also because in reproductive-age women obesity is associated with polycystic ovary syn- drome (PCOS), insulin resistance and decreased fertility. The RNA-binding protein Sam68 is expressed in granulosa cells and the knockout female mice are subfertile, with ovulation problems. Since we have previously found that Sam68 may be recruited to insulin and leptin receptors, we planned to study the participation of Sam68 in signaling and action of insulin and leptin re- ceptors in granulosa cells. In addition, we aimed to study the expression of Sam68 in granulosa cells in response to leptin and insulin in vitro. Materials and methods: Signaling was studied by immunoprecipitation and immunoblot of the phosphorylated proteins. The expression of Sam68 is in- hibited by antisense strategy. The expression level of Sam68, and leptin and insulin receptors are quantified by qPCR and immunoblot. Results: We have found that Sam68 is tyrosine phosphorylated by insulin or leptin stimulation in granulosa cells, recruiting Sam68 to signaling complexes and decreasing its RNA binding capacity. In addition, both insulin and leptin increase the expression of Sam68 in granulosa cells. Finally, full expression of Sam68 is required for the activation of PI3K and MAPK signaling pathways by insulin or leptin in granulosa cells. Conclusion: Sam68 is recruited to leptin and insulin receptor signaling, its expression is induced by both hormones, and Sam68 is necessary for the full activation of insulin and leptin receptor signal transduction in granulosa cells. Sam68 may be a new important element in the ovarian insulin resist- ance and the decreased fertility found in obese women. Supported by: PI12/01172 ISCIII 572 Insulin regulates glycogen synthesis in endometrial epithelial cells through an increase in glycogen synthase G. Choe, C.A. Flannery, F. Saleh, H.S. Taylor; Obstetrics, Gynecology, & Reproductive Sciences, Yale School of Medicine, New Haven, USA. Background and aims: Endometrial gland secretion of glucose, lipids, and protein supports blastocyst nutrition prior to establishment of placental circulation. The endometrial glands fill with glycogen during the secretory phase in preparation for implantation, and regulation of glycogen synthesis has been attributed to progesterone. However, we hypothesized that insulin regulates glycogen synthesis in endometrial epithelial cells as it does in tradi- tional insulin targets such as liver and muscle. Materials and methods: We isolated endometrial epithelial and stromal cells from eight healthy women, and treated the cells separately with insulin, medroxyprogesterone (MPA), or vehicle. After 48hours, the cells were lysed, Diabetologia (2014) 57:[Suppl1]S1–S564 S 238 1 C and glycogen was isolated and quantified. Total and phosphorylated glycogen synthase (GS) and glycogen synthase kinase 3α/β (GSK3α/β) were evaluated by western blot after a 30min or 48hr hormone stimulation. Endometrial tis- sue from 45 women distributed across the menstrual cycle was assessed for insulin receptor β (IRβ) by PCR and IHC. Results: We found that IRβ had higher expression during the secretory phase of the menstrual cycle, with an increase in expression in both epithelial and stromal cells between day 16 and 24 of the menstrual cycle. In vitro MPA increases IRβ expression over 25-fold in primary stromal cells (p S 239 1 C insulin in old mice on HFD fail to suppress glucagon release, indicating in- sulin resistance of glucagon producing alpha-cells. This was paralleled by in vitro findings with islets from old mice under HFD. Supported by: Else Kröner-Fresenius-Stiftung 575 Different proprotein convertases are involved in the maturation of the two insulin receptor isoforms F. Peiretti, I. Kara, B. Bonardo; UMR1062, Inserm, Marseille, France. Background and aims: The insulin receptor (IR) is expressed as two iso- forms, IRA (exon 11-) and IRB (exon 11+), which originate from the alterna- tive splicing of exon 11 and differ from 12 amino acids located at the carboxyl terminus of the extracellular a-subunit. The most significant difference be- tween the two IR isoforms is that IRA is more responsive to activation by IGF2 than IRB. Thus, it is suggested that IRA and IRB are both responsible for metabolic effects (triggered by insulin) whereas, only IRA is responsible for the mitogenic/proliferate effects (triggered by IGF2). The two IR isoforms are matured in the Golgi apparatus by the proteolytic activity of the proprotein convertase furin. We analyzed if the absence or presence of exon 11 could affect the maturation of IR isoforms. Materials and methods: IR isoforms were overexpressed in different cell lines. Proprotein convertases activity was modulated by pharmacological inhibition, specific knockdown and overexpression. IR maturation was ana- lyzed by western blot. Results: In furin-defective LoVo cells, IRB is more matured than IRA. In furin-expressing cells, both IR isoforms are equally matured and furin knock- down is more efficient in reducing the maturation of IRA than that or IRB. The global inhibition of proprotein convertase activity abolishes the furin- independent maturation of IRB, indicating the involvement of a propotein convertase other than furin in this maturation. We demonstrate that in the absence of furin activity the proforms of IR are exposed at the cell surface where IRB (but not IRA) is matured by PCSK6. This difference in the matu- ration of IRA and IRB was used as an opportunity to modulate the balance metabolic/mitogenic signals emanating from IR. Furin inhibition reduces the signals triggered by IGF2 (emanating from IRA) without unduly affecting those triggered by insulin (emanating from IRB). Conclusion: Our results are the first to reveal a difference in the maturation of IR isoforms. The maturation of IRA is almost strictly under the control of furin whereas that of IRB involves furin and PCSK6. Furin inhibition reduces the amount of mature IRA without drastically altering that of IRB, which reduces the signals triggered by IGF2 but not those triggered by insulin. 576 FKBP5 gene polymorphisms and expression in human adipose tissue are associated with insulin resistance and type 2 diabetes M.J. Pereira1, J. Palming2, M.K. Svensson2, M. Rizell3, J. Dalenbäck4, M. Hammar5, T. Fall1, C.O. Sidibeh1, P.-A. Svensson6, J.W. Eriksson1; 1Department of Medical Sciences, Uppsala University, 2Department of Molecular and Clinical Medicine, , 3Sahlgrenska University Hospital, Gothenburg, 4Frölunda Specialist Hospital, Gothenburg, 5AstraZeneca R&D, Mölndal, Sweden, 6Department of Molecular and Clinical Medicine, University of Gothenburg, Sweden. Background and aims: Central obesity is associated with a cluster of meta- bolic alterations, which include insulin resistance (IR), dyslipidemia and cardiovascular disease. Glucocorticoid excess is associated with redistribu- tion of fat from peripheral to central depots and with IR and development of diabetes. To identify potential novel mechanisms for IR we investigated dex- amethasone-induced changes of gene expression in human adipose tissue. Materials and methods: Subcutaneous and omental adipose tissue, obtained from 25 non-diabetic subjects (28-60 yrs; 20.7-30.6 kg/m2), was incubated without or with dexamethasone (0.003-3 μM) for 24 h. Gene expression was assessed by microarray and real time-PCR. Protein levels were assessed by immunoblotting. Results: FKBP5 (FK506 binding protein 5) was one of the genes responding most to dexamethasone. Gene expression increased up to 7-fold in a dose-de- pendent manner in both subcutaneous and omental fat depots (p S 240 1 C 578 Physical inactivity and high fat diet synergistically enhance the accumulation of intramyocellular diacylglycerol and induce insulin resistance in murine soleus muscle S. Kakehi1, Y. Tamura2, S.-I. ikeda3, R. Kawamori3, H. Watada2; 1Sportology Center, 2Metabolism and Endocrinology, 3Sportology Center, Juntendo University, Tokyo, Japan. Background and aims: Intramyocellular lipids (IMCL), especially intramy- ocellular diacylglycerol (IMDG) has been reported as one of the causes of insulin resistance in skeletal muscle. It has been demonstrated that physical inactivity and high fat diet (HFD) increase IMCL levels and induce insulin resistance, respectively. However, the combined effect of these two factors on insulin resistance has not been clarified yet. Materials and methods: To elucidate them, C57BL6J mice were randomly assigned to four groups; control group, 24h hind-limb cast immobilization (HCI) group, short-term (2wk) HFD group, and 24h HCI after 2wk HFD group. Then, we evaluated ex-vivo insulin-stimulated 2-deoxy glucose uptake (2DG-uptake), insulin signal and intracellular fat composition in skeletal muscle. Results: Twenty four hours HCI significantly decreased insulin-stimulated 2DG-uptake by 40%, while 2wk HFD did not alter 2DG-uptake. On the other hand, 24h HCI after 2wk HFD dramatically decreased insulin-stimulated 2DG-uptake by 74%. In parallel with decreased insulin-stimulated 2DG- uptake, we observed decreased insulin stimulated serine phosphorylation of Akt and tyrosine phosphorylation of insulin receptor substrate (IRS)-1 after 24h HCI, which were more exacerbated in 24h HCI after 2wk HFD group, while only 2wk HFD did not change these phosphorylation states. Concern- ing, it has been hypothesized that IMDG accumulation and PKC activation impairs insulin signal transduction. Consistent with this hypothesis, we found that 24h HCI increased IMDG in soleus by 192%, while the amount of intramyocellular triacylglycerol (IMTG) was not changed. In addition, whereas IMDG and IMTG were not changed by 2wk HFD feeding, 24h HCI after 2wk HFD dramatically increased IMDG and IMTG by 330% and 144%, respectively. In parallel with IMDG accumulation, we observed increased PKCε activity in 24h HCI group and 24h HCI after HFD group by 151% and 210%, respectively. Associated with the IMDG accumulation, expression level of diacylglycerol acyltransferase (DGAT)-1, which converts diacylglyc- erol to triacylglycerol, was decreased in both HCI group and HCI after 2wks HFD group. In addition, activity of Lipin1, which converts phosphatidic acid to diacylglycerol, was increased in 24h HCI group and 24h HCI after HFD group by 141% and 210%, respectively Conclusion: These results suggested that physical inactivity and HFD syner- gistically induce IMDG accumulation and insulin resistance in soleus mus- cle. Increased Lipin1 activity and decreased DGAT-1 expression might be involved in the mechanism of IMDG accumulation after physical inactivity and HFD. PS 036 Mechanisms of insulin resistance in vivo 579 Impact of hepatic lipid accumulation and composition on glucose tolerance and insulin sensitivity: a longitudinal study in male and female mice A.F. Soares1, H. Lei2, R. Gruetter1,2; 1Laboratory of Functional and Metabolic Imaging, 2Biomedical Imaging Research Center, École Polytechique Fédérale de Lausanne, Switzerland. Background and aims: Metabolic disruptions characterized by high hepatic lipid content (HLC) are associated with impairments in whole body glucose homeostasis. To gain insight on the role of hepatic lipids in the metabolic performance in the absence artificial metabolic stresses we measured non- invasively and longitudinally the HLC and profile in mice during adult devel- opment by Magnetic Resonance (MR) Spectroscopy in vivo. In parallel, mice were challenged with insulin and glucose tolerance tests. Materials and methods: Male (N=10) and female (N=10) C57Bl/6J mice were studied at 3 (3Mo), 7 (7Mo) and 10 months (10Mo) of age. Mice were scanned in a 14.1 T magnet with a 1H quadrature surface coil over the abdo- men. Localized 1H spectra were acquired from a 8 µl volume with stimulated echo acquisition mode sequence and the HLC expressed as the percent of to- tal 1H MR signal, with corrections for spin-spin relaxation effects. Additional spectra were acquired from the same volume with suppression of the water signal to enable the detection and quantification of all the lipid protons. The lipid profile was characterized by the following indices: saturated component (SC); unsaturated fatty acyl chains (UFA); mean number of double bonds per fatty acyl chain (ndb/FA), mean number of poly-unsaturated double bonds per fatty acyl chain (PUdb/FA) and per UFA (PUdb/UFA); mean chain length (MCL). OGTTs (1.5 g/Kg) and i.p. insulin tolerance tests (ITTs) were per- formed after a 6h-fast. Plasma insulin was determined by ELISA and insulin sensitivity estimated with the quantitative insulin check index (QUICKI) as the inverse of the log10 sum of fasting insulin (µIU/ml) and fasting glucose (mg/dl). Data are expressed as mean ± SEM. Statistical significance was ac- cepted for a P < 0.05 (one-way ANOVA with Newman-Keuls post test) and correlations assessed by the Pearson r coefficient. Results: In males, the HLC at 3Mo was 1.35 ± 0.15%, increasing to 3.06 ± 0.38% at 7Mo, not different from 2.70 ± 0.31% at 10Mo. Females had higher HLC at 3Mo (2.63 ± 0.19%) but no further changes henceforward (2.31 ± 0.20% at 7Mo; 2.36 ± 0.20% at10 Mo). In males, the SC and MCL of hepatic lipids increased with age, with a trend for decreased PUdb/FA and PUdb/ UFA with no changes in ndb/FA or UFA content. Females showed the same trends. Glycemia 3h-post ITT and 2h-post OGTT was lower in females, while QUICKI was higer. These scores were preserved until 10Mo in females. In males, glycemia 2-h post OGTT increased with age and the area above the curve (AAC) for the ITT decreased. In males, but not females, higher body weight correlated with hepatic lipid accumulation (r = 0.7); worse ITT scores correlated with higher body weight (r = -0.6) and HLC (r = -0.7) and lower Pudb/UFA (r = 0.5); worse OGTT scores correlated with higher HLC (r = 0.4). Conclusion: In male mice, loss of insulin sensitivity correlated with weight gain, HL accumulation and lower poly-unsaturation. Glucose intolerance was specifically associated with HLC, suggesting a deleterious effect of lipids on the adaptation of hepatic metabolism to the fed state. This behaviour was not observed in females even if they showed similar HLC. In fact, the poly- unsaturation of HL in females didn’t change with HLC, suggesting a positive effect of PUFA on preserving the hepatic metabolic performance. Supported by: CIBM 580 Liver specific G0/G1 switch gene 2 (G0S2) overexpression exacerbates hepatic insulin resistance by exacerbating hepatic steatosis but ameliorates hepatic fibrosis Y. Sugaya, H. Satoh, T. Watanabe; Department of Nephrology, Hypertension, Diabetology, Endocrinology, and Metabolism, Fukushima Medical University, Japan. Background and aims: Hepatic steatosis is strongly associated with insulin resistance. Recently it has been reported that G0/G1 switch gene 2 (G0S2) inhibited the lipolysis activity of adipose triglyceride lipase. Moreover, we Diabetologia (2014) 57:[Suppl1]S1–S564 S 241 1 C confirmed that G0S2 protein content was increased in the livers of high fat diet (HFD) fed rats. However, the precise physiological role of hepatic G0S2 is still unknown. In the current studies, we investigated the effect of hepatic G0S2 on insulin sensitivity in HFD-fed male Wistar rats by overexpressing G0S2 proteins using an adenovirus (Ad) encoding mouse G0S2. Materials and methods: Male Wistar rats were fed with 60% HFD for a total of 4 weeks. After 3-week feeding, the rats were injected with control Ad-GFP or Ad-G0S2. On day 7 post injection, glucose tolerance test and euglycemic- hyperinsulinemic clamp studies were performed after an 8-hour fast. Results: There were no significant changes in the body weight and fasting Glucose levels between Ad-GFP and Ad-G0S2 rats. However, after the glu- cose load, the glucose levels were significantly higher in the Ad-G0S2 rats at 15 and 30 min. During the clamp studies, the glucose infusion rate required to maintain euglycemia was significantly decreased by 16% in Ad-G0S2 rats. Clamp hepatic glucose output was significantly increased by 31% in Ad-G0S2 rats, but there was no significant changes in insulin-stimulated glucose dis- posal rate between two groups. Furthermore, the Oil Red O staining revealed that overexpression of G0S2 significantly increased lipid accumulation in the liver. However, Masson trichrome staining revealed that overexpression of G0S2 significantly ameliorated fibrosis in the liver. Consistent with histologi- cal data, expression of TGF-β and Smad2 were significantly decreased in the livers of Ad-G0S2 rats Conclusion: Overexpression of hepatic G0S2 protein exacerbates hepatic insulin resistance by the exacerbation of hepatic steatosis but ameliorates he- patic fibrosis. 581 Fasting and postprandial hepatic energy metabolism in insulin resistant states M.J. Fritsch1,2, C. Koliaki1, R. Livingstone1, E. Phielix1, A. Bierwagen1, M. Meisinger1, T. Jelenik1, J.-H. Hwang1, J. Szendroedi1,3, M. Roden1,3; 1Institute for Clinical Diabetology, German Diabetes Center, Düsseldorf, Germany, 2Department of Pediatrics and Adolescent Medicine, Medical University of Vienna, Austria, 3Department of Endocrinology and Diabetology, Heinrich-Heine University, Germany. Background and aims: Alterations of hepatic energy metabolism are dis- cussed as a mechanism contributing to hepatic steatosis and insulin resist- ance. However, it remains unclear whether meal ingestion differently affects hepatic energy metabolism in insulin sensitive and insulin resistant humans. Materials and methods: Young lean insulin sensitive (body mass index; 23.2±0.5 kg/m2), age-matched obese insulin resistant nondiabetic (OBE: 34.3±0.5 kg/m2) and elderly type 2 diabetes patients (T2D: 32.0±0.8 kg/m2) were examined (n=10 per group). Hepatic phosphorous compounds and fat content were quantified in vivo with 31P/1H magnetic resonance spectroscopy before, 160 min and 240 min after ingestion of a high caloric mixed meal (652 kcal, 55% CHO, 15% protein, 30% fat). Whole body insulin sensitiv- ity (M-value) was assessed with the hyperinsulinemic-euglycemic clamp and skeletal muscle oxidative capacity was assessed ex vivo with high-resolution respirometry applied in biopsy samples. Results: OBE and T2D patients were similarly insulin resistant (M-value; OBE: 3.6±0.4 mg.kg-1.min-1; T2D: 2.5±0.7 mg.kg-1.min-1). OBE and T2D had 14fold (p=0.01) and 20fold (p=0.002) higher hepatic fat contents than lean persons. Fasting hepatic inorganic phosphate (Pi) was 42% lower in OBE (p=0.002) and 32% lower T2D (p=0.02) than in lean persons and correlated positively with M-value (r=0.63, p=0.001). After meal ingestion, hepatic ATP and Pi increased by 27% (p=0.02) and 33% (p=0.05) only in OBE. Circulating lipid peroxides transiently increased by 24% (p S 242 1 C medium of murine or human macrophages treated with low doses of PGE2 stimulated the phosphorylation of STAT3 and induced SOCS3 in rat and hu- man hepatocytes, while insulin-dependent glucokinase induction in hepato- cytes was attenuated. These effects were abolished after parallel treatment of macrophages with the COX2-inhibitor indomethacin or incubation with an OSM-neutralizing antibody. In summary, in addition to directly interrupting insulin signalling pathway in hepatocytes, PGE2 may further impair hepatic insulin sensitivity in an autocrine feed-forward vicious cycle by inducing cy- tokines like OSM and its own synthesis in macrophages. Conclusion: The study provides first evidence that in addition to established mechanisms prostaglandins have an impact on the development of hepatic insulin resistance. The enzymes involved in the prostaglandin synthesis and the receptors to which they bind on the different liver cell types might there- fore be potential new drug targets for the treatment of hepatic insulin resist- ance. Supported by: DGE 584 Mitochondrial respiration in skeletal muscle, liver and fat does not differ in morbidly obese patients with and without type 2 diabetes M.T. Lund1,2, M. Hansen1, A. Floyd2, J.W. Helge1, F. Dela1; 1Biomedical Sciences, University of Copenhagen, 2Department of Surgery, Koege Hospital, Denmark. Background and aims: After an oral glucose load the uptake in skeletal muscle and liver accounts for ~80-90% of the ingested glucose. In patients with type 2 diabetes the ability to take up glucose in both tissues are severely impaired leading to decreased whole body glucose uptake and prolonged hy- perglycemia. It has been speculated that the impaired glucose uptake partly is the result of an underlying mitochondrial dysfunction limiting the abil- ity to oxidize glucose and fat. We have now compared the maximal coupled mitochondrial respiration in morbidly obese patients with and without type 2 diabetes by studying three different tissues (skeletal muscle, liver and fat) obtained from all patients. Materials and methods: 16 morbidly obese patients (5M/11F; 7 with (T2D) and 9 (OB) without type 2 diabetes) reported to the lab twice 2 months prior to their scheduled Roux-En-Y gastric bypass operation. Body composition was assessed by a dual Energy X-ray Absorption scan and maximal oxygen uptake (VO2max) was measured during a graded bicycle test. On a sepa- rate day skeletal muscle (m. vastus lateralis) and subcutaneous fat samples (abdomen) were acquired by a Bergstrom needle biopsy and subsequently a hyperinsulinaemic euglycemic clamp (80 mU/m2/min) was performed to measure whole body insulin mediated glucose uptake. Finally, a liver biopsy was obtained during the gastric bypass procedure. Prior to respiration analy- sis the skeletal muscle fibers were split and permeabilized in Saponin (50µg/ ml) for 20 min. Liver and fat tissue were dissected in 2 x 2 mm tissue samples before transferred to the respirometer. 2 µl digitonin (2.5 mg/ml) were ini- tially added to the respirometer to permeabilize the fat tissue. Substrates were added consecutively: Malate (2 mM (liver: 6 mM)) and glutamate (10 mM); ADP (5 mM (liver: 1 mM)) and magnesium (20 mM (liver: 4 mM)); octanoyl carnitine (1.5 mM); succinat (10 mM). Results: Weight (133±6 vs. 122±9 kg), BMI (43±2 vs. 41±1 kg/m2), age (36±3 vs. 40±6 years) and VO2max (20±1 vs. 23±2 ml/min/kg) were similar in OB and T2D, respectively. Whole body glucose uptake tended to be higher in OB compared to T2D (4.8±0.5 vs. 3.6±0.3 mg/min/kg, respectively, P = 0.05). Maximal coupled respiration in skeletal muscle, liver and fat tissue are shown in Figure 1. Conclusion: The main finding in the current study was the similar maximal coupled respiration in tissue samples from two different groups of patients with 25% different insulin sensitivity. Differences in body weight, age or aero- bic fitness are not confounding factors. Thus, the mitochondria in patients with type 2 diabetes are able to consume oxygen at similar rates as healthy obese patients and this speaks against an underlying mitochondrial dysfunc- tion in insulin resistance. These data do not rule out that substrate sensitivity at physiological concentrations is altered in patients with type 2 diabetes. Supported by: The Danish Council for Strategic Research 585 Failure of the skeletal muscle to contribute to diurnal glucose homeostasis in type 2 diabetes M.E.O. Macauley, F.E. Smith, P. Thelwall, R. Taylor; Institute of Cellular Medicine, Newcastle Magnetic Resonace Centre, Newcastle University, Newcastle upon Tyne, UK. Background and aims: In health, food carbohydrate is stored as glycogen in skeletal muscle and liver, preventing a deleterious rise in osmotically active plasma glucose after eating. Glycogen concentrations increase sequentially after each meal to peak in the evening. Skeletal muscle accounts for the larger part of this diurnal buffering capacity. Approximately 30% of meal carbohy- drate is stored in skeletal muscle compared with 20% in liver after a single meal. The effectiveness of this diurnal mechanism has not been previously studied in type 2 diabetes. We have quantified the changes in skeletal mus- cle and liver glycogen concentration before and after 3 meals consumed at 4 hours interval. Materials and methods: We studied 40 (25M;15F) well controlled type 2 diabetes subjects on metformin only (HbA1c 6.4±0.07% or 47±0.8 mmol/ mol) and 14 (8M;6F) glucose tolerant controls matched for age, weight and BMI.13C Magnetic resonance spectroscopy at 3.0- Tesla, was used to quantify calf muscle and liver glycogen concentration in the fasting state (0830h) and at 2000h after 3 defined meals (2700kcal(M) and 2200kcal(F); 60%carbohy- drate, 20%protein, 20%fat). Results: Mean fasting plasma glucose and insulin concentrations were high- er in the type 2 diabetes group (7.7±0.2 vs. 5.1±0.1mmol/l; P S 243 1 C 586 Skeletal muscle develops insulin resistance before subcutaneous adipose tissue. Study on obese post-menopausal women A. Sultan1,2, O. Fabre3, C. Amouzou3, C. Breuker3,1, K. Lambert3, C. Fedou1, A. Dupuis1, J. Cristol1, F. Galtier4, A. Avignon1,3, J. Mercier3, C. Bisbal3; 1CHU Lapeyronie, 2INSERM Unit 1046, 3INSERM Unit 1046, 4CIC, Montpellier, France. Background and aims: Obesity has deleterious effects on human health by enhancing the development of chronic diseases such as insulin resistance (IR) and type 2 diabetes (T2D). Nowadays, the hypothesis commonly admitted is that, during obesity, there is a hypertrophy and a hyperplasia of subcutaneous adipose tissue leading to its dysfunctions and increase lipolysis. Consequent- ly to this adipose tissue dysfunctions, subcutaneous adipose tissue cannot expand more .Then free fatty acid (FFA) concentration increases in systemic circulation and lipids are ectopically stored in visceral adipose tissue, liver and skeletal muscle which affect importantly their function. Moreover, mac- rophages infiltration of adipose tissue induces a local chronic low-grade in- flammation which develops to become systemic. This inflammation impairs the insulin response of insulin sensitive tissues (muscle, liver and fat) leading to peripheral IR. However, some obese people (25-30%) do not develop IR, and are metabolically healthy. In order to identify the early mechanisms lead- ing from obesity to IR, we conducted a translationnal study both at systemic and tissue levels Materials and methods: 31 post-menopausal women 50-65 years old, with no personal or familial history of type 2 diabetes were involved in this study.: (i)10 were leans (BMI 30kg/m2) and insulin sensitive (HOMA-ir 3): OIR. Skeletal muscle and subcutaneous adipose tissue were obtained from biopsies respectively in the vatus lateralis and abdominal subcutaneous adipose tissue. We analyzed the insulin response, inflammatory state at systemic and tissues level, intramuscular fatty acid accumulation, angiogenesis and adipocytes size of subcutaneous adipose tissue. This protocol was approved by the local Ethic Committee (N°:2011.01.04, ANSM: B110204-20) and informed written consent was obtain from all participants. Results: Ours results show a significant decrease in insulin sensitivity of OIR skeletal muscle compared to control (p=0.03) and OIS (p=0.04) skeletal mus- cle while subcutaneous adipose tissue remain insulin sensitive in all three groups of volunteers. Further, there is an increase of fatty acid accumulation in type I muscular fiber of OIR compare to control (p=0.02). Histological analysis show that macrophages infiltration is low in OIR’s subcutaneous adi- pose and absent from OIS and CT volunteer’s subcutaneous adipose tissue. Adipocytes size remains the same among three groups. However angiogen- esis is significantly decreased in subcutaneous adipose tissue of OIR (p=0.02) and OIS (p=0.007) compared to control. Conclusion: All these results suggest that skeletal muscle seem to be affected by IR before subcutaneous adipose tissue during IR development. The only one impairment which occurs in subcutaneous adipose tissue is the decrease in adipose tissue angiogenesis Clinical Trial Registration Number: 2010-A01522-37 Supported by: CHU Montpellier/SFD 587 Serum myostatin and insulin resistance in postmenopausal women V. Avila-Rubio, C. Novo-Rodriguez, B. Garcia-Fontana, A. Garcia-Martin, R. Reyes-Garcia, S. Morales-Santana, M. Muñoz-Torres; Bone Metabolism Unit of Endocrinology division of University Hospital Cecilio, Granada, Spain. Background and aims: Myostatin is a secreted growth factor expressed in skeletal muscle, which negatively regulates skeletal muscle mass. Recent ani- mal studies suggest an additional role for myostatin in glucose metabolism. We evaluated the possible associations between circulating levels of myosta- tin and glucose metabolism parameters in non obese women with postmeno- pausal osteoporosis. Materials and methods: Prospective study including 28 women with post- menopausal osteoporosis, matched for age, BMI and bone mineral density (BMD), divided into two groups according to the antiosteoporotic drug: A) 60 mg Denosumab subcutaneous semiannually (n = 17) and B) 20 μg Teri- paratide subcutaneous daily (n = 11), both for 3 months. The monitoring was conducted by the Bone Metabolism Unit of our hospital. We performed fasting plasma determinations of myostatin, glucose, peptide C, insulin, insu- lin resistance index (HOMA2-IR), insulin sensitivity (HOMA2-%S), insulin secretion (HOMA2-%β) and HbA1c at baseline and one week, one month and three months after treatment. Results: Both groups were comparable at baseline in the analyzed param- eters without significant differences between them. Serum myostatin levels remained unchanged regardless of treatment group. A global analysis of all determinations (108) shows a positive correlation between myostatin and peptide C (r = 0.231, p = 0.02), insulin (rs = 0.215, p = 0.02), HOMA2-IR (rs = 0.200, p = 0.04), HOMA2-%β (rs = 0.210, p = 0.03) and HbA1c (r = 0.209, p = 0.03) and an inverse correlation with HOMA2-%S (rs = -0197, p = 0.04). No correlations were observed between myostatin and other related carbohy- drate metabolism parameters as undercarboxilated osteocalcin, osteocalcin, IGF-1, vitamin D and BMI. Conclusion: The findings of the present study show significant association between circulating myostatin and insulin resistance in non obese women with postmenopausal osteoporosis. This study supports the idea that myosta- tin may exert a negative effect on glucose metabolism acting as an independ- ent mechanism in the pathogenesis of type 2 diabetes mellitus. Diabetologia (2014) 57:[Suppl1]S1–S564 S 244 1 C PS 037 Mechanisms of insulin action and inter-organ cross-talk 588 Intracerebroventricular administration of vaspin triggers a brain-liver circuit to improve hepatic glucose homeostasis via the hepatic branch of the vagus W. Zhu1,2, L. Li1,2, K. Li3, G. Yang3; 1College of Laboratory Medicine, 2Key Laboratory of Diagnostic Medicine (Ministry of Education), 3Endocrinology, The Second Affiliated Hospital, Chongqing Medical University, China. Background and aims: Visceral adipose tissue-derived serpin (vaspin) is a member serpin A12 of the serine protease inhibitor family, which plays an important role in the modulation of glucose metabolism and insulin sensitiv- ity. More and more findings raise the possibility that, vaspin regulates glucose metabolism and energy balance through its act on the central (hypothalam- ic) site. However, the regulatory role of vaspin in the brain in the control of liver glucose fluxes is unknown. In this study, we investigated the effects of vaspin signal conveyed by the hypothalamus on liver glucose fluxes in nor- mal-chow-diet (NCD) or high fat diet(HFD)-fed male rats with or without hepatic branch vagotomy. We reported here a novel role of central vaspin in triggering a brain-liver molecular signaling pathway and neuronal network to control glucose production in vivo Materials and methods: We established a model of central administration of vaspin. Hyperinsulinemic-euglycemic clamp and hepatic branch vagotomy were used to assess the effects of central vaspin on glucose metabolism and changes in liver and signaling pathway. [3-3H] glucose radioactivity was de- termined by scintillation counter. mRNA and protein expressions were meas- ured by qRT-PCR and Western blot, respectively. Results: We showed that central infusion of vaspin in HFD-fed animals significantly increased glucose uptake in peripheral tissues and decreased HGP. These changes were accompanied by a significant decrease in the he- patic glucose-6-phosphatase (G6Pase) and phosphoenolpyruate carboxyki- nase (PEPCK) expression. In agreement with this, we also found that central vaspin in HFD rats activates the insulin receptor (InsR) →insulin receptor substrate-1(IRS-1) →Akt kinase (Akt) →forkhead box-containing protein of the O subfamily 1(FoxO1) signaling cascade in the liver leading to the in- creased insulin sensitivity and improved glucose metabolism. Conclusion: Our findings suggest that hypothalamus is a site of vaspin ac- tion, where vaspin triggers a brain-liver circuit via the hepatic branch of the vagus nerve to inhibit glucose production and improve insulin resistance. Supported by: National Natural Science Foundation of China (81100567, 81300702, 81300670) 589 Abnormal hepatic energy metabolism in type 1 diabetes associates with long-term metabolic control S. Gancheva1, A. Bierwagen1, P. Begovatz1, J. Lundbom1, P. Nowotny1, G. Giani2, B. Hoffmann3, J. Szendroedi1,4, M. Roden1,4; 1Institute for Clinical Diabetology, German Diabetes Center, 2Institute for Biometry and Epidemiology, German Diabetes Center, 3Leibniz Research Institute for Environmental Medicine, 4Department of Endocrinology and Diabetology, University Hospital Duesseldorf, Duesseldorf, Germany. Background and aims: Insulin resistant patients with type 2 diabetes exhibit lower hepatic energy metabolism, which has been explained by increased lipid availability and/or impaired glycemic control. This study set out to ex- amine the role of insulin resistance and glycemic control for hepatic energy metabolism in type 1 diabetes mellitus (T1D) patients with good glycemic control. Materials and methods: Thirty-seven newly diagnosed T1D patients (age: 34±9 years; body mass index (BMI): 24.4±3.3 kg/m²) with known diabetes du- ration of less than one year underwent euglycemic-hyperinsulinemic clamps to assess whole-body (M-value) and hepatic insulin sensitivity (insulin-me- diated percent suppression of endogenous glucose production). Hepatic fat content (HCL), ɣ-adenosine triphosphate (ɣATP) and inorganic phosphate (Pi) concentrations were quantified via magnetic resonance spectroscopy (31P/1H-MRS) on a 3-T clinical MR scanner (Philips, Best, The Netherlands). BMI- and age-matched healthy controls (CON; n=27, age: 38±9 years; BMI: 24.4±3.4 kg/m²) were also studied with 31P/1H-MRS. Results: Well-controlled T1D patients (HbA1c: 6.5±1.1%) presented high whole-body and moderate hepatic insulin sensitivity (7.8±2.9 mg*kg-1*min-1 and 62±24%). HCL was comparable between T1D and CON (1.7±3.1% and 1.9±3.9%). Hepatic ɣATP concentrations were 16% lower in T1D (2.22±0.54 vs CON: 2.61±0.55 mmol/l, P S 245 1 C 591 The association of serum fibroblast growth factor 21 with insulin secretory function and metabolic parameters in type 2 diabetes M.-J. Kim1, M.-K. Kim2, I. Song2, H. Kim2, Y.-K. Choi3, K.-H. Bae3, K.-G. Park3, I.-K. Lee3, J.-E. Lee4, E. Jung5, E. Kim6; 1Pooren Meerae Internal Medicine Clinic, 2Department of Internal Medicine, Keimyung University School of Medicine, 3Kyungpook National University School of Medicine, 4CHA Gumi Medical Center, 5Department of Internal Medicine, Catholic University of Daegu School of Medicine, 6Daegu Fatima Hospital, Republic of Korea. Background and aims: Fibroblast growth factor 21 (FGF21) is a cytokine produced by the liver, adipose tissue, skeletal muscle, and the pancreas. Re- cently, FGF21 is suggested as a key metabolic regulator in glucose and lipid metabolism and its level in serum is higher in the patients with metabolic syndrome, type 2 diabetes and non-alcoholic steatohepatitis. On the other hand, diminished expression of FGF21 is shown in type 1 diabetes and la- tent autoimmune diabetes in adult (LADA). Type 2 diabetes is a heterogenic disease entity and patients with type 2 diabetes have diverse insulin secre- tory dysfunction. So, we investigated the association of serum FGF21 and insulin secretory function in patients who admitted our institute for poorly controlled blood glucose and needed insulin therapy. Materials and methods: We recruited 154 patients (men 52; women 152) with type 2 diabetes who were not treat with insulin prior to informed con- sent from January 2011 to July 2013, who admitted Keimyung University Dongsan Medical Center. We reviewed personal history including drug his- tory, anthropometric parameters, other metabolic parameters, and measured serum adiponectin and FGF21. Results: Mean age was 55.4±14.0 in men and 61.4±14.1 in women and body mass index was 24.2±4.7 kg/m2 in men and 25.0±4.2kg/m2 in women. There was no significant difference in hemoglobin A1c (10.4% in men; 10.3% in women). Fasting C-peptide, insulin and homeostasis model assessment of insulin secretion (HOMA-IS) were used as assessment of insulin secretory function. Serum FGF21 level shows correlations with HOMA-IS and fasting insulin in men, but not in women. In addition, serum FGF21 level shows correlations with serum triglyceride, insulin and homeostasis model assess- ment of insulin resistance (HOMA-IR) in men, however, shows correlations with urine albumin/creatinine and HOMA-IR in women. As a result of multi- variate linear regression analysis, only HOMA-IS has a significant correlation with serum FGF21 level. Serum adiponectin level shows negative correlations with duration of diabetes, BMI, total cholesterol and low density lipoprotein in men and shows negative correlation with BMI, waist circumference, total cholesterol and triglyceride, high density lipoprotein in women. Conclusion: Consequently, this study exhibits a positive correlation of serum FGF21 level with an index of insulin secretory function in men with type 2 DM and supports the result of another study with type 1 diabetes and LADA. 592 Novel paracetamol glucuronide derivative for quantifying gluconeogenesis using the deuterated water method J. Jones1,2, S. Kahl3,4, F. Carvalho1, C. Barosa1, M. Roden3,4; 1Metabolic Control Croup, Center for Neurosciences, Coimbra, Portugal, 2Portuguese Diabetes Association (APDP), Lisbon, Portugal, 3Institute of Clincal Diabetology, Deutsches Diabetes-Zentrum, 4Department of Endocrinology and Diabetology, Heinrich-Heine University, Düsseldorf, Germany. Background and aims: In fasted humans, the analysis of plasma glucose 2H- enrichment in position 5 relative to position 2 (H5/H2) following ingestion of deuterated water (2H2O) is widely used to measure the fractional contri- bution of gluconeogenesis to endogenous glucose production (EGP). This information can be noninvasively obtained via urinary Paracetamol glucu- ronide (PG): a product that is more stable and abundant compared to blood glucose. However, current methods for derivatizing urinary PG to a form that can be analyzed for positional 2H-enrichment by NMR or MS are laborious and ill-suited for high-throughput studies. We developed a novel procedure where urinary PG is derivatized to 5-O-acetyl monoacetone glucuronic lac- tone (MAGLA). The chemical transformation and purification steps are ro- bust and amenable to automation and parallel processing and yield resolved signals for H5/H2 analysis by 2H NMR (see Figure). We applied this proce- dure to a study where 2H2O and Paracetamol were given to overnight fasted healthy subjects and compared H5/H2 of plasma glucose with that of urinary glucuronide. Materials and methods: Eleven subjects were admitted to the clinical re- search unit on the evening before the study and provided a standard supper at 17.30. Each ingested 0.5 g 2H2O/kg body water in three equally divided doses at 20:00, 22:00, and 24:00. At 05:00 the following day, Paracetamol (0.5g) was given and a primed infusion of [6,6-2H2]glucose was initiated. Urine was sampled from 06:00-08:00 and blood was sampled at 07:40. Plasma glucose and urinary PG enrichments of positions 5 and 2 were measured by 2H NMR following their derivatization to monoacetone glucose and MAGLA, respec- tively. The fractional gluconeogenic contributions to EGP were calculated from plasma glucose H5/H2. These were compared with estimates obtained from urinary glucuronide H5/H2. Results: The overall yield of MAGLA from urinary glucuronide was 20-40% and the preparations gave well-resolved 2H NMR signals for quantifying H5/ H2. Analysis of MAGLA yielded identical estimates of fractional gluconeo- genic contributions to EGP to that of plasma glucose (54 ± 2% versus 55 ± 3%, respectively). Furthermore, a Bland-Altman analysis indicated agree- ment at the 95% confidence level between the sets of plasma glucose and urinary MAGLA measurements. Conclusion: The conversion of urinary Paracetamol glucuronide to MAGLA is a relatively simple and robust procedure for obtaining estimates of gluco- neogenesis using 2H2O, or indeed any gluconeogenic tracer. For overnight- fasted healthy subjects, 2H NMR analysis of MAGLA provides identical esti- mates of fractional gluconeogenesis to that of plasma glucose. 593 BTG2 regulates hepatic gluconeogenesis through Nur77 in diabetes S.-S. Im1, Y.-D. Kim1, B.-C. Oh2, J.-H. Bae1, D.-K. Song1; 1Physiology, Keimyung University School of Medicine, Deagu, 2Lee Gil Ya Cancer and Diabetes Institute, Inchon, Republic of Korea. Background and aims: Beta cell translocation gene 2 (BTG2) is a member of a family that is involved in cellular functions. Herein, we demonstrate that BTG2 regulates glucose homeostasis via up-regulation of Nur77 in diabetic mice. Materials and methods: Diabetes was induced in wild-type (WT) and db/ db mice by i.p injections of streptozotocin (80 mg/kg body weight). GTT and ITT were performed with db/db mice. Mice were infected with adenoviral vector expressing Btg2(1×109 pfu) by tail vein injection. Gene expression and induction level were mseasured by qPCR and immunoblotting. Results: Hepatic BTG2 gene expression was elevated by fasting and forsko- lin. Overexpression of BTG2 increased the expression of hepatic gluconeo- genic genes, glucose output, and subsequently impaired glucose and insulin tolerance. Up-regulation of the transcriptional activity of Nur77 and gluco- neogenic genes, and glucose production by forskolin were observed by BTG2 transduction, but not in BTG2 knockdown. BTG2-stimulated glucose pro- duction and G6pc promoter activity was attenuated by dominant-negative Nur77, but not in Pck1 promoter. Chromatin immunoprecipitation assays showed that BTG2 induced Nur77 occupancy on the glucose-6-phosphatase gene promoter via a physical interaction. BTG2 gene expression was in- creased in streptozotocin-treated and db/db mice. Finally, impairment of Diabetologia (2014) 57:[Suppl1]S1–S564 S 246 1 C glucose homeostasis such as the increase of blood glucose, glucose intoler- ance and insulin intolerance was elevated in diabetic mice, whereas this phe- nomenon was abolished in knockdown of BTG2. Conclusion: These data suggest that BTG2 participates in the regulation of hepatic glucose homeostasis, which involves that BTG2 might serve as a po- tential therapeutic target for combating metabolic dysfunction. Supported by: KHIDI (A120864) to SSI 594 New diabetes drug targets from structural characterisation of liver glycogen R.G. Gilbert1,2, M.A. Sullivan2,1, S.A. Aroney2, S. Li2, P.O. Powell2, F.J. Warren2; 1Tongji School of Pharmacy, Huazhong University of Science and Technology, Wuhan, China, 2Centre for Nutrition and Food Science, University of Queensland, Brisbane, Australia. Background and aims: Glycogen is a highly branched polymer of glucose, functioning as a blood-glucose buffer. It comprises relatively small beta par- ticles, which may be joined as larger aggregate alpha particles. The molecular size distributions from size-exclusion chromatography of liver glycogen from non-diabetic and diabetic mice show that diabetic mice have impaired alpha particle formation in liver glycogen, with diabetic mice unable to form as many large glycogen particles as non-diabetic controls. There is also reason to believe that the degradation of larger particles back to glucose will be slower, and hence under better control, than smaller beta particles. Further analysis into the nature of “healthy” glycogen has provided evidence that these alpha particles are held together via an unknown glue that is most likely proteina- ceous. This study probes the structure of glycogen alpha particles by ana- lyzing their digestion during acid hydrolysis and by comparing the in vitro digestion rate of glycogen phosphorylase for glycogen of various sizes. This greater insight into the internal nature of alpha particles is an important step into finding potential drug targets that would allow diabetic sufferers to form these larger alpha particles and hence have better blood-sugar control. Materials and methods: The digestion of glycogen by acid hydrolysis was investigated by analyzing the size distributions, obtained using size-exclusion chromatography, of pig-liver glycogen and phytoglycogen at various time points in a low-pH buffer. In addition, glycogen from the livers of wild-type mice were extracted and fractionated into different sizes using sucrose den- sity centrifugation. The size distributions of these fractions were compared using size-exclusion chromatography. The in vitro kinetics of the initial rates of phosphorylase degradation were examined and compared for each frac- tion using an assay that utilized the production of NADPH in the reaction mixture, which has an absorbance at 340 nm. Results: The acid hydrolysis experiments show that the alpha particles in the liver glycogen degrade via a different mechanism to that of the smaller beta particles, indicating a different type of chemical bonding (most likely pro- teinaceous). The in vitro kinetics analysis of glycogen phosphorylase on frac- tionated glycogen showed that larger glycogen particles indeed had a lower initial rate of phosphorylase degradation. Conclusion: The difference in kinetics of the acid hydrolysis of alpha and beta particles indicates that they have a different mechanism of biosynthesis. The hypothesis that larger glycogen alpha particles are more slowly degraded by glycogen phosphorylase, presumably due to a lower surface area to vol- ume ratio, was supported by the present data. This shows that diabetic mice, which have been shown to lack the larger alpha particles, may be vulnerable to having a fast, uncontrolled release of glucose. Because alpha particles have been shown here to have a separate mechanism of formation to smaller beta particles, the activation of this process in diabetic sufferers is now a potential drug target. 595 Post-prandial and ChREBP mediated PANDER expression W.A. Danse, M.N. Kuehl, A.C. Chechele, B.R. Burkhardt; Cell Biology, Microbiology and Molecular Biology, University of South Florida, Tampa, USA. Background and aims: Pancreatic-Derived Factor (PANDER, FAM3B) is a novel endocrine pancreatic secreted protein mainly characterized from this tissue. PANDER regulates hepatic insulin sensitivity and overall glyce- mic levels. However, recent research now suggests that PANDER is poten- tially expressed within the liver as well. Our previous data have shown the PANDER promoter to be highly active and glucose-responsive in a murine liver-derived cell line. To further investigate the regulation and expression of hepatic derived PANDER, we examined hepatic PANDER levels in mice and characterized the minimal element of the PANDER promoter in a liver- derived cell line. Materials and methods: Western analysis was performed to evaluate the presence of PANDER protein within murine tissues. Promoter deletion analysis and co-transfection experiments were performed using PANDER promoter-luciferase constructs and ChREBP expression plasmids in a mu- rine liver-derived cell line, BNL-CL2. Results: Initial western analysis performed on 15 different murine-derived tissues revealed robust PANDER expression within the murine liver and to a lesser extent in the brain and ovary. Furthermore, hepatic PANDER lev- els were measured in the fasted and fed state of our PANDER transgenic and control mice and revealed increased levels in the post-prandial state. The minimal element was identified between -293 to -3 bp of the hepatic transcriptional start site. The region between -193 to -93 bp of the TSS was demonstrated as being necessary for basal promoter activity. This region con- tains three E-box elements which may serve as a binding site for the critical hepatic transcription factor carbohydrate responsive element binding protein (ChREBP). Co-transfection reporter studies revealed that PANDER promot- er activity is significantly upregulated (p S 247 1 C (11.8±6.8), pre-diabetic (14.2±8.7) and diabetic (13.8±11.0 µU/mL) groups. However, HOMA2-IR values [median (IQR)] were 1.37 (0.99-2.12) for con- trol, 1.71 (1.27-2.60) for pre-diabetic and 1.95 (1.13-3.23) for diabetic groups (p S 248 1 C 598 Impact of adiposity on glycaemic variability with type 2 diabetes T. Mayu, H. Makino, T. Tamanaha, Y. Ohata, R. Koezuka, A. Hishida, I. Kishimoto; Diabetes, National cerebral and Cardiovascular Center, Suita City, Japan. Background and aims: High glycemic variability generates oxidative stress and contributes to the development of endothelial dysfunction and cardio- vascular complications in patients with type 2 diabetes. However, the factors affecting glycemic variability have not been fully studied. In the present study, we examined the effect of body status, body composition and abdominal fat area on the parameters of glycemic variability determined by continuous blood glucose monitoring system (CGMS) in subjects with type 2 diabetes. Materials and methods: Among patients referred to our department between August 2012 and February 2014, 41 patients with type 2 diabetes (mean age 68.7 years; 70 % men; HbA1c 8.4 %; BMI 24.2; energy intake; 22-32 kcal/ ideal body weight (60 % carbohydrate)) who underwent CGMS were select- ed. Their antihyperglycemic therapy consisted of oral hypoglycemic agents (n=31) and/or insulin (n=32) (None were diet alone). Based on CGMS data, we evaluated the following criteria for variability of glycemic control from data of the intermediate 48 hours (midnight to midnight): average glucose levels, standard deviation (SD) values, mean amplitude of glycemic excur- sions (MAGE), percentage of time spent in hyperglycemia (> 140 mg/dL) and percentage of time spent in hypoglycemia (< 70 mg/dL). We investigated the effect of body status (body mass index (BMI)) for these glycemic variability indices. Furthermore, the impact of body composition (body fat mass (kg), skeletal muscle mass (kg); n=19) and of abdominal fat area (subcutaneous fat area (SFA, m2), visceral fat area (VFA, m2); n=31) on glycemic variability is also evaluated in the subgroup analyze. Results: Among all participants, the mean ± SD of average glucose levels, SD of glucose variability and MAGE were 140 ± 27 mg/dl, 32 ± 13 mg/dl and 81 ± 28 mg/dl, respectively. When patients were divided into two cat- egories by median (74 mg/dl) split of MAGE, higher MAGE category had a significantly lower BMI (p=0.004) than those with lower MAGE category. In univariate analysis, BMI was significantly negatively correlated with SD (p S 249 1 C with HOMA-IR was abolished (-0.077, p=0.121). Furthermore, when HO- MA-IR was corrected for, the inverse association between android % fat and IGFBP1 was retained (-0.241, p=9.2x10-7). By contrast, gynoid % fat was in- versely associated with HOMA-IR after controlling for total fat mass (-0.154, p=1.8x10-3) and the association with IGFBP1 was abolished (0.095, p=0.054). These findings could not be replicated using anthropometric measures of regional fat depots. For example, after adjusting for total fat mass, waist cir- cumference was equally associated with HOMA-IR (0.116, p=1.9x10-2) and IGFBP1 (-0.113, p=2.3x10-2). In females, after controlling for total fat mass, android and gynoid % fat also displayed opposite directions of association with HOMA-IR (android: 0.124, p=1.4x10-2; gynoid: -0.197, p=9.0x10-5) and IGFBP1 (android: -0.145, p=4.2x10-3; gynoid: 0.161, p=1.4x10-3) however the strength of association was similar for both depots. Conclusion: Our findings highlight the advantage of quantifying specific fat depots by DXA over anthropometric measures when assessing insulin resist- ance. In males, hepatic insulin resistance (assessed by IGFBP1) was inde- pendently associated with greater android fat content whereas whole-body insulin resistance (HOMA-IR) was associated with diminished gynoid fat content. These findings lend further support to the paradoxical relationship between diabetes risk and body fat distribution. Supported by: NIHR Biomedical Research Centre 601 A study of the mechanisms through which the biogenesis of adiponectin is regulated in white adipose tissue L. Zhang1, M.M. Li1, M. Corcoran1, S. Zhang1, G.J.S. Cooper1,2; 1School of Biological Sciences, University of Auckland, New Zealand, 2Manchester Biomedical Research Centre, University of Manchester, UK. Background and aims: Adiponectin is an adipose tissue-derived hormone with anti-diabetic, anti-atherogenic and anti-inflammatory functions. De- fects in adiponectin multimerization are associated with decreased adiponec- tin levels, obesity and insulin resistance. In this study, we investigated the en- zyme systems responsible for adiponectin multimerization and production, and other factors also contributing to this process. Materials and methods: OB mice (10- or 27-week) and age-matched C57 controls were used in this study. Fully differentiated 3T3-L1 adipocytes were also used. Quantitative real time-PCR was performed to measure the mRNA expression levels of a number of genes in white lipid tissue (WAT) from mice and 3T3-L1 adipocytes. ELISA/western blots were used to measure the ex- pression of relative protein levels. Statistical analysis was performed by t-test or analysis of variance (ANOVA) with appropriate post-hoc tests. Results: In fully-differentiated 3T3-L1 adipocytes, we observed that de- creased insulin signalling caused by blocking the insulin receptor (InsR) with a blocking anti-InsR antibody, increased extracellular adiponectin lev- els, whereas coexisting hyperinsulinaemia counteracted this effect. Blocking the adenosine monophosphate-activated protein kinase (AMPK) pathway by using an inhibitor (compound C) significantly decreased extracellular adiponectin levels. Furthermore, we demonstrated the expression of ly- syl hydroxylases (LHs), prolyl hydroxylases (PHs) and glycosyltransferase 25-domain-containing proteins 1&2 (Glt25D1&2) in the WAT of mice. Ex- pression of LH3 was markedly increased in the WAT of OB mice compared with controls. In differentiated 3T3-L1 adipocytes, non-specific inhibition of LHs and PHs by the hydroxylase inhibitor, dipyridyl markedly suppressed adiponectin production, particularly of the higher molecular-weight iso- forms, but increased the low molecular-weight isoform. Specific inhibition of prolyl-4-hydroxylase (P4H) with ethyl-3,4-dihydroxybenzoate had effects comparable to dipyridyl: however, it had different effects on the intracellular composition of adiponectin isoforms. Specific inhibition of LHs with minoxi- dil also suppressed adiponectin production, especially of the high molecular- weight (HWM) isoform. In addition, transient gene knock-down of LH3 (Plod3) suppressed adiponectin production, especially of the HMW isoform. Conclusion: Our results indicate that: hyperinsulinaemia and decreased insulin signalling exert countervailing effects on adiponectin production; AMPK pathway signalling also regulates adiponectin production; both PHs and LHs are required for adiponectin production; and the hydroxylation/gly- cosylation of lysyl residues and hydroxylation of prolyl residues are essential for formation and secretion of higher-order adiponectin isoforms. Supported by: Health Research Council of New Zealand 602 The plasma levels of adipokines and incretins in patients with Alzheimer’s disease B. Bendlová1, J. Vcelak1, D. Vejrazkova1, P. Lukasova1, O. Bradnova1, K. Dvorakova1, G. Vacinova1, R. Rusina2, I. Holmerova3, M. Vankova1; 1Molecular Endocrinology, Institute of Endocrinology, 2Neurology, Thomayer Hospital, 3Faculty of Humanities, Charles University, Prague, Czech Republic. Background and aims: Alzheimer’s disease (AD) is linked to type 2 diabe- tes (T2D). T2D is a risk factor for vascular dementia but also for a progres- sive neuron damage, which is probably associated with insulin resistance in central nervous system. AD is therefore sometimes called “diabetes of the brain”. The aim of the study was to compare the plasma levels of adipokines, incretins and other biomarkers associated with T2D in AD patients and in non-diabetic subjects without pathological changes in the brain. Materials and methods: The study included 38 females (19 age and BMI matched pairs AD patient and control; age 70±8 years and BMI 26±4 kg/m2) and 24 males (12 age and BMI matched pairs AD patient and control; age 67±7 years and BMI 27±3 kg/m2) who underwent neuropsychological ex- amination and magnetic resonance imaging of the brain. The characteristics of fasting glucose and lipid metabolism and parameters of body composition including body adiposity index (BAI) were determined. Multiplex methods for evaluation of fasting plasma biomarkers were used: Bio-Plex ProHuman Diabetes 10-Plex Assay - C-peptide, ghrelin, GIP, GLP-1, glucagon, insulin, leptin, total PAI-1, resistin, visfatin and Bio-Plex ProHuman Diabetes Adip- sin and Adiponectin Assays (Bio-Rad). Statistical analyses were performed using NCSS 2004 software (ANOVA). Results: Compared groups of AD patients and controls (men/women sepa- ratelly) did not differ in either age, BMI, WHR, waist and abdomen circum- ferences, body adiposity index (BAI) or insulin resistance (HOMAR). In women, AD patients had significantly higher levels of visfatin (p=0.0006), ghrelin (p=0.03), GLP-1 (p=0.006) and glucagon (p=0.00002) compared with control women. Similar results were found in men, they had higher levels of visfatin (p=0.02) and ghrelin (p=0.02) compared with control men, GLP-1 and glucagon did not reach statistical significance. Conclusion: Alzheimer´s disease is associated with increased levels of plas- ma visfatin, ghrelin, GLP-1 and glucagon. Plasma visfatin is a pro-inflamma- tory cytokine that increases risk of endothelial cell deterioration due to the inflammation and oxidative stress. Visfatin is a marker of ageing and age-de- pendent diseases. Ghrelin, GLP-1 and glucagon are considered as neuropro- tective hormones. Dysregulation of these hormones in AD could be caused by impaired secretion/degradation and/or resistance of the target tissues. Supported by: IGA MH CZ NT/13543/4, NT/13544/4, MH CZ 00023761 603 Anti-TNF therapy has a possible favorable effect on insulin sensitivity in non-diabetic, non-obese patients with inflammatory bowel disease S.A. Paschou1, F. Kothonas2, A. Myroforidis2, V. Loi1, T. Terzi1, O. Karagianni1, A. Poulou2, K. Goumas2, A. Vryonidou1; 1Department of Diabetes and Endocrinology, 2Department of Gastroenterology, Hellenic Red Cross Hospital, Athens, Greece. Background and aims: Insulin resistance is very common in autoimmune systemic diseases and recently it was also found in children and adults with inflammatory bowel disease (IBD). Inflammation and insulin resistance are closely linked, and inflammatory cytokines such as tumor necrosis factor al- pha (TNFa) may inhibit insulin signaling and promote insulin resistance. The aim of this study was to investigate the effect of anti-TNF therapy on glucose and lipid metabolism in non-diabetic, non-obese patients with IBD. Materials and methods: We studied 41 patients with IBD (25M/16F, 36.4 ± 11 (19-64) years old, 28 with Crohn’s disease and 13 with ulcerative co- litis), without known history of diabetes. Eighteen patients (9M/9F, 33.6 ± 8.8 years) were on anti-TNF therapy for more than 1 year, while the other 23 patients (16M/7F, 38.7 ± 12.5 years) were treated with azathioprine and mesalazine (Aza/Mes). Nine of the patients from the second group were then treated with anti-TNF and studied again 6 months after. Fasting glucose, in- sulin, c-peptide, HbA1c, lipids, and CRP levels were determined and HOMA- IR index was calculated, in all patients. Statistical analysis of the data was performed using SPSS 16.00. Results: Three of the patients were diagnosed with overt diabetes and were excluded from the analysis. Patients from the two therapy groups were Diabetologia (2014) 57:[Suppl1]S1–S564 S 250 1 C matched for age (anti-TNF: 33.6 ± 8.8 years vs Aza/Mes: 38.7 ± 12.5 years, p>0.05) and BMI (anti-TNF: 23.3 ± 3.4 vs Aza/Mes: 23.1 ± 1.7, p>0.05), and were not obese. We did not find any statistical differences between the pa- tients from the two therapy groups in the levels of fasting glucose (anti-TNF: 88 ± 10.7 vs Aza/Mes: 93.4 ± 14.9 mg/dl, p>0.05), insulin (anti-TNF: 10.9 ± 7.9 vs Aza/Mes: 12.1 ± 6.6 mIU/ml, p>0.05), c-peptide (anti-TNF: 1.9 ± 0.9 vs Aza/Mes: 2.2 ± 1.4 ng/ml, p>0.05), HbA1c (anti-TNF: 5.2 ± 0.3 vs Aza/Mes: 5.3 ± 0.4 %, p>0.05), total cholesterol (anti-TNF: 168.6 ± 32.7 vs Aza/Mes: 162.8 ± 34.3 mg/dl, p>0.05), HDL (anti-TNF: 57.5 ± 15.7 vs Aza/Mes: 53.8 ± 20.3 mg/dl, p>0.05), LDL (anti-TNF: 95.8 ± 28.7 vs Aza/Mes: 90.7 ± 24.4 mg/ dl, p>0.05), triglycerides (anti-TNF: 75.8 ± 37.6 vs Aza/Mes: 90.8 ± 61.3 mg/ dl, p>0.05), CRP (anti-TNF: 3 ± 5.4 vs Aza/Mes: 4.9 ± 6.1, p>0.05) and in the HOMA-IR index (anti-TNF: 2.77 ± 2 vs Aza/Mes: 3.1 ± 1.9, p>0.05). In pa- tients who were treated for 6 months with anti-TNF, a statistically significant decrease in insulin (before: 15.4 ± 5.8 vs after: 10.2 ± 2.7 mIU/ml, p=0.049) and c-peptide (before: 2.4 ± 1.2 vs after: 1.4 ± 0.4 ng/ml, p=0.038) levels as well as the HOMA-IR index (before: 4.1 ± 2.1 vs after: 2.3 ± 0.7, p=0.047) was observed, without any statistically significant changes in weight, BMI, glucose, HbA1c, lipids and CRP levels (in all comparisons p>0.05). Conclusion: These preliminary data indicate that anti-TNF therapy may have a favorable effect on insulin sensitivity in non-diabetic, non-obese patients with inflammatory bowel disease. 604 Dual character of osteoprotegerin as a metabolic biomarker in a healthy/ prediabetic population B. Buday1, F.P. Pach2, B. Literati-Nagy1, Z. Vecsei1, M. Vitai1, L. Koranyi1; 1Department of Metabolism, Drug Research Center, Balatonfured, 2Department of Medical Informatics, Bay Zoltán Nonprofit Ltd. for Applied Research, Budapest, Hungary. Background and aims: Osteoprotegerin (OPG) is a member of the tumor necrosis factor receptor superfamily involved in bone metabolism (OPG/ RANK/RANKL axis), cardiovascular (CV) system and the bone - energy ho- meostasis. It protects against vascular calcification in animal models while its high level has been linked to diabetic complications, CV death, inflammation and poor glycemic control in humans. However some conflicting data show that in normal glucose tolerant (NGT) subjects, these connections do not exist or show opposite results. To clarify this issue and the possible sex differ- ence we investigated the metabolic connections of OPG in a cohort of healthy or prediabetic male and female subjects. Materials and methods: We included 155 males and 154 females (aged 49.18± 9.43 vs.41.62±12.95 years) in our study. They were diagnosed as NGT or glucose intolerant (GI = IGT, IFG or drug naïve 2DM) by OGTT, with no known diabetic/CV complication. IVGTT and a hyperinsulinemic euglyce- mic clamp were done to determine insulin secretion and sensitivity (IS=Mm muscle glucose uptake). Serum glucose, insulin, free fatty acid (FFA), adipo- cytokines, bone markers, lipid subfractions (SF: small-dense (SD): LDL-3,4,5, large-buoyant (LB): LDL-1-2.), apoproteins, anthropometric data (body fat percent, abdominal circumference (AC)) were measured. Serum OPG levels were determined by Biomedica ELISA kits. Results: OPG levels did not differ significantly between the NGT and GI groups in either gender. In both genders it showed a significant (p S 251 1 C PS 039 Glucagon 606 Fasting hyperglucagonaemia in patients with non-alcoholic fatty liver disease with and without type 2 diabetes A.E. Junker1, L. Gluud2, F.K. Knop1, J.J. Holst3, T. Vilsbøll1; 1Diabetes Research Division, Hellerup, 2Department of Gastroenterology, Hvidovre, 3Department of Biomedical Science, Copenhagen, Denmark. Background and aims: Non-alcoholic fatty liver disease (NAFLD) and hy- perglucagonemia seems to influence glucose metabolism and both features are associated with type 2 diabetes (T2D). We investigated glucagon respons- es during oral and intravenous glucose administration in biopsy-verified NAFLD patients with and without T2D. Materials and methods: We performed a 50 g-OGTT, and isoglycaemic intravenous glucose infusion (IIGI) in participants with I) normal glucose tolerance (NGT) and no NAFLD (N=10; age (median±interquartile range): 58±17 years; BMI: 29±1 kg/m2: HbA1c: 34±6 mmol/mol (5.5±0.2%)), II) NGT+NAFLD (N=9; age: 54±22 years; BMI: 28±7 kg/m2: HbA1c: 33±4 mmol/ mol (5.2±0.3%)), III) T2D and no NAFLD (N=8; age: 59±18 years: BMI: 27±3 kg/m2: HbA1c: 45±9 mmol/mol (6.2±1.1%)), or IV) T2D+NAFLD (N=8; age: 65±3 years; BMI: 30±4 kg/m2; HbA1c: 53±19 mmol/mol (7.0±1.4%)). Results: In all groups, isoglycaemia was achieved during oral and intrave- nous glucose administrations. Fasting plasma glucagon was elevated in par- ticipants with NAFLD and NGT (7.0±2.7 pM) and T2D (6.7±2.3 pM), com- pared to participants without NAFLD and NGT (2.8±4.0 pM, P S 252 1 C expression was assessed by qPCR analysis and SGLT1/2 proteins by histology and ELISA. RNA interference or SGLT inhibitors such as phlorizin or dapa- gliflozin was used to inhibit SGLT2 in alpha cells, human islets or healthy mice. Results: Here, we demonstrate that SGLT1/2 proteins (encoded by the SLC5A1/2 genes) are specifically expressed in human pancreatic alpha cells. SLC5A2 gene expression is lower in islets isolated from T2D patients (*p S 253 1 C role of the atypical orphan receptor GPR17, which is activated by uracil nu- cleotides and can be partially antagonized by ATP, a well known autocrine/ paracrine signal in islets. We first verified the GPR17 expression in endocrine cell lines and in human islets, and then we investigated its possible role in islet physiology Materials and methods: The expression of GPR17 was analysed by RT- PCR analysis, western blot and immunofluorescence techniques in cell lines (αTC3, βTC3, the δ cell line RIN14B), in human isolated islets of Langerhans, and human pancreatic sections. The possible involvement of GPR17 in the control of hormone release and cell proliferation/viability was studied in cell lines and in human isolated by means of ELISA assays. Results: By means of RT-PCR, we detected GPR17 mRNA in human islets, in α and δ cell lines, but not in β cell lines (n=5). By means of western blot experiments, we confirmed GPR17 protein expression in α and δ cell lines and in human islets, where its abundance was upregulated by hyperglycemia (2.13+0.15 fold increase over normoglycemic conditions; p S 254 1 C PS 040 Exercise physiology I 615 Improvement in insulin sensitivity following exercise training is independent of mitochondrial changes in older adults F. Amati1, C. Greggio1, N.T. Broskey1, W. Janssens1, A. Boss2, R. Kreis2, C. Boesch2; 1University of Lausanne, 2University of Bern, Switzerland. Background and aims: Previous studies reported lower skeletal muscle mi- tochondrial oxidative capacity as one of the links between type 2 diabetes and aging. This gave rise to the interesting indication that mitochondrial deficien- cy may be an important factor in the pathogenesis of insulin resistance. We have recently shown that exercise training improves mitochondrial content and function in older sedentary adults. The purpose of this study was to link these improvements with improvements in insulin sensitivity. Materials and methods: Twenty sedentary older (age range 61 to 72) men and women (11M, 9F), with average BMI 28±4 kg/m2 and either impaired or normal glucose tolerance (11 IGT, 8 NGT), participated in a four months supervised endurance exercise intervention with a pre/post-intervention design. Investigations included glucose clamps to measure insulin sensitiv- ity, dual X-ray absorptiometry for overall body composition, magnetic reso- nance imaging for visceral adipose tissue (VAT) and graded exercise testing for maximal oxygen consumption (VO2max). Maximal ATP production ca- pacity (ATPmax) was used as marker for mitochondrial function and was es- timated by phosphocreatine recovery following exercise using 31P magnetic resonance spectroscopy. Vastus lateralis muscle biopsies specimen were used to measure mitochondrial content by electron microscopy and western blot- ting. Transcription factors involved in mitochondrial biogenesis were meas- ured in a subset of volunteers by quantitative RT-PCR. Results: The following changes were observed with intervention (data are mean change±SEM): insulin sensitivity (glucose infusion rate, GIR) +41±6% (P S 255 1 C of these training adaptations have not yet been explored in older subjects at risk for type 2 diabetes. The purpose of this study was to examine the changes in exercise efficiencies and exercise fat oxidation in older sedentary adults undergoing an exercise intervention and compare these values to athletes of the same age. Materials and methods: In this pre/post intervention design, 11 sedentary obese (age 67.0±1.1) and 12 sedentary lean (age 64.6±1.0) men and wom- an underwent four months of endurance exercise intervention. Volunteers trained 3 supervised sessions per week at 60-75% of their maximal heart rate. Fifteen athletes matched by age (age 68.5±1.5) were also used for comparison purposes. Peak oxygen consumption (VO2peak) was determined by a graded exercise testing (GXT). Gross efficiency (GE, %) and exercise fat oxidation (EFO, µmol/min) were determined during one-hour of submaximal (55% of VO2peak) cycle ergometry exercise in tightly controlled conditions (fasting and 72 hours from last exercise bout). Delta efficiency (DE, %) was calculated from the GXT. Leg mass, used to normalize efficiency data, was measured by dual X-ray absorptiometry. Statistical comparisons between groups at base- line were done with ANOVA. Intervention effects were measured by Paired T Test. Results are presented in mean ± SEM. Results: At baseline, athletes had a higher GE than lean sedentary, which had a higher GE than the obese sedentary (p S 256 1 C cise (IT) for 10 weeks (a 2-week conditioning phase + 8-week exercise pro- gramme, 3 times/week) or no activity (SED) for the same time period. Fasting plasma glucose, insulin, HbA1c, triglycerides (TG), total cholesterol, HDL and LDL cholesterol and body composition (DXA) were measured at base- line and at study end. Results: Baseline characteristics of the 3 groups were comparable. After 10 weeks, there were no significant differences from baseline in HbA1c% [pre vs. post: 6.7 (6.5-7.1) vs. 6.6 (6.5-6.9); 6.6 (6.5-7.0) vs. 6.5 (6.5-6.6); 6.7 (6.1- 7.5) vs. 6.6 (6.4-7.2) in the CL, IT and SED groups, respectively; p for time and interaction=ns) or cholesterol levels. Significant improvements in fasting plasma glucose (p=0.03), TG (p=0.02) and insulin resistance estimated with the HOMA-IR index (p=0.04) were observed in the IT group only, despite similar and significant changes in total body composition in both exercise groups (fat mass: -4.3% and -4.9%, fat free mass +1.5% and +1.7% in the CL and IT groups, respectively; both p0.05). During the ride, mean IG was significantly greater in T1rec and significantly lower in T1pro compared to CON (7.6±2.8 vs. 5.2±2.7 vs. 6.3±01.0 mmol.l-1; p0.05). During recovery time before sleep T1pro displayed similar mean IG to CON (6.4±2.8 vs. 6.6±0.9 vs. mmol.l-1; p=0.553), which was significantly less than T1rec (9.3±1.7 mmol.l-1; p=0.013). Similar patterns were observed during sleep, whereby T1pro displayed similar mean IG to CON (T1pro 6.7±3.4, CON 6.1±0.9 mmol.l-1; p=0.412), which were both sig- nificantly less than T1rec (8.7±2.6 mmol.l-1; p S 257 1 C PS 041 Exercise physiology II 623 An acute bout of exercise is not able to lower liver fat content in overweight middle-aged men B. Brouwers1, L. Bilet1, P.A. van Ewijk1, M.K.C. Hesselink2, M.E. Kooi3, P. Schrauwen1, V.B. Schrauwen-Hinderling3, NUTRIM, School of Nutrition, Toxicology and Metabolism; 1Human Biology, 2Human Movement Sciences, 3Radiology, Maastricht University, Netherlands. Background and aims: Elevated hepatic lipid content (IntraHepatic Lipid, IHL) markedly increases the risk of metabolic complications, including in- sulin resistance and cardiovascular events. Although prolonged exercise training has been shown to lower IHL, it is unknown if acute exercise has the same effect. Next to IHL, hepatic ATP content may be related to insulin resistance and IHL. Therefore, we aimed to investigate if acute exercise leads to changes in IHL and if these changes are accompanied by changes in hepatic ATP content. Materials and methods: IHL content was determined by proton magnetic resonance spectroscopy (MRS) at baseline, directly after 2h of exercise (50 % of the maximal output) and again 4h post-exercise and expressed as percent- age of the water resonance. Hepatic ATP content was measured by phospho- rus MRS at baseline and 4h post-exercise in 8 subjects and is given relative to the total phosphorus signal. Blood was sampled and substrate oxidation was measured by indirect calorimetry. Since exercise leads to an increase in plasma FFA that may affect liver fat content, subjects performed the study not only in the fasted condition, but also while ingesting glucose. Results: Twenty-one overweight middle-aged men with a wide range of IHL, including nonalcoholic fatty liver disease (age 54.8±7.2 years, BMI 29.7±2.2 kg/m²) participated in this study. IHL was unchanged directly after exercise (8.27±1.80 to 8.34±1.85 %). Similar effects were observed when exercising with glucose supplementation (IHL: 8.26±1.88 to 8.40±1.89 %). 4h after ex- ercise, IHL was elevated compared to baseline in the fasted condition (from 8.3±1.8 to 8.7±1.8 %; p=0.010), while it did not change when glucose was sup- plemented (from 8.3±1.9 to 8.3±1.9 %; p=0.789). Hepatic ATP content tended to decrease post-recovery in the fasted condition compared to baseline values (0.164±0.009 vs. 0.146±0.009 γ-ATP/total phosporus; p=0.086), whereas it did not change when glucose supplementation was given (0.168±0.008 vs. 0.163±0.005 γ-ATP/total phosporus; p=0.582). Plasma FFA concentrations increased with time during exercise and recovery in the fasted condition (p S 258 1 C 5d/wk) while same number, age-matched, control animals were sedentary. Prior and upon completion of the training period, fed plasma glucose and immunological soluble factors were monitored. Results: After 12 weeks of training, control mice were all diabetic (n=5) whereas only 2 out 5 mice on-exercise became diabetic. An exercise-induced weight loss (-9%, p S 259 1 C and 25.6%high. Low PA was more frequent in: women, ≥75 years, smokers, hypertensive, patients with higher BMI and waist risk circumference, higher Stroke scores, CRD, depression, and elevated lipid profile. Independent risk factors associated with low PA in the multivariate analysis were (Odds ratios, CI 95%): Hypertension (3.1, 1.5-6.2), CRD (1.8, 1.04-3.2), depression (2.7, 1.2-5.9), longer time of T2DM onset (1.04, 1.0-1.09), higher levels of triglyc- erides (1.01, 1.0-1.01) and ldl (1.01, 1.0-1.02). By contrast, have a good con- trol of glycaemia (0.5, 0.3-0.8), of blood pressure (0.4, 0.3-08) and of triglyc- erides (0.4, 0.2-1.0) were protective factors for sedentary. Female had a higher risk for sedentary in the limit of the significance (OR: 1.5; CI95%: 0.9-2.5). Conclusion: 24% of our T2DM population was sedentary. PA improves gly- caemia control, lipid profile, CV risk factors and CRD, all of them very im- portant to prevention of diabetes morbidity. Inactivity is also related with a risk of depression, probably associated to the quality of life. Supported by: Mutua Terrassa Fundation 629 Lowered muscle endurance in association with neuropathy in patients with type 2 diabetes T. Gysel1, J. Ruige2, P. Boon2, D. Cambier1, P. Calders1; 1Dept. of Rehabilitation Sciences and Physiotherapy, 2Ghent University, Ghent, Belgium. Background and aims: Type 2 diabetes is characterized by a reduced sensi- tivity of the tissues to insulin (insulin resistance), and a reduced formation of insulin by the pancreas. This is often accompanied by damage to blood ves- sels and nerves (neuropathy). The treatment of diabetes includes medication, nutritional advice and exercise therapy. Strength training has a positive effect on insulin resistance and muscle function. What kind of strength training is most optimal, remains unknown. To investigate this, first of all, it should be examined to what extent different types of strength (maximal-, explosive-, endurance,⋯) are influenced by the severity of diabetes, taking into account additional conditions, such as peripheral neuropathy. Since the negative in- fluence of diabetes and neuropathy on maximal strength has already been investigated, the purpose of this study was to investigate the association be- tween neuropathy and arm/leg muscle endurance. Materials and methods: Data from an isokinetic dynamometer, a clinical neurological examination + quantitative sensory examination and nerve conduction studies were obtained in 35 type 2 diabetic men, with (n=24) and without neuropathy (n=11). Seven patients had small fiber sensory neuropa- thy, 17 had sensorimotor neuropathy. All men performed 25 elbow- and 30 knee maximal concentric contractions on an isokinetic dynamometer at a given angular velocity of 180°/sec. Elbow- and knee flexor fatigue was meas- ured with the fatigue index, calculated as the ratio of the average peak torque of the last third to the average peak torque of the first third. A neuropathy rank sum score (NRSS) was calculated for each patient to quantify the pe- ripheral nerve function, adding the rank scores for the Michigan Neuropathy Screening Instrument (MNSI) questionnaire, the neuropathy disability score (NDS), the vibration perception thresholds (VPT’s), the motor and sensory nerve conduction velocities (MNCV’s and SNCV’s), as well as the amplitudes of the compound muscle - and sensory nerve action potentials (CMAP’s and SNAP’s). The VPT’s were ranked on the basis of the sum of the two percen- tiles obtained at right and left great toe. The EMG-parameters (MNCV & CMAP of peroneal, tibial and ulnar nerve, SNCV & SNAP of radial and sural nerve) were ranked for all nerves according to the calculated z-score. In order to calculate these z-scores, MNCV + CMAP and SNCV + SNAP were meas- ured in 16 healthy subjects in an age-matched comparison group. Finally, the mean sum score for the conduction studies was used for the total rank sum score. If no action potential could be obtained, the highest rank score was given. A high NRSS thus reflected severe nerve dysfunction. To estimate the correlations between the neuropathy rank sum score and the fatigue indices, pearson correlation analysis was used. Results: A significant negative correlation was indicated between the isoki- netic fatigue index of knee flexors and the NRSS (r= -0.356 , p=0.036). For elbow flexors there was also a tendency towards a negative correlation with the NRSS (r= -0.318 , p=0.062). Conclusion: This finding suggests that not only maximal strength, but also endurance in diabetic people with neuropathy is worse than in diabetics without neuropathy. 630 Characteristics of patients with type 2 diabetes who achieved better glycaemic control through resistance training: a meta-analysis H. Ishiguro1, S. Kodama2, S. Matsunaga3, C. Horikawa3, Y. Heianza3, N. Ohara4, T. Yamada3, A. Suzuki3, O. Hanyu3, H. Sone3; 1Tsugawa Hospital, Niigata prefecture, 2MIto kyodo Hospital, Ibaraki prefecture, 3Niigata University Faculty of Medicine, 4Nagaoka Red Cross Hospital, Niigata prefecture, Japan. Background and aims: Recently, an increasing number of studies have sug- gested that resistance training (RT) has been effective in improving glyce- mic control (GC) in patients with type 2 diabetes (T2D). Our aim of this meta-analysis is to determine the antihyperglycemic effect of RT and analyze characteristics of patients for whom RT could be especially indicated for im- provement in GC. Materials and methods: Electronic-based literature search was conducted for intervention studies examining the metabolic effect of RT in T2D patients using MEDLINE and EMBASE. Studies were included if 1) both an RT-exer- cise group and non-RT control group were examined and 2) the net change in hemoglobin A1C (A1C) (i.e., difference in mean change in A1C from before to after the study between RT and non-RT groups) and its corresponding standard error as a study weight could be estimated. We included studies that described other interventions such as aerobic training if those other interven- tions were similar between the RT and non-RT groups. Results: Twenty-seven studies were eligible. Net change in A1C (95% confi- dence interval [CI]) was -0.3 [-0.46--0.14]%. A significantly larger reduction in A1C was observed in studies of subjects with the following characteristics: 1) short duration of T2D (>= 6 years vs. =7.5% vs. S 260 1 C PS 042 Consequences of hypoglycaemia 631 Severe hypoglycaemia requiring medical assistance in patients with diabetes is associated with simultaneous prolongation of QTc interval M. Mylona1, G. Anastasiadis2, K. Zerva2, V. Vasgiourakis2, C. Liaskos1, S. Liatis1; 1Diabetes Center, First Department of Internal Medicine, 2Department of Cardiology, Laiko Hospital, Athens, Greece. Background and aims: It has been shown that hypoglycemia induces mul- tiple pre-arrhythmic changes, mainly increase pre-existent QT prolongation, produce intracellular Ca2+ overload, and decrease serum K+. As a conse- quence, hypoglycemia is considered a state that may lead to sudden death, as a result of cardiac arrhythmia. Few data, however, directly link episodes of severe hypoglycemia with pre-arrhythmic conditions, such as QT prolonga- tion. Aim of the present study is to explore the association between severe hy- poglycemic events requiring medical assistance and QT interval, in patients with diabetes. Materials and methods: Eight hospitals (nine clinics), in five cities of Greece, participated in a 16-month prospective survey of documented iatrogenic hy- poglycemia at the emergency departments (ED). According to the protocol, a 12-lead ECG was obtained simultaneously or immediately after the manage- ment of hypoglycemia and no later than 30 minutes after the administration of glucose. The ECGs obtained from hypoglycemic patients were compared to those from a control group consisted of patients with diabetes, matched for age and gender, visiting the outpatient diabetes clinics of the same hos- pitals, during the same time period. QT and RR intervals were measured blindly by three independent cardiologists. QTc was calculated according to the Bazett formula: QTc = QT interval/square root of the R-R interval. QTc measurements of≥440msec were considered abnormally prolonged and those of≥500msec as highly prolonged. Patients receiving medications possibly af- fecting the QTc interval and those with hypokalaemia (serum potassium < 3.5mEq/l) were excluded from the analysis. Results: During a median follow-up of 16 months, 295 episodes of iatrogenic hypoglycemia in 294 patients with diabetes were identified and 223 ECGs were obtained according to the pre-specified criteria. However, 46 ECGs were excluded from the analysis, due to the presence of the above mentioned cri- teria. Finally, 177 ECGs from hypoglycemic patients were analyzed (mean age 72.7±15.7 years, 48.6% women, 9% with type 1 diabetes) and com- pared to 91 age and gender matched controls. Mean QTc interval was sig- nificantly prolonged in patients compared to controls (440.4±45.1 msec vs. 413.9±32.5msec, p S 261 1 C Results: In the diabetic group, HR increased maximally from 72±10 to 77±10 bpm (p=0.04) after 30 minutes of hypoglycaemia (T90), but fell to 71±8 bpm despite maintained hypoglycaemia (T120)(see Fig). High frequency HRV de- creased up to T90 (logHF 1.89±0.61 vs 1.61±0.63, p =0.02), indicating acute vagal withdrawal, but returned towards baseline at T120. There was a rise in normalised low frequency power indicating increased sympathetic contribu- tion. In the control group, the maximal increase in HR occurred earlier at T60 during hypoglycaemia (from 62±9 to 66±13 bpm, p=0.04) but HR re- mained elevated at 65±11 bpm after hypoglycaemia for 1 hour. HF power de- creased from baseline (logHF 2.36±0.44 to 2.14±0.59, p=0.05) and remained unchanged despite continued hypoglycaemia, indicating inhibition of vagal tone throughout. No significant changes in HRV occurred in either group during euglycaemic clamps. Conclusion: Cardiac autonomic regulation during hypoglycaemia appears time-dependent and differs between patients with type 2 diabetes and non- diabetic individuals. In type 2 diabetes, the initial heart rate increment to hypoglycaemia is delayed. Following early vagal withdrawal, there is reactiva- tion of vagal activity during more sustained hypoglycaemia which did not occur in nondiabetic individuals. These mechanisms may relate to autonomic dysfunction and could contribute to tachy- and bradyarrhythmias during clinical hypoglycaemic episodes. Clinical Trial Registration Number: DRN686 Supported by: NIHR 634 Survey of outpatients with type 2 diabetes among older adults in Japan M. Fukuda, K. Doi, M. Sugawara, Y. Naka, K. Mochizuki; Japan Physicians Association, Tokyo, Japan. Background and aims: The number of older adults with Type II diabetes (T2DM) is continuously increasing in Japan, and care needs to be provided according to the individual needs of these patients, which require more ur- gent attention in Japan than in the West. Despite this situation, few studies have investigated the prevalence of hypoglycemia among older T2DM in Ja- pan. This study describes the results of a survey conducted by the Japan Phy- sicians Association (JPA) to elucidate the prevalence of hypoglycemia among older T2DM in Japan. The survey collected data on patient knowledge about hypoglycemic symptoms as well as their clinical background. Materials and methods: A total of 15,892 older adults aged ≥65 years (8319 men and 7516 women) who were undergoing regular outpatient treatment for their diabetic conditions were included in this study. The survey included questions pertaining to 28 items concerning hypoglycemic symptoms and age-related syndromes experienced in the last month. The JPA also collected patients’ medical records from their doctors. Results: The average age of patients was 74.2 years (men, 73.6 years; wom- en, 75.0 years), the mean disease duration was 12.8 years (men, 13.1 years; women, 12.5 years), and the mean body mass index (BMI) was 24.1 (men, 24.0; women, 24.2). Medications used for diabetes treatment, including con- comitant prescriptions, were as follows: DPP-4 inhibitors, 53.3%; SU, 39.4%; α-GI, 23.4%; metformin, 22.8%; pioglitazone, 14.0%; glinide, 4.4% and insu- lin, 17%. Although the incidence of hypoglycemic symptoms as estimated by the doctors on the basis of their diagnoses in the previous month was 7.8%, the actual incidence indicated by the patient responses was 10.4%. The re- sponses indicated that the incidence of hypoglycemia among patients treated with either SU or insulin was 20.9%, whereas that among patients treated without neither drugs was lower at 2.5%. Tiredness/feeling languid (32.8%), loss of balance/dizziness (32.4%), cold sweats (30.4%), and eye blink (24.8%) were the most frequently experienced hypoglycemic symptoms. Among the patients who used SU or insulin, 80.1% were aware of their hypoglycemic symptoms and 64.4% knew how to treat these symptoms. However, only 38.1% patients who were aware of their hypoglycemia always carried some form of glucose with them. Conclusion: The results of our survey indicated that the symptoms of hypo- glycemia were atypical and vague in older T2DM patients, and the actual inci- dence of hypoglycemia among the patients was considerably higher than that estimated by the doctors on the basis of their diagnoses. Although there was satisfactory awareness of hypoglycemia among the older adults, two-thirds of them did not always carry some form of glucose with them. Therefore, more careful treatment of T2DM and more positive measures to promote the awareness of hypoglycemia are required for older adults with T2DM. 635 Impaired awareness of hypoglycaemia in insulin-treated patients with type 2 diabetes mellitus N. Sakane, K. Kotani, K. Tsuzaki, K. Takahashi, Y. Sano; Kyoto Medical Center, Japan. Background and aims: Severe hypoglycemia (SH) is associated with a high- er risk of cardiovascular disease-related mortality and mortality. Impaired awareness of hypoglycemia (IAH) in insulin-treated patients with type 2 dia- betes mellitus (T2DM) has been reported to be associated with the presence of SH. However, little is known regarding the actual states of hypoglycemia, including IAH, in Japan. The aim of the study was therefore to survey the states of hypoglycemia and patient-physician communication among insu- lin-treated patients with T2DM. Materials and methods: A questionnaire survey on hypoglycemia and pa- tient-physician communication was conducted in 331 patients with insulin- treated T2DM patients at 16 hospitals and clinics. Inclusion criteria were: 1) over 20 years, 2) T2DM, 3) insulin treatment, 4) patients regularly attended the surveyed hospitals and clinics. Exclusion criteria were: 1) children and young patients under 20 years old, and 2) type 1 diabetes mellitus. Clinical data such as the insulin regimen and diabetic complications were collected from their physicians. Hypoglycemia was defined as blood sugar ≤50 mg/ dL (2.8 mmol/L) or symptoms of dizziness, blurry vision, confusion, and/or sweating that the patient was able to resolve without assistance. Similar symp- toms that required external assistance were defined as SH. Favorable glyce- mic control to prevent diabetic complications was defined as HbA1c S 262 1 C 636 Hypoglycaemia unawareness in emerging adults transferred from a paediatric to an adult diabetes unit M.C. Careaga1, M. Vidal1, M. Jansa1, C. Yoldi2, R. Cardona2, A. Gomez2, M. Gimenez1, E. Esmatjes1, I. Conget1; 1Diabetes Unit. Endocrinology and Nutrition Department, Hospital Clinic and University of Barcelona, 2Pediatric Endocrinology, Hospital Sant Joan de Deu, Barcelona, Spain. Background and aims: Repeated hypoglycemia and the syndrome of hypo- glycemia unawareness (HU) are major barriers to achieving normoglycemia over a lifetime of using intensive insulin therapy and thus precludes eugly- cemia’s long-term benefits. Type 1 Diabetes (T1D) management in young patients is troublesome specially when they are transferred to adult Diabetes Units. The aim of our study was to evaluate the impact of the presence of HU on the results of a transition therapeutic education program (TEP) spe- cifically addressed to the transfer T1D patients from a pediatric to an adult diabetes Unit. Materials and methods: We included 56 patients with T1D (age 18.0 ± 0.3 years, 26 girls, disease duration 8.0 ± 4.0 years, HbA1c 8.0 ± 1.2%) consec- utively transferred (2009-2011). The TEP included a co-ordinated transfer between the Units, individual visits and group sessions. At baseline, we reg- istered data on BMI, insulin dose, HbA1c and the frequency of hypoglyce- mia. For evaluation of the quality of life (QoL) and knowledge in T1D man- agement, self-report questionnaires were assessed. Mother tongue (spanish and catalan versions) validated Clarke’s test was used to evaluate HU.After 12-months, all subjects were re-evaluated. Results: Nine out of 56 patients had HU (16%). Age, gender, duration of dis- ease, type of treatment, HbA1c, the different aspects of QoL and the degree of knowledge of the disease were not different in those subjects afected by HU in comparison to the unaffected group. Only the number of severe hypo- glycemia episodes was significantly different between groups (0.33±0.50 vs. 0.09±0.28, episodes-subject-year with and without HU respectively,p 5 non severe episodes/week was close to significantly different (66 vs 34 %, p=0.13). After 12 momths, in the group with HU, 4 patients improved awareness, there was a reduction in the number of non severe hypoglycemia episodes/ week , but this was not the case of severe episodes. The HbA1c was not significantly different between the two groups (8.2 ± 1.1 vs. 7.6 ± 0.9% with and without HU, respectively) Conclusion: The presence of HU is far from uncommon when emerging adults with T1D are transferred from a pediatric to an adult diabetes unit. While the use of an specific TEP improves metabolic prognosis in the mid- term, the reduction of hypoglycemia episodes, specially severe, remains an elusive issue 637 Timing of hypoglycaemia is associated with increased mortality and length of stay among patients with diabetes admitted to internal medicine departments J. Wainstein1, M. Boaz2,3, E. Leibovitz4,5; 1Diabetes clinic, Wolfson Medical Center, 2Research and Epidemiology, Wolfson Medical Center, Holon, 3Epidemiology, Tel Aviv University, 4Interal Medicine E, Wolfson Medical Center, Holon, 5Sackler school of Medicine, Tel Aviv University, Israel. Background and aims: Hypoglycemia among hospitalized patients with dia- betes is associated with poor hospital prognosis. The aim of the study was to examine the effect of hypoglycemia timing on hospital prognosis Materials and methods: In this retrospective analysis of electronic medical records, we included all 3941 patients with diabetes (Mean age 71.7±12.9 years, 49.3% males) discharged from internal medicine departments during 2009. All glucose measurements were computerized using an institutional glucometer. Patients were categorized into 3 groups according to hypogly- cemia timing: Group1, no hypoglycemia, Group 2, hypoglycemia only upon admission and Group 3, hypoglycemia during hospital stay. Results: Included in the analysis were 3413 (86.4%) patients with diabetes who had glucose measurements performed during hospitalization. A total of 157 patients (4.6%) had at least 1 hypoglycemia event during the hospital stay (42 patients upon admission and 115 patients during hospitalization). Patients in Group 3 had significantly increased in-hospital mortality and length of hospital stay (27.0%, 18.4±24.7 days), compared to Groups 1 (3.4% mortality, 5.1±8.8 hospital days) and Group 2 (4.7% f mortality, 3.2±3.3 hos- pital days). Mortality rates were associated with the severity and number of hypoglycemia events. Patients in Group 3 had higher creatinine and lower albumin levels compared to Groups 1 and 2. Conclusion: Hypoglycemia during hospitalization is associated with in- creased rates of in-hospital mortality and prolonged hospital stay among diabetic patients admitted to internal medicine departments. Poor hospital outcome was associated with timing of hypoglycemia, as well as the number and severity of the events. 638 Baseline insulin deficiency is associated with increased mortality and severe hypoglycaemia in the ACCORD study E.R. Seaquist1, H. Chen2, M. Miller2, L. Chow1; 1Medicine, University of Minnesota, 2Biostatistics, Wake Forest School of Medicine, Winston Salem, USA. Background and aims: Intensive glycemic control is associated with in- creased mortality, but the role of hypoglycemia remains unclear. Hypogly- cemia is associated with insulin deficiency. Whether insulin deficiency and/ or islet autoimmunity may affect mortality in patients with type 2 diabetes remains unknown. Using the ACCORD (Action to Control Cardiovascular Risk in Diabetes) cohort, we tested the hypothesis that baseline insulin de- ficiency and islet autoimmunity in type 2 diabetes would be associated with severe hypoglycemia and mortality. Materials and methods: A nested case-control study (86 cases, 344 controls) was used. A participant who died during ACCORD with at least 1 episode of severe hypoglycemia, defined as hypoglycemia requiring assistance, was clas- sified as a case. Participants who did not die during ACCORD and did not have severe hypoglycemia were classified as controls. Each case was matched to 4 controls [age (+/- 5 yrs), BMI (+/- 2.5), glycemic intervention arm, race]. Insulin deficiency was defined as fasting C-peptide < 0.45 nmol/L. Islet auto- immunity was determined by antibodies against glutamic acid decarboxylase (GAD), tyrosine phosphatase-related islet antigen 2 (IA-2A), insulin (IAA), and zinc transporter (Zn-T8). Conditional logistic regression was used. Results: Death during ACCORD with at least 1 episode of severe hypoglyce- mia was associated with baseline insulin deficiency (OR 4.8, 95% CI 2.1-11.1, p S 263 1 C During the year following basal insulin initiation, unadjusted achieved A1C decreased -1.0% from baseline in NPH and -1.2% in GLA. Mean number of hypoglycemic events in the year was 0.74 in NPH and 0.50 in GLA after ad- justment of baseline characteristics. Overall, number of hypoglycemic events increased as achieved A1C during the year approached the target (≤7%) and remained higher in NPH relative to GLA (p+1SD) on high HbA1c value (>=7 %) and occurrence of severe hypoglycemia within 30 days. Materials and methods: Data from 1927 French patients with insulin-treated diabetes (1008 patients with type 1 diabetes mellitus [T1DM] and 919 with type 2 diabetes mellitus [T2DM]) completing the HFS-II questionnaire in the DIALOG observational prospective study was used. The correlated factors were analyzed using a multivariate logistic regression model. Results: Factors that were correlated to a HbA1c>=7 % were separately inves- tigated in patients with T1DM and T2DM. For patients with T1DM, the main factors contributing to an HbA1c>=7 % were the HFS class >+1SD (odds ratio [OR] 3.5; 95 % confidence interval [95CI] 1.9-6.4) and the class between -1 and +1SD (OR 2.8; 95CI 1.7-4.7) versus the lowest class (10 years (OR 1.8; 95CI 1.3-2.5). In patients with T2DM, the HFS score was no longer significant in the multivariate model. On the other hand, the main associated factors contributing to occurrence of a severe hypoglycemic event in both T1DM and T2DM patients were history of hypoglycemia (OR 3.8; 95CI 1.4-10.6), T1DM (OR 1.8; 95CI 1.3-2.5) and an HFS score >+1SD versus S 264 1 C 642 A simple algorithm to predict the risk of hypoglycaemia in case of physical activity (PA) in patients with type 1 diabetes under pump therapy G. Charpentier1, J. Boudoux2, M.-H. Petit3, A. Doudi3,1, C. Simon4, A. Pochat3,5, H. Laroye3, N. Brunel2, S. Franc1; 1CHSF, Corbeil, France, 2ENSIIE, 3CERITD, Evry, 4CHU of Lyon, 5CHSF, Evry, France. Background and aims: The aim of our study was to develop a model to pre- dict the occurrence of hypoglycemic events from data of Continuous Moni- toring Systems (CGMs), collected during physical activity (PA) sessions, in patients with type 1 diabetes (T1D), under pump therapy. Materials and methods: Twenty adult patients with type 1 diabetes (11 males, mean age 45.3±11.9 yrs, HbA1c 7.9±0.9%, VO2max 33±10 ml/kg/ min) performed 5 basal sessions (i.e. 3 hours after lunch) in a random order: 4 PA sessions (30-min cycling sessions on a constant-power electronic cycle ergometer) at 2 levels of PA and with 2 kinds of adjustment of the basal rate (BR) of their pump (50% VO2max with a BR reduction of 50 or 80%, 75% VO2max with again a BR reduction of 80 or 100%) and 1 rest session. Patients wore a CGMs from the beginning of each session and until the next morning. Classification methods were used, based on the patients’ profile and the kind of sessions, to classify patients according to the occurrence or not of hypo- glycemia. Several models of prediction were then elaborated. A number of explanatory variables had been previously selected and were used to compare the effectiveness of each model to predict the risk of hypoglycemia. Results: From one model to another, variables that proved not to be essential or that were not available for the patient in a realistic context, were progres- sively removed. Finally, we considered variables available during the first ten minutes of a PA: weight, age, gender, VO2max, BR reduction, glucose level at the beginning of PA, glycemic gradient during the first 10 min of PA. Among these variables, we used statistical model selection methods in order to select the most relevant variables for the prediction of hypoglycemic events which were weight, and glycemic gradient during the first 10 min of PA reported to the initial glucose level (GG10). The value of the Area Under the Curve (AUC) of the ROC curve based on these variables was 0.74 which means a rather acceptable prediction but that can still be improved. We used a lo- gistic regression for modeling the probability p of occurrence of hypoglyce- mic events and the estimated model is: logit(p) = -4.13 - 438.7 x GG10 + 0.5 weight (kg), where logit(p)=log(p/(1-p)) is an increasing function of p. The model, based on these 3 variables easy to collect in clinical practice, provides a prediction rate exceeding 80% (concordance between the predicted and the actual situation) with less than 20% of mismatch data. Conclusion: A simple model of prediction mainly based on the CGM gradi- ent of the first 10 min of a 30-min physical activity session, allows to predict the occurrence of hypoglycemia with a relative good efficiency. This kind of model that we can imagine coupled with CGMs to predict the risk of hypo- glycemic event in clinical practice, (open-loop system), may, in an improved version, reveal to be of particular interest in the development of closed-loop systems. Clinical Trial Registration Number: NCT01277965 Supported by: EFSD, Medtronic 643 The early warning indicators for nocturnal hypoglycaemia in type 2 diabetes Y. Wang, H. Su, Y.M. Xue; Department of Endocrine and Metabolism, The First People’s Hospital of Yunnan Province, Kunming City, China. Background and aims: The existing research on hypoglycaemia is largely done in special environments such as use of a specific medicine or focus- ing on a certain age period, its practicality in broad clinical contexts is thus limited. Moreover, most causes of hypoglycaemia like endogenous insulin inadequacy and the diabetic process cannot be altered. This study is to ex- plore quantitative indicators warning of nocturnal hypoglycaemia by using Continuing Glucose Monitoring System (CGMS) on a large sample in rou- tine clinical environments to seek some basis for preventing hypoglycaemia in clinics. Materials and methods: First, 1147 patients (male 723 and female 424) of type 2 diabetes were sampled, whose ages being 59.21±11.31(14~91 years old) and Mean Blood Glucose (MBG) being 7.41±2.98 mmol/L. Those whose BGs turned stable after treatment, except those who took medicine like hypnotics or sedative before sleep, were monitored for 65 to 82 hours by CGMS.. Second, another 479 patients whose BG 3 hours after supper was ≤ 4.7mmol/L were sampled and randomly divided into the intervention group and the non-intervention group, measures like insulin dosage reduction be- fore sleep and/or eating some 50 to 100 grams of protein or carbonhydrate were adopted for the intervention group. Statistical analyses were made by correlation regression (with r below representing correlation coefficient), rel- ative risk (RR) calculation and Chi-square test whereas comparison between the two groups was made by t test. Results: The occurrence rate of hypoglycaemia, asymptomatic hypoglycae- mia and nocturnal hypoglycaemia is respectively 37.23% or 427 cases,22.75% or 261 cases, and 18.31% or 210 cases. Among 965 person-times showing hypoglycaemia, 60.8% or 697 person-times were asymptomatic whereas 49.2% or 470 person-times showed nocturnal hypoglycaemia. 1. MBGs are correlated negatively with incidence rates of hypoglycemia on an exponen- tial curve (r=-0.995, Ye=102.550-0.139X)and negatively with incidence rates of nocturnal hypoglycaemia (r=-0.953, Ye=104.135-3.445lgX)and of asymptomatic hypoglycaemia(r=-0.963, P S 265 1 C Statistically significantly higher hormonal counter-regulatory responses (adrenaline, growth hormone and cortisol) to hypoglycaemia were observed as compared with normoglycaemic night. Conclusion: Nocturnal hypoglycaemia in patients with type 2 diabetes caused a decrease in awakening response following the event. These find- ings underscore the risks associated with nocturnal hypoglycaemia since these events potentially affect the subject’s ability to wake up and respond adequately to hypoglycaemia. Clinical Trial Registration Number: NCT01780272 Supported by: Novo Nordisk 645 Impact of nocturnal hypoglycaemia on the risk of hypoglycaemia the next day G. Reihill, L. Green, S.A. Amiel, P. Choudhary; Diabetes Research Group, King’s College London, UK. Background and aims: In people with type 1 diabetes [T1D], nocturnal hypoglycaemia (NH) is reported to occur on about 30% of nights. Fear of [NH] can lead to suboptimal glucose control. Antecedent hypoglycaemia, including specifically at night, blunts symptomatic and hormonal responses to subsequent hypoglycaemia, which may lead to a vicious cycle of recurrent hypoglycaemia, ultimately in some patients leading to impaired awareness of hypoglycaemia [IAH]. We aimed to study the impact of biochemical NH on the risk of subsequent daytime hypoglycaemia [DH], and conversely also the impact of antecedent DH on the risk of subsequent NH in adults with T1D using continuous glucose monitoring [CGM]. Materials and methods: 172 consecutive CGM traces obtained using Medtronic iPRO2 met the following criteria: 48 hours of CGM data starting from midnight, ≥ 3 valid calibrations/24 hours, mean absolute error S 266 1 C events 180 mg/dL (p=0.057) in case of PBR compared to BRR. Only 1 hypoglycemia occurred during PA (BRR). During the afternoon, there was a trend towards less hypoglycemic events in case of PBR vs BRR (3 events in 3 patients vs 10 in 6, p=0.07) with a mean number of events 3.5 times lower (0.16 ± 0.37 vs 0.56 ± 0.92). There was no significant hyperglycemic rebound before dinner in both cases and no significant difference in the CGM level at bedtime (176 ±79 mg/dl vs 160±85, p=NS). The average glucose level re- mained stable and at a similar level at night in both cases. There was no differ- ence in the timing and in the level of nocturnal nadir between PBR and BRR [77 ± 35 vs 77 ± 34 mg/dL, p=NS]; no differences were observed in the total AUC, the time spent in or out the range [70;180] or [80;140]. At night, the number of hypoglycemic events did not differ in case of PBR vs BRR (7 events in 5 patients vs 10 events in 7 patients, p=0.39) and the mean number did not significantly differ (0.37 ± 0.68 vs 0.56 ± 0.86). There was no hyperglycemic rebound the next morning, even in case of hypoglycemia at night. Conclusion: In case of PA 90 min after a meal, PBR led to a higher sustain- able glucose level in the afternoon than in case of BRR, limiting therefore the risk of hypoglycemia. PBR can thus be considered as the security option while BRR would lead to a lower glucose level and potentially to an increased risk of hypoglycemia. However, in both cases the practice of PA will first lead to an initial drop in glucose level of around 70 mg/dL. If a BRR strategy is choosen, then the BRR should be anticipated at least 30 min earlier to limit the risk of hypoglycemia. Clinical Trial Registration Number: NCT01277965 Supported by: EFSD, Medtronic 649 Effects of automatic insulin pump interruption and bedtime glucose levels on nocturnal hypoglycaemic events in the ASPIRE In-Home Study D. Klonoff1, A. Ahmann2, S. Garg3, R. Slover3, T. Bailey4, B. Bode5, J. Thrasher6, R. Brazg7, F. Schwartz8, M. Mao9, R. Weiss10, ASPIRE In-Home Study Group; 1Mills-Peninsula Health Services, San Mateo, 2Oregon Health and Sciences University, Portland, 3University of Colorado Denver, Aurora, 4AMCR Institute, Inc., Escondido, 5Atlanta Diabetes Associates, 6Arkansas Diabetes and Endocrinology Center, Little Rock, 7Rainier Clinical Research Center, Renton, 8Ohio University, Athens, 9Medtronic Diabetes, Northridge, USA, 10Human Metabolism and Nutrition, Hebrew University, Jerusalem, Israel. Background and aims: The ASPIRE In-Home Study was designed to evalu- ate the Threshold Suspend (TS) feature of sensor-augmented pump (SAP) therapy, which automatically interrupts insulin delivery at a pre-set sensor glucose (SG) threshold. We evaluated the risk of nocturnal hypoglycemia (NH) with regard to bedtime sensor glucose (SG) values and use of the TS feature. Materials and methods: Nocturnal hypoglycemia (NH) was defined as an event lasting >20 min with SG values ≤65 mg/dL that started between 10:00 PM and 8:00 AM. Subjects with type 1 diabetes with ≥2 NH episodes during a 2-week run-in phase were eligible. A total of 247 patients were randomized to SAP therapy with the TS feature (TS group, n=121) or without the TS feature (Control group, n=126). During the 6-month study phase, each patient-night was classified according to the patient’s treatment group assignment and the 10:00 PM (“bedtime”) SG value. Four categories of bedtime SG values were defined as 200 mg/dL. The percentage of nights with a NH event was calculated for the TS and Con- trol groups within each bedtime SG value category. Results: The figure shows the percentage of nights with a NH event according to treatment group and bedtime SG value, and the number of patient-nights in each SG category. Higher bedtime SG values tended to decrease the risk of NH in the Control group (p S 267 1 C Conclusion: In subjects with type 1 diabetes and bedtime SG values S 268 1 C of fT3, the levels of FBG(P S 269 1 C 655 The relationship between vitamin D and hypogonadism in patients with type 2 diabetes mellitus H. Dogan1, G. Sargin2, E. Harman3; 1Department of Internal Medicine, Bozyaka Training and Research Hospital, Izmir, 2Department of Internal Medicine, Canakkale State Hospital, Canakkale, 3Division of Endocrinology and Metabolism, Yesilyurt Training and Research Hospital, Izmir, Turkey. Background and aims: Association has been reported between abnormal glucose metabolism, serum vitamin D levels and gonadal dysfunction. 25 (OH) D levels are lower in patients with Type 2 diabetic male patients com- pared with non-diabetic male patients. Type 2 diabetic patients were sepa- rated into three groups according to vitamin D levels ( S 270 1 C NO generating system, a known negative modulator of CS release. Systemic antioxidant treatment not only normalized oxidative stress parameters and NOS activity, but also corrected the observed effects of STZ-induced diabetes on CS levels. Compared to control animals, STZ-diabetic rats showed lower baseline circulating levels of ACTH, probably resulting from exaggerated ba- sal CS output. Antioxidant therapy restored basal and ACTH-induced gluco- corticoid release without modifying plasma ACTH levels, further supporting the role of local regulatory signals, such as NO, in the dysregulation of adre- nal steroidogenesis observed in this animal model of diabetes. Supported by: PICT 2008 N1034 658 Progressive reduction in beta cell mitochondrial function leads to diabetes S. Weksler-Zangen1, G. Aharon-Hananel1, A. Kogot-Levin1, A. Saada2, I. Raz1; 1Department of Internal Medicine, Diabetes Unit, 2Department of Human Genetics and Metabolic Diseases, Hadassah University Hospital, Jerusalem, Israel. Background and aims: Pancreatic β-cells couple glucose stimulated insulin secretion (GSIS) with oxidative phosphorylation via the mitochondrial res- piratory enzyme, cytochrome c oxidase (COX). Emerging evidence suggests interplay between mitochondrial dysfunction and the development of diabe- tes. We investigated the relation between COX activity and GSIS in isolated pancreatic islets of Cohen diabetes sensitive rats and the role it plays in dia- betes development Materials and methods: Cohen diabetic sensitive and Cohen diabetic resist- ant rats were maintained on a diabetogenic-diet for periods of 4, 11, 20 and 30 days. Blood glucose and insulin concentrations were measured before and during OGTT performed at different time periods on the diabetogenic-diet. Islet-COX activity was determined spectrophotometrically and normalized to citrate-synthase (a ubiquitous mitochondrial matrix enzyme). Results: Islets of normoglycemic Cohen diabetic sensitive rats had 50% of COX activity and GSIS compared to islets of Cohen diabetic resistant rats supporting the existence of a congenital COX impairment. Following expo- sure to a diabetogenic-diet, Cohen diabetic sensitive rats exhibited a pro- gressive increase in blood glucose levels and decrease in GSIS, developing overt hyperglycemia within 20 days. The decline in GSIS coincided with a significant decrease in islets-COX activity. The reduction in islet-COX activ- ity positively correlated (R2 = 0.9317) with GSIS and inversely correlated (R2 =0.760) with blood glucose levels. Cohen diabetic resistant rats maintained normoglycemia during the various time periods on diabetogenic-diet and ex- hibited a partial reduction in GSIS after only 30 days. Islets of Cohen diabetic resistant rats retained more than 50% of their baseline islets-COX activity after 30 days on diabetogenic-diet while in Cohen diabetic sensitive rats less than 20% of islets-COX activity remained. Conclusion: These results show a link between islets COX activity, reduced GSIS and the development of diabetes in the Cohen diabetes sensitive rats. On the diabetogenic-diet, COX activity in islets of normoglycemic-Cohen diabetes resistant rats was reduced by 50%, while in the hyperglycemic-Co- hen diabetes sensitive rats the activity was below 20%. Thus, we suggest that COX activity needs to be above a certain threshold to sustain adequate β-cell function and prevent diabetes development. Supported by: Lower Saxony, Hannover, Germany and Israeli Science founda- tion 659 Reduced mitochondrial biogenesis in skeletal muscle is associated with obese phenotype in androgen receptor deficient mice (ARKO) K. Kajita1, T. Usui1, Y. Kitada1, K. Taguchi1, I. Mori1, T. Ikeda1, H. Okada1, H. Morita1, T. Sasaki2, T. Kitamura2, T. Sato2, T. Ishizuka1; 1General Internal Medicine, Gifu University Graduate School of Medicine, 2Institute for Molecular and Cellular Regulation, Gunma University, Maebashi, Japan. Background and aims: Androgen, including dehydroepiandrosterone (DHEA) and testosterone, reduces adiposity, although the underlying mecha- nisms are unknown. Here, we examined the effect of androgen on heat pro- duction and mitochondrial biogenesis. Materials and methods: C57/Black male mice at 8 w of age were fed CE2 powder with (testosterone group) or without (control group) 0.4% testoster- one ad libitum for 4wk. In addition, we assessed mitochondrial biogenesis in ARKO at 8wk and 24wk of age. ARKO was provided by Dr. Shigeaki Kato. O2 consumption, CO2 production and locomotor activity were measured by in- direct calorimetry. The expression levels of PGC1α, ATP5B and Cox4 protein were measured with western blot. The mRNA expression levels of PGC1α, , nuclear respiratory factor-1 (NRF-1), nuclear respiratory factor-1 (NRF-2), mitochondrial transcription factor A (Tfam) and cytochrome C,and mito- chondrial DNA (mitDNA) were evaluated with real time PCR. Results: Testosterone group exhibited weight reduction, decreased fat mass and elevated heat production without reduction of food consumption. Obese phenotype was prominent in ARKO at 24wk, but not 8wk, of age. Serum testosterone level was elevated up to approximately 12-fold, while it was decreased to 10.4% in ARKO. The expression level of PGC1α, ATP5B and Cox4 protein in skeletal muscle, but not those in brown adipose tissue (BAT) and liver, were elevated in testosterone group, compared with control group, whereas they were suppressed in ARKO at 24wk compared with wild type littermate. Amount of mitDNA and expression level of mRNA involved mi- tochondrial biogenesis, such as PGC1α, NRF-1, NRF-2, and Tfam in skeletal muscle were increased in testosterone group. Expression levels of these genes were reduced in skeletal muscle isolated from ARKO at 24wk, however sig- nificant reduction was not observed at 8wk, showing that late-onset obesity in ARKO is caused by decreased mitochondrial biogenesis in skeletal muscle. Treatment with 10 nM testosterone significantly increased mRNA levels of PGC1α, NRF-1, NRF-2, Tfam and cytochrome C in C2C12 myotubes, which was abolished by flutamide, an AR inhibitor. Conclusion: These results demonstrate that the testosterone-induced in- crease in energy expenditure is derived from elevated expression level of PGC1α, and subsequent mitochondrial biogenesis in skeletal muscle, but not BAT. Therefore, skeletal muscle is the major organ contributing to testoster- one-induced weight reduction. Diabetologia (2014) 57:[Suppl1]S1–S564 S 271 1 C PS 045 Lipid metabolism in humans 660 Obese but healthy: Is it really possible? E. Hamo-Tchatchouang, E. Cosson, S. Chiheb, C. Cussac-Pillegand, I. Banu, M. Nguyen, M. Nguyen, P. Valensi; AP-HP, Jean Verdier Hospital, Bondy, France. Background and aims: Patients displaying the metabolically healthy but obese (MHO) phenotype have an intermediate cardio-metabolic prognosis as compared with normal weight healthy and metabolically unhealthy obese (MUO) patients. We aimed to evaluate the prevalence and characteristics of definite MHO individuals. Materials and methods: Definite MHO phenotype was defined as having none of the International Diabetes Federation metabolic syndrome criteria but waist circumference criterion which is considered as non discrimina- tory, in 1159 patients including 943 women admitted in stable conditions for obesity (body mass index (BMI) 38.4±6.3 kg/m2) without known diabetes. Patients were characterized for cardio-metabolic disorders. Results: The 202 (17.4%) MHO individuals, younger and with lower BMI than the 957 MUO patients in the total cohort, were matched for gender, age and BMI with 404 MUO patients. In addition to the features of metabolic syndrome, the definite MHO patients as compared with the MUO ones were fewer with an homeostasis model assessment of insulin resistance index ≥ 3 (23.6 versus 38.9%, p S 272 1 C 663 A highly potent and specific PPAR alpha agonist, K-877, improves lipid profiles and insulin sensitivity in dyslipidaemia subjects; an integrated analysis of 3 Phase 2/3 trials E. Araki1, S. Ishibashi2, S. Yamashita3, H. Arai4, K. Yokote5, H. Suganami6, T. Kodama7; 1Department of Metabolic Medicine, Kumamoto University Graduate School of Medicine, 2Division of Endocrinology and Metabolism, Jichi Medical University, Tochigi, 3Department of Cardiovascular Medicine, Osaka University Graduate School of Medicine, Suita, 4Department of Human Health Sciences, Kyoto University Graduate School of Medicine, 5Department of Clinical Cell Biology and Medicine, Chiba University Graduate School of Medicine, 6Clinical Data Science Department, Kowa Company, Ltd., Tokyo, 7Laboratory for Systems Biology and Medicine, The University of Tokyo, Japan. Background and aims: K-877 is a specific PPAR alpha agonist with a poten- tially improved pharmacological profile with respect to efficacy and safety compared with other PPAR alpha agonists. In subjects with atherogenic dys- lipidemia, insulin resistance often co-exist, and as such, the treatments which concurrently improve plasma lipoprotein profiles and insulin sensitivity are highly demanded. Therefore, this integrated analysis set out to assess retro- spectively the effect of K-877 on efficacy, safety and measures of insulin resist- ance over 12 weeks in 3 phase 2/3 trials. Materials and methods: This is an integrated analysis of two Phase 2 tri- als (one monotherapy trial and one add-on to stable pitavastatin 2 mg/day therapy) and one Phase 2/3 trial (monotherapy). All subjects had hypertri- glyceridemia and low HDL-C and/or high non-HDL-C. All these were 12 weeks, randomized, multicenter, double-blinded, placebo controlled trials. The efficacy and safety of K-877 0.05, 0.1, 0.2 and 0.4 mg/day (twice daily) were examined in all trials. Results: A total of 676 subjects (K-877; 37, 0.05 mg/day; 127, 0.1 mg/day; 215, 0.2 mg/day; 172, 0.4 mg/day: placebo; 125) were evaluated in this inte- grated analysis. The subject baseline characteristics were comparable among all groups. From baseline to Week12, a significant reduction in TG and an increase in HDL-C were observed for K-877 groups (TG; -40.6 to -52.7%, HDL-C; +12.0 to +24.0%, p S 273 1 C 666 Lower intensified target LDL-C level of statin therapy results in a higher risk of incident diabetes: a meta-analysis of randomised statin trials R. Cai, Y. Yuan, Y. Zhou, W. Xia, P. Wang, H. Sun, Y. Yang, R. Huang, S. Wang; Affiliated Zhong Da Hospital of Southeast University, Nanjing, China. Background and aims: A recent meta-analysis has reported that intensive- dose statin drug increases the risk of incident diabetes. However, doubling of the statin dose generates only a further 6% decrease in low-density lipopro- tein cholesterol (LDL-c) on average. This study aimed to determine whether statin therapy with lower intensive-target LDL-c level contributes to higher risk of new-onset diabetes. Materials and methods: Medline, Embase, and the Cochrane Central Reg- ister of Controlled Trials were searched for randomized controlled endpoint trials of statins conducted from 1966 to 2012. We included trials with more than 1000 participants who were followed up for at least 2 years. The included trials were stratified by the target LDL-c level. I² statistic was used to measure heterogeneity between trials. We further calculated risk estimates with ran- dom-effect meta-analysis. Meta-regression was used to identify the potential risk factors of statin-induced diabetes. Results: Fourteen trials with a total of 95 102 non-diabetic participants were included. The risks elevated by 33% [odds ratio (OR) = 1.33; 95% confidence interval (CI) 1.14-1.56; I² = 7.7%] and 16% (OR = 1.16; 95% CI 1.06-1.28; I² = 0.0%) when the intensified target LDL-c levels were ≤1.8 mmol/L and 1.8- 2.59 mmol/L, respectively. Incident diabetes did not increase when the target LDL-c level was ≥2.59 mmol/L. Apart from age, female, and baseline level of total cholesterol, meta-regression analysis showed that the target and base- line levels of LDL-c and relative LDL-c reduction were predictors of statin- induced diabetes. Conclusion: A lower intensified target LDL-c level of statin therapy resulted in higher risk of incident diabetes. 667 This work was partially supported by the National Natural Science Foundatio Intensified LDL-C target of statin therapy and cancer risk: a meta-analysis S. Haixia, Y. Yuan, P. Wang, Y. Zhou, R. Cai, W. Xia, Y. Yang, R. Huang, S. Wang; The Affiliated ZhongDa Hospital of Southeast University, Nanjing, China. Background and aims: Statin therapy and cancer incidence have no rela- tionship according to a recent system review and meta-analysis. However, whether an intensified low-density lipoprotein-cholesterol (LDL-c) target of statin therapy is associated with cancer incidence has been rarely reported. Materials and methods: PUBMED, EMBASE, and Cochrane Central Regis- ter of Controlled Trials (CENTRAL) data as of July 2013 were searched for randomized controlled trials (RCTs) on statins. An intensified LDL-c target of S 274 1 C PS 046 Brain and glucose metabolism 669 MAPK signalling is required for brain insulin to suppress lipolysis in rodents T. Scherer1,2, E. Zielinski2, C. Lindtner2, J. O’Hare2, C. Buettner2; 1Department of Medicine III, Medical University of Vienna, Austria, 2Department of Medicine, Mount Sinai School of Medicine New York, USA. Background and aims: Insulin is the major anti-lipolytic hormone and is se- creted postprandially. We recently showed that brain insulin action is able to suppress lipolysis by reducing sympathetic nervous system outflow to white adipose tissue. However, the relative importance of brain insulin versus the well-characterized peripheral insulin effects in regulating whole body lipoly- sis is unknown. Materials and methods: Here, we studied the role of brain insulin versus pe- ripheral insulin action using two inducible insulin receptor knock-out mouse models and their respective controls. The first model lacks the insulin recep- tor (IR) in the whole body (IRdeltaWB), while the second model lacks the IR in all peripheral tissues (IRdeltaPER) except the brain. All mice were subject- ed to 2-hour hyperinsulinemic euglycemic clamp studies with stable isotope glycerol tracer to assess whole body lipolysis. To further identify the signal transduction pathway mediating brain insulin’s effects on lipolysis, we co-in- fused insulin with either the Pi3K inhibitor LY294002 or the MAPK inhibitor U0126 directly into the mediobasal hypothalamus (MBH) of male Sprague Dawley rats (n ≥ 4 per group) using stereotaxically implanted cannulae, and assessed lipolytic flux using the rate of appearance (Ra) of the glycerol tracer. Results: Male IRdeltaPER knock-out mice subjected to a hyperinsulinemic clamp showed a comparable % suppression of Ra glycerol per unit of insulin (4.7 ± 1.7 vs. 6.0 ± 1.4 %; P=0.6) to littermate controls despite being extremely insulin resistant (glucose infusion rate 21 ± 5 vs. 97 ± 10 mg*kg-1*min-1; P = 0.001). In contrast, male IRdeltaWB knock-out mice were unable to ad- equately suppress lipolytic flux (0.4 ± 0.2 vs. 6.9 ± 2.2 %, P=0.017) compared to control mice when exposed to high doses of insulin during a clamp. MBH insulin co-infusion with LY294002 suppressed Ra glycerol to a similar extent than MBH insulin alone (17 ± 2 vs.15 ± 6 vs. 32 ± 7 (controls) µmol*kg- 1*min-1; P=0.7 MBH LY+Insulin vs. MBH Insulin) in rats. However, the co- infusion of the MAPK inhibitor U0126 blocked the ability of insulin to sup- press lipolytic flux (60 ± 8 µmol*kg-1*min-1; P=0.006 vs. MBH LY+Insulin; P S 275 1 C Glud1-/- mice was confirmed by measurements of elevated incorporation of 13C from [U-13C]-glucose into those amino acids representative for oxi- dative metabolism of glucose products in the TCA cycle (alanine, aspartate, glutamate). CnsGlud1-/- mice were glucose intolerant with normal insulin production. However, they displayed increased endogenous glucose produc- tion and decreased muscle glucose uptake. Fat turnover was also modified with higher circulating triglycerides and ketone bodies. Furthermore, plasma concentrations of most amino acids were lower in CnsGlud1-/- mice in com- parison with control mice, suggesting increased mobilization of metabolites in CnsGlud1-/- skeletal muscles for provision of brain-specific energy sub- strates. Lower plasma pancreatic polypeptide and higher catecholamines levels highlighted repressed parasympathetic activity with a concomitant in- crease in the sympathetic input favouring energy mobilization. Conclusion: Brain-specific GDH deletion induced impairment of glutamate oxidation in the CNS leading to central hypoglycaemia; in turn mobilizing energy substrates from the periphery. Supported by: Swiss National Foundation 672 Acyl ghrelin acts in the brain to control liver function and peripheral glucose homeostasis R. Stark1, A. Reichenbach1, S.H. Lockie1, C. Pracht2, Q. Wu3, A. Tups2, Z.B. Andrews1; 1Physiology, Monash University, Melbourne, Australia, 2Animal Physiology, Philips University, Marburg, Germany, 3Traditional Chinese Medicine, Peking Union Medical College Hospital, Beijing, China. Background and aims: The gut hormone, ghrelin, acts on neuropeptide Y neurons in the arcuate nucleus of the hypothalamus and stimulates food intake to control energy balance. However, recent evidence suggests that peripheral ghrelin regulates glucose metabolism. Here, we designed experi- ments to examine how central acyl ghrelin infusion affects peripheral glucose metabolism under pairfed or ad libitum feeding conditions. Materials and methods: Three groups of mice received either ICV infusion of aCSF, ghrelin and allowed to eat ad libitum (Gh-lib) or ghrelin and pairfed to the average of the aCSF group (Gh-pf). Mini pumps delivered acyl ghrelin at a dose of 0.25 µg/hour at 0.5 µl/hour for 7 days. Results: There was no difference in daily fed blood glucose, insulin, glucagon, triglycerides or non-esterified fatty acids between the groups. Body weight gain and food intake was significantly higher in Gh-lib mice compared to aCSF and Gh-pf. However, both Gh-lib and Gh-pf groups exhibited heavier white adipose mass, independent of body weight and food intake. Gh-pf mice exhibited a state of negative energy balance, as hypothalamic NPY and AgRP mRNA expression was increased in Gh-pf mice relative to aCSF or Gh-lib. Gh-pf mice exhibited better glucose tolerance than aCSF or Gh-lib mice dur- ing a GTT, although both Gh-lib and Gh-pf increased insulin release dur- ing the GTT. Central acyl ghrelin infusion and pairfeeding also increased breakdown of liver glycogen and triglyceride, and regulated genes involved in hepatic lipid and glucose metabolism. Gh-pf mice had an increase in plasma blood glucose during a pyruvate tolerance test in relative to Gh-lib or aCSF mice. Our results suggest that under conditions of negative energy (Gh-pf mice) central acyl ghrelin engages a neural circuit that influences hepatic glu- cose function. Conclusion: Metabolic status affects the ability of central acyl ghrelin to reg- ulate peripheral glucose homeostasis. 673 Increasing central exposure of an MC4 agonist potentiates food intake and body weight lowering transiently in lean mice M. Fritsch Fredin1, A. Lindblom1, P.R. Ceuppens2, E. Lundborg3, M. Ahlqvist3, M.H. Larsson1; 1Bioscience, CVMD, AstraZeneca R&D, Mölndal, Sweden, 2Discovery Statistics, AstraZeneca, Alderley Park, UK, 3DMPK & Bioanalytical Chemistry, CVMD, AstraZeneca R&D, Mölndal, Sweden. Background and aims: It is well known that Melanocortin-4 receptors (MC4r) are widely expressed inside (hypothalamus) and outside (circum- ventricular organs, brain stem) the blood brain barrier (BBB) and that hy- pothalamus play a central role in regulation of food intake (FI) and body weight (BW). Interestingly, MC4r agonistic peptides which do not penetrate the BBB, decrease FI and BW in rodents both after peripheral and central administration. This study investigated whether increased CNS penetration potentiates the effects on FI and BW lowering of an orally administered small molecule MC4 agonist, cpd X. Materials and methods: Female wildtype (WT) and Mdr1a/b-Bcrp triple targeted mutation (KO) mice on FVB background were dosed orally twice daily for 10 days with vehicle, or cpd X at 0.5 μmol /kg (only KO) or 15 μmol /kg. Individual FI and BW were recorded daily, including 2 base line days, while plasma and brain PK profiling were performed after 1 and 10 days of dosing. Lean and diet induced obese (DIO) mice on C57Bl /6 background were dosed twice daily with either vehicle or 30 μmol /kg cpd X. Analysis of Covariance, and 5% significance level tests based on Student´s t-test were used for statistical validation (n=5-8). Results: Cpd X dosed at 15 μmol /kg resulted in plasma exposure (AUC0-24h, day 1) of 5.7 and 3.3 h*μmol/L in KO and WT mice, respectively, while the unbound brain /plasma ratio was more than 40 times higher in KO compared to WT mice. Cpd X significantly reduced FI and BW over the first 24hr of treatment in WT FVB (15 μmol /kg; FI; -20.2 ± 6.9%, BW; -2.3 ± 1.2%) and KO mice (FI; 0.5 μmol /kg -21.7 ± 6.3% and 15 μmol /kg -48.5 ± 8.9%, BW; 0.5 μmol /kg -0.9 ± 0.8% and 15 μmol /kg -5.1 ± 1.0%) relative to vehicle. The effect on FI was transient and returned towards vehicle levels after 48 hours especially in the KO mice. In contrast, BW lowering was sustained for the remainder of the study, especially for the WT, but also for the high dose KO mice. Cpd X (30 µmol /kg) reduced BW 10.3 ± 1.7%) in DIO mice but had no effect in lean C57Bl /6 mice after 10 days treatment. Conclusion: Higher brain exposure of cpd X was associated with greater transient reductions in FI and BW over the first 24 hours only. The initial effects support the hypothesis that CNS penetration of MC4r agonists poten- tiates FI/BW lowering. The lack of durable efficacy may be due to the FVB genetic background or the fact that the mice used were not obese. The latter is supported by the lack of BW reduction in lean versus DIO mice on C57Bl/6 background. Thus, lean FVB mice are unlikely to be predictive for the long- term effects on FI/ BW lowering in C57Bl/ 6 DIO mice. 674 Lack of alpha-MSH in POMC neurons controls hepatic glucose production M. Schneeberger1,2, A. Garcia1,3, D. Sebastian4,3, M. Imbernón5,6, C. Castaño1,3, Y. Esteban1,3, P. Garcia-Roves1,3, R. Gomis1,2, R. Nogueiras5,6, A. Zorzano4,3, M. Claret1,3; 1Laboratory of Diabetes and Obesity, IDIBAPS, Barcelona, 2Department of Endocrinology and Nutrition, Universitat de Barcelona, 3CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CBIERDEM), Barcelona, 4Departament de Bioquimica i Biologia Molecular, IRB Barcelona, 5Departamento de Fisiología, Universidad de Santiago de Compostela, 6CIBER de la Obesidad y Nutrición (CIBERobn), Santiago de Compostela, Spain. Background and aims: Anorexigenic proopiomelanocortin (POMC)-ex- pressing neurons in the arcuate nucleus of the hypothalamus constitute a fundamental nexus in the neuronal hierarchy controlling whole-body energy balance. In particular, POMC neurons are able to sense and integrate a range of nutrient and hormonal cues informing about the energy status of the or- ganism. Mitochondrial fusion, the merge of two mitochondria into one single compartment, is a key process implicated in mitochondria quality control and also in the bioenergetic adaptations to changes in environmental nutri- ent availability. A key protein mediating this fusion processes is Mitofusin 2 (Mfn2). We tested the hypothesis that Mfn2 in POMC neurons may be an important component of the sensing machinery implicated in the regulation of systemic glucose homeostasis. Materials and methods: Mice lacking Mfn2 in POMC neurons (POMCM- fn2KO) were generated and detailed phenotyping performed. Results: Young 6-week old POMCMfn2KO mice showed unaltered body weight and adiposity when compared to control counterparts. However, they exhibited mild glucose intolerance and insulin resistance. In vivo glucose- stimulated insulin release and pancreatic morphometric parameters were unaltered in POMCMfn2KO mice. Mutant mice showed increased glucose levels in response to pyruvate test and upregulation of key hepatic gluconeo- genic genes, enzymatic activity and proteins, consistent with enhanced he- patic glucose production (HGP). POMCMfn2KO mice also showed reduced insulin signaling in the liver indicating insulin resistance. These alterations were related to a dramatic reduction in alpha melanocyte stimulating hor- mone (αMSH), the main anorexigenic peptide released by POMC neurons. Diabetologia (2014) 57:[Suppl1]S1–S564 S 276 1 C ICV αMSH administration was able to reverse most of the glucose metabo- lism alterations observed in POMCMfn2KO mice. Conclusion: Deletion of Mfn2 in POMC neurons leads to defective glucose homeostasis independent of alterations in body weight. Our data indicate that the underlying cause is reduced αMSH production that eventually leads to enhanced HGP. Supported by: EFSD/Lilly, ISCIII, UB, RecerCaixa 675 Impaired cognition coexists with hippocampal remodelling obstacles in LDL receptor knock-out mice Y. Huang1, W.-Q. Xia2, S.-H. Wang2; 1Endocrinology, Yancheng First Peoples’ Hospital, 2Endocrinology, Zhongda Hospital of Southeast University, Nanjing, China. Background and aims: Evidences from clinical studies support that the ab- normal cholesterol metabolism in the brain lead to the progress of cognitive dysfunction. Low-density lipoprotein receptor (LDLR) is well known for its role in regulating brain cholesterol homeostasis. We investigated whether LDLR plays roles in cognition and the related potential mechanisms. Materials and methods: Twelve-month-old Ldlr-/-mice (n =12) and wild- type littermates C57BL/6J (n =14) maintained on a standard lab chow diet were subjected to the morris water maze test, sucrose consumption test, and open-field test when they were 4, 7, and 12 month-old, respectively. All ani- mals were killed for the hippocampal remodeling studies, including the hip- pocampal neural stem cells (NSCs) proliferation, survival and differentiation studies, as well as synapse and apoptosis studies one week after the comple- tion of all behavioral tests. Results: The plasma cholesterol concentrations of Ldlr-/- mice increased moderately than C57BL/6J (P S 277 1 C mance was strongly correlated to hippocampal glutamine (r=0.62, P S 278 1 C PS 047 Adipose tissue function in humans 679 Obesity decreases the expression of sirtuin proteins in adipose tissue of monozygotic twins, independent of genetic factors J.S. Jukarainen, S. Heinonen, M. Tummers, E. Rappou, J. Rämö, M. Mundiany, A. Hakkarainen, N. Lundbom, J. Kaprio, K. Pietiläinen; University of Helsinki, Finland. Background and aims: Sirtuins 1 and 3 (SIRT1, 3) are reported to be down- regulated in the subcutaneous adipose tissue (SAT) of obese individuals. Some studies associate human SAT SIRT1 and 3 expression with a favorable obesity phenotype. How much all of these associations reflect acquired obe- sity or are confounded by genetics and other factors is unclear. The objec- tive of this study was to assess the roles of SAT SIRT1, 3 and 5 expression in obesity, insulin resistance and inflammation, controlling for genetic and early environmental factors by studying monozygotic (MZ) twins. Materials and methods: Global gene mRNA expression, body composition and insulin sensitivity were assessed with Affymetrix U133 Plus 2.0 arrays, magnetic resonance imaging (MRI), DEXA scans and 75g oral glucose tol- erance tests (OGTT) in healthy MZ twin pairs (N = 40) of which 26 pairs were discordant for BMI (BMI difference > 3 kg/m2). For statistical analyses, Δ-variables depicting intrapair differences in different traits were calculated by subtracting the value of the leaner co-twin from the value of the heavier co-twin. These Δ-variables were used to control for genetic and early environ- mental factors that the twins share. Results: SIRT1, SIRT3 and SIRT5 were significantly downregulated in the SAT of heavier co-twins of BMI-discordant pairs. Out of adiposity variables, Δintra-abdominal fat volume was the best predictor of ΔSIRT1 expression (r = -0.81 P < 0.001). Where SIRT3 and 5 were correlated with adiposity in the individual level (r = -0.30 to -0.55, P < 0.007), these associations disap- peared in the intrapair differences control setting (r = -0.01 to -0.25, P > 0.114). ΔSIRT1 and 5 expression was associated with insulin sensitivity dur- ing OGTT (r = -0.32 to -0.47, P < 0.047). ΔSIRT1 and 5 were significantly cor- related with ΔCRP levels, ΔCD14 expression and various inflammatory gene expression pathways in SAT (r = -0.29 to -0.52, P < 0.042). Contrary to earlier research, ΔSIRT1 mRNA was not correlated with Δadiponectin mRNA (r = 0.12, P = 0.469). Conclusion: Adiposity (especially IA fat) was strongly related to SAT SIRT1 independent of genetic or early environmental influences, whereas SIRT3 and 5’s association with adiposity seems to be confounded by genetic or early environmental factors. SAT SIRT1 and 5 are negatively associated with in- sulin resistance and adipose tissue inflammation, even after controlling for twin-shared factors. Supported by: Academy of Finland (grants 266286, 272376), Novo Nordisk Foundation, Helsin 680 Obesity is associated with lower expression of mitochondrial respiratory chain and mitoribosomal transcripts in adipose tissue in monozycotic twins S. Heinonen1, J. Buzkova2, R. Kaksonen1, A. Hakkarainen3, J. Lundbom3, N. Lundbom3, J. Kaprio4,5, A. Rissanen1, A. Suomalainen2, K. Pietiläinen1, University of Helsinki; 1Obesity Research Unit, Diabetes and Obesity Research Programs Unit, 2Research Program of Molecular Neurology and Department of Neurology, 3Helsinki Medical Imaging Center, 4FIMM, Institute for Molecular Medicine, 5Finnish Twin Cohort Study, Department of Public Health, Hjelt Institute, Helsinki, Finland. Background and aims: Low mitochondrial number and activity are sug- gested to be underlying factors in obesity and metabolic syndrome. However the functions of mitochondrial ribosomes, which are responsible for the syn- thesis of respiratory chain proteins, in acquired obesity are largely unknown. Materials and methods: Rare young adult MZ twin pairs discordant (n =26, ΔBMI > 3 kg/m2) and concordant for obesity (n=14, ΔBMI S 279 1 C sponse, ChREBP expression was knocked down. Although siChREBP alone did not affect glucose metabolism, we observed a significant decreased glu- cose transport, glucose oxidation and DNL in cells exposed to dual inhibi- tion (siHSL/ChREBP) compared to simple inhibition (siLHS), suggesting an involvement of ChREBP in the effects mediated by HSL down regulation. Cellular lipidomic analyses demonstrated a clear qualitative change in fatty acid composition of phospholipids (PL) and triglycerides (TG) between the different siRNA conditions. The knock down of HSL led to an increase pro- portion of oleate (in PL and TG), in parallel to a decreased proportion of pal- mitoleate, while the knock down of ChREBP had the opposite. An intermedi- ate profile was observed with dual inhibition (siHSL/ChREBP). These results could partly be explained by the changes in the expression of ELOVL6, an elongase responsible for the shift between C:16 to C:18 series. Interestingly, these changes in PL composition could contribute to the observed changes in lipid rafts associated to insulin signaling. Conclusion: These data concur to a prominent role of DNL in the improved glucose metabolism observed when adipocyte HSL expression is diminished. Supported by: ANR ObeLip 682 High fat diet induced adipose tissue inflammation by suppression of anti-inflammatory phospholipids M.A. Osterhoff1,2, T. Frahnow1, A.-C. Seltmann1, A.S. Mosig3,4, K. Neunübel3, S. Sales5, J.L. Sampaio5, S. Hornemann1, M. Kruse1, A.F.H. Pfeiffer1,2; 1Clinical Nutrition, German Institute of Human Nutrition, Nuthetal, 2Endocrinology, Diabetes and Nutrition, Charité - University Medicine, Berlin, 3Molecular Haemostaseology, Jena University Hospital, 4Center for Sepsis Control and Care, Jena University Hospital, 5Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany. Background and aims: Little is known about the interdependency of com- plex gene-networks regulating the metabolism or signaling function of spe- cific lipids in humans. Lipid metabolism and signaling might be either con- trolled genetically or by environmental factors. The aim of the study was to correlate lipidomic and genomic data of human subjects to identify specific gene-modules responsible for the regulation or connected with the function of specific lipid metabolites. Materials and methods: In the NUGAT-Study (NUtriGenomic Analysis in Twins) all subjects (monozygous and dizygous twins) first received a car- bohydrate-rich low-fat diet for 6 weeks (Clinical Investigation Day 1, CID1 thereafter), immediately followed by a high-fat diet for 1 week (CID2) and additional 5 weeks (CID3). At each CID periumbilical fat biopsies were taken for RNA isolation and Agilent 8x40K gene micro Arrays. Plasma was meas- ured for lipid metabolites and cytokines (ELISA). Weighted gene Co-Expres- sion Network Analysis (WGCNA) was used for identification of co-expressed gene-networks and their correlation with lipidome data. MetaCore was used to find potential roles of identified gene-sets. Results: By analysis of the 5000 strongest regulated genes 10 gene-modules were identified whereof 1 revealed a high correlation (p S 280 1 C 685 Human adipose tissue microvascular endothelial cells (MVEC) regulate both tissue lipid in-/efflux and adipose cell PPARγ activation S. Gogg, U. Smith; Department of Molecular and Clinical Medicine, University of Gothenburg, Sweden. Background and aims: Animal studies have shown that specific deletion of PAPRγ in endothelial cells (EC) leads to dyslipidemia and endothelial dys- function. We have previously established a method to isolate and culture human primary MVEC from adipose tissue biopsies. We here asked if the MVEC and (pre)adipose cells cross-talk in the regulation of the bidirectional transport of fatty acids between the blood and the adipose cells. We also ex- amined the putative cross-talk between MVEC and PPARγ activation in the adipocytes which, in turn, is required to store excess fat in the adipose cells Materials and methods: MVEC were extracted from human subcutaneous adipose tissue biopsies using immunomagnetic separation with anti-CD31 Dynabeads. Selected MVEC were cultured and expanded in EC growth me- dium and the remaining preadipocytes in DMEM-F12 prior to differentia- tion. Subsequently, these cells were co cultured. To promote lipid accumula- tion, the MVEC were incubated with 300 µM oleic acid (OA) and/or 5µM Rosiglitazone (ROSI) for 48 h. Accumulated lipids in the cells were visual- ized with the fluorescent long-chain fatty acid analogue BODIPY-500/506 C1 C12 and microphotographs were taken with fluorescence or confocal micro- scopes. PPARγ activation was identified with the cell-based PPARγ reporter assay-GeneBLAzer UAS-bla HEK 293Hcells. Results: Both lipid transporters CD36 and FABP4 were highly expressed in MVEC and they were further markedly increased by the addition of the PPARγ agonist ROSI. Interestingly, a similar effect was seen by just add- ing oleic acid (OA) which in turn, also increased PPARγ gene activation in MVEC and there was no additive effect of combining OA with ROSI. Fur- thermore, the stimulatory effect of OA was significantly inhibited by the PPARγ inhibitor GW9662, indicating that OA directly activated PPARγ in MVEC. We also examined the cross-talk between MVEC and adipose cells in co-cultures and confirmed the presence of a regulated bidirectional lipid transport (flux and efflux) between these cells. We then investigated the effect of OA in cultured adipocytes and, in contrast to MVEC. OA did not directly increase CD36, FABP4 or PPARγ expression. However, in co-cultures, the addition of the MVEC insert increased the expression of both lipid transport- ers as well as that of PPARγ suggesting that MVEC secreted a PPARγ agonist to which the adipocytes responded. To further verify this, we incubated the PPARγ reporter cells (reporter assay) with MVEC conditioned media or with equivalent naïve media. Indeed, PPARγ activation was significantly increased when the cells were incubated with MVEC conditioned media as compared with naïve media. This was consistent for both basal culture medium and was further increased by the OA-containing media. Conclusion: These findings support the presence of an intimate cross-talk between adipose cells and MVEC in the regulation of the bidirectional trans- port of fatty acids. Furthermore, we can here also show for the first time that MVEC secrete a PPARγ ligand targeting both the MVEC and the adipose cells and that this process is further increased in the MVEC by fatty acids. 686 The effect of 2 days very-low calorie diet on cytokines in adipose tissue and in plasma in obese women M. Siklova, V. Mayerova, L. Rossmeislova, J. Kracmerova, E. Czudkova, V. Stich; Department of Sport Medicine, 3rd Faculty of Medicine, Charles University in Prague, Czech Republic. Background and aims: A number of bioactive molecules produced by adi- pose tissue (AT), such as cytokines, chemokines and acute phase proteins, have been suggested as a possible link between obesity and insulin resistance. It has been shown that circulating adipokines are elevated in obese and dia- betic subjects. Very-low-calorie diets (VLCD) are used as dietary interven- tion to achieve a rapid weight loss in obesity treatment and were shown to improve insulin sensitivity (IS) as soon as after 2days of intervention. The aim of our study was to investigate the response of pro- and anti-inflammatory cytokines in AT and plasma after 2 and 28 days of VLCD and its possible relationship to IS improvement. Materials and methods: 16 obese pre-menopausal women (BMI 32.5 ± 3.6 kg/m2) followed 800 kcal/d VLCD for 28 days. Anthropometric meas- urements, blood sampling and biopsy of subcutaneous abdominal AT were performed before diet and at the day 2 and 28 of VLCD. To evaluate insu- lin sensitivity, euglycemic-hyperinsulinemic clamp was performed at each phase and Homeostatic model assessment of insulin resistance (HOMA-IR) was calculated. IL-6, IL-1β, IL-8, IL-10, MCP-1, and TNFα, were analyzed in plasma and in conditioned media of AT explants using multiplex immunoas- say at Luminex 100. AT mRNA expressions of cytokines and lipogenic genes DGAT2, SDC, and FASN were analyzed. Results: At day 2 IS increased as evaluated by HOMA-IR but not by glu- cose disposal rate. At day 28 both indices of IS increased. Fat mass was not changed at day 2, while it decreased at day 28 (38±8.5 kg vs. 32±7 kg). At day 2, levels of IL-1β, IL-6 and IL-10 were increased in plasma, while the secretion and mRNA expression of all cytokines in AT was unchanged. In contrast, at day 28, secretion and mRNA expression of all cytokines in AT was increased, while their plasma levels were not different from the pre-diet condition. Expression of lipogenic genes DGAT2, SCD and FASN decreased at day 2 (DGAT2, p=0.04; SCD, p=0.005; FASN, p=0.05) and their decrease was further enhanced at day 28 (p=0.001 for all genes). Conclusion: Improvement of IS after 2days of VLCD was not associated with changes in expression and secretion of cytokines in AT. The concomitant in- crease in plasma cytokines suggests a contribution of non-adipose tissues. The lipogenesis in AT was down-regulated at the early phase of the diet. Supported by: IGA NT 14486, PRVOUK 687 The number of microRNA target sites identifies those transcripts that will experience the most dramatic downregulation in adipose tissue from obese subjects after weight loss F.J. Ortega1, J.M. Mercader2, J.M. Moreno-Navarrete1, L. Nonell3, E. Puigdecanet3, G. Xifra1, M. Sabater1, M. Moreno1, N. Fuentes-Batllevell1, D. Mayas4, N. Moreno-Castellanos5, W. Ricart1, F.J. Tinahones4, M. Malagón5, J.M. Fernández-Real1; 1Department of Diabetes, Endocrinology and Nutrition (UDEN), Institut d’Investigació Biomédica de Girona (IdIBGi), 2Joint IRB-BSC program on Computational Biology, Barcelona, 3Institut Hospital del Mar d’Investigacions Mèdiques (IMIM), Barcelona, 4Service of Endocrinology and Nutrition, Hospital Clínico Universitario Virgen de Victoria de Malaga, 5Department of Cell Biology, Physiology and Immunology, Instituto Maimonides de Investigaciones Biomedicas de Cordoba (IMIBIC)/Reina Sofia University Hospital, University of Cordoba, Spain. Background and aims: Bariatric surgery is an excellent therapeutic approach to elucidate the pathophysiology of obesity-associated metabolic distur- bances. The identification of abnormal function and gene expression patterns regulated by multiple microRNAs (miRNAs) may be useful in this context. In this study we aimed to identify changes in the genome wide adipose tis- sue (AT) transcriptome linked to the miRNA profile after surgery-induced weight loss. Materials and methods: Whole genome and miRNA expression patterns were assessed in subcutaneous AT of 16 morbidly obese women before and after surgery-induced weight loss. Validation of both genome wide and miRNA microarrays was made by quantitative real-time PCR using both longitudinal and cross-sectional cohorts. Three alternative datasets were downloaded from public repositories and analyzed to check for reproduc- ibility. Analyses in macrophages and differentiated human adipocytes were performed in vitro to try to comprehend the associations found in human AT. Results: Five thousand and eighteen different probe sets identified significant variations in gene expression. Only 15 miRNAs differed in the comparison before-after surgery-induced weight loss. Functional analysis revealed sig- nificant changes in cell cycle, development and proliferation, in lipid me- tabolism, and the inflammatory response, as further confirmed by in vitro results. Interestingly, when transcriptomes were analyzed taking into account the presence of miRNA target sites, the mRNAs that experienced the most dramatic changes were precisely those with more miRNA target sites (p=10- 195), being significantly down-regulated after surgery-induced weight loss. Conclusion: Current findings suggest that surgery-induced weight loss is as- sociated with an improvement in post-transcriptional modulation by miR- NAs, in parallel to changes in inflammatory status and lipid metabolism of obese AT. Supported by: EFSD/Lilly, FIS 2011-00214, CIBERobn, ISCIII Diabetologia (2014) 57:[Suppl1]S1–S564 S 281 1 C 688 Relationship between muscle/adipose tissue morphology, insulin sensitivity and beta cell function in diabetic and nondiabetic obese patients: effects of bariatric surgery S. Camastra1, A. Vitali2, M. Anselmino3, A. Gastaldelli4, R. Bellini3, R. Berta3, S. Baldi1, B. Astiarraga1, E. Barsotti1, M.C. Zingaretti2, S. Cinti2, E. Ferrannini1; 1Dept. Clinical & Experimental Med., University of Pisa, 2Univ Politecnica Marche, Ancona, 3Bariatric Surg. Unit, AOUP, Pisa, 4Inst Cl Phys, CNR, Pisa, Italy. Background and aims: Obesity is associated with low-grade inflammation in adipose tissue (AT), high lipolytic activity, ectopic fat deposition and insulin- resistance (IR). Bariatric surgery leads to major fat mass loss and improve- ments in IR and ß-cell function (ß-GS). Aim of this study was to relate the improved metabolic status after Roux-en-Y gastric bypass (RYGB) to the morphological changes of subcutaneous (SAT) and visceral AT (VAT) and skeletal muscle. Materials and methods: 14 non-diabetic (ND) and 14 type 2 diabetic (T2D) obese patients (BMI = 50 ± 2 and 52 ± 2 kg/m2) received a euglycaemic clamp study (to measure IR) with 2H5-glycerol infusion (to measure lipolysis) and a mixed meal test (to measure insulin secretion and ß-GS) before and one year after RYGB. During RYGB, VAT (omental), SAT and rectus abdominis samples were excised for light (LM) and electron microscopy (EM) analysis. Results: Before surgery, both T2D and ND patients showed marked IR and enhanced lipolysis, T2D also had impaired ß-GS. On LM, perivascular and interfibrillar muscle fat content was similar in ND and T2D; intramyocel- lular fat was more abundant in T2D than ND patients (1.0 [1.5] vs 0.1 [0.5] units, p=0.008). In SAT, adipocyte area and density of crown-like structures (CLS) were similarly increased in T2D and ND. In VAT, adipocyte area (5806 [1793] vs 5056 [1320], p=0.01) and CLS density (3.60 [7.31] vs 0.00 [1.75], p=0.0002) were higher in T2D than ND. ß-GS was inversely related to VAT adipocyte area (r = -0.58, p=0.004) and CLS density (r = -0.50, p=0.02). On EM, SAT and VAT adipocytes showed necrotic material, fibrosis, thickened capillary basal membrane, degenerating adipocytes with thin cytoplasm with extrusion of free lipids in interstitium and small mitochondria. In T2D pa- tients, VAT and SAT blood capillaries contained neutrophils. After RYGB (33% weight loss), IR and lipolysis were markedly improved, equally in ND and T2D (p S 282 1 C tochondrial architecture may represent adaptive changes to bioenergetics needs. Conditions requiring high mitochondrial ATP synthesis capacity and/ or efficiency are associated with mitochondrial elongation. Whereas condi- tions of excess energy supply and low ATP demand acutely induce mitochon- drial fragmentation. However it remains unclear if diversity in subcellular mitochondrial architecture may play a role in generating functional speciali- zation of subpopulations of mitochondria within the cell. We rationalized that the brown adipocyte may shed light on this question as its mitochondria are required for various and competing tasks, such as lipogenesis and beta oxidation as well as ATP synthesis and uncoupling, raising the possibility that the brown adipocyte may harbor a diverse set of mitochondria which are structured to fit different functions. Materials and methods: Brown pre-adipocytes were harvested from 3-week- old male C57BL6/J mice and differentiated in culture. Mitochondrial mem- brane potential, motility and morphology were measured using TMRE and PAGFP respectively with Zeiss LSM 710 confocal microscope. NADH au- tofluorescent was used to measure the NADH content of the mitochondria. Protein import and turnover were measured using MitoTimer probe. Results: Mitochondria in brown adipocyte are divided to two different popu- lations, mitochondria with submicron proximity to the lipid droplet or Peri- droplet (PD) mitochondria, and mitochondria that are located at least 5μm away from the vicinity of a lipid droplet or Cytoplasmic (C)mitochondria. Photo-conversion of matrix targeted photoactivatable GFP shows that these two populations rarely mix and that mitochondria remain faithful to the li- pid droplet they adhere to and do not share their matrix content with other mitochondria, nor do they switch their affiliation from one group to another. These studies also show that PD-mitochondria are more elongated while the C-mitochondria tend to be smaller in size suggesting a more coupled respira- tory function and involvement in ATP synthesis. In addition, using TMRE, we found that PD-mitochondria in brown adipocytes but not in other lipid- containing cells are in different energetic state than C-mitochondria. Ratio- metric imaging of membrane potential and NADH/NAD ratio show that PD- Mitochondria have higher NADH/NAD ratio and more polarized membrane potential. Protein import and turnover studies using MitoTimer also indicate that PD-mitochondria have a higher rate of protein import and biogenesis, in agreement with the hyperpolarized state. Altogether these observations sup- port a hypothesis that PD-mitochondria have a higher TCA cycle activity. Conclusion: These data suggests that peridroplet mitochondria represent a functionally exclusive subpopulation of mitochondria in terms of biogenesis and function. This may also demonstrate that the unique architectural char- acteristics of cells in diverse bioenergetics states may apply subcellularily to a diverse set of mitochondria within the cell, fulfilling different metabolic functions. 691 Diet-induced interleukin-15 promotes obesity by inhibiting adaptive thermogenesis in adipose tissues S. Ramanathan1, G. Lacraz1, Y. Cepero-Donates1, M. Mayhue1, M.-F. Langlois2, M. Rola-Pleszczynski1, S. Ilangumaran1; 1Pediatrics, Immunology Division, 2Medicine, Endocrinology Division, University of Sherbrooke, Canada. Background and aims: Inflammatory cytokines are implicated in the patho- genesis of obesity. Interleukin-15 (IL-15) is an inflammatory cytokine secret- ed by many cell types. IL-15 is also produced during physical exercise by skel- etal muscle and has been reported to reduce weight gain in mice. Contrarily, our findings on IL-15 knockout (KO) mice indicate that IL-15 promotes obe- sity. The aim of this study is to investigate the molecular and cellular mecha- nisms underlying the pro-obesity role of IL-15 in adipose tissues. Materials and methods: C57BL/6 wildtype and IL-15 KO mice were main- tained on high fat diet (HFD) or normal control diet. After 16 weeks, body weight, adipose tissue and skeletal mass, serum lipid levels and gene expres- sion in the adipose tissues were evaluated. Primary cultures of brown and white adipocytes were studied using the Seahorse cell metabolism analyzer to evaluate oxygen consumption rates. IL15 gene expression and the expression of inflammation markers were studied in visceral adipose tissues obtained from patients undergoing bariatric surgery. Results: We show that IL-15 deficiency prevents diet-induced weight gain and accumulation of lipids in visceral and subcutaneous white adipose tis- sues. Circulating levels of cholesterol and non-esterified fatty acids were el- evated in wildtype mice but not in IL-15 KO mice. The adipose tissues of IL-15 KO mice showed decreased expression of pro-inflammatory cytokines, chemokines and macrophage markers CD68 and F4/80. Gene expression analysis also revealed elevated expression of genes associated with adaptive thermogenesis in the brown and subcutaneous adipose tissues of IL-15 KO mice. Accordingly, oxygen consumption was increased in the brown adi- pocytes from IL-15 KO mice. Adipose tissues from obese diabetic patients showed a tendency towards increase in the expression of IL-15. Conclusion: Our results clearly show that IL-15 plays a pathogenic role in obesity. High fat diet-induced IL-15 promotes accumulation of fat in the white adipose tissues by inhibiting lipid utilization via adaptive thermogen- esis. IL-15 also promotes inflammation in adipose tissues that could sustain chronic inflammation leading to obesity-associated metabolic syndrome. Supported by: CIHR, CMDO 692 Effect of the n-3 LC-PUFA EPA on white-to-brown transition of primary human adipose-derived stem cells M. Elsen1, M.H. Schoemaker2, E.A.F. van Tol2, T. Romacho1, J. Eckel1; 1German Diabetes Center, Duesseldorf, Germany, 2Mead Johnson Nutrition, Nijmegen, Netherlands. Background and aims: Promoting the induction of brite adipocytes and thereby increasing energy expenditure may protect against the development of obesity. We have recently shown that both bone morphogenetic protein (BMP)4 and BMP7 induce browning of primary human adipose-derived stem cells (hASCs). Long-chain polyunsaturated fatty acids (LC-PUFAs) and their metabolites are potential PPARγ ligands and chronic activation of PPARγ is known to promote browning of white adipocytes. Furthermore, dietary n-3 LC-PUFAs have been shown to increase oxidative metabolism in white adipose tissue (AT) as well as adaptive thermogenesis in brown AT, suggesting a possible role for n-3 LC-PUFAs in the white-to-brown transi- tion. Therefore, we investigated the direct effects of the n-3 LC-PUFA eicosa- pentaenoic acid (EPA) on browning of hASCs in comparison to oleic acid (OA) as non-essential fatty acid and BMP4. Materials and methods: Primary hASCs were isolated from the subcuta- neous depot of different donors and challenged with EPA (20µM) or OA (20µM) during adipocyte differentiation (12 days). Cells treated with the known inducer of browning BMP4 (50 ng/ml) served as a positive control. Diabetologia (2014) 57:[Suppl1]S1–S564 S 283 1 C After 12 days of differentiation, lipid accumulation was measured by Oil Red O staining and gene expression was assessed by qRT-PCR. Results: Chronic treatment of hASCs with EPA but not OA increased lipid accumulation to similar extents as BMP4. The BMP4-mediated enhancement of PPARγ and C/EBPα expression was absent in OA-treated hASCs, while the n-3 PUFA EPA slightly increased the expression of PPARγ and C/EBPα. Neither EPA nor OA affected the expression of the white-specific marker Tcf21, which was significantly reduced by BMP4. Remarkably, UCP1 expres- sion was strongly enhanced (7-fold) after EPA exposure, while OA did not induce UCP1 expression. This EPA-mediated effect on UCP1 expression was comparable to that of BMP4. Conclusion: In conclusion, chronic exposure to physiological concentrations of the n-3 LC-PUFA EPA induces browning of hASCs derived from white AT, providing a potential mechanism for the beneficial effects of dietary n-3 LC-PUFAs on metabolism. The non-essential fatty acid OA has no effect on white-to-brown transition. The underlying mechanisms of the effect of the essential n-3 fatty acid EPA on browning will be investigated in future studies. Supported by: Mead Johnson Nutrition, FP7-EU Marie Curie-IEF (ADDIO- PIEF-2012-328793) 693 PPARγ and PPARα agonists induce white-to-brown conversion of human white adipocytes along with a metabolic shift from glucose to fatty acid oxidation V. Barquissau1, D. Beuzelin1, M. Giroud2, G. Béranger2, A. Mairal1, B. Roussel1, D. Pisani2, J.-C. Chambard2, E.-Z. Amri2, D. Langin1; 1UMR1048-I2MC, INSERM Toulouse, 2CNRS, Nice, France. Background and aims: Since it was demonstrated that adult humans possess active thermogenically-competent brown adipose tissue, fighting obesity and its complications through increasing adipose tissue expenditure has emerged as a putative strategy. “Browning” of white adipose tissue in rodents is sug- gested to be protective against obesity-induced insulin resistance. However, the capacity of human white adipocytes to acquire a brown/beige/brite me- tabolism has not yet been demonstrated. We aimed at identifying the molecu- lar and metabolic changes associated with the white-to-brown conversion of human mesenchymal adipose-derived stem (hMADS) cells following treat- ment by PPARγ (rosiglitazone, Rosi) or PPARα (GW7647, GW) agonists. Materials and methods: hMADS cells were differentiated into white adipo- cytes within 14 days. Thereafter, Rosi or GW were added for 4 additional days before molecular (microarrays, qPCR and Western blot) and metabolic measurements using radiolabelled tracers were carried out. Results: Both PPARγ and PPARα agonists promoted browning of white hMADS adipocytes, as evidenced by higher expression of UCP1 mRNA (29-fold and 18-fold respectively, p S 284 1 C immunoselected MSCA1+ cells exhibited brite precursor characteristics: high mitochondrial content (2-fold more compared to other progenitor subsets, p S 285 1 C 698 Dopaminergic effects on brown adipose tissue F. Holleman, L. Bähler, M.R. Soeters, J.B.L. Hoekstra, H.J. Verberne; Academic Medical Center, Amsterdam, Netherlands. Background and aims: Bromocriptine is a centrally acting dopamine recep- tor agonist that improves insulin sensitivity in obese subjects. Yet, no expla- nation has been found for this effect of bromocriptine. Brown adipose tissue (BAT), a tissue that converts calories into heat, might be involved in this pro- cess. Since the central sympathetic nervous system is the primary activator of BAT, we hypothesized that dopamine plays a role in the activation of BAT. Therefore, the aim of this study was to investigate the influence of bromocrip- tine on BAT activity in lean, healthy males. Materials and methods: We included 8 lean (BMI 23[21-25] kg/m2), healthy Caucasian males (20.9[19-23] years). All subjects were studied before and after using bromocriptine (1st week 1,25mg/day, 2nd week 2,5mg/day in the evening) in a climate room at 21°C after an overnight fast. On these 2 study visits we measured metabolic BAT activity, defined as maximal standardized uptake value (SUVmax), using 18F-Fluorodeoxyglucose Positron Emission Tomography CT scans. Furthermore we investigated glucose metabolism with a 7 point oral glucose tolerance test, energy expenditure (EE) using indi- rect calorimetry, weight and body temperature. Subjects recorded their eating behavior in the 4 days before the study visits. Results: The use of bromocriptine did not significantly alter metabolic BAT activity (SUVmax before 11.97[4.3-15.8]; after 10.3[2.7-18.2]), EE (before 2103 Kcal/day [1340-2486]; after 1915 [1784-2437]), body temperature (be- fore 36.0 °C [35.6-36.4]; after 36.2 [36.0-36.7]) or weight (80 kg[72.1-82.2]; after 80 [72.1-81.8]). Unexpectedly, subjects became significantly less insulin sensitive after bromocriptine use. The area under the curve for glucose in- creased (before 652 [539-752]; after 857 [772-992] (p=0.02)). But the area under the curve for insulin also increased (before 27x103 [26x103-37x103]; after 44x103 [41x103-65x103] (p=0.03)). There were no changes in diet be- tween the 2 measurements that could explain the change in insulin sensitivity. Conclusion: We conclude that bromocriptine does not activate BAT and does not increase EE in lean, healthy males. PS 049 Animal models of obesity and type 2 diabetes 699 Taurine treatment prevents obesity and diabetes through normalisation in circadian rhythms A.L. Castillo Figueroa1, H.A. Figueiredo1,2, S. Rebuffat1,2, R. Gomis1,2, E. Vieira1,2; 1Diabetes and Obesity Laboratory, IDIBAPS, 2CIBERDEM, Barcelona, Spain. Background and aims: Obesity and type 2 diabetes are associated with dis- ruption of circadian rhythms. It has been suggested that several nutrients are potential candidates to correct alterations in circadian rhythms and prevent these diseases. More over the aminoacid taurine has positive effects on meta- bolic dysfunction in obesity and diabetes in animal models of high fat diet. However, there is no study whether taurine can regulate circadian rhythms in mice. Therefore, we aim to study whether long term taurine treatment can ameliorate disturbances in circadian rhythms caused by high fat diet feeding Materials and methods: Male C57BL/6 mice were divided in 4 groups. Con- trol (C): mice fed with chow and Control+ taurine (C+T): mice fed with chow and 2.0% (w:v) of taurine in the drinking water. Obesity and diabetes were induced by HFD 45% of fat, and HFD + taurine group (HFD+T) treated with 2.0% (w:v) taurine in the drinking water. Glucose tolerance test and insu- lin tolerant test were performed in the last week of intervention. After 10 weeks of taurine treatment mice were sacrificed at different times of the day (6:00, 12:00, 18:00, 24:00). Plasma insulin levels were measured by ELISA. The expressions of clock genes in isolated pancreatic islets were measured by RT-PCR. Results: Mice treated with HFD increased food intake (p S 286 1 C neal glucose tolerance tests (OGTT/IPGTT). Also, tissue and blood samples were collected for biochemical analyses and gene expression data. Results: LPS-infused mice had an increased daily food intake compared to controls (185.1 ± 3.0 vs. 170.5 ± 2.9 g/kg, P < 0.01), which was reversed by RSV (175.2 ± 2.2 g/kg, P < 0.05). Despite this, there were no differences in average body weight gain after 28 days of treatment between the groups. Fasting glucose was reduced by RSV after 28 days of treatment (7.7 ± 0.2 vs. 8.9 ± 0.4 mM, P < 0.01). During an OGTT, LPS-treated mice had a reduced hyperglycemia 15 min after glucose administration compared to control animals (14.9 ± 0.6 vs. 17.5 ± 0.8 mM, P < 0.05). RSV had no effect on the glucose metabolism during OGTT. Neither RSV nor LPS had any significant effect on transcriptional expression of glucagon-like peptide-1 (GLP-1), pep- tide YY (PYY), glucose-dependent insulinotropic peptide (GIP), neuroten- sin, TNF-α, prohormone convertase 1/3 (PC1/3) or dipeptidyl peptidase-4 (DPP4) in the ileum of the small intestine. Conclusion: LPS-infusion for 28 days induced hyperphagia without affect- ing the body weight. RSV reversed this LPS-induced hyperphagia. To our surprise, LPS-treated mice did not show a worsening in glucose clearance as has earlier been suggested. Actually, LPS-treated mice had a small reduction in hyperglycemia 15 min following oral glucose challenge. Supported by: The A.P. Møller Foundation, The Danish Council for Strategic Research 701 T-bet regulation of the gut microbiota impacts insulin sensitivity E. Stolarczyk1, C. Bailey1, A.W. Walker2, N. Garrido-Mesa3, E. Marks3, J. Willis3, I. Jackson3, N. Powell3, J.B. Canavan3, G.M. Lord3, J.K. Howard1; 1Division of Diabetes and Nutritional Sciences, King’s College London, 2University of Aberdeen, 3Division of Transplantation Immunology and Mucosal Biology, King’s College London, UK. Background and aims: The gut microbiota plays a role in energy harvest, storage, and expenditure. Its composition differs in lean and obese humans and animals and transfer of microbiota from conventionally-raised to germ- free mice increases body fat content and insulin resistance. The immune sys- tem has co-evolved with the gut microbiota. We have recently highlighted the role of the immune cell transcription factor T-bet as a metabolic regulator uncoupling adiposity from insulin resistance. We hypothesised that the en- hanced insulin sensitivity of T-bet (Tbx21) deficient mice was associated with an altered gut microbiota composition. Materials and methods: Microbiota composition of T-bet deficient and con- trol wild-type (WT) mice was analysed by 16S-pyrrosequencing. Subsequent analyses of colonic metabolites (short chain fatty acids, SCFA) were assessed by gas chromatography-mass spectrometry. The impact of genotype on adi- posity and glucose homeostasis was determined by glucose and insulin toler- ance testing following microbiota transfer to antibiotic-treated mice. Results: T-bet deficient mice showed a decrease in the Firmicutes and De- ferribacteres phyla and a significant increase in the Bacteroidetes and TM7 phyla compared to wild-type mice. Furthermore, they exhibited an increase in the proportion of Clostidium species together with an increase in butyrate production (217.6±30.6 vs 336.6±32.9 μg/g feces, WT vs T-bet-/-, p=0.02). Transfer of T-bet deficient microbiota to microbiota-depleted WT hosts im- proved insulin sensitivity (256.6±7.2 vs 221.0±8.8, WT vs T-bet-/- area under the curve insulin tolerance test, p=0.02). This was accompanied by an in- crease in SCFA production, especially butyrate (216.3±76.0 vs 1581.1±636.8 μg/g feces, WT vs T-bet-/- microbiota, p=0.02). Conclusion: Our results suggest that deficiency of the immune cell transcrip- tion factor, T-bet is associated with an altered gut microbiota composition that can enhance systemic insulin sensitivity. Supported by: MRC grant MR/K002996/1 702 TIP27 increases insulin sensitivity through PI3-kinase/Akt pathway and regulates glucose homeostasis in mice L. Yuan1,2, L. Li1,2, G. Yang3; 1College of Laboratory Medicine, 2The Key Laboratory of Laboratory Medical Diagnostics in Ministry of Education, 3Department of Endocrinology, The Second Affiliated Hospital, Chongqing Medical University, China. Background and aims: TIP27 (Juxtaposed with another zinc finger protein1) is a 27kDa transcription factor containing three putative zinc finger motifs and its expression is associated with diabetes mellitus and prostate cancer. However, little is known about its role in regulation of metabolism. In this study, we investigated the effects of TIP27 overexpression on glucose homeo- stasis and insulin signaling pathway in high fat diet (HFD)-fed TIP27 trans- genic (Tg) mice and diabetic db/db mice. Materials and methods: Mainly,we use TIP27-Tg mice as research object. Also,we established a gain-of-function model of TIP27 in db/db mice using an adenoviru-mediated cDNA.Using these models,We assessed the effects of TIP27 overexpression in both TIP27-Tg mice and db/db mice on glu- cose metabolism and changes in insulin sensitivity during glucose tolerance tests (GTT) and insulin tolerance tests (ITT).Hyperinsulinemic-euglycemic clamp experiments was performed in Tip27-Tg mice.Glucose rates of ap- pearance (GRa) were determined with 3-[3H] glucose. Whole body GRa and glucose uptake (GRd) were calculated using the non-steady-state equation. mRNA and protein expressions were measured by qRT-PCR and Western blot, respectively. Results: We showed that TIP27 overexpression in both TIP27-Tg mice and db/db mice led to reduced total cholesterol, fasting plasma insulin levels, enhanced glucose tolerance and Insulin sensitivity.Hyperinsulinemic-eugly- cemic clamp experiments also demonstrated that TIP27 overexpression in TIP27-Tg mice enhanced insulin sensitivity. In addition, the expression levels of PEPCK and Glucose-6-phosphatase (G-6-Pase) mRNA as well as proteins were significantly decreased in TIP27-Tg mice, whereas the phosphorylations of insulin-receptor(IR), IRS-1, Adenosine Monophosphate Activated Protein Kinase(AMPK) and Akt were significantly increased in the insulin target tissues. Finally, inhibition of phosphatidylinositol 3-kinase(PI3-kinase)/Akt signaling by LY294002, a PI3-Kinase inhibitor, abolished the suppressive ef- fect of TIP27 overexpression on PEPCK and G-6-Pase expression. Conclusion: TIP27 plays an important role in glucose homeostasis by the regulation of hepatic glucose metabolism and insulin sensitivity and this regulation requires activation of the PI3-kinase/Akt pathway. Supported by: National Natural Science Foundation of China (81270913,81300702, 81300670) 703 LEOPARD syndrome-associated SHP2 mutation confers leanness and protection from diet-induced obesity M. Tajan1,2, T. Cadoudal1,2, A. Batut2,1, S. Deleruyelle1,2, S. Le Gonidec1,2, C. Saint-Laurent1,2, M. Vomscheid1,2, K. Treguer1,2, A. Nédélec1,2, M.-A. Marques1,2, P. Raynal2, H. Cavé3, T. Edouard4, P. Valet1,2, A. Yart1,2; 1Institute of Cardiovascular and Metabolic Diseases, U1048, 2Paul Sabatier University, Toulouse, 3Robert Debré Hospital, Paris, 4University Hospital Center, Toulouse, France. Background and aims: LEOPARD syndrome (LS) is a rare autosomal domi- nant disorder associating various developmental defects notably cardiopa- thies, dysmorphism and short stature. It is mainly caused by mutations of the PTPN11 gene that catalytically inactivate the ubiquitous tyrosine phos- phatase SHP2. SHP2 plays essential roles in transducing canonical signal- ing pathways (PI3K, MAPK) in response to a wide range of growth factors Diabetologia (2014) 57:[Suppl1]S1–S564 S 287 1 C and hormones. Recent studies using targeted invalidation of SHP2 in vari- ous tissues and organs pinpointed key roles for SHP2 in regulating energy metabolism. However, the metabolic outcomes of LS mutations have never been examined. Therefore, we aimed to evaluate the metabolic status associ- ated with LS using an original LS mouse model expressing one of the most common LS-associated SHP2 mutations. Beyond LS pathophysiology, since this mutation results in systemic partial inhibition of SHP2, this study also allowed assessing its role in the regulation of whole body metabolism. Materials and methods: We performed a comprehensive analysis of the met- abolic impact of LS-associated SHP2 mutations, by combining an extensive metabolic exploration of LS mice and of their WT littermates (body com- position, glucose/insulin tolerance, energy expenditure,⋯) with functional analysis in cellulo. Animals were handled in accordance with the principles and guidelines established by the local Ethics Committee. Results: Our results reveal that LS mice gain less weight than their WT lit- termates (p S 288 1 C Conclusion: Together, these data provide the first evidence in any organism that obesity can be reversed by acutely reducing the production and circula- tion of insulin. Our results have profound implications for nutritional guide- lines and therapeutic efforts to combat obesity. Supported by: CIHR 706 BACE2 suppression ameliorates beta cell dysfunction induced by human islet amyloid polypeptide (IAPP) overexpression G. Alcarraz-Vizán1,2, C. Castaño1,2, L. Cadavez1,2, M. Visa1,2, J. Montane1,2, J.M. Servitja1,2, A. Novials1,2; 1Diabetes & Obesity Laboratory, IDIBAPS, 2CIBERDEM, Barcelona, Spain. Background and aims: BACE2 (β-site APP-cleaving enzyme 2) is a β-protease that has been found in the brain, where it is thought to play a role in the development of Alzheimer’s disease (AD). It has also been localized in the pancreas, where it seems to play a physiological role since BACE2-defi- cient mice elicit better glucose tolerance than control littermates. Amyloido- genic diseases, including AD and type 2 diabetes (T2D), have been reported to share the accumulation of abnormally folded and insoluble proteins that interfere with cell function. In the case of T2D, amylin (IAPP) deposits have been shown to be a key feature of the disease. The aim of the present study was to investigate the effect of BACE2 modulation on β-cell alterations in- duced by IAPP overexpression. Materials and methods: Heterozygous-hIAPP mice, BACE2-KO mice and their respective controls were used to analyze their phenotype after 16 weeks with high-fat diet (HFD) feeding. Afterward, these two models were crossed in order to analyze the impact of BACE2 suppression on β-cell alterations observed in Tg-hIAPP mice. Insulin tolerance test (ITT) and glucose toler- ance test (GTT) were performed to evaluate the metabolic phenotype, and the area under the curve (AUC) of the GTT was calculated as a measure of glucose homeostasis. The ability to secrete insulin in response to glucose (GSIS) was quantified with an ELISA kit. Proliferation was analyzed by Ki67 immunostaining and β-cell mass by insulin immunostaining. Results: The GTT of Tg-hIAPP mice after 16 weeks of regular chow diet re- vealed glucose intolerance with respect to the wild type animals. These ani- mals showed a 1.3-fold increase (p S 289 1 C creased in female (p S 290 1 C to knock down SIRT1, and transfected cells were then treated with exendin-4 for 24 h. Results: The animal study showed that both body weight and epididymal fat weight were sharply decreased after exenatide treatment compared with saline control in WT+HFD group. However, the decrease in SIRT1+/-+HFD group after exenatide treatment was much milder. While lipids profile showed the same change. Both RT-PCR and western blot results indicated that the ex- pression of SIRT1 was up-regulated in WT+HFD group with exenatide treat- ment as compared with saline control; however, no significant difference was observed in SIRT1+/-+HFD mice after exenatide treatment. In differentiated 3T3-L1 adipocytes, exendin-4 up regulated SIRT1 and p-AMPK expression in a dose-depended way. Lipolytic related factors including p-ACC, ACC and ATGL were all induced by exendin-4. After knocking down SIRT1 with SIRT1 RNAi, these factors were down regulated dramatically even with ex- endin-4 treatment. Conclusion: Our data demonstrated that GLP-1 receptor agonist exenatide could reduce visceral fat by inducing lipolysis in adipocytes, and this effect is mediated by SIRT1 both in vivo and in vitro. It might be a new mechanism for GLP-1 and its mimetics in the role of losing weight. Supported by: 973 project (2012CB517506) 711 Activation of adipose tissue cannabinoid receptors 1 (CB1R) alters antilipolytic action of insulin and increases lipolysis in mice P. Degrace, T. Muller, S. Troy-Fioramonti, L. Demizieux, J. Gresti, M. Bouam, B. Vergès; Equipe Physiopathologie des Dyslipidémies, UMR866 INSERM-Université de Bourgogne, Dijon, France. Background and aims: Recent data indicate that activation of peripheral en- docannabinoid system (ECS) in tissues such as liver, muscle or adipose tis- sue may directly influence carbohydrate and lipid metabolism. The existence of cannabinoid receptors 1 (CB1R) and ECS enzymatic machinery has been demonstrated in mature adipocytes, nevertheless studies regarding their role in lipogenesis and lipolysis control are often conflicting. This study was designed to examine the consequence of ECS activation by anandamide on lipolysis activity and related regulation pathways. Materials and methods: Lipolysis activity was estimated measuring for 45 min plasma glycerol release in response to β3-adrenergic receptor agonist (BRL37344) in wild type, DIO or CB1R-/- mice after acute peripheral anan- damide injection compared with vehicle. Additional in vitro experiments were conducted to test direct effects of anandamide on glycerol release and signaling pathways on adipose tissue explants exposed to various concentra- tions of insulin and norepinephrine. Results: While anandamide alone had no remarkable effects on basal lipoly- sis, ECS activation potentiated the effect of BRL37344 on glycerol release in wild type mice. The effect of anandamide on stimulated lipolysis was strongly reduced in CB1R-/- while it was amplified in obese animals whose adipose tissue CB1R mRNA expression was much higher than in lean mice. In control mice, the stimulatory effect of BRL37344 on glycerol release was totally coun- teracted by insulin injection (0,025UI/kg) while it was partially maintained in the presence of anandamide suggesting that ECS activation may be associ- ated with an alteration of the inhibitory action of insulin on lipolysis. These findings were also observed in cultured explants exposed to anandamide in which inhibition of glycerol release by insulin was abrogated. Further, anan- damide treatment increased protein levels of the active form of hormone- sensitive lipase and reduced level of Akt and Pi3K phosphorylation compared to control in accordance with a decrease in the activity of insulin-dependent signaling cascade. Conclusion: All together, these data showed that activation of ECS in adi- pose tissue increases lipolysis by altering the antilipolytic action of insulin. This suggests that antagonism of CB1R may constitute a new strategy to limit ectopic fat deposition associated with obesity. Supported by: SFD research grant 712 Increased BMP4 improves insulin sensitivity and increases beige/brown adipogenesis in adult mice J.M. Hoffmann1, J.R. Grünberg1, I. Elias2, F. Bosch2, A. Hammarstedt1, S. Hedjazifar1, U. Smith1; 1Department of Molecular and Clinical Medicine/Diabetes, The Lundberg Laboratory for Diabetes Research, Gothenburg, Sweden, 2Department of Biochemistry and Molecular Biology, CBATEG (Center of Animal Biotechnology and Gene Therapy), Barcelona, Spain. Background and aims: Bone Morphogenic Proteins (BMPs) belong to the transforming growth factor beta superfamily and regulate a wide range of developmental and cellular processes. BMP4 has been shown to be critical for adipose precursor cell commitment and differentiation to the white adipose lineage. However, recent data also suggest a role of BMP4 in beige/brown adi- pose cell differentiation. A previous study has examined the developmental effect of BMP4 in mice genetically engineered to overexpress BMP4 in the adipose tissue. However, to be a potential therapeutic target it is important to examine if BMP4 also is able to increase beige/brown adipogenesis in adult animals. This is further underscored by our recent findings that adipose pre- cursor cells in obesity exhibit a resistance to the effects of BMP4. To examine the effect of BMP4, we injected adult mice with AAV8-BMP4 or AAV8-null virus vectors targeting the liver and, thus, increasing circulating BMP4 levels. Materials and methods: Six weeks old male C57BL6/N mice received a retro- orbital injection of AAV8 expressing BMP4 or control (null) virus. The mice were fed a high fat diet (HFD) for 16 weeks and they underwent ip GTT, ip ITT and DEXA. Adipose tissue and muscle were taken for gene and protein analyses and immunohistochemistry was performed on paraffin-embedded, formalin fixed tissue samples. All reported results are statistically significant. Statistical analyses were performed using Students t-test or Mann-Whitney non-parametric U-test. Results: HFD-BMP4 mice gained less weight than the HFD-null mice in spite of having the same food intake. They also had reduced total adipose tissue mass and smaller adipose cells. Furthermore, they had an improved glucose tolerance and were more insulin-sensitive than the HFD-null mice. In addition to the reduced fat mass they also had increased lean body mass. Both gene and protein expression of beige/brown adipose markers (TMEM26, CD137 and UCP1) were increased in the subcutaneous adipose tissue. Markers of mitochondrial biogenesis, like PGC1α, were significantly increased in both skeletal muscles and the adipose tissue. Since obesity also has been shown to lead to increased adipose tissue fibrosis and BMP4 may exert an anti-fibrotic effect, we examined markers of this process. Interest- ingly, both α-SMA and CTGF mRNA levels were significantly reduced in the obese BMP4 mice. Conclusion: We here demonstrate that increased circulating levels of BMP4 in obese and adult mice lead to clear positive phenotypic changes with re- duced body fat, increased lean body mass together with increased insulin sensitivity and improved glucose tolerance. These effects are probably due to the marked induction of beige/brown adipose cells in the subcutaneous adipose tissue as well as increased mitochondrial biogenesis also seen in the skeletal muscles. BMP4 also reduced markers of adipose tissue fibrosis known to be increased in obesity. 713 Glyoxalase-1 overexpression reduces body weight and adipokine expression, and improves insulin sensitivity in high-fat diet-induced obese mice D. Maessen1, O. Brouwers1, T. Miyata2, C. Stehouwer1, C. Schalkwijk1; 1Internal Medicine, Maastricht University Medical Centre, Netherlands, 2Centre of Translational and Advanced Research, Tohoku University, Sendai, Japan. Background and aims: The development of obesity, and especially the ex- pansion of visceral adipose tissue (VAT), is associated with the development of inflammation and insulin resistance. We previously demonstrated an ac- cumulation of advanced glycation endproducts (AGEs) in obese VAT, lead- ing to dysregulated expression of adipokines. In addition, the increased AGE levels were associated with the development of insulin resistance. Based on previous work, we hypothesized that inhibition of the major AGE precur- sor methylglyoxal (MGO) attenuates the impaired adipokine profile and im- proves insulin sensitivity in obesity. Because MGO can be detoxified by the Diabetologia (2014) 57:[Suppl1]S1–S564 S 291 1 C glyoxalase-1 (GLO-1) enzyme, we used a GLO-1 overexpressing transgenic mouse model which was fed with a high-fat diet (HFD). Materials and methods: After a run-in period of 4 weeks, both wild-type (WT) mice and GLO-1 transgenic mice (10 weeks old) were either fed with a low-fat diet (LFD, 10% kcal% fat) or a HFD (45% kcal% fat) for 15 weeks. In the last two weeks of the study we performed an intraperitoneal glucose tolerance test (ipGTT) and insulin tolerance test (ipITT). All mice were eu- thanized after 15 weeks of LFD or HFD and multiple organs were collected for ex vivo examination. Results: Endogenous GLO-1 mRNA levels of the VAT were significantly lower in WT mice of the HFD group compared to the WT mice of the LFD group (p S 292 1 C Materials and methods: In two independent cohorts [cohort 1, participants with different degree of obesity (n= 105) and cohort 2, morbidly obese par- ticipants (n=47)], DBC1 gene expression was investigated in adipose tissue according to obesity and insulin resistance. DBC1 mRNA and protein levels during adipocyte differentiation and the effects of DBC1 knockdown (using Dbc1-targeted and control shRNA lentiviral particles) in the early stage of this process were also evaluated. Results: DBC1 mRNA was detected at substantial levels in both visceral (VAT) and subcutaneous (SAT) adipose tissue (mainly in the adipocyte frac- tion), decreased in obese subjects, and directly associated with the expression of lipogenic and lipolytic genes in two both independent cohorts. Strikingly, DBC1 was linked to TNF mRNA levels in obese and morbidly obese partici- pants and the administration of macrophage conditioned medium resulted in increased DBC1 gene expression in adipocytes. In line with these find- ings, in the early stages of adipocyte differentiation, Dbc1 knockdown led to decreased lipid accumulation and adipogenic gene expression in parallel to the increase of Sirt1 and AMPK activity. On the other hand, Dbc1 gene knockdown also led to a significant reduction in the expression of inflam- matory genes (Tnf, Il6, Stamp2, Lbp and Mcp1) at the end of adipocyte dif- ferentiation. Conclusion: Altogether these findings suggest that DBC1 exerts a bifunc- tional role in adipocyte physiology, being essential for adipocyte develop- ment, but also involved in adipocyte inflammation in the late stage of adipo- cyte differentiation. Supported by: ISCIII 717 Antenatal antipsychotic exposure induces multigenerational and sex-specific programming of body weight and diabetes susceptibility in adult mouse offspring E. Courty1, P. Gobalakichenane1,2, M. Garcia1, T. Ledent1, D. Mitanchez2, M. Buyse1,2, B. Fève1,2; 1INSERM U938, 2APHP, Paris, France. Background and aims: Atypical antipsychotic medications have revolution- ized the treatment of severe mental diseases, but their metabolic side effects have been pointed, including obesity and diabetes. Among them, Olanzapine use has been reported during pregnancy and breastfeeding, but so far no clin- ical trials assessed the safety of in utero olanzapine exposure to fetuses and infants, or long term health outcome. In this study we analysed the transgen- erational impact on metabolic homeostasis of mouse fetal antipsychotic ex- posure throughout gestational period. Materials and methods: At 7d gestation, pregnant F0 CD1 mice received olanzapine (olz4 or olz8: 4 or 8 mg/kg/d) or haloperidol (hld: 2 mg/kg/day) or vehicule (control) through mini-osmotic pump until they gave birth natu- rally. F1 female offspring exposed in utero to olz (mat-olz) and hld (mat-hld) were bred to produce F2 offspring. Offsprings were weight every 3 days until the age of 3 weeks then weekly until 3 months of age. Glucose metabolism was investigated by insulin tolerance test (ITT), oral glucose tolerance test (OGTT) and plasma insulin measurement at the age of 3 months and 1 year for F1 offsprings, and at the age of 3 months for F2. Part of the F1 offspring cohort was sacrificed at 3 months and 1 year to analyse body and tissue com- position and measure biochemichal and hormonal parameters. Results: The litter size of F1 mice was not influenced by the maternal treat- ment. However birth weight of mat-antipsychotic offspring groups of both sexes was significantly decreased compared to mat-control (p S 293 1 C PS 051 The role of the liver in glucose and lipid metabolism 718 Fat accumulates preferentially in the right liver lobe in non-diabetic subjects: implications for the diagnosis of NAFLD H. Bian1,2, A. Hakkarainen3, Y. Zhou2, N. Lundbom3, X. Gao1, H. Yki-Järvinen2; 1Department of Endocrinology and Metabolism, Zhong Shan Hospital, Fudan University, Shanghai, China, 2Minerva Foundation Institute for Medical Research, Helsinki, 3Department of Radiology, Helsinki Medical Imaging Center, Helsinki University Central Hospital, Finland. Background and aims: The right lobe of the liver preferentially receives blood flow from superior mesenteric vein, while the left lobe mainly drains the spleen. In obese subjects, the rate of visceral lipolysis is increased com- pared to non-obese subjects and can account for up to 50% of free fatty acids (FFA) delivery to the liver. Many obese subjects accumulate fat in the liver due to non-alcoholic causes (NAFLD) and therefore possibly an increased flux of FFA from visceral fat to the superior mesenteric vein. Studies using Doppler ultrasonography in healthy volunteers have shown that in response to a meal, intrahepatic portal vein blood flow increases more in the right than the left lobe. FFA delivery from both the meal and from visceral lipolysis might preferentially increase liver fat content more in the right than the left lobe. Metabolic consequences of hepatic insulin resistance in NAFLD such as hyperinsulinemia and hypertriglyceridemia could therefore also be better correlated with fat in the right than the left lobe. We aim to examine the dis- tribution of liver fat (LFAT) in non-diabetic subjects and to test whether the fat in the right as compared to the left lobe correlate better with components of the metabolic syndrome or not. Materials and methods: We determined LFAT by 1H-MRS in the right lobe (LFAT%MRS), and by MRI (LFAT%MRI) in four regions of interest (ROIs 1-4, ROI1-2 in the left and ROI 3-4 in the right lobe) in 97 carefully phe- notyped non-diabetic subjects (age range 22-74 yrs, BMI 18-41 kg/m2) and compared the accuracy of LFATMRI in the different ROIs in diagnosing non- alcoholic fatty liver disease (NAFLD) using areas under the receiver operator characteristic (AUROC) curves. Results: 38% of the subjects had NAFLD (LFAT%MRS). LFAT%MRI in ROIs 1, 2, 3 and 4 averaged: 4.8±0.5%, 5.5±0.5%, 5.8±0.5% and 5.7±0.5%. The LFAT%MRI was significantly lower in ROI 1 than in ROI 2 , ROI 3 and ROI 4 . LFAT%MRI was significantly higher in the right (5.7±0.5%) than the left (5.1±0.4%) lobe (p S 294 1 C isoprenoid biosynthesis only in PKCε KO. Taken together, these further sug- gest that PKCδ and PKCε exert partly overlapping and differential effects on liver metabolism in response to fat oversupply. Conclusion: These studies therefore provide a detailed comparison of the ef- fects of fat feeding and deletion of PKC isoforms at the protein level, which will aid further understanding of the function of these PKCs, given their as- sociation with defective glucose homeostasis. Supported by: NHMRC Australia 721 The hepatokine betatrophin is increased in nonalcoholic fatty liver disease and may affect insulin secretion in prediabetes A. Peter1, K. Kantartzis1, A. Königsrainer2, I. Königsrainer2, H. Staiger1, F. Machicao1, H.-U. Häring1, A. Fritsche1, N. Stefan1; 1Department of Internal Medicine IV, Institute of Diabetes Research and Metabolic Diseases of the Helmholtz Center Munich at the University of Tübingen and German Center for Diabetes Research, University of Tübingen, 2Department of General, Visceral and Transplant Surgery, Germany. Background and aims: Animal data suggest that the liver-secreted protein betatrophin is elevated in insulin resistant states and improves glucose toler- ance by expanding beta cell mass. We now investigated whether betatrophin is a hepatokine that may be increasingly produced in nonalcoholic fatty liver disease (NAFLD) and may affect insulin secretion in humans. Materials and methods: In liver tissue samples from 92 donors the rela- tionship of hepatic betatrophin mRNA expression with liver triglyceride (TG) content, and in a subgroup of 29 donors the relationship of betatro- phin mRNA expression with circulating betatrophin was studied. The asso- ciations of the cholesterol-regulating single nucleotide polymorphism (SNP) rs2278426 in the betatrophin-encoding gene DOCK6, with insulin secretion was studied in a cross-sectional study (N=2136) and during 9 month of a lifestyle intervention (N=344). Results: Betatrophin mRNA expression correlated positively with liver TG content (r=0.31, p=0.0025). In addition, a positive relationship was found be- tween betatrophin mRNA expression and betatrophin plasma levels (r=0.41, p=0.03). The minor T allele of the SNP rs2278426, that also associated with lower plasma cholesterol levels in the present study (p=0.002), associated with higher insulin secretion, adjusted for sex, age and insulin sensitivity, in subjects with prediabetes (N=701, p=0.044), but not in subjects with nor- mal glucose regulation (N=1435, p=0.78). The minor T allele of the SNP rs2278426 also predicted a larger increase in adjusted insulin secretion dur- ing the lifestyle intervention in subjects with prediabetes (p=0.01) and associ- ated with higher arginine-stimulated insulin secretion (p=0.04) in 97 subjects during a hyperglycaemic clamp. Conclusion: We provide novel information that the liver-secreted protein betatrophin is increasingly expressed in NAFLD and that it may not only be involved in cholesterol metabolism, but also in glucose metabolism in sub- jects who are at increased risk for type 2 diabetes. Supported by: BMBF German Center for Diabetes Research; DFG KFO 114 and STE 1096/3-1 722 Acute elevation of triglycerides and free fatty acids does not modulate plasma betatrophin levels in healthy individuals D. Tripathy1,2, T. Arakawa1, A. Chavez-Velazquez1, Z. Perez-Cadena1, R. DeFronzo1; 1Diabetes, Medicine, University of Texas Health Science Center, 2Audie L Murphy Veterans Affair Hospital, San Antonio, USA. Background and aims: Betatrophin (ANGPTL8) is a liver-derived protein, and its overexpression in liver has been shown to cause beta cell proliferation, increased insulin secretion and hypertriglyceridemia in mice. ANGPTL8 knockout leads to marked reduction in postprandial triglyceride levels in mice. However, the relationships between plasma triglycerides, free fatty ac- ids (FFA), and betatrophin have not been studied in humans. Materials and methods: We examined the effect of acute elevation of plasma triglycerides and FFA on plasma betatrophin levels in 9 healthy NGT subjects (7 male/2 female, age = 42 ± 4 yrs, BMI = 26.3 ± 1.0 kg/m2, FPG = 90±3 mg/ dl, 2h PG = 135 ± 11 mg/dl), without family history of T2DM. Subjects re- ceived on two different occasions either a lipid (intralipid 60 ml/h) or saline (0.9% NaCl 60 ml/h) infusion for 6 hours followed by euglycemic insulin clamp (80 mU/m2.min) for 2 hours. Plasma triglyceride, FFA, and betatro- phin levels were measured at baseline, 6 hours, and 8 hours. Results: Plasma triglyceride levels during the lipid infusion increased from 187 ± 39 mg/dl at baseline to 455 ± 59 mg/dl at 6 hours and declined to 398 ± 62 mg/dl at 8 hours, and during the saline infusion decreased from 183 ± 60 mg/dl to 129 ± 28 mg/dl at 6 hours and 106 ± 25 mg/dl at 8 hours. Plasma FFA levels during the lipid infusion increased from 370 ± 39 µM to 846 ± 68 µM at 6 hours and declined to 496 ± 66 µM at 8 hours. Plasma FFA concen- trations did not change during the saline infusion. Insulin-mediated glucose disposal (Rd) was reduced during the lipid versus saline infusion (9.08 ± 1.5 vs 11.3 ± 1.4 mg/kgFFM.min, p = 0.02). Plasma betatrophin concentrations during the lipid infusion were 22 ± 15, 27 ± 16, and 27 ± 17 ng/ml at 0, 6, and 8 hours, respectively (p=ns). The corresponding values of betatrophin during the saline infusion were 21 ± 6, 18 ± 6, and 24 ± 5 ng/ml, respectively. There was no relationship between plasma triglyceride, FFA, betatrophin concen- trations, or insulin mediated whole body glucose disposal. Conclusion: Short-term physiologic increase in plasma FFA, and triglycer- ides induce insulin resistance in skeletal muscle, but does not affect plasma betatrophin levels in healthy subjects. The association between betatrophin and lipids reported in animals is not observed in humans. 723 ZBTB20 regulates lipid metabolism W. Zhang1, L. Zhou1, L. Li1, Y. Zhang1, X. Jiang1, R. Chen1, D. Zou2; 1Department of Pathophysiology, Second Military Medical University, 2Department of Endocrinology, Changhai Hospital, Shanghai, China. Background and aims: We aims to investigate the potential role of novel zinc finger protein ZBTB20. Materials and methods: We established liver-specific Zbtb20 knockout mice (LZB20KO) by Cre/LoxP, analyzed phenotypes, determined gene expression by RT-PCR and Western blot, and examined protein/DNA interaction by ChIP and reporter assay. Results: Specific deletion of ZBTB20 in the liver resulted in a marked reduc- tion in plasma lipid levels and hepatic triglyceride contents, and a significant improvement of high carbohydrate diet-induced hepatic steatosis and insu- lin resistance. Hepatic lipogenesis was substantially impaired in the absence of ZBTB20, which was associated with decreased expression of ChREBP-α and its target genes critical for glycolysis and lipogenesis, and could in part be restored by overexpression of ChREBP-α. Furthermore, liver ZBTB20 ex- pression was significantly increased in ob/ob mice. Strikingly, liver-specific deletion of ZBTB20 in ob/ob mice could dramatically suppress hepatic lipo- genesis, and improve obesity-induced hyperglycemia, dyslipidemia, hepatic steatosis, and insulin resistance. Conclusion: Together, these date indicate that liver ZBTB20 is essential for lipid etabolism and may serve as a therapeutic target for metabolic syndrome. Supported by: National Natural Science Foundation of China 724 The role of starvation-induced autophagy in gluconeogenesis and ketogenesis M. Kondo, A. Takagi, S. Kume, S. Ugi, T. Uzu, H. Maegawa; Department of Medicine, Shiga University of Medical Science, Japan. Background and aims: During starvation, fatty acids released from adipose tissue are mainly used for hepatic gluconeogenesis and ketogenesis to main- tain systemic energy homeostasis in animals. Autophagy, a major intracel- lular degradation system, is induced by starvation. Systemic autophagy-de- ficient mice die in neonatal starvation period, suggesting that autophagy is a critical survival response against starvation. However, the exact role of star- vation-induced autophagy in systemic energy homeostasis remains unclear. Materials and methods: To determine the importance of autophagy in each organ, we examined starvation-induced gluconeogenesis and ketogenesis in liver specific Atg5 knockout mice (L-Atg5-/-), kidney proximal tubular epi- thelial cells specific Atg5 knockout mice (K-Atg5-/-), and their control mice (Atg5lox/lox). Results: Under ad-libitum feeding condition, there were no significant dif- ferences in plasma ketone levels. But under 36-h fasting condition, increase in plasma ketone levels was impaired in L-Atg5-/- mice but not in K-Atg5-/- mice. The impairment of ketogenesis in L-Atg5-/- mice was incomplete. In Diabetologia (2014) 57:[Suppl1]S1–S564 S 295 1 C addition, the oil-red-O-stained lipid accumulation as a main source of ke- togenesis and elevation of expression of HMG-CoA synthase, a rate-limiting enzyme of ketogenesis were observed in the kidney likely in the liver. We then hypothesized that the kidney compensated for the absence of ketone in L-Atg5-/- mice. To examine this hypothesis, we generated the mice lack- ing Atg5 in both the liver and kidney. These mice showed further decline in plasma ketone concentrations compared to L-Atg5-/- mice during a 36-h fast. Starvation-induced lipid droplets formation was impaired in Atg5-deficient organs. No significant differences in blood glucose levels and the expression of some enzymes for gluconeogenesis were observed among each type of mice, even with starvation. Conclusion: These results suggest that starvation-induced autophagy is es- sential for ketogenesis relating with lipid accumulation from free fatty acid, and that the kidney potentially has a function to generate ketone bodies. Supported by: KAKENHI Grants-in-Aid PS 052 Mitochondrial function, oxidative stress and glucose metabolism 725 The Cohen diabetic rat: a matter of mitochondrial disorder? A. Kogot-Levin1, C. Mantzur1, A. Saada2, I. Raz1, S. Weksler-Zangen1; 1Department of Internal Medicine, Diabetes Unit, 2Department of Human Genetics and Metabolic Diseases, Hadassah University Hospital, Jerusalem, Israel. Background and aims: The Cohen diabetic rat model consists of two strains: the Cohen diabetic sensitive rat (CDs) developing hyperglycemia and de- creased glucose stimulated insulin secretion (GSIS) on the diabetogenic- high-sucrose diet (HSD) and the Cohen diabetic resistant rat (CDr) which remains normoglycemic on HSD. Although underlying mechanisms causing inhibited GSIS in CDs remains yet elusive, we recently showed reduced ac- tivity of the mitochondrial respiratory-chain enzyme, cytochrome c oxidase (COX) and ATP content in islets of hyperglycemic-CDs, tightly linking in- hibited GSIS to COX deficiency. We examined whether CDs has a ubiquitous COX deficiency, thus suggesting a model of mitochondrial disorder related diabetes. Materials and methods: CDs and CDr males, eight weeks old were fed a regular-diet (RD) or a diabetogenic-diet containing 72% sucrose for 30 days. Activity of COX normalized to citrate-synthase (CS) was determined in pan- creatic-islets, lymphocytes, pancreas, heart and liver homogenates spectro- photometrically. Expression levels of COX subunits 1 & 2 (COX1 & COX2) and their chaperons Sco1, COX11 and COX17 were examined in RNA ex- tracted from islets by qRT-PCR and expressed as a fold-change of the expres- sion in CDr fed RD. Mitochondrial DNA (mtDNA) content was analyzed in DNA from islets by qRT-PCR. COX1 protein levels were determined in islets by Western-blot analysis. Results: An initial significant decrease in COX activity was observed in islets, lymphocytes, pancreas, heart and liver of normoglycemic-CDs demonstrat- ing 85, 40, 30, 40 and 30% respective reduction (p S 296 1 C mitochondrial fission factor (Mff) and the dynamin related protein 1 (Drp1) control fission. Recent studies indicate that obesity deteriorates mitochondri- al dynamics, and thus, mitochondrial function. In this study we investigated regulation of mitochondrial dynamics in liver, muscle and adipose tissue of ob/ob mice after calorie-restricted feeding compared to ad-libitum feeding. Materials and methods: Three month old male obese ob/ob (B6.V-lebob/ob) mice were fed for twelve weeks with a calorie-restricted diet, whereas the con- trol group had unrestricted access to feed. Age matched ob/+ (B6.V-lebob/+) mice served as lean controls. Finally mice were killed and tissues were ana- lyzed for Mfn1/2, Opa1, Fis1, Mff and Drp1 expression by quantitative PCR analyses. Fis1 protein expression was investigated using western blot and im- munofluorescence analyses. The mitochondrial network was visualized by Mito Tracker DeepRed staining and fat accumulation by Red Oil staining. Results: At the age of three month ob/ob mice showed a significantly higher body weight than ob/+ mice (46±2 vs. 31±1 g). After twelve weeks of calorie- restricted diet obese ob/ob mice showed a significantly lower body weight compared to ad-libitum feed mice (39±2 vs. 62±2 g). In addition a lower level of fat accumulation in liver and muscle was observed in ob/ob mice after calorie-restricted diet. Gene expression of all fusion and fission proteins was lower in liver, muscle and adipose tissue of ob/ob mice compared to lean con- trol mice, indicating reduction of mitochondrial dynamics. Calorie-restricted diet resulted in ob/ob mice in a significantly higher expression of all genes involved in mitochondrial fission and fusion compared to ad-libitum feed- ing. The strong increase of Fis1 expression in calorie-restricted feed ob/ob mice could be confirmed in liver, muscle and adipose tissue on the protein level. Obese ob/ob mice showed formation of mitochondrial cluster in liver and fatty tissue whereas a more homogenous network of mitochondria was observed in diet feed ob/ob mice. Conclusion: We observed a correlation between reduced body weight and lower fat accumulation and improved mitochondrial dynamics in liver, mus- cle and adipose tissue of ob/ob mice. Thus, a calorie-restricted diet has a ben- eficial effect on mitochondrial morphology in obese subjects. Mitochondrial function in liver, muscle and adipose tissue is significant to maintain insulin sensitivity and glucose tolerance. Our study provides further evidence that obesity induced mitochondrial dysfunction contributes to the development of type 2 diabetes. 727 An mtDNA mutation in the cytochrome c oxidase changes mitochondrial dynamics and favours development of type 2 diabetes mellitus J. Niemann, C. Johne, J. Schultz, M. Tiedge, S. Baltrusch; Institute of Medical Biochemistry and Molecular Biology, University Rostock, Germany. Background and aims: Mutations in the mitochondrial genome (mtDNA) affect subunits of respiratory chain complexes and can impair their func- tion. It has been postulated that mtDNA mutations which accumulate with age contribute to the pathogenesis of type 2 diabetes mellitus (T2DM). We have shown that the conplastic mouse strain C57BL/6NTac-mtNOD/AtJ (mtNOD) with a mtDNA point mutation in the subunit 3 of cytochrome c oxidase (OXPHOS complex IV) has a higher ROS production with age com- pared to controls. In this study we investigated mitochondrial morphology and mitochondrial fission-fusion processes in hepatocytes of this conplastic mouse strain. Materials and methods: Primary hepatocytes were isolated from 3, 6, 9 and 12 month old mtNOD and C57BL/6NTac (control) mice. Hepatocytes were stained with MitoTrackerGreen. Thereafter mitochondrial morphology was investigated by fluorescence microscopy and the median mitochondrial elon- gation level was calculated using AutoQuant software. Gene expression of the mitochondrial fusion proteins OPA1, MFN1 and MFN2 and of the fission proteins FIS1 and DNM1L was determined by quantitative PCR analyses. Results: The median mitochondrial elongation level in 3 month old hepato- cytes of mtNOD mice was significantly lower compared to control mice. The median mitochondrial elongation level decreased with age in hepatocytes of control mice. In contrast, an increase of the median mitochondrial elongation level was observed in hepatocytes of mtNOD mice. Gene expression of the mitofusin MFN1 and MFN2 declined during ageing in control mice, whereas the expression of both genes increased in mtNOD mice. At the age of 12 month the expression was in mtNOD mice twice as high as in control mice. In addition, the expression of the fission protein DNM1L was significantly lower in 12 months old mtNOD compared to control mice. Conclusion: The observed mitochondrial phenotype correlated with the gene expression profile of mitochondrial fission and fusion genes at the investigated points of age. Our results indicate that a mtDNA mutation in the cytochrome c oxidase triggers a higher mitochondrial elongation level of hepatocytes with age. This process could be mediated by an interaction between the higher ROS production and the up-regulation of MFN1 and MFN2 in aged mtNOD mice. Thus, the C57BL/6NTac-mtNOD/AtJ mouse strain represents an interesting model to study trigger effects of age depend- ent mitochondrial phenotypes in the pathogenesis of T2DM. Supported by: Federal Ministry of Education and Research (BMBF) 728 Dual effect of Nd2 gene mutation in complex I of the respiratory chain: ROS generation and induction of mitoprotective effects in the course of ageing M. Wietzke, E. Meyer, F. Koch, J. Niemann, H. Weiss, S. Baltrusch, M. Tiedge; Institute of Medical Biochemistry and Molecular Biology, Rostock, Germany. Background and aims: The mitochondrial genome of the diabetes-resistant B6-mtALR mouse strain is characterized by a mt-DNA encoded Nd2 poly- morphism in complex I of the respiratory chain. In metabolically active or- gans such as the liver disorders of the mitochondrial electron transfer with increased ROS generation are linked to accelerated aging. It was the aim of the study to investigate the production of reactive oxygen species (ROS), the expression of antioxidant enzymes and the expression of the aging marker p53 in liver tissue of conplastic B6-mtALR mice compared to B6-mtAKR control strain. Materials and methods: Conplastic B6-mtAKR (AKR) and B6-mtALR (ALR) mice were examined over a period of 2 years. AKR and ALR animals were killed and liver tissue was collected at the age of 3, 6, 9, 12, 18 and 24 months. The gene expression of Nd2, Sod1, Sod2, catalase, UCP2, TFAM and mTOR was quantified by real-time RT-PCR and the p53 protein expression was quantified by of Western blot analyses. The measurement of mitochon- drial ROS production in the liver was performed by fluorescence microscopy after injection of the ROS sensor Mitosox. Results: The gene expression of mt-Nd2 in the ALR strain showed no differ- ences compared to the control strain AKR. The ROS production in the liver at the ages of 6 and 12 months was significantly increased in AKR mice (2- 18-fold, p S 297 1 C of p66Shc in human adipose tissue and the characteristics of visceral adipose (VAT) and subcutaneous adipose tissue (SAT) in p66Shc-/-Ob/Ob mice at 18 and 30 weeks of age. Materials and methods: 4 different mouse models were used: C57/BL6 wt mice, Ob/Ob mice, p66 -/- mice, p66 -/- Ob/Ob mice. 2 time points: 18 and 30 weeks of age. Both visceral (VAT) and sub-cutaneous (SAT) white adipose tissue samples were analysed. Results: p66Shc-/-Ob/Ob mice develop obesity, but at 30 weeks, their body weight was significantly lower than in Ob/Ob controls. The size of visceral adipocytes was reduced in p66Shc-/- mice at 18 weeks, but not significantly different compared to control Ob/Ob mice at 30 weeks. The number of apop- totic cells in VAT was increased in Ob/Ob mice compared to wt, but signifi- cantly reduced by p66Shc deletion. The number of macrophages in VAT was also reduced in p66Shc-/-Ob/Ob compared to Ob/Ob mice. Based on ITT and GTT, p66Shc-/-Ob/Ob mice at 18 weeks are less glucose intolerant and insulin resistant compared to Ob/Ob mice, but become insulin resistant at 30 weeks of age. In a sample of 46 individuals undergoing bariatric surgery, the expression of p66Shc linearly correlates with BMI in VAT, but inversely cor- relates in SAT, only in young subjects. Conclusion: These data support the hypothesis that p66Shc modulates the processes occurring during the development of obesity and metabolic disor- ders, affecting adipogenesis and inflammation of visceral adipose tissue. The role of p66Shc appears to be influenced by age, most likely in relation to the role of this protein during aging. 730 Regulation of skeletal muscle lipolysis and oxidative metabolism by G0/G1 switch gene 2 C. Laurens, P.-M. Badin, K. Louche, D. Langin, V. Bourlier, C. Moro; Institute of Metabolic and Cardiovascular Diseases, Inserm UMR1048, Toulouse, France. Background and aims: Intramyocellular triacylglycerol (IMTG) pool is an important energy source in skeletal muscle, especially during exercise. Recent data suggest that the adipose triglyceride lipase (ATGL) plays a key role in providing energy substrate from IMTG and perturbations of its expression/ activity are linked to altered energy metabolism in skeletal muscle. However, little is known about its regulation. The aim of this study was to investigate the role of the protein G0/G1 Switch Gene 2 (G0S2), recently described as an inhibitor of ATGL in white adipose tissue, in the regulation of lipolysis and oxidative metabolism in skeletal muscle. Materials and methods: G0S2 was overexpressed and knockdowned in hu- man primary myotubes obtained from rectus abdominis biopsies in healthy volunteers and transduced with adenoviruses and lentiviruses, respectively. ATGL activity, IMTG content, fatty acid mobilization and oxidation, glucose oxidation and glycogen synthesis were assessed by using radiolabeled sub- strates. Metabolic studies were supplemented with the investigation of mito- chondrial phenotype using fluorescent probes and RT-qPCR (for the analysis of key gene expression). Results: Our results show that G0S2 is expressed in human myotubes and inhibits ATGL activity (-47% ; P S 298 1 C PS 053 Metabolic effects of dietary composition, caloric restriction and bariatric surgery 732 Associations of dietary fibre intake and risk factors of the metabolic syndrome. The lipid and glucose under prospective surveillance (LUPS) study J. Haas1, H. Pinnschmidt2, M. Toeller1, A. Partenheimer1, K. Wegscheider2, D. Mueller-Wieland1; 1Department of General Internal Medicine and Institute for Diabetes Research, Asklepios Clinic St. Georg, Asklepios Medical School, Faculty of Medicine of Semmelweis Medical University, 2Department of Medical Biometry and Epidemiology, University Medical Center, Hamburg, Germany. Background and aims: The metabolic syndrome is classified by a clustering of metabolic and cardiovascular risk factors, in particular, elevated waist cir- cumference, serum triglycerides, blood pressure, fasting glucose and reduced HDL-cholesterol. The exact etiology remains unclear, but it is known to be a complex interaction between genetic, metabolic, and environmental factors. Amongst exogenous factors nutrition has a central role in the development of the metabolic syndrome, and dietary fibre intake is supposed to be an im- portant factor in this respect. Therefore we aimed to evaluate associations of dietary fibre intake and the prevalence of risk factors of the metabolic syn- drome in the LUPS cohort. Materials and methods: Baseline data from the prospective LUPS cohort (1887 healthy workers, 1274 men and 613 women, aged 25-60 years) were investigated for associations between dietary fibre intake and risk factors of the metabolic syndrome (classification criteria) using logistic regression and Jonckheere-Terpstra test. Fibre intake was assessed by a standardized semi- quantitative food frequency questionnaire with 85 items covering typical food choices and portion sizes in Germany. Results: Mean dietary fibre intake (g/1000 Kcal/day) was 11.9 (95% CI: 11.6; 12.3) in women and 9.9 (95% CI: 9.7; 10.1) men. The highest prevalence of risk factors for the metabolic syndrome was seen for increased blood pressure in 70% of men and 40% of women. The prevalence of the other risk factors of the metabolic syndrome in the cohort decreased from elevated waist cir- cumference (28%), high triglycerides (20%), reduced HDL-cholesterol (19%) to increased fasting blood glucose (10%). Only elevated waist circumference showed a higher prevalence in women (37%) than in men (25%). Significant negative associations were found between fibre intake and waist circumfer- ence (-6,36 standardized JT-statistics with continuous variable, odds ratio 0,952 for grouped variable; waist circumference for men ≥ 102 cm and ≥ 88 cm for women) as well as for fibre intake and fasting triglyceride levels (-5,57 standardized JT-statistics, odds ratio 0,927; triglycerides ≥ 150 mg/dl), while positive associations were seen between fibre intake and HDL-cholesterol (8,66 standardized JT-statistics, odds ratio 0,948 ; HDL-cholesterol < 40 mg/ dl for men and < 50 mg/dl for women) (all p 0.41 ; TLR4: pBH > 0.31). Conclusion: Although extensive adaptation occurs in our healthy partici- pants after 6 weeks on a high-fat diet, parts of the metabolism remain dis- turbed. The decreasing lysophospholipids seem to accompany the activation of the immune system as well as higher inflammatory processes. Our analysis Diabetologia (2014) 57:[Suppl1]S1–S564 S 299 1 C shows novel pathways by which saturated fats increase inflammation inde- pendent of the known TLR4-related mechanisms of inflammation. Clinical Trial Registration Number: NCT01631123 Supported by: BMBF, DZD 735 The effect of meal frequency on fatty acid composition in serum phospholipids in patients with type 2 diabetes H. Kahleova1, H. Malinska1, L. Kazdova1, L. Belinova1, M. Hill2, T. Pelikanova1; 1Diabetes Centre, Institute for Clinical and Experimental Medicine, 2Institute of Endocrinology, Prague, Czech Republic. Background and aims: Fatty acids are important cellular constituents that may affect many metabolic processes relevant for the development of diabetes and its complications. We demonstrated previously the superior effect of two meals a day, breakfast and lunch (B2) on body weight, hepatic fat content and insulin sensitivity compared to the same diet divided into six smaller meals a day (A6). The aim of this secondary analysis was to explore the effect of frequency of meals on the fatty acid composition of serum phospholipids in subjects with type 2 diabetes (T2D). Materials and methods: In a randomised, crossover study, we assigned 54 patients with T2D to follow two regimens of a hypocaloric diet (-500 kcal/ day), each for 12 weeks: six meals (A6), and two meals a day, breakfast and lunch (B2). The diet in both regimens had the same macronutrient and en- ergy content. The procedures were performed at weeks 0, 12 and 24. The fatty acid composition of serum phospholipids was measured by gas liquid chro- matography. Insulin sensitivity was derived as an oral glucose insulin sensi- tivity (OGIS) index. Results: Saturated fatty acids (mainly the myristic and palmitic acids) de- creased (p S 300 1 C 738 The effects of bariatric surgery on pancreatic lipid metabolism in morbidly obese subjects H. Honka1, J. Koffert1, J.C. Hannukainen1, J.J. Tuulari1, H.K. Karlsson1, V. Oikonen1, P. Salminen2, M. Soinio3, A. Mari4, P. Iozzo5, P. Nuutila1,3; 1Turku PET Centre, University of Turku, 2Division of Digestive Surgery and Urology, Turku University Hospital, 3Department of Endocrinology, Turku University Hospital, Finland, 4Institute of Biomedical Engineering, National Research Council, Padua, 5Institute of Clinical Physiology, National Research Council, Pisa, Italy. Background and aims: As a treatment modality for morbid obesity, bari- atric surgery leads to a sustained weight loss and a high remission rate of type 2 diabetes. However, the exact mechanisms behind the improved glu- cose tolerance are not fully elucidated. In the present study we investigated whether bariatric surgery is able to improve pancreatic lipid metabolism in obese subjects. Materials and methods: Totally, 24 morbidly obese subjects (BMI 41±4.0 kg/ m2, age 42±9.5 years) eligible for bariatric surgery, of which ten had type 2 diabetes, and 15 healthy age-matched subjects were recruited in the study. Pancreatic fatty acid (FA) uptake was measured at fasting state using PET/ CT and a palmitate analogue [18F]FTHA. Pancreatic total volume, and fat and parenchymal volumes were measured using a CT based approach as recently described. Oral glucose tolerance test (OGTT) was performed to quantitate glucose tolerance and empirical/model-derived β-cell function parameters. Obese subjects were studied preoperatively and six months after bariatric surgery procedure (either Roux-en-Y gastric bypass or sleeve gastrectomy). Results: Before surgery, obese subjects had higher rate of pancreatic FA up- take (1.4±0.6 versus 0.7±0.3 µmol/min, P < 0.001) and fat percent (19±22 versus 4.0±4.0 %, P < 0.01), and had greater amount of ectopic fat (17±21 versus 3.0±3.1 ml, P < 0.01) than healthy controls. After bariatric surgery, obese subjects lost 24% of their weight (P < 0.0001), but were still obese (BMI 32±4.2 kg/m2). Fasting plasma glucose and HbA1c were lower (both by 12%, P < 0.0001), whereas no difference was found in plasma FFA levels (0.82±0.22 versus 0.76±0.18 mmol/l, NS). In subjects with type 2 diabetes preoperatively, parameters of β-cell function (insulinogenic index, glucose sensitivity, rate sensitivity) were improved (all P < 0.02), and in eight out of ten diabetes was in remission. In the obese group, pancreatic FA uptake (1.4±0.6 versus 1.0±0.4 µmol/min, P < 0.01), fat percent (19±22 versus 14±13 %, P < 0.01), and ectopic fat volume (17±21 versus 11±12 ml, P < 0.01) were decreased, whereas parenchymal volume was unchanged (79±29 versus 72±22 ml, NS) after operation. Diminishment of ectopic fat was independent of weight loss and FA uptake (NS). Decrease in total pancreatic volume was associated with improved glucose tolerance, as measured with mean glucose during OGTT (r = 0.59, P < 0.005), and insulinogenic index (r = -0.45, P = 0.04). Conclusion: The present study shows that, pancreatic lipid accumulation in obesity is reversible and modified by bariatric surgery. The results of the study imply that decreased pancreatic FA metabolism and loss of ectopic fat play a role in the improved glucose tolerance seen after bariatric surgery. Clinical Trial Registration Number: NCT01373892 Supported by: Academy of Finland, Finnish Cultural Foundation 739 The metabolic signature of RYGB is similar to that of equivalent caloric restriction I. Lingvay1, L. Golici1, M. Poduri1, L.K. Mitchell1, A. Coster2, T. Zhao3; 1Endocrinology, 2Pharmacology, 3Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, USA. Background and aims: Diabetes improves or remits within days following gastric by-pass surgery. It is debated whether this dramatic effect is due to the strict caloric restrictions imposed in the peri-operative period or due to surgery-specific changes in metabolism. Materials and methods: Ten patients with type 2 diabetes underwent a strictly-observed 9-day inpatient caloric restriction intervention. Following re-equilibration to baseline characteristics (average 3.3 months), they were readmitted for a second identical intervention (identical daily oral intake) during which they also underwent gastric bypass. All measurements (6-hrs Mixed Meal Challenge Test) were performed before and after each study peri- od (total of four measurements/patient). This precise study design allowed us to separate the effects of diet (first period) from the effects of surgery (second period-first period) on the hormonal changes observed after surgery. Paired t-test was used to compare changes observed during each period. Results: Patients were 54.1+/-8.7 years old, 3 males, 70% minorities, BMI 50.8+/-9 kg/m2, with advanced diabetes (9.2+/-8 years duration, treated with 1.5 oral agents, additionally 7 were also treated with insulin, median dose 0.8 u/kg). There was comparable and significant improvement in glycaemia, insulin sensitivity and beta-cell function. The diet only period fared better in regards to changes in ghrelin, GIP, and hepatic glucose output. Surgery induced large increases in GLP-1 and PYY (figure). Conclusion: The severe caloric restriction employed immediately post gas- tric bypass surgery is primarily responsible for the rapid improvement in glycaemia. The increase in incretin hormones is specific to surgery, but does not seem to have a direct effect on glycaemia, insulin resistance, nor beta- cell function. Their effect, if any, is likely mediated through central mecha- nisms which control appetite and promote the long term adherence to such restricted caloric intake. Clinical Trial Registration Number: NCT01153516 Supported by: NIH 3UL1RR024982-05S1, K23RR024470 740 Immediate post-operative effects after a mixed-meal test and long-term effects on hormone expression in the alimentary limb in gastric-bypassed patients A. Lindqvist1, J. Berggren2, B. Nergård3, L. Groop1, J. Hedenbro4, N. Wierup1; 1Clinical Sciences, Malmö, 2Department of Surgery, Kalmar, 3Aleris Obesity, Lund, 4Department of Surgery, Lund, Sweden. Background and aims: Gastric bypass surgery (GBP) is the most effective treatment for morbid obesity. Also, GBP results in rapid improvement in gly- caemia long before significant weight loss. In the present study we aimed to 1) investigate the influence of pre-surgery caloric restriction with that of the immediate effect of GBP on glycemia and 2) assess long-term effects of GBP on gut hormone expression in the alimentary limb. Materials and methods: Obese women (47±1 years) were subjected to a mixed meal test (MMT) four weeks before (MMT-4w), 4h before (MMT- 4h) and one day after (MMT+1d) Roux-en-Y gastric bypass surgery. MMT- 4w was performed before initiation of the low-calorie diet regimen and MMT+1d constituted the first meal the patients received after surgery. Blood was collected at -10, -5, 0, 5, 10, 15, 30, 60 and 90 minutes upon initiation of MMT-ingestion. Glucose, insulin, glucose-dependent insulinotropic pep- tide (GIP) and active glucagon-like peptide 1 (GLP-1) were analyzed for the MMTs. In another study, female patients were recruited and biopsies from the alimentary limb were obtained during surgery and via gastroscopy 12 months after GBP. Samples were analyzed for all major enteroendocrine cell populations using immunocytochemistry and morphometry. Results: While there was difference in glucose levels between the vari- ous MMTs, the insulin response was markedly increased immediately at MMT+1d, compared to both the MMT-4w and MMT-4h (2.4-fold and 2.8- fold, respectively). Active GLP-1 levels were similar in all MMTs, whereas the GIP-response was higher at the MMT+1d, compared to MMT-4w and MMT- 4h (1.6-fold and 1.4-fold, respectively). The elevation in insulin and GIP re- sulted in increased insulin AUC-to-GIP AUC-ratio (p S 301 1 C design, GIP is a strong contributor to the enhanced insulin secretion. Fur- thermore, we report for the first time the effects of GBP on expression of gut hormones in the alimentary limb one year after GBP. Our data showing that GLP-1 and serotonin immunoreactive cells are markedly increased in the alimentary limb 12 months after GBP provide anatomical prerequisites for sustained hormonal changes after GBP. Supported by: ERC, VR, ALF 741 Searching for biomarkers determining the early efficacy of bariatric surgery in the treatment of type 2 diabetes mellitus M. Ciborowski1, P. Samczuk1, M. Luba2, A. Citko1, J. Godzien3, H. Razak Hady2, J. Dadan2, C. Barbas3, M. Gorska4, A. Kretowski4,1; 1Clinical Research Centre, Medical University of Bialystok, 21st Clinical Department of General and Endocrine Surgery, Medical University of Bialystok, Poland, 3Center for Metabolomics and Bioanalysis (CEMBIO), Universidad CEU San Pablo, Madrid, Spain, 4Department of Endocrinology, Diabetology and Internal Medicine, Medical University of Bialystok, Poland. Background and aims: Bariatric surgery is an effective treatment for obesity and in 80-90% of subjects with type 2 diabetes improves significantly glycae- mic control, independently of weight loss. Recent studies suggest different mechanisms that affects metabolic changes after surgery: incl. alterations in gut hormones secretion (ex. GLP-1), neurohumoral pathways modulation, changes in bile acid concentrations in the blood and/or altered intestinal microbiota. The recovery rate of the T2DM is also related to the method of bariatric surgery with the higher T2DM remission rate observed for a Roux- en-Y gastric bypass (RYBG) in comparison to a laparoscopic adjustable gas- tric banding (LAGB). The aim of this study was to search for the preoperative and postoperative metabolic biomarkers, which could determine the efficacy of the bariatric surgery and predict an early remission failure in the treatment of T2DM. Materials and methods: The study was performed in the group of 49 obese patients with T2DM, who underwent bariatric procedures (laparoscopic sleeve gastrectomy - LSG or Roux-Y gastric bypass - RYGB) and were evalu- ated for an early remission failure (1 month after the surgery) of T2DM. The fasting serum samples were fingerprinted by LC-QTOF-MS and data were collected in ESI (+) mode (50-1,000 m/z). Chromatograms were aligned and quality assurance of obtained data was performed. Statistical analysis was conducted to compare patients before and one month after the surgery (paired t-test), and to compare (before the surgery) patients who recovered from T2DM in one month post-surgery with those who did not. Patients be- fore the surgery were compared by use of t-test or Mann-Whitney test de- pending on the normality of variables distribution. Identification of signifi- cant features was performed based on custom library with accurate mass and retention time of more than 100 metabolites. Results: We found that before bariatric surgery lower serum levels of lipid compounds (incl. lysophosphatidylcholines, phosphatidylethanolamines, sphingomielynes, ceramides) and lower vitamin D3 metabolites concentra- tions significantly predicted an early T2DM remission failure (p S 302 1 C associated with higher insulin sensitivity reflected by reduced HOMA-IR and increased IGFBP-1. Conclusion: Adv36 infection associated with lower occurrence of type 2 dia- betes and better insulin sensitivity in adults, particularly among females. Supported by: VINNOVA, Sw. Res. Council, FAS, Sw. Diab. Ass., E Persson Found., Obetech Res. Center 744 AKT/JNK switch induces beta cell apoptosis as part of TLR3 signalling during Coxsackievirus infection N. Busse1, F. Paroni1, E. Domsgen2, J. Kerr-Conte3, A. Dotzauer1, K. Maedler1; 1Center for Biomelecular Interactions, University of Bremen, Germany, 2Center for Infection Medicine, Karolinska Institutet, Stockholm, Sweden, 3University of Lille, France. Background and aims: Type 1 diabetes results from a sophisticated interplay between genetic factors, immune system and environmental factors. Among environmental factors, viruses, mainly enteroviruses such as members of the Coxsackie B virus family, have been suspected since long time to trigger dia- betes. Intra-islet viral particles have been detected in pancreata from patients with T1DM and T2DM, and known to cause β-cell dysfunction in vitro, suggesting that a direct infection contributes to onset of diabetes. Nonethe- less the mechanisms of a correlation between virus infection and diabetes progression are poorly understood. Clarification of the viral triggered β-cell dysfunction and apoptosis may provide an indication to develop preventive therapies. We previously identified Toll-like receptor 3 (TLR3) and Protein kinase R (PKR) as first defensive defensive line during coxsackie virus infec- tion. In this study we asked the question of what is the contribution of PKR and TLR3 in β-cell apoptosis during CVB3 infection. Materials and methods: Isolated human islets and CM cell line were infected with coxsakievirus serotype B3. Replication of CVB was confirmed by im- munostaining of viral protein 1 (VP1) and titration of islet lysate. shRNA and siRNA were used to silence TLR3 and PKR in both human islets and CM cells. Plasmids encoding for TLR3 and TBK1 were used for overexpres- sion experiments. Either dominant negative expression constructs or specific inhibitors were used to study the role of JNK and AKT in human islets. Islet protein expression and phosphorylation were analyzed by western blot. Bind- ing of CVB ssRNA and dsRNA was investigated by immunoprecipitation-RT- PCR coupled assays. Results: A time course analysis of CVB3 infection for downstream signaling showed AKT activation at 1h post-infection, which lasts for several hours. AKT was down-regulated at 24h post infection, when activation of pJNK, Caspase-3 and VP1 appeared in human islets. A possible correlation between AKT activation and PKR/TLR3 signaling was investigated. Knock down of PKR resulted in increased virus infection (VP1), JNK phosphorylation and activation of caspase 3, while AKT phosphorylation was PKR-independent. Lack of TLR3 had a pro-survival effect in CVB3 infected human islets leading to a reduced reduced JNK phosphorylation, caspase activation and VP1 as well as TBK1 phosphorylation. In contrast, overexpression of TLR3 enhanced AKT activation together with VP1 induction. Overexpression of TLR3/TBK1 further enhanced phosphorylation of AKT, viral replication and caspase ac- tivation suggesting pivotal role of AKT, enhanced by TLR3. AKT inhibition reduced viral replication, activation of caspase 3 and phosphorylation of JNK. Overexpression of DN-JNK led to a decrease in apoptosis indicated by the decreased level of cleaved caspase 3 but not of VP1. Treatment of human is- lets with the JNK inhibitor confirmed the protective effects of JNK inhibition. Conclusion: In summary, our data show that CBV infections have a direct deleterious effect on β-cell survival, resulting from virus-induced apoptosis, and potentiated by the AKT/TLR3 signaling pathway. With our present data we provide novel targets towards protecting the β-cell during virus infection. Supported by: JDRF 745 DPP-4 inhibitors: anti-inflammatory effects on M1 activation E. Derlindati, V. Curella, I. Zavaroni, R.C. Bonadonna; Clinical and Experimental Medicine, University of Parma, Italy. Background and aims: Macrophages are a heterogeneous cell population which, in response to the cytokine milieu, differentiate in either classically activated macrophages (M1) or alternatively activated macrophages (M2). This plasticity is essential to regulate inflammation, immune response and tissue remodeling and it may be a novel therapeutic target in inflammatory- based diseases such as atherosclerosis. Some recent evidence suggests that dipeptidyl peptidase 4 may play a role in in cardiovascular diseases and that its inhibition might be beneficial in curbing inflammation and atherogenesis. The aim of this study was to evaluate the anti-inflammatory effects of a DPP-4 inhibitor on classical macrophage activation . Materials and methods: Monocytes were isolated from healthy donors’ buffy coats. After differentiating into macrophages, cells were stimulated with IFNγ and LPS in order to induce M1 activation, in presence or in absence of 2µM of DPP-4 inhibitor (KR-62436, Sigma Aldrich) for 24 hours. The working con- centration of DPP-4 inhibitor was selected after dose finding experiments. Resting macrophages (RM) -no added stimuli- were used as a control. The expression of DPP-4 was evaluated in each experimental conditions, both by real time PCR and cytofluorimetric assay. Key pro-inflammatory genes and chemokines (IL6, IL8, TNF-alpha, MCP-1, COX2, SOD2, SOCS1) were evaluated by real time PCR. Protein expression of IL-6 and TNF-alpha were evaluated by ELISA. Results were obtained by means of six biological repli- cates. DPP-4 activity is presently being assessed. Results: M1 macrophages showed a significant up-regulation of DPP-4 mRNA compared to RM (p=0.015), which was unchanged by DPP-4 inhibi- tion. These data were confirmed by cytofluorimetric analysis. M1 significant- ly up-regulated also the expression of pro-inflammatory genes SOCS1, IL-6, COX2 (by one hundred-fold), IL8, TNF-alpha (by fifty-fold), and MCP1, SOD2 (by ten-fold) compared to RM (p S 303 1 C [7.63, 34.18], p≤0.01) , however this difference was abolished after HBOT (6.24 μM [5.08, 6.71] vs. 9.68 μM [6.47, 19.24], NS.). Leukocyte tyrosine nitration was elevated due to diabetes with (23.78±3.43% vs. 41.2±3.49%, p≤0.01) or without HBOT (20.44±4.08% vs. 38.1±4.12%, p≤0.01). The plas- ma levels of pro-inflammatory cytokines were increased in untreated diabe- tes (CINC-1: 69.1 ng/L [63.85, 74.63] vs. 137.1 ng/L [66.71, 343.8], p≤0.01; LIX: 282 ng/L [141.9, 382] vs. 430.8 ng/L [327, 604.4], p≤0.05), however due to HBOT this difference was ceased (CINC-1: 66.96 ng/L [61.3, 83.5] vs. 96.9 ng/L [70.73, 147.5], NS.; LIX: 245.3 ng/L [172.6, 340.3] vs. 346.8 ng/L [276.9, 774.6], NS.). The anti-inflammatory factor TIMP-1 was not altered by dia- betes in the lack of HBOT (24.49 μg/L [21.48, 26.41] vs. 27.79 μg/L [22.18, 29.73], NS.), on the other hand after HBOT the plasma level of TIMP-1 was significantly higher in diabetic animals (20.76 μg/L [18.43, 25.62] vs. 32.31 μg/L [21.65, 38.43], p≤0.01). Conclusion: According to our results two weeks after HBOT, the oxidative stress and pro-inflammatory cytokine production induced by experimental diabetes were abolished. Also the anti-inflammatory factor TIMP-1 is elevat- ed as a response to HBOT in diabetic animals. On the other hand nitrative stress is not altered by the therapy. These findings suggest that HBOT is not just safe to use in diabetic patients, but may also have beneficial effect on their chronic subclinical inflammation. Supported by: Bolyai János Research Fellowship of the Hungarian Academy of Sciences 747 The anti-inflammatory protein NUPR1 (p8) generally enhances viability of pancreatic islets but does not prevent lipotoxicity B. Mihic-Necic1, A.E. Mehana1,2, C. Liu1, L. Parsons1,3, J. Straetener1, J. Baumann1, N. Perakakis1, K. Laubner1, J. Seufert1, G. Päth1; 1Department of Internal Medicine II, Division of Endocrinology and Diabetology, University Hospital of Freiburg, 2University of Freiburg, Faculty of Biology, Germany, 3School of Biosciences, Cardiff University, UK. Background and aims: Type 2 diabetes is closely related to obesity. Accord- ing to the so-called “nutrition overflow hypothesis”, type 2 diabetes results from nutrition-induced local tissue inflammation leading to insulin resist- ance, insulin secretory dysfunction and finally pancreatic beta cell loss. Free fatty acids acting through toll-like receptors are key mediators of this lipotox- ic inflammatory response. (1) We previously demonstrated that cocultured human bone marrow-derived mesenchymal stromal cells (hMSC-TERT) protect rat INS-1E beta cells from alloxan- and streptozotocin-induced in- jury. (2) We further demonstrated that mice with beta cell-specific overex- pression of the intracellular protein NUPR1 maintained insulin secretion and storage during high fat diet-induced local low grade inflammation and intense inflammatory stress by insulitis. Here, we investigated the anti-lipo- toxic potential of (1) cocultured hMSC-TERT and (2) beta cell-specific Nupr1 overexpression. Materials and methods: For coculture, INS-1E were seeded in wells and hMSC-TERT were seeded in inserts with 1 µm pores to allow soluble factors but not cells to pass the membrane. Primary mouse islets were obtained from transgenic (Tg) mice with beta cell-specific Nupr1 overexpression under the control of the rat insulin gene 1 promoter (RIP1). Non-transgenic littermates served as WT controls. Lipotoxicity was induced by palmitate. Viability of cells and islets was measured by MTS assay. Results: We first evaluated by kill curve experiments that 24 h exposure to 0.13 mM palmitate inhibits the viability of INS-1E by 50%. Experiments were performed with 0.1 mM palmitate. (1) 24 h exposure to 0.1 mM palmitate reduced viability of INS-1E by 42%. Cocultured hMSC-TERT could not pre- vent this loss of viability. (2) Tg islets with ectopic NUPR1 demonstrated 1.5- fold enhanced viability compared to WT controls. However, ectopic NUPR1 did not reduce lipotoxicity since 24 h exposure to 0.1 mM palmitate reduced viability by 50% in both WT and Tg islets. Conclusion: Despite their protective effects in other experimental settings, cocultured hMSC-TERT and also ectopic Nupr1 did not prevent lipotoxicity. Ectopic NUPR1 overproduction generally enhances viability of Tg islets and thereby contributes to enhanced beta cell function during inflammatory tis- sue stress in Tg mice compared to WT controls. Supported by: BMBF, FNS 748 Protection of pancreatic beta cells in vitro and in vivo by NUPR1 (p8) protein during inflammatory diabetogenic stresses A.E. Mehana1,2, I. Pilz1, B. Dufner1, C. Jäger1, S. Sojka1, J. Baumann1, M. Alt1, C. Liu1, L. Parsons1,3, K. Laubner1, N. Perakakis1, G. Päth1, J. Seufert1; 1Division of Endocrinology and Diabetology, Department of Internal Medicine II, University Hospital of Freiburg, 2University of Freiburg, Faculty of Biology, Germany, 3Cardiff University, School of Biosciences, UK. Background and aims: Type 2 diabetes is closely associated with obesity. Chronically enhanced circulating free fatty acids activate toll-like receptors and induce local inflammation leading to insulin resistance, insulin secre- tory dysfunction and finally pancreatic β-cell loss. However, many obese subjects do not develop diabetes. This observation suggests the existence of molecular defence systems that prevent β-cell degradation in non-susceptive individuals. Therefore, identification of molecular mechanisms that provide anti-inflammatory protection and control β-cell turnover are of key interest. The intracellular protein NUPR1 has been shown to reduce tissue damage during acute pancreatitis in Nupr1-/- mice. The protein is also expressed in the endocrine pancreas. To investigate its potential protective role and cel- lular mechanisms in the under diabetogenic stress, we generated transgenic mice with β-cell-specific Nupr1 overexpression under the control of the RIP1 promoter (Tg). Materials and methods: We induced low-grade inflammation by feeding mice with high fat diet (HFD) and/or insulitis by multiple low dose STZ in- jections. Glucose tolerance was evaluated by measuring non-fasting random blood glucose, intraperitoneal glucose tolerance test (ipGTT), intraperito- neal insulin tolerance test (ipITT) and serum insulin levels (ELISA). β-cell mass was determined morphologically. Islet inflammation was quantified by numbers of infiltrating CD45+ lymphocytes and NF-kB activation (Western blot). Ex vivo isolated islets from untreated wild type (Wt) and Tg animals and from animals after 20 weeks of feeding with HFD were cultured for 16 days without treatment and subsequently exposed for 24hrs to 10ng/ml IL-1β or 0.33mM STZ were evaluated by glucose stimulated insulin secretion and content (GSIS), cleaved PARP (Asp214) for apoptosis (ELISA) and BrdU for cell proliferation. Results: Immunohistochemically, we observed an islet enriched expression of NUPR1 in non-Tg mice. In vivo, Tg mice displayed improved glucose tol- erance, improved insulin secretion and increased β-cell mass during HFD and/or insulitis as compared to Wt controls. NUPR1-Tg islets also showed reduced lymphocyte infiltration and reduced NF-kB activation. Ex vivo, Tg islets displayed greatly reduced apoptosis while insulin secretion and content was significantly enhanced compared to Wt islets during 16 days of culture or in response to 24 h exposure to IL-1β or STZ. Further, Wt islets showed a massively reduced proliferation rate in response to diabetogenic injuries, whereas this was completely maintained in isolated NUPR1-Tg islets. Conclusion: NUPR1 protein may be an important molecular mediator in the stress defence system of pancreatic β-cells. It exhibits potent protection of insulin biosynthesis, secretion and content during inflammatory pancreatic β-cell stress by reducing activation of NF-kB and apoptosis and enhancing proliferation capacity as a possible mechanism to compensate for the diabe- togenic decrease in pancreatic β-cell mass. 749 Siglec F knockout impairs glucose stimulated insulin secretion in isolated mouse islets K. Stolz, G. Dharmadhikari, K. Maedler; Centre for Biomolecular Interactions, University of Bremen, Germany. Background and aims: Type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM) result from a decline in β-cell function and survival partly as a consequence of islet inflammation. The inflammatory signaling and the activation of immune cells are caused by secreted stimulators or via cell-cell interactions. A group of cell surface adhesion and signaling molecules, the Siglecs (sialic acid-binding immunoglobulin (Ig)-like lectins) mediate such interactions. In humans, Siglecs are differentially expressed in α-and β-cells and show protective effects on β-cell function and survival upon overexpres- sion. In this study we investigated, if the knockout (KO) of the functionally relevant paralog Siglec F shows an effect on β-cell function in mice. Materials and methods: To investigate the effect of the knockout isolated mouse islets were treated with diabetogenic conditions (2ng/ml IL-1β/1000 U/ml IFN-γ or 22.2mM Glucose/0.5mM Palmitate) for 3 days and subjected Diabetologia (2014) 57:[Suppl1]S1–S564 S 304 1 C Glucose stimulated Insulin secretion. Expression studies were carried out by FACS analysis and quantitative RT-PCR. For depletion of infiltrated mac- rophages isolated mouse islets were treated with Chlodronate containing liposomes for 48 hours. Results: Knockout of Siglec F shows no effect under diabetogenic treatment, but significant lower Insulin release upon Glucose stimulation under control conditions (p S 305 1 C PS 055 Inflammation in adipose tissue and muscle 752 Lycopene attenuates the inflammatory adipose tissue response to high fat feeding by regulating both macrophage recruitment and M1/M2 status M. Nagashimada, Y. Ni, N. Nagata, L. Xu, F. Zhuge, S. Kaneko, T. Ota; Brain/Liver Interface Medicine Research Center, Kanazawa University, Japan. Background and aims: Obesity activates the innate immune system with subsequent recruitment of immune cells such as macrophages and T cells, which contribute to the development of insulin resistance. In particular, adi- pose tissue macrophages (ATM) recruitment and polarization are considered pivotal in obesity-induced inflammation and insulin resistance. However, promising treatment modalities targeting ATM for insulin resistance and type 2 diabetes remain limited. Here, we show that lycopene, an antioxidant carotenoid compound, ameliorates adipose tissue inflammation and whole- body insulin resistance in high-fat diet (HFD)-induced obese (DIO) mice by regulating both macrophage recruitment and M1/M2 status. Materials and methods: C57BL/6J mice were fed a HFD or a HFD contain- ing lycopene (HFD+LY; 12 mg/kg body weight) for a total of 8 weeks. The histology of epididymal white adipose tissue (eWAT) and insulin sensitivity were examined. Next, we quantified immune cells in stromal vascular frac- tion of eWAT, peripheral blood, and bone marrow by fluorescence-activated cell sorter (FACS). Results: After 8 weeks of feeding, lycopene improved HFD-induced glucose intolerance, and hyperinsulinemia (HFD 4.0±0.2 vs HFD+LY 2.0±0.1 ng/ml, p < 0.01; fed state), and also enhanced insulin signaling assessed by IRβ and Akt phosphorylation in eWAT of DIO mice. HFD+LY mice had decreased macrophage infiltration and crown-like structure formation in eWAT com- pared with HFD mice even though weight and adiposity were similar. In ad- dition, lycopene administration decreased expression of mRNA for MCP-1 by 50% and pro-inflammatory cytokines, such as TNFα and interleukin-1β, by 62% and 55%, respectively (all p < 0.01) in eWAT of DIO mice. DIO mice had 2.3-fold increase in the levels of thiobarbituric acid reactive substances (TBARS) in eWAT compared to wild type (WT) mice, whereas lycopene reduced lipid peroxidation by 50% (p < 0.01) in DIO mice. These findings were associated with reduction of ER stress (CHOP/GRP78), and attenu- ation of JNK/p38MAPK and NF-kB activation in eWAT. To further assess the impact of lycopene on adipose tissue inflammation, FACS analysis was performed on stromal vascular cells isolated from eWAT. ATMs identi- fied as CD45+CD11b+F4/80+cells were increased in DIO mice by 11.4-fold compared with WT mice. In addition to reduction of total ATM content, HFD+LY mice had 35% fewer CD11c+CD206-(M1) ATMs whereas 60% more CD11c-CD206+(M2) ATMs than HFD mice, resulting in predominance of M2 over M1 ATM population. However, the predominance of the Ly6C- over Ly6Chi monocyte population was not observed in both peripheral blood and bone marrow of HFD+LY mice. Moreover, the numbers of either CD3+, CD4+, or CD8+ T cells were decreased by 36%, 52%, or 38% respectively (all p S 306 1 C associated with measures of impaired glucose metabolism. There was an in- creased wnt-signaling activity in the subcutaneous adipose tissue of the FDR and the gene expression of CCND2, Tcf7L2 and FN1 were significantly up- regulated. Cytokines involved in inflammation (TNFalpha, IL1beta and IL10) showed an increased expression in adipose tissue among the FDR, as well as markers of macrophage infiltration (CD68 and MCP1). There were also signs of adipose tissue remodeling (MMP2) and fibrosis (CTGF and ACTA2) among the FDR. Conclusion: Genetic predisposition for type 2 diabetes is associated with im- paired glucose metabolism and adipose tissue dysfunction, including inflam- mation, differentiation and remodeling. Reduced ability to maintain adipose tissue functionality may be a major susceptibility factor for later development of type 2 diabetes. Supported by: EFSD/Lilly, SFF, Novo Nordisk Foundation 755 Evidence for the presence of the phagocytic machinery in adipose tissue L. Boutens, J. van der Reest, A. Helmke, R. Stienstra; Wageningen University, Netherlands. Background and aims: The chronic inflammatory state that develops in adi- pose tissue (AT) during obesity is generally considered to promote insulin resistance (IR) and Type 2 Diabetes (DMII). Obese, inflamed AT is character- ized by the presence of dead hypertrophic adipocytes surrounded by inflam- matory macrophages, forming so-called Crown-like structures (CLS). Con- trary, in lean AT hardly any dead adipocytes and inflammatory macrophages can be identified. Immunologically silent clearance of dead cells by mac- rophages is known to be crucial for maintaining homeostasis in many tissues. The presence of CLS in obese, inflamed AT suggests that the clearance of dead adipocytes is disrupted during the development of obesity and might con- tribute to AT inflammation. We therefore aim to unravel the metabolic and immunologic (dys)regulation within AT macrophages (ATM) related to their phagocytic trait, in order to find new leads for therapeutic targets and strate- gies for obesity-induced AT inflammation and subsequent IR and DMII. Materials and methods: In vivo, C57Bl/6 mice were fed a Low Fat diet (LFD) or a High fat diet (HFD) to promote IR and DMII. Phagocytosis was studied ex vivo using a co-culture system of labelled macrophages and dead adipo- cytes followed by staining for lipids and subsequent analysis using Flow Cy- tometry. The effect of AT-derived factors on the phenotype of the macrophag- es was examined using a co-culture system with AT explants. Quantitative PCR, Western Blotting, ELISA and Microarray were used to determine the immunological and metabolic state of macrophages. Results: Microarray analysis of AT depots isolated from mice on a LFD or HFD revealed differential regulation of multiple genes that are known to play crucial roles in phagocytosis (Mertk, Gas6, Mfge8, Lrp, Gulp1, Elmo1 and Ucp2) upon the development of obesity, accompanied by changes in the expression of apoptosis- (Caspases, Bax) and inflammation-related (Il-1β, Nlrp3) genes. In vitro analysis demonstrated a profound increase in lipids in macrophages that were co-cultured with dead adipocytes (using Flow Cy- tometry), suggestive of phagocytosis of adipocytes by macrophages. Upon the phagocytosis of dead adipocytes, genes involved in fatty acid oxidation (FAO; Pparα, Pgc1α) were higher expressed and the macrophages were skewed towards a predominant anti-inflammatory phenotype, marked by higher expression and secretion levels of IL-10 and decreased expression and secretion of TNFα. Interestingly, macrophages that were co-cultured with AT explants from lean mice had a twofold higher phagocytic capacity than macrophages that were cultured without AT explants. An increased expres- sion of genes involved in phagocytosis (Mertk, Gas6, Ucp2) in these mac- rophages was accompanied by higher expression levels of genes involved in FAO (Cpt1β, Acad, Pgc1β) and anti-inflammatory cytokines (IL-10); closely mimicking the phenotypic changes that were found in macrophages that have phagocytised dead adipocytes. Conclusion: Our data suggest that the phagocytic machinery is present in AT and is differentially regulated in obese, inflamed AT. Interestingly, mac- rophages are capable of phagocytising dead adipocytes ex vivo and their phagocytic capacity is enhanced by AT-derived factors. Supported by: NWO Vidi (ZonMw) 756 Local biocommunication between peripancreatic adipose tissue and pancreatic beta cell: characterisation of factors involved in beta cell dysfunction E. Sidot, C. Guzman, M. Tournier, A.-D. Lajoix, S. Peraldi-Roux; EA 7288, Montpellier, France. Background and aims: In metabolic syndrome, pancreatic beta-cell is able to compensate for the insulin resistance by increasing beta-cell mass and func- tion which allows to maintain the normoglycaemia. During the transition from pre-diabetes/obesity to type 2 diabetes (T2D), beta-cell is subjected to metabolic and inflammatory stress, that gradually lead to secretory dysfunc- tion. In obesity, factors secreted by visceral adipose tissue may influence beta- cell mass and function. Moreover, increased visceral adiposity is associated to the development of peripancreatic adipose tissue (PPAT), which penetrates the pancreas and establishes direct contact with islets of Langerhans. Re- buffat et al (Endocrinology, 2012) showed in a model of obesity and insulin resistance that factors secreted by the PPAT modulate the proliferation of beta-cells. This is in favor of a local biocommunication between PPAT and beta-cell. Our goal is to characterize the PPAT in terms of immune cells and inflammatory factors secreted during the onset of type 2 diabetes (T2D) and determine how these factors are able to influence beta-cell compensation and decompensation in response to insulin resistance and metabolic syndrome. Materials and methods: Our study was performed in the Zucker fa/fa rats/ ZDF rats model of obesity and T2D. Three groups of animals were studied, pre-obese rats (6 weeks old) obese/pre-diabetic rats (10 weeks old) and dia- betic rats (12 week olds). The presence of immune cells infiltrating the PPAT was investigated by immunohistochemistry using the macrophage marker ED1. The expression profile of peripancreatic fat-derived factors was ana- lyzed by antibodyarray (Chemiarray, RD System) and proteomic technology. A comparison of each group of rats allowed us to identify modifications in adipose tissue-derived factors and to correlate these changes to the metabolic status of the animals. Results: An increase in macrophages infiltrating the PPAT was observed in obese/pre-diabetic animals (10 weeks old). Using chemiarray technology, we observed changes in cytokines and chemokines profiles secreted by PPAT in the ZDF rat model. While in the early stages of the disease we observed an increase in pro-inflammatory cytokines and chemokines secreted by PPA (such as, CINC- 1, 2, 3, IL-1β, LIX, TIMP-1, and MIP- 1α), surprisingly and unexpectedly, the expression profile of these factors changed and their levels decreased, when animals become diabetic (12 weeks ZDF rats). Proteomic profile also varies during the onset of the disease. Conclusion: Therefore, our results showed that in obese/pre-diabetes rats, PPAT adipose tissue expression of inflammatory cytokines and chemokines are modified and thus could (i) induce and perpetuate inflammation by at- tracting immune cell in tissue and (ii) contribute to progressive beta-cell dys- function occurring in T2D. 757 The metabolic and anti-inflammatory effects of long-term resveratrol intake; a randomised clinical trial T.N. Kjær, M.M. Poulsen, M.J. Ornstrup, J. Jørgensen, B. Richelsen, S.B. Pedersen; Aarhus University Hospital, Denmark. Background and aims: Low-grade inflammation due to macrophage infiltra- tion and hypoxia in adipose tissue (AT) is thought to be the link between obesity and the metabolic syndrome (MS). Resveratrol (RV) is a natural anti- inflammatory compound found in e.g. grape-skin that reduces inflammation and normalizes both insulin sensitivity and lifespan of diet induced obese rodents. Sirt1 enzymes, thought to mediate the RV-effects in rodents, are pre- sent in human AT. In vitro, RV stimulates human AT to increase lipolysis, reduce secretion of the macrophage-attracting MCP1 and reduce inflamma- tion. Furthermore, hypoxia induced inflammation in AT is completely re- verted when exposed to RSV. In clinical trials the effects of RV are inconclu- sive. A previous short term cross over study has shown that RV lowered sys- tolic blood pressure, hepatic liver fat, and plasma triglycerides, leptin, TNF-α, leukocytes, alanine transaminase (ALT) and HOMA index in healthy obese humans. However, we and others have not been able to verify these findings. We aim to investigate the anti-inflammatory effects of long term RV intake and it’s potential to ameliorate aspects of MS. Diabetologia (2014) 57:[Suppl1]S1–S564 S 307 1 C Materials and methods: A placebo controlled double blind clinical trial. 76 males with MS were randomized to treatment with high dose RV (HD) 500 mg x 2 or low dose RV (LD) 75 mg x 2 daily or placebo (PL) for 4 months. Be- fore and after treatment, biopsies from skeletal muscle, adipose tissues were obtained for analysis of gene expression. MR and DEXA-scans were used to quantify changes in body fat amount and body composition respectively. Clinical blood pressure (BP) measurement, blood tests and urine samples were obtained at day 0, 30, 60 and 120 to follow BP, inflammation, lipid pro- file, liver function, insulin sensitivity (HOMA). Results: 76 were randomized for the trial and 66 completed the four months of treatment. At baseline the subjects were well matched for BMI, age and waist circumference. However lean mass was significantly greater in LD com- pared to HD (p=0.048) and systolic BP was significantly higher in the PL group compared to LD group (p=0.05). Plasma biochemistry: At baseline there was a positive correlation between severity of metabolic syndrome and IL-6 level (p70% inhibition of insulin-stimulated 2-DG uptake by these cells and a similar decline in the expression of GLUT4 protein and mRNA. However, how the environment affect Prep1 gene remains unknown. The aim of this study is to evaluate whether and how a diabetic condition such as high glucose can dysregulate Prep1 gene expression by inducing epigenetic changes through NF-κB recruitment to its promoter. Materials and methods: We mimicked diabetic condition by culturing L6 skeletal muscle cells in High Glucose (25 mM) relative to Normal glucose (5.5 mM). Prep1 and NF-κB expression were analyzed by Real-Time PCR and western blot analysis. Chromatin Immunoprecipitation assays have been performed to identify histone marks and histone-associated proteins. Results: L6 cells cultured in HG for 72 hours show a significant increase of Prep1 mRNA and protein levels compared to NG cultured cells. Time couse experiments of HG showed an increase of Prep1 expression both at mRNA and protein levels yet at four hours of HG incubation with the maximum effect reached at 48 and 72 hours of high glucose incubation. This effect was specific to glucose since equimolar amounts of xylose (osmolarity control) and 2-DG had no effect. As expected Glut4 mRNA levels decreased in re- sponse to HG. In parallel, HG incubation induced nuclear translocation and binding to Prep1 promoter of the transcription factor NF-κB p65 that in turn enhanced the binding of the histone methyltransferase SET7/9 and histone acetyl-transferase p300 to Prep1 promoter. This event coincided with an in- crease in lysine 4 dimethylation and lysine 9 and 14 acetylation on histone H3 (two marks of active transcription). Co-incubation of L6 cells with HG plus JSH-23, an NF-κB inhibitor, prevented HG-induced p65 binding, histone modifiers recruitment and histone modifications on Prep1 promoter and re- stored HG-induced Prep1 overexpression. Conclusion: Elevated glucose concentrations alter the epigenetic state of Prep1 gene resulting in an increase of its expression levels. These findings might have clinical relevance as preliminary evidence in our lab indicates that Prep1 gene is overexpressed in euglycemic offspring of type 2 diabetic patients, suggesting that Prep1 overexpression may provide an early contri- bution to disease in these individuals. Clarify Prep1 epigenetic regulation can yeld new insights into the physiopatology of glucose homeostasis and un- cover potential targets for diabetes treatment and prevention. 759 Palmitate challenged muscle cells attract monocytes through TLR4-dependent release of ATP N.J. Pillon1, Y.E. Li1, P.J. Bilan1, J.T. Brozinick2, A. Klip1; 1Cell Biology Program, The Hospital for Sick Children, Toronto, Canada, 2Eli Lilly and Company, Indianapolis, USA. Background and aims: Obesity and type 2 diabetes are associated with chronic activation of the immune system, and insulin-resistance arises by both lipotoxicity and low-grade inflammation. In response to high fat diet, the fat tissue attracts immune cells that cause low, sustained inflammation re- sponsible for making the body resistant to insulin. Although inflammation is clearly recognized in adipose tissue, recent studies show that high fat-feeding is also responsible for a significant increase in pro-inflammatory cytokine expression in muscle. Yet, it is unknown whether there is a cause-effect re- lationship between muscle inflammation and whole body insulin resistance. Strikingly, muscle contains resident macrophages and we and others have re- cently detected infiltrating pro-inflammatory macrophages in muscle from type 2 diabetic individuals or from high fat-fed mice. Beyond this, little is known about the interplay between immune and muscle cells in the context of hyperlipidic environment, and a key unresolved issue is whether and how muscle cells themselves are capable of attracting monocytes. Materials and methods: Skeletal muscle is composed primarily of muscle fibers and satellite myoblasts, but also encompasses blood and lymph vessels, nerves and immune cells. As each of these cell types can potentially respond to a high fat environment in vivo, we chose a cell culture approach to investigate the specific crosstalk between muscle cells and macrophages in the context of fatty acid exposure. This strategy enables us to control individual variables, to determine vectorial communication, and to explore separately the responses of each cell type. Medium collected from L6 myotubes (conditioned media, CM) exposed to saturated or unsaturated fatty acids (palmitate vs palmi- toleate) was tested for its ability to attract monocytes. Secreted factors in the CM and inflammatory pathway activation in myotubes were analyzed using western blot, qPCR, nucleotide degradation strategies and gene knockdown. Results: CM from L6 myotubes treated with palmitate - but not palmitoleate - induced THP1 monocyte migration across transwells. Palmitate caused both myotube lipotoxicity and inflammation. Although ceramide levels rose, they were not required for monocyte chemoattraction. Palmitate activated the TLR4-NFκB pathway in myotubes and elevated cytokine expression, but the monocyte chemoattracting agent in CM was not a polypeptide. In- stead, nucleotide degradation eliminated the chemoattracting properties of CM. Moreover, palmitate induced the expression and activity of pannexin-3 channels in myotubes, mediated by TLR4-NFκB. Importantly, TLR4-NFκB inhibition or pannexin-3 knockdown in myotubes prevented monocyte che- moattraction by CM. Conclusion: These findings constitute proof of concept that high levels of sat- urated fats may promote monocyte migration towards skeletal muscle in vivo, promoting the macrophage infiltration of this tissue that arises with fatty diets. They also predict that targeting chemokine production may be insufficient to reduce macrophage infiltration of muscle, as other factors such as nucleotides may significantly contribute to immune cell chemoattraction. Moreover, this study identifies pannexin-3 as an interesting target to taper the arrival and contribution of tissue inflammatory macrophages to metabolic disease. Supported by: Canadian Diabetes Association and Canadian Institutes of Health Research Diabetologia (2014) 57:[Suppl1]S1–S564 S 308 1 C PS 056 Novel adipokines 760 Serum omentin is associated with lipid levels, but not with glucose tolerance and type 2: KORA F4 study C. Herder1,2, M. Carstensen1,2, D.M. Ouwens3,4, B. Kowall5, A. Peters6,7, W. Rathmann5, M. Roden1,2, B. Thorand6,7; 1Institute for Clinical Diabetology, German Diabetes Center, Düsseldorf, 2German Center for Diabetes Research, Düsseldorf, 3Institute for Clinical Biochemistry and Pathobiochemistry, German Diabetes Center, Düsseldorf, 4Department of Endocrinology, Ghent University Hospital, Belgium, 5Institute for Biometrics and Epidemiology, German Diabetes Center, Düsseldorf, 6Institute of Epidemiology II, Helmholtz Zentrum Muenchen, Neuherberg, 7German Center for Diabetes Research, Neuherberg, Germany. Background and aims: Circulating omentin levels are positively correlated with adiponectin. Although several studies suggested that omentin - like adiponectin - may be inversely associated with obesity and insulin resistance (HOMA-IR), data from population-based studies are not available so far. Therefore, we investigated in the population-based KORA F4 survey from Augsburg/Germany (i) whether serum levels of omentin were associated with prediabetes, type 2 diabetes (T2D) and continuous T2D risk factors and (ii) whether associations between omentin, prediabetes, T2D and related risk factors were mediated by BMI. Materials and methods: Serum levels of omentin were measured by ELISA in 1089 participants (48% male; age range 61-82 years; BMI 29±4 kg/m²; NGT: n=576, IFG: n=56, IGT: n=186, IFG/IGT: n=46; newly diagnosed T2D: n=66, known T2D: n=159). Associations between serum omentin, glucose tolerance and continuous risk factors of T2D were assessed using logistic and linear regression models, respectively. Results: We found no inverse association between serum omentin and pre- diabetes or diabetes (P>0.05 for unadjusted comparison of categories of glu- cose tolerance) and no significant associations with glucose tolerance status after adjustment for age, sex, BMI, lifestyle factors, lipids, hypertension and history of myocardial infarction. After adjustment for multiple confound- ers, serum levels of omentin were associated with blood lipids (HDL choles- terol: β=0.04, P S 309 1 C Results: In visceral adipose tissue, expression of Sfrp4 was 2.3-fold higher in patients with type 2 diabetes either versus patients with morbid obesity only or controls (P S 310 1 C Materials and methods: Immunostaining for FABP5 and GIP was performed in murine samples of small intestine for confirmation of specific expression of FABP5 in K-cells. To evaluate the acute GIP secretory response in FABP5- /- mice, lard (10ml/kg) and glucose were injected orally; plasma glucose and serum levels of GIP, GLP-1 and insulin were measured. Using K-cells isolated from newly-generated GIP-GFP knock-in hetero (GIPgfp/+)-FABP5+/+ and GIPgfp/+-FABP5-/- mice, quantitative real-time PCR (qPCR) for GIP expres- sion and measurement of GIP content were performed to determine whether or not FABP5 is involved in GIP biosynthesis. To assess potential effects of FABP5-deficiency on body weight and composition, mice were fed a high fat diet (HFD) for 10 weeks. Whole body CT scans of HFD-fed wild-type (WT), FABP5-/-, GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5-/- mice were com- pared. Using WAT taken from HFD-fed WT and FABP5-/-mice, mRNA ex- pression levels of lipid metabolism-related genes were evaluated by qPCR. Results: Immunostaining of intestinal mucosa showed that GIP-positive cells were totally merged with FABP5-positive cells and that 90% of FABP5- positive cells were merged with GIP-positive cells. Although plasma GIP lev- els after lard injection were significantly lower in FABP5-/-mice compared to those in WT mice, there were no significant differences in the results of OGTT. Plasma glucose, insulin and GLP-1 levels after both glucose and lard administration were similar in WT mice and FABP5-/- mice. There was no significant difference in GIP mRNA expression or GIP content in K-cells from GIPgfp/+-FABP5+/+ and GIPgfp/+-FABP5-/-mice. Under HFD feed- ing conditions, FABP5-/- mice exhibited significantly decreased body weight gain compared to WT control, but there was no significant difference in body weight between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp -FABP5-/- mice, in which GIP expression is genetically deleted. Whole body CT scan showed that body fat mass was significantly reduced in FABP5-/- mice compared to that in WT mice and that body fat mass in GIPgfp/gfp -FABP5+/+ and GIPgfp/gfp -FABP5-/- mice was comparable. qPCR of WAT revealed no significant change in mRNA expression levels of hormone sensitive lipase, PPARδ, aP2 or GIP-receptor between HFD-fed WT and FABP5-/-mice. Conclusion: Our results show that FABP5 is involved in fatty acid-induced acute GIP secretion and that it contributes to the development of HFD-in- duced obesity in a GIP-dependent manner. 767 NOV/CCN3: a new adipokine involved in obesity-associated insulin resistance? B. Feve1, C. Martinerie1, M. Fesatidou1, P.-O. Marchal1, C. Kazazian1, M. Buyse1, H. Do Thi Tu1, B. Antoine1, T. Ladent1, M.-P. Garcia1, H. Koseki2, C. Chadjichristos3, R. Denis4, S. Luquet4; 1Faculté de Médecine Pierre et Marie Curie, CDR St-Antoine UMR_S938 INSERM/UPMC, Paris, France, 2Chiba University Graduate School of Medicine, Chiba, Japan, 3Hôpital Tenon, UMR_S 702 INSERM/UPMC, 4Université Paris Diderot-Paris7, CNRS EAC 4413, Paris, France. Background and aims: The pathophysiology of obesity is far to be elucidat- ed. Identification of new adipokines represents a major challenge to decipher the mechanisms of adipose tissue development and metabolism. Our recent results suggest that NOV/CCN3, a multifunctional matricellular protein, is synthesized and secreted by adipose tissue, with plasma levels correlated with body mass index and fat mass, and down-regulated following bariatric sur- gery (Pakradouni et al.). NOV has been involved previously in tissue repair, fibrotic and inflammatory diseases, and cancer. However, its role in adipose tissue and energy homeostasis remains unknown. Materials and methods: The aim of our work was to investigate the meta- bolic phenotype of NOV knockout mice fed a standard or high fat diet (HFD) Results: Weight gain was similar between wild type (WT) and NOV-/- mice fed a chow diet. Strikingly, the weight of NOV-/- mice fed a HFD was mark- edly lower than that of WT animals. This was related to a sharp decrease in fat mass, whereas lean mass was preserved. Accordingly, NOV-/- mice fed a HFD displayed an improved glucose tolerance and insulin sensitivity. No significant changes were detected for food intake, energy expenditure, or locomotor activity. Importantly, the absence of NOV was associated with a marked decrease in adipose tissue expression of several proinflammatory cy- tokines and chemokines. Conversely, exposure of 3T3-L1 mature adipocytes to NOV led to an induction of these cyto-/chemokines. Conclusion: Altogether, these results show that NOV is a new adipokine that could be involved in obesity-associated insulin resistance Supported by: Ypsomed-SFD 768 The mRNA expression and release of (pro)-inflammatory and angiogenic factors in perivascular fat cells is influenced by fetuin-A via different signalling pathways D.I. Siegel-Axel1,2, S. Ullrich1,2, N. Stefan1,2, K. Rittig1, F. Gerst1,2, U. Schmidt1, B. Schreiner1, H.-E. Schaller3, H.-U. Haering1,2; 1Department of Internal Medicine, Medical Clinic IV, Division of Endocrinology, Diabetology, Angiology, Nephrology and Clinical Chemistry, 2Institute for Diabetes Research and Metabolic Diseases of the Helmholtz Centre Munich at the University of Tuebingen and German Center for Diabetes Research 3Department of Plastic, Hand and Reconstructive Surgery, BG-Trauma-Center, Tuebingen, Germany. Background and aims: Our previous in vitro and in vivo studies showed that fetuin-A exerts effects on perivascular fat cells (PVFC) and vascular cells which may influence cardiovascular diseases, predominantly effects on (pro)- inflammatory and angiogenic factors. Analysis of underlying pathways demonstrated that fetuin-A/palmitate (fet/palm) stimulates IL-6 and -8 ex- pression via TLR4 receptor signalling. In contrast, the inhibition of the potent angiogenic protein HGF was mediated via the well known insulin-dependent receptor tyrosine kinase pathway. In order to elucidate pathways downstream TLR4 receptor activation, in this study fet/palm induced signalling pathways were studied in PVFC more detailed using several kinase inhibitors. Materials and methods: PVFC were isolated from fat specimens around arm arteries of patients from our trauma center. Cells were treated with or with- out the JNK inhibitor SP500125 (10 μmol/L), the p38MAPkinase inhibitor SB203580 (10 μmol/L), the MEK1/2 inhibitor PD98059 (10 μmol/L) and the NFκB inhibitor BAY11-7082 (1 μmol/L) 2 h prior to the addition of fet (600 μmol/L) and palm (50 μmol/L). After 24 h protein release was quantitiated by ELISA technology and mRNA expression by realtime PCR. Results: The expression of IL-8, IL-6 and MCP-1 was upregulated by fet and this effect was potentiated significantly by the addition of palm. Additional incubation with the NFκB inhibitor abrogated these effects significantly. JNK inhibition caused a partial blockade of the fet/palm induced stimulatory ef- fects on interleukins whereas MEK1/2 and p38MAPkinase inhibition had only slight influence on the potent stimulatory effects of fet/palm. In contrast to the TLR4-mediated stimulatory effects on interleukins, the expression of the strong angiogenic factor HGF was potently inhibited by fet/palm and this effect could not be abrogated by the NFκB inhibitor and only partially by the other MAPkinase inhibitors. However, the angiogenic factor bFGF was not inhibited but slightly stimulated by fet/palm and this was predominantly influenced by the JNK inhibitor and slightly by NFκB inhibitor. Finally, fet/ palm exerted nearly no effects on the angiogenic factor VEGF. Conclusion: Fet/palm caused strong stimulatory effects on (pro)inflamma- tory proteins. However, the observed effects on several potent angiogenic factors were diverse: bFGF was stimulated slightly, VEGF remained nearly unchanged and HGF was even inhibited significantly. Treatment with sev- eral signalling pathway inhibitors showed that fet acts via different pathways downstream TLR4 receptor activation, such as JNK or NFκB signalling but to a lesser extent on other MAPkinases, such as MEK1/2 and p38. Neverthe- less, the diverse inhibitory effect of fet/palm on the potent angiogenic factor HGF seems to be mediated at least partially by another - TLR4-independent - pathway. Supported by: German Federal Ministry of Education and Research to the DZD and REGINA 769 NAMPT knockdown attenuates atherosclerosis by promoting macrophage reverse cholesterol transport M. Li1,2, L. Li1,2, C. Wang3, G. Yang3; 1College of Laboratory Medicine, Chongqing Medical University, 2The Key Laboratory of Laboratory Medical Diagnostics in the Ministry of Education, 3Department of Endocrinology, the Second Affiliated Hospital, Chongqing Medical University, Chongqing, China. Background and aims: Nicotinamide phosphoribosyltransferase (NAMPT), also known as visfatin, was claimed to be secreted predominantly from vis- ceral adipose tissue and to exert insulin-like action. It has been reported that NAMPT induced the release of inflammatory cytokines, cholesterol accumu- lation and the formation of macrophage foam cells in cultured cells. So there may be some association with atherosclerosis due to its role in inflammation and lipid accumulation. However, the impact of NAMPT on atherosclerosis Diabetologia (2014) 57:[Suppl1]S1–S564 S 311 1 C remained to be systematically investigated. To determine the role of NAMPT in the development of atherosclerosis, we have determined whether NAMPT knockdown can improve insulin resistance and formation of atherosclerotic lesions by using a loss-of-function approach (adenovirus-mediated RNAi) to inhibit NAMPT expression in the liver in C57BL/6J and apoE KO mice on a high fat diet. More importantly, we disscussed the molecular mechanisms of NAMPT-knockdown in the regulation of cholesterol metabolism and the development of atherosclerosis. Materials and methods: In vivo, NAMPT-knockdown C57BL/6J and apoE KO mice by a adenovirus-mediated RNAi was used to assessed the effects of NAMPT knockdown on metabolic parameters on a standard chow diet or a high fat diet; Euglycemic- hyperinsulinemic clamps were performed to assess glucose kinetics by tracer dilution methodology. An in vivo reverse cholesterol transport (RCT) assay that traces 3H-cholesterol derived from macrophages loaded with cholesterol ex vivo was performed. In vitro, as- says testing effects of Ad-GFP or Ad-shVisfatin treatment on cholesterol efflux from 3H-cholesterol-loaded RAW264.7 cells to lipoprotein acceptors was conducted. The expressions of hepatic genes related to cholesterol me- tabolism were analyzed by qRT-PCR and western blots; MK886, a selective inhibitor of PPARα was used to analyze the mechanisms by which NAMPT knockdown leads to increased cholesterol efflux and RCT. Results: HFD-fed C57BL/6J mice with reduced hepatic NAMPT content showed no alterations in all tested metabolic parameters, except for elevated HDL-C and slightly decreased LDL-C levels without statistical significance compared to littermates fed HFD ;However, plasma HDL-C levels in HFD- fed ApoE KO mice treated with Ad-shNAMPT were significantly increased and hepatic TC contents were decreased compared to littermates fed HFD ; Analysis of lipoproteins by FPLC showed an increase in cholesterol content in the HDL fractions of Ad-shNAMPT-treated mice compared with control mice; RCT assay showed that higher HDL levels in response to NAMPT knockdown increased cholesterol transport from peripheral macrophages to the liver . In cultured RAW264.7 cells, HDL-mediated 3H-cholesterol ef- flux was increased by NAMPT knockdown; NAMPT knockdown upregu- lated mRNA and protein expression involved in cholesterol efflux and RCT; MK886 abolished NAMPT knockdown-induced ABCA1/G1 and LXRα ex- pressions in Hepa1-6 and RAW264.7 cells. Conclusion: NAMPT knockdown exerted antiatherogenic effects by pro- moting cholesterol efflux and macrophage RCT through the PPARα- LXRα- ABCA1/G1pathway in vitro or in vivo. Supported by: National Natural Science Foundation of China (81270913, 81070640, 81100567) PS 057 Nutrition and diet I 770 Dietary fat intake in low-carbohydrate diets and subsequent mortality and weight change in type 2 diabetes J.W.J. Beulens1, M.J.E. Campmans-Kuijpers1, I. Sluijs1, U. Nöthlings2, H. Freisling3, K. Overvad4, H. Boeing5, A.M. May1, on behalf of the EPIC consortium; 1Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Netherlands, 2Department of Food Sciences and Nutrition, University of Bonn, Germany, 3International Agency for Research on Cancer (IARC-WHO), Lyon, France, 4Department of Epidemiology School of Public Health, Aarhus University, Denmark, 5Department of Epidemiology, German Institute of Human Nutrition, Potsdam-Rehbrücke, Germany. Background and aims: The replacement of carbohydrates with fat in a diet for type 2 diabetes patients is still debated. This study aimed to investigate the association between dietary carbohydrate intake and replacement with different types of fat with all-cause and cardiovascular (CVD) mortality risk and 5-year weight change in patients with type 2 diabetes. Materials and methods: The study included 6,192 patients with type 2 dia- betes from 15 cohorts of the European Prospective Investigation into Cancer and Nutrition (EPIC). Dietary intake was assessed at baseline with country- specific food-frequency questionnaires. Cox regression was used to estimate whether replacement of carbohydrates with different types of fat was associ- ated with cardiovascular mortality adjusted for diabetes risk factors and di- etary factors. Linear regression was used for associations with weight change. Results: After a mean follow-up of 9.2 y ±; SD 2.3 y, 791 (13%) participants had died, of which 268 (4%) due to CVD. The risk of all-cause mortality in- creased when 10 gram or 5 energy % of carbohydrates was replaced by total fat (HR 1.07 [1.02-1.13]), or SFAs (HR 1.25 [1.11-1.40]) and decreased when replaced by MUFAs (HR 0.89 [0.77-1.02]). CVD mortality risk increased when carbohydrates were substituted with SFAs (HR 1.22 [1.00-1.49]) or PUFAs (HR 1.29 [1.02-1.63]). The 5-year weight change decreased when car- bohydrates were substituted with total fat or MUFAs, and increased when substituted by SFAs. After adjusting for 5-year weight change, the associa- tions for (CVD) mortality lost significance. Conclusion: In diabetes patients, substitution of carbohydrates with total fat or SFAs increased mortality risk. Replacement of carbohydrates with MUFAs was associated with lower mortality risk and weight change. Dietary guide- lines should focus on replacement of carbohydrates by fat-subtypes, espe- cially replacement of SFAs by MUFAs seems beneficial. 771 Differential acute postprandial effects to isocaloric meals with different macronutrient content in subjects with type 2 diabetes and healthy controls L. Belinova1, H. Kahleova1, H. Malinska1, O. Topolcan2, J. Vrzalova2, O. Oliyarnyk1, L. Kazdova1, M. Hill3, T. Pelikanova1; 1Department of Diabetes, Institute for Clinical and Experimental Medicine, Prague, 2University Hospital in Pilsen, 3Institute of Endocrinology, Praha, Czech Republic. Background and aims: Dietary guidelines for treatment of diabetes focus mainly on carbohydrate counting. Evidence suggests that postprandial meta- bolic response is also modified by other macronutrient content of meals and their effect on various gastrointestinal hormone release. We aimed to investi- gate the effects of two standardized isocaloric meals: a processed hamburger meat meal rich in protein and saturated fat (M-meal) and a vegetarian sand- wich meal rich in carbohydrates (V-meal) in terms of postprandial glucose, lipids and insulin levels, gastrointestinal hormone (GIH) response and oxida- tive stress markers. Materials and methods: In a randomized crossover study, 50 patients with type 2 diabetes (T2D) and 50 healthy subjects underwent two 3-hour meal tolerance tests. For statistical analyses, repeated-measures ANOVA was used. Results: The M-meal resulted in a higher postprandial increase in lipids in both groups (p S 312 1 C was lower (p S 313 1 C 775 Two weeks of adaptation to a high-protein hypercaloric high-fat-diet changes the postprandial carbohydrate and fat response after a mixed meal A. Rietman1, J. Schwarz1, E. Siebelink1, F.J. Kok1, D. Tomé2, M. Mensink1; 1Human Nutrition, Wageningen University, Netherlands, 2Nutrition Physiology and Ingestive Behavior, AgroParisTech, INRA, UMR914, Paris, France. Background and aims: We and others previously showed changes in fasting lipid metabolism after adaptation to a high-protein diet. To further explore these changes we evaluated the effect of a two week short-term adaptation to a high protein diet on postprandial carbohydrate and lipid metabolism in healthy human subjects, by applying a meal challenge (MC). Materials and methods: A randomized crossover trial with a parallel control group was performed. After a 2-week-run-in period, participants were as- signed to either the control-diet for 4 weeks (CD-group; n=10; 27.8 En% fat; 16.9 En% protein; 55.3 En% carbohydrates), or a hypercaloric high-fat diet (HD-group; n=17; + 2 MJ per day), with in random order two periods of 2 weeks, with either a high protein (HP; 37.7 En% fat; 25.7 En% protein; 36.6 En% carbohydrates) or a normal protein content (NP; 39.4 En% fat; 15.4 En% protein; 45.2 En% carbohydrates). The MC (Liquid meal, 40 En% fat; 15 En% protein and 45 En% carbohydrate) was performed after 2 weeks and 4 weeks of intervention. Blood samples were drawn regularly during 6 hours post- prandial for metabolite analyses. Results: No significant differences were observed between NP and HP in fasting glucose (5.04 ± 0.08 vs. 5.05 ± 0.09 mmol/L) and insulin (4.21 ± 0.62 vs. 3.95 ± 0.63 µU/L) levels. However, during the MC, subjects who were adapted to the HP diet showed a significantly higher iAUC of glucose as compared to subjects adapted to the NP diet (NP: 377.4 ± 14.1 vs. HP: 408.2 ± 13.2 mmol/L*60min; p=0.03), without clear differences in insulin iAUC (NP: 2583.9 ± 383.7 vs. HP: 2984.8 ± 496.9 mU/L*60 min; p=0.38). Fasting triglyceride (TG) concentration tended to be lower on the HP compared to the NP condition (NP: 0.77 ± 0.05 vs. HP: 0.65 ± 0.03 mmol/L; p=0.07). Also during the MC subjects adapted to the HP diet showed a lower iAUC of TG as compared to subjects adapted to the NP diet (NP 67.1 ± 5.6 vs. HP 57.9 ± 4.4 mmol/L*60min; p=0.11). Conclusion: Adaptation to a high protein hypercaloric high-fat diet resulted in a decreased clearance of carbohydrates after the MC as indicated by the in- creased glucose iAUC, without changes in plasma insulin response compared to adaptation to a normal protein hypercaloric high-fat diet. Additionally, adaptation to the HP diet was associated with an improved lipid metabolism, as indicated by a lower level of circulating TG in fasted state as well as during the MC. Additional metabolomics data are anticipated to be present at the time of the meeting. Clinical Trial Registration Number: NCT01354626 Supported by: NZO 776 Differential islet, incretin and glucose prandial responses after high caloric breakfast vs isocaloric dinner in type 2 diabetic patients D. Jakubowicz1, O. Froy2, Y. Bar-Dayan1, H. Rabinovitz2, Z. Landau1, J. Wainstein1; 1Diabetes Unit, E.Wolfson Medical Center, Holon, 2Institute of Biochemistry, Food Science and Nutrition, Rehovot, Israel. Background and aims: It was shown in type 2 diabetic (T2D) individuals that a high-calorie (HC) that a diet with high-calorie (HC) breakfast (B) with reduced dinner (D), resulted in a significant decrease in HbA1c, and was more beneficial than isocaloric diet with HCD and reduced B in improving glucose, lipids and hunger scores. Whether it is related to a potential diurnal pattern of incretin secretion and β-cell responsiveness to incretins after Iso HCB vs HCD has however, not been studied in T2D. We assessed in T2D postprandial early AUC30, late AUC 60-180 and AUC180 min response for plasma insulin, glucose, intact (i) and total (t) glucagon-like peptide-1 (GLP- 1) and hunger scores after HCB vs isocaloric HCD Materials and methods: In a randomized crossover design 18-T2D (8 males), aged 57 ± 83 yrs; BMI: 28.11 ± 2.92 kg/m2; HbA1c: 7.55 ± 0.42%, un- dergone HCB meal test on day with HCB diet: B: 700 kCal; % of CH: protein: fat: 50:30:20%; lunch (L) (500 kCal, 20:45:25%) and D (200 kCal, 13:40:47%). Another day HCD meal test was done on HCD diet: B :(200 kCal,13:40:47%), L:(500 kCal, 20:45: 25%) and D (700 kCal, 50:30:20%). Results: Glucose peak was lower after HCB: 230 ± 24 vs. HCD: 285 ± 25 mg/ dl, (p S 314 1 C therapeutic dose of caffeine that restores insulin sensitivity and normalizes blood pressure in diet-induced IR rats. The time required for the reversion of IR was also evaluated. Materials and methods: All experiments were performed in 8-13 weeks Wistar rats of both genders (250-450g). Two groups of rats were used: a high- sucrose diet-group (HSu) and a control group. The HSu insulin-resistant model was obtained by submitting the animals to a 35% sucrose diet during 28 days. A dose-response curve for chronic caffeine intake (0.25-1g/L) on insulin sensitivity was performed both in HSu and control groups. Caffeine was administered in drinking water after the first 28 days of control or HSu diet and maintained during 13 weeks. The insulin sensitivity was evaluated by an insulin tolerance test in conscious animals. Plasma glucose and insulin were monitored each 2 weeks. At the end of the experimental period, rats were anaesthetized with pentobarbitone (60mg/Kg) and blood pressure was measured in the femoral artery. Blood was collected by heart puncture and skeletal muscle collected for Glut4, Insulin receptor and AMPKα1 quantifica- tion by Western-blot. Results: Administration of HSu diet during 28 days decreased insulin sensi- tivity to 2.81±0.29 %glucose/min (KITT control= 4.70±0.28 %glucose/min). Chronic administration of 0.5 and 1g/L of caffeine did not modify insulin sensitivity. In the HSu group, 3 weeks after chronic administration of 1g/L of caffeine, insulin sensitivity was restored (KITT =4.70±0.299 %glucose/min). The insulin sensitivity in this group was maintained during the following weeks. In the HSu rats submitted to chronic administration of 0.5g/L caffeine insulin sensitivity start to increase (30%) after 5 weeks of administration and it was completely restored after 9 weeks (KITT = 4.75±0.669 %glucose/min). Conclusion: Chronic caffeine administration in all concentrations tested re- stores insulin sensitivity. The period of time to restore insulin sensitivity is positively related with caffeine concentration. Supported by: EXPL/NEU-SCC/2183/2013 779 A fish-based diet intervention improves endothelial function in postmenopausal women with type 2 diabetes mellitus: a lipidomics approach to understand the mechanisms K. Kondo1, K. Morino1, Y. Nishio2, M. Kondo1, K. Nakao1, F. Nakagawa3, A. Ishikado4, O. Sekine1, S. Ugi1, H. Maegawa1; 1Department of Medicine, Shiga University of Medical Scienece, Otsu, 2Department of Diabetes and Endocrine Medicine, Kagoshima University, 3JCL Bioassay Corporation, Suita, 4R&D Department, Sunstar Inc, Takatsuki, Japan. Background and aims: The beneficial effects of fish and n-3 polyunsaturated fatty acids (PUFA) consumption on atherosclerosis have been reported in nu- merous epidemiological studies. However, to the best of our knowledge, the effects of a fish-based diet intervention on endothelial function have not been investigated. Therefore, we studied these effects in postmenopausal women with type 2 diabetes mellitus (T2DM). Materials and methods: Twenty-three postmenopausal women with T2DM were assigned to two four-week periods of either a fish-based diet (n-3 PUFA ⋯ 3.0 g/day) or a control diet in a randomized crossover design. Endothelial function was measured with reactive hyperemia using strain-gauge plethys- mography and compared with the serum levels of fatty acids and their me- tabolites. Endothelial function was determined with peak forearm blood flow (Peak), duration of reactive hyperemia (Duration) and flow debt repayment (FDR). Serum lipid compositions and eicosanoids were measured with tan- dem mass spectrography (LCMS). Results: A fish-based dietary intervention improved Peak by 63.7%, Dura- tion by 27.9% and FDR by 70.7%, compared to the control diet. Serum n-3 PUFA levels increased after the fish-based diet period and decreased after the control diet period, compared with the baseline period (1.49 vs. 0.97 vs. 1.19 mmol/l, p < 0.0001). There was no correlation between serum n-3 PUFA levels and endothelial function. Eicosanoids were measured with LCMS and 7 / 86 metabolites were significantly changed after a fish-based diet interven- tion. Interestingly, the ratio of epoxyeicosatrienoic acid (EET, known as the vasodilator) / dihydroxyeicosatrienoic acid (the inactive metabolites of EET) was increased after a fish-based diet intervention. Conclusion: A fish-based dietary intervention improves endothelial func- tion in postmenopausal women with T2DM. The improvement of endothelial function was not associated with serum n-3 PUFA concentration. These data suggests that the vasoactive metabolites such as EET, may contribute to this phenomenon. Clinical Trial Registration Number: UMIN000002277 PS 058 Nutrition and diet II 780 Incretin hormones are released in healthy subjects by both diet oil and olive oil while short chain fatty acids have no effect on secretion M.J. Mandøe1,2, K.B. Hansen3, B. Hartmann2, J.J. Holst2, H.S. Hansen4; 1Department of Clinical Physiology and Nuclear Medicine, Glostrup Hospital, 2NNF Center for Basic Metabolic Research, Department of Biomedical Sciences, University of Copenhagen, 3Department of Endocrinology, Rigshospitalet, 4Department of Drug Design and Pharmacology, University of Copenhagen, Copenhagen, Denmark. Background and aims: Diet oil (1,3-di-butyryl-2-oleoyl-glycerol) consists of two short chain fatty acids and 2-oleoyl-glycerol (2-OG), the latter being a naturally occurring ligand for the G protein-coupled receptor GPR119. Intes- tinal 2-OG causes release of the incretin hormone, glucagon-like peptide-1 (GLP-1), in healthy subjects and diet oil does the same in patients with type 2 diabetes. This study evaluated tributyrin, which in mouse studies may in- crease GLP-1 release, and a new form of diet oil with medium chain fatty ac- ids (1,3-di-octanoyl-2-oleoylglycerol or “C8-diet oil”) on GLP-1 and glucose- dependent insulinotropic polypeptide (GIP) secretion. Octanoic acid may not significantly activate known fatty acid receptors. Materials and methods: C8-diet oil, tributyrin, olive oil and carrot were ad- ministered orally to 12 Caucasian healthy male subjects (age: 24±0.6 years (mean±SEM), body mass index: 22.±0.6 kg/m2; HbA1c: 30.45±0.67 mmol/ mol) in a randomised, single-blinded cross-over study. The subjects were giv- en four meals on four different days: 200 g grated carrot + 6.53 g tributyrin; 200 g grated carrot + 13.15 g diet oil; 200 g grated carrot + 19 g olive oil or 200 g grated carrot. All lipids amounted to 0.0216 mol, and theoretically, the C8- diet oil and olive oil regimens both result in formation of 0.0216 mol 2-OG during digestion. Primary outcomes were total GLP-1 and GIP in plasma. Results: C8-diet oil and olive oil resulted in (p≤0.01) greater postprandial GLP-1 and GIP responses [GLP-1: incremental area under curve 583 and 538 pmol/l×120 min, p=0.733; GIP: 1674 and 3293 pmol/l×120 min, p=0.002] compared to both tributyrin (GLP-1: 94 pmol/l×120 min and GIP: 403 pmol/ l×120 min) and the carrot meal (GLP-1: 126 pmol/l×120 min and GIP: 423 pmol/l×120 min Conclusion: C8-diet-oil and olive oil enhanced secretion of GLP-1 and GIP significantly compared to tributyrin and carbohydrate. Tributyrin had no ef- fect. Olive oil liberates not only 2-OG but in addition 2 oleic acid molecules, which may also stimulate incretin secretion. However, the previous reported effect of diet oil and the present effect of C8-diet oil giving the same full GLP- 1 response as olive oil, suggests that it is mediated by 2-OG and not by fatty acids. Supported by: The Lundbeckfoundation, The Augustinusfoundation and Fog’s Foundation 781 Glucose spikes of eating vegetables before carbohydrates with snacks at mid-afternoon are lower than that of eating carbohydrates before vegetables without snacks S. Imai1, M. Fukui2, N. Ozasa3, S. Kajiyama4,5; 1Graduate School of Comprehensive Rehabilitation, Osaka Prefecture University, 2Graduate School of Medical Science, Kyoto Prefectural University of Medicine, 3Graduate School of Medicine, Kyoto University, 4Kajiyama Clinic, Kyoto, 5Kyoto Prefectural University of Medicine, Japan. Background and aims: Postprandial plasma glucose and glucose spikes are associated with cardiovascular diseases in patients with diabetes. An appro- priate diet is needed to reduce postprandial hyperglycemia and minimize hy- poglycemic events in addition to improving mean blood glucose. However, a compliance of diet in patients with diabetes is low, partly as a result of eating snacks. We evaluated whether eating order without snacks and with snacks at different time of the day could change the postprandial glucose values as- sessed by continuous glucose monitoring system (CGMS) in Japanese pa- tients with type 2 diabetes mellitus (T2DM). Materials and methods: This was a randomized, controlled, four-treatment, crossover study. Thirteen patients with T2DM (69.1 ± 6.7 yrs, HbA1c 7.5 ± 1.2%: mean ± SD) were assigned to use CGMS (Medtronic Minimed Gold) for 8 days (4-day each) either (1) eating vegetables first or (2) carbohydrates first and (3) eating vegetables first with additional snacks (biscuits 18 g, 75 Diabetologia (2014) 57:[Suppl1]S1–S564 S 315 1 C kcal, carbohydrate 13.5 g, protein 1.2g, fat 1.8 g) at 12:30 and (4) 15:30. The test meals consist of rice/bread, meat/fish, 500 g of vegetables, and contain 21 g of dietary fiber: the energy ratio was 60, 16, and 24% from carbohydrates, protein, and fat, respectively. First, at 12:00 of the 1st day, each participant wore a CGMS at the clinic, and consumed the test meals at 7:00, 12:00, and 19:00 at home on the 2nd and the 3rd day, and at noon of the 4th day a CGMS was removed at the clinic. The patients consumed the first dish of vegetables for 5 min, then the main dishes for 5 min, and rice or bread for 5 min of the test meals or eating vice versa. Next, the patients ate vegetables before car- bohydrates of the test meals at the same time and consumed snacks at 12:30 and 15:30 on the 2nd and the 3rd day. The daily glucose parameters were compared between four treatments. Results: Mean blood glucose increased 0.67 mmol/L by adding snacks. Eat- ing vegetables before carbohydrates with snacks at 15:30 resulted in lower in- cremental glucose peak (IGP) and incremental area under the glucose curve (IAUC) 0-3h of lunch and dinner compared with eating carbohydrates before vegetables without snacks (Figure). Conclusion: Glucose spikes of eating vegetables before carbohydrates with snacks at mid-afternoon were significantly lower than that of eating carbohy- drates before vegetables without snacks in patients with T2DM. Clinical Trial Registration Number: UMIN-9465 Supported by: Japanese Ministry of ESC and Osaka Prefecture University 782 Dietary carbohydrates predict maternal hyperinsulinaemia and cord blood visfatin levels in pregnancy compared to total daily caloric intake, fat and proteins G. Valsamakis1, D. Papatheodorou1, Z. Papadimitriou1, G. Landis2, E. Kapantais3, A. Thomopoulos1, G. Mastorakos1; 1Endocrine Unit, Aretaieion University Hospital, 2Endocrine Unit, Evgenideion University Hospital, 3Diabetes Obesity Metabolism, Metropolitan Hospital, Athens, Greece. Background and aims: Maternal diet affects insulin resistance and fetal growth-metabolism during pregnancy. Hyperinsulinemia is a marker of in- sulin resistance. There is still ongoing debate regarding the effect of maternal diet composition and caloric intake into maternal hyperinsulinemia and fetal metabolism. The effect of maternal dietary nutrients (carbohydrates, pro- teins, fat) and total daily caloric intake into maternal hyperinsulinemia , fetal metabolism and birthweight was assessed during normal pregnancy Materials and methods: A prospective, non interventional study, 50 preg- nant non diabetic women (mean±SD) pre-pregnancy BMI 24.2±4 kg/m2 and age 29.3±4.3 years were randomly assessed with validated quantitative food frequency questionnaire and a 7 day dietary diary for nutrient and total daily caloric intake (kcals) during the second trimester (24th-26th week). A 75gr oral glucose tolerance test was performed and maternal glucose and insulin levels measured at 0,5,15,30,60,90,120 min. 3 cases diagnosed with gesta- tional diabetes were removed from the study. The delta area under the curve (ΔAUCΙnsulin) for maternal hyperinsulinemia was calculated by trapezoidal rule. During delivery cord blood levels for visfatin, adiponectin, leptin, glu- cose and insulin were measured. Birth weight was measured. Results: Dietary analysis revealed a diet composition of 20% protein, 40%car- dohydrates and 40% fat. Mean birthweight was 3132±125 gr. Mean weight gain was 16±3 kg. There was no association between mean maternal total caloric intake and dietary nutrients with birth weight or maternal weight. Carbohydrates and caloric intake were negatively and positively associated with cord blood visfatin (r=-0.91, p=0.002) (r=0.72, p=0.037) respectively. Backwards multiple regression analysis showed carbohydrates (p=0.015, beta=-1.28) being the best predictors of maternal ΔAUCΙnsulin of the second trimester and cord blood visfatin levels (p=0.009) compared to dietary fat, proteins and mean total daily caloric intake. Conclusion: Maternal carbohydrate intake is the best predictor of hyperinsu- linemia among protein, fat and total caloric daily intake. Furthermore carbo- hydrate intake seems to affect the insulinomimetic secretion of adipocytokine visfatin of the fetus. Further studies need to assess the effect of maternal diet composition into fetal metabolism and growth. 783 Vitamin D supplementation might increase the benefits of weight loss on insulin sensitivity in obese individuals with hypovitaminosis D C. Conte, C.M.A. Cefalo, C. Policola, V.A. Sun, S. Moffa, G.P. Sorice, T. Mezza, A. Giaccari; Catholic University of the Sacred Heart, Rome, Italy. Background and aims: Hypovitaminosis D is highly prevalent in the general population, and particularly among obese individuals. It has been suggested that, besides being associated with obesity, vitamin D insufficiency might have an independent role in the pathogenesis of insulin resistance. Healthy dietary habits and weight loss are the mainstay for the prevention of type 2 diabetes mellitus in overweight/obese individuals. Vitamin D supplemen- tation might increase the benefits of these lifestyle interventions. Here we present the preliminary results of a currently ongoing double blind, placebo controlled randomised study. The aim of the study is to assess whether vita- min D supplementation in conjunction with a weight-loss dietary regimen has a greater impact on insulin sensitivity as compared to weight loss alone in obese subjects with hypovitaminosis D. Materials and methods: 12 obese volunteers (8/4 F/M; mean age±SD: 40.4±12.4 yrs; BMI: 38.3±6.3 kg/m2) with hypovitaminosis D [serum 25(OH) D: 14.8±7.8 ng/ml] were enrolled in this study after providing signed in- formed consent. Subjects were randomised (1:1) to hypocaloric diet + either oral cholecalciferol 25,000 I.U./week (VIT) or placebo (PLA) for 3 months. Anthropometric assessments, a standard 2-hr OGTT and a hyperinsulinae- mic euglycaemic clamp (HEC) for the measurement of insulin sensitivity (M value) were performed at baseline and at study end. Results: Baseline anthropometric characteristics, insulin sensitivity and 25(OH)D levels were comparable in the two groups. After 3 months of hy- pocaloric diet + cholecalciferol or placebo, body weight was significantly reduced in both groups (-6.3% and -9.3% for VIT and PLA, respectively; p S 316 1 C tential health promoting effects. Aim: To determine if a high dietary intake of bitter and strong tasting vegetables have a beneficial impact on insulin resist- ance related to type 2 diabetes (T2D) when compared to equivalent intake of modern mild and sweet tasting vegetables. Materials and methods: The study was a 3-month randomized controlled parallel intervention study involving 77 participants aged 35-70 years with T2D. The participants were randomized into 3 different diets; 1) consum- ing daily 500 g of bitter and strong tasting (BST) vegetables 2) 500 g daily of sweet and mild tasting (SMT) vegetables and 3) normal diet (control). Both vegetable diets (group 1 and 2) consisted of root vegetables and cabbages, but the seeds for the vegetables in the BST group originate from the Nordic gene bank collection. Additionally the vegetables were ensured a bitter and strong taste through modulation of the agronomic and postharvest storage practice. Results: Both diets high in vegetables did significantly reduce the partici- pants BMI (p S 317 1 C 787 Relationships of serum 25-OH vitamin D and parathyroid hormone levels with all-cause mortality in type 2 diabetes F.H. Nystrom1, P. Jennersjö1, H. Guldbrand1, S. Björne1, T. Länne1, M. Fredriksson2, C. Östgren1; 1Department of Medical and Health Sciences, Faculty of Health Sciences, 2Department of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping, Sweden. Background and aims: New and clinically useful markers of risk are of es- sence in high risk states such as Type 2 diabetes. Materials and methods: We analyzed baseline data from 741 patients (489 men and 252 women) who participated in “Cardiovascular Risk factors in Patients with Diabetes - a Prospective study in Primary care” (CARDIPP). Participants of this prospective observational study were between 55-66 years old when recruited between the years 2005-2008 and they all had Type 2 diabetes. 25-OH-vitamin D and parathyroid hormone levels (PTH) were analyzed using electrochemo-luminiscense on a Cobas® e602 unit (Roche Di- agnostics Scandinavia AB, Bromma, Sweden). Carotid-femoral pulse wave velocity (PWV) was measured with applanation tonometry (SphygmoCor® system, AtCor Medical, Sydney, Australia) and intima-media thickness of the carotid arteries (IMT) was evaluated using B-mode ultrasound (ATL HDI 5000, Bothell, WA, USA). The participants were followed for all-cause mor- tality until December 31, 2012, using the Swedish Cause of Death Registry. Results: There was a negative linear relationship between vitamin D levels with waist circumference (r = -0.147, p < 0.0001) in the total cohort. During the follow up period 24 men and 9 women died. When entering major risk factors in Cox regression vitamin D levels were negatively related to all-cause mortality in men, independently of age, PTH, HbA1c, waist circumference, mean 24-hour systolic ambulatory blood pressure (ABP) and apolipoprotein B levels (p= 0.049). This finding of increased mortality related to low levels of vitamin D in men was also statistically significant when PWV and IMT were added to the equations in men (p= 0.028 Hazard ratio of 0.97, CI from 0.95 to 0.997 for each nmol/l of vitamin D). The figure below shows mortality in men of the lower and higher tertiles of vitamin D levels, respectively. In women, on the other hand, levels of PTH (p= 0.016 without PWV and IMT) were posi- tively related with all-cause mortality when performing the corresponding analyses (p= 0.006 with PWV and IMT, Hazard ratio of 1.049, CI from 1.014 to 1.085 for each ng/l of PTH) while vitamin D levels were without statistical significance in the models (p =0.9). Conclusion: Our data show that information about the levels of serum vita- min D in men and of PTH in women add prognostic information in terms of total-mortality that goes beyond traditional risk factors and that also adds information independently of the levels of systolic ABP, IMT and PWV. Supported by: ALF, FORSS, HLF, VR 788 Oral glucose has no effect on appetite and satiety sensations compared to isoglycaemic i.v. glucose in healthy subjects M. Gyldenløve1, A.E. Junker2, L. Skov1, T. Vilsbøll2, F.K. Knop2; 1Department of Dermato-Allergology, 2Diabetes Research Division, Department of Medicine, Gentofte Hospital, University of Copenhagen, Hellerup, Denmark. Background and aims: Oral glucose ingestion is thought to elicit release of satiety-promoting gastrointestinal (GI) hormones. However, the effect of glu- cose stimulation of the GI tract on appetite per se has been difficult to separate from the satiety-inducing effect of elevated plasma glucose concentrations. We aimed to evaluate the GI-independent effect of glycaemic excursions fol- lowing OGTT on appetite and satiety sensations in healthy volunteers. Materials and methods: Hunger, satiety, fullness, prospective food consump- tion, thirst, well-being, and nausea were assessed by 100-mm visual analogue scales in 20 healthy subjects (7 females; age (mean±SEM): 54±3 years; BMI: 27.0±0.6 kg/m2; HbA1C 5.4±0.1%) every 20 minutes during a 3h 50g-OGTT and an isoglycaemic i.v. glucose infusion (IIGI), respectively. Overall appetite score (OAS) was calculated as (satiety + fullness + (100 - hunger) + (100 - prospective food consumption)) / 4. Gastrointestinal-mediated glucose dis- posal (GIGD) was calculated as the difference in glucose amounts used dur- ing OGTT and IIGI related to the amount used during OGTT (GIGD = 100% × (glucoseOGTT - glucoseIIGI) / glucoseOGTT). Results: Isoglycaemia during the oral and i.v. glucose stimuli was achieved in all subjects. As expected, glucose stimulation of the GI tract had a strong impact on glucose disposal from the blood reflected by a mean±SEM GIGD of 57±2%. No significant differences in baseline appetite and satiety sensa- tions were observed between the two experimental days. Overall, appetite sensations increased whereas satiety sensations decreased during both glu- cose stimuli. There were no significant differences in hunger, satiety, fullness, prospective food consumption, thirst, well-being, nausea or OAS. Conclusion: Our findings show that glucose stimulation of the GI tract has little or no effect on appetite and satiety sensations independently of plasma glucose levels. This suggests that the collective effect of GI factors triggered by oral glucose is of minor importance for appetite and satiety sensations in healthy subjects. Diabetologia (2014) 57:[Suppl1]S1–S564 S 318 1 C PS 059 Metformin 789 Effectiveness of add-on glucose-lowering drugs for early glycaemic control in users of metformin: population-based cohort study R.W. Thomsen1, L.M. Baggesen1, L. Pedersen1, E.S. Buhl2, C.L. Haase2, E. Svensson1, M.L. Svendsen1, M. Søgaard1, H. Nørrelund1, S.P. Johnsen1; 1Department of Clinical Epidemiology, Aarhus University Hospital, 2Novo Nordisk Scandinavia AB, Copenhagen, Denmark. Background and aims: We examined the association of different add-on glucose-lowering drugs with achieved quality of early glycaemic control in a large population-based cohort of incident metformin users. Materials and methods: We included all individuals with a first-time pre- scription for metformin in Northern Denmark between 2000 and 2012 who had their treatment intensified with another glucose-lowering drug within the first year. We obtained individual-level data on glycated hemoglobin (HbA1c) measurements, comorbidities, and medications from population- based medical databases. Results: We identified 25,631 T2D patients with a first metformin prescrip- tion. Of these, 3,898 (15%) received an add-on glucose-lowering drug within the first year following metformin start, of which 3,089 had a baseline HbA1c measured before their add-on drug. A number of 1,888 patients (61%) re- ceived a sulfonylurea (SU) add-on, 648 (21%) received a DPP4 inhibitor, 132 (4%) a GLP-1 receptor analogue, 125 (4%) another oral hypoglycemic agent (OHA), and 296 (10%) started insulin in addition to metformin. Me- dian patient age at time of add-on was 60 years, 59% were males, and the median HbA1c before add-on was 7.9% (inter quartile range 7.0-9.2%). Pa- tients with SU add-on had a median age of 61 y, a median HbA1c of 7.8% at baseline, and 23% had hospital-diagnosed comorbidity. The corresponding values were similar for patients with DPP4 inhibitor add-on (59 y, 7.7%, and 21%); for GLP-1 add-on (55 y, 8.0%, and 15%); and for other OHA add-on (59 y, 7.9%, and 18%). Insulin add-on patients tended to have worse glycae- mic control and more comorbidity (56 y, 9.6%, and 29%). Within 3-6 months after add-on therapy start, 2,357 T2D patients (76%) had a control HbA1c measurement recorded. Overall, the median HbA1c dropped from 7.9% at baseline to 6.7%, corresponding to a reduction of 1.2 percentage points (pp). Absolute reductions were 1.2 pp (from 7.9 to 6.7) for sulfonylurea add-on, 1.0 pp (7.7 to 6.7) for DPP4 inhibitors, 1.5 pp (8.0 to 6.5) for GLP-1 recep- tor analogues, 1.1 pp (8.0 to 6.9) for other OGDs, and 2.7 pp (9.7 to 7.0) for insulin add-on. Overall, 61% of all metformin users who had add-on therapy attained an HbA1c target of S 319 1 C Materials and methods: A total of 220 newly diagnosed type 2 diabetes patients taking oral metformin were recruited, genotyped and divided into three groups by SLC22A2 genotypes (G/G, G/T, T/T), as well as three groups by SLC47A1 genotypes (G/G, G/A, A/A). Six to ten patients in each group were randomly selected and undertaken metformin pharmacokinetic study. A one-year follow up randomized cohort study was performed to clarify met- formin pharmacodynamics. Results: After one year, the falling range of HbA1c level was significantly greater in subjects with heterozygous variant genotype (GT and AA) than wild type homozygote (-1.28% in GT vs -1.01% in GG, -2.32% in AA vs -1.07% in GG, both PA variants through delaying its transportation and renal clearance in Chi- nese type 2 diabetes populations. Supported by: the National Natural Science Fund of China 792 Long term effects of metformin on plasma concentrations of C-peptide in insulin-treated type 2 diabetes patients: a placebo controlled trial W.M.C. Top1,2, A. Kooy1,2, P. Lehert3, C.D.A. Stehouwer4; 1Internal Medicine, Bethesda Hospital Care Group Leveste Middenveld, Hoogeveen, 2Bethesda Diabetes Research Centre, Hoogeveen, Netherlands, 3Statistical Department, Louvain University, Belgium, 4Internal Medicine, Maastricht University Medical Centre, Netherlands. Background and aims: Type 2 diabetes (T2D) is characterized by beta cell failure and varying degrees of insulin resistance, both being targets of treat- ment. While metformin is known to improve insulin sensitivity, its effects on beta cell function are assumed to be neutral. Because beta cell mass decreases over time, improvement of beta cell function may be more difficult to obtain in advanced T2D. In this study, the effects of metformin on beta cell func- tion were evaluated in patients with advanced, insulin-treated T2D during a follow-up of 4.3 years. Materials and methods: To evaluate the effect of metformin on endogenous insulin secretion, we analysed the plasma concentrations of C peptide (im- muno-assay) at baseline, after 4 months and yearly for 4 years in a placebo- controlled, randomized trial in insulin-treated patients with T2D (N=390, mean age 61.3 years, mean diabetes duration 13.1 years, mean baseline HbA1c 7.9%, 54% men). A linear model was performed on logtransformed values because of a skewed distribution of the data, and with adjustment for baseline values. Results: Unadjusted mean C-peptide levels decreased from 0.52(+/- 0.38) to 0.47(+/- 0.34) nmol/l in the placebo group, and increased from 0.47(+/- 0.40) to 0.55(+/- 0.47) nmol/l in the metformin group. For the final visit, regression analysis showed a significant treatment effect (difference, 5.8% (95% CI, 1.0- 11.5), P=0.034). The summary mean in the placebo group was 0.52(+/- 0.35) and in the treatment group 0.53 (+/- 0.40) nmol/l. No statistically significant treatment effect was found on the summary mean (difference, 2.7% (95% CI, -0.5-5.9), P=0.101). As a measure for steady state beta cell function, we calcu- lated the fasting C-peptide/fasting glucose index (C-peptide/glucose x 100). This ratio decreased from 5.4 (+/- 4.1) to 5.0 (+/- 3.9) in the placebo group, and increased from 4.9 (+/- 4.0) to 5.9 (+/- 4.7) in the metformin group. Regression analysis showed a significant treatment effect both for the final visit (difference 0.9 (95% CI, 0.2-1.6), P=0.010) and for the summary mean (difference 0.9 (95% CI, 0.5-1.4), P S 320 1 C 794 Effects of metformin on fuel metabolism, energy expenditure and body composition in human subjects and rats I. Tokubuchi1, Y. Tajiri1, S. Iwata1, K. Hara1, K. Yamada1, H. Mifune2; 1Division of Endocrinology and Metabolism, 2Department of Animal Experiment, Kurume University School of Medicine, Japan. Background and aims: Metformin is widely used to treat obese diabetics be- cause of its beneficial effects on body weight, energy intake, and glucose reg- ulation. Metformin phosphorylates AMP-activated protein kinase (AMPK) and exerts its effects on hyperglycemia and insulin resistance through a blockade of gluconeogenesis in the liver and an enhancement of fat oxida- tion in the skeletal muscle. However, results regarding effects on energy ex- penditure (EE), respiratory quotient (RQ) and body composition have been rather controversial. In the present study, we investigated this unsolved issue separately during either fasting or post-prandial state in both human subjects and rats. Materials and methods: All data were expressed as means±S.D. In human study, metformin hydrochloride (1500 mg/day) was administered either 23 healthy subjects (group C: 16 males, 26±3 years old, BMI 22.0±3.4) or 13 type 2 diabetic patients (group DM: 7 males, 44±17 years old, BMI 30.9±6.2) for 2 weeks. Meal tolerance test (592 kcal, 75 g of carbohydrate, 28.5 g of fat) was performed in the morning before and after 2 weeks administration of metformin. Blood samples were collected before and 1, 2, 3 hours after the ingestion of meal, and plasma concentrations of glucose (G), triglyceride (TG), lactate (LA) and pyruvate (PA) were measured. Simultaneously expired gas analysis was performed for the measurements of EE and RQ. In animal study, male Sprague Dawley rats (13 weeks old) were fed 10 g of standard chow twice daily (at 8AM and 7PM), and drinking water either with or without metformin hydrochloride (2.5 mg/ml) was given for 2 weeks. At 15 weeks old, EE and RQ were measured in a chamber equipped with expired gas analysis system, body compositions were measured by CT scan, then they were sacrificed in either fasting or post-prandial state and blood samples were collected. Results: In human study, by the administration of metformin for 2 weeks sig- nificant decreases of G (P S 321 1 C practices throughout Germany. Practices anonymously report all diagnoses (ICD-10), prescriptions (ATC), hospital admissions, and laboratory test re- sults on an ongoing basis. The study sample included 22,556 patients who received a diagnosis of type 2 diabetes during the index period used for this study (January 2000 to December 2012), and who did not have cancer before diagnosis of diabetes. The outcome measure was cancer incidence (ICD-10: all, breast, colon, and prostate cancer). The median follow-up time was 4.8 years. 1,446 (6.4%) patients developed cancer. In an intention-to-treat type analysis, patients with sulfonylurea (or insulin, or other antidiabetic medi- cations) as their first diabetes medication were compared to patients with metformin as their first antidiabetic drug prescription. In a second type of analysis, patients with a monotherapy of sulfonylurea, insulin, or other medi- cations, respectively, were compared to patients with a metformin monother- apy. For these analyses, time 0 was the time one year (three years in a sensi- tivity analysis) after the first prescription of the respective medication. Cox regression models were fitted, and hazard ratios (HR) and 95% confidence intervals (CI) were calculated to compare users of sulfonylurea, insulin, and other medications with users of metformin. Results: Compared to patients with metformin as first diabetes medication, no increased risks of cancer of all sites were found in patients with other first diabetes medication. In models adjusted for age, sex, obesity, hyperten- sion, hyperlipidemia, duration of diabetes, and Charlson comorbidity index, the corresponding HRs were 1.02 (95%-CI: 0.89 - 1.17) for sulfonylurea, 1.09 (95%-CI: 0.86 - 1.37) for insulin, and 0.91 (95%-CI: 0.74 - 1.12) for other dia- betes medication. In the analogous analyses for the comparison of monother- apies, similar results were obtained. In additional analyses with a latency time of three years, and in analyses for some selected cancer sites (breast, colon, and prostate cancer), no decreased cancer risk was found in metformin users. Conclusion: In a retrospective database analysis, taking into account po- tential time related biases, no different cancer risk was found in metformin, sulfonylurea and insulin users. This finding is in line with few other observa- tional studies using time-dependent analyses. To clarify the association be- tween antidiabetic drug use and cancer risk, there is a strong need for further well-designed observational studies and RCTs. 797 The suppression mechanism of metformin on the tumour death ratio accelerated by a high calorie diet in the RasH2 mouse: a model for evaluating tumour prevention or regression O. Shuichi; Department of Internal Medicine, Kurume University, Kurume, Japan. Background and aims: The objective of the current study was to investigate whether Metformin intake changed the rate of tumor death under different diet conditions in the rasH2 transgenic mouse, a model for evaluating and designing tumor prevention or regression therapies. Materials and methods: Male rasH2 transgenic mice at 8 weeks of age were divided into 2 groups (n=4 per group): one fed a high calorie diet of 592 kcal/100 g and drinking water ad libitum; and the other fed the same high calorie diet with Metformin added to the drinking water from initiation until 32 weeks. Metformin was administered at a small dose of 25 mg/kg body weight per day. Then all mice were sacrificed and examined under microar- rays analysis (SurePrint G3 Mouse GE microarrays kit 8x60K, Agilient) and then quantitative PCR. Results: Body weights of the rasH2 transgenic mice fed a high calorie diet with or without Metformin were almost the same from the age of 10 weeks until 32 weeks (76.5 ± 17.6 g vs. 76.5 ± 17.6 g). The human ras gene expres- sion of liver, lung, kidney, spleen and visceral fat were not significantly dif- ferent between the 2 groups. The lifespan of mice fed a high calorie diet with Metformin (n = 15; 75.3 ± 17.4 weeks) was longer than those fed a high calo- rie diet without Metformin (p < 0.05), regardless of no differences in weekly body weights. However, At autopsy, multiple occurrences of tumors in the lung, liver, spleen, digestive organs and kidney were found in all groups and malignancy was considered as the cause of death.Based on the results of mi- croarrays analysis, we identified that Ddit4 gene, which plays an important role in responses to cellular energy levels and cellular stress, including re- sponses to hypoxia and DNA damage, was a possible cause for this phenom- enon. Our quantitative PCR experiments showed that the Ddit4 gene expres- sion of mice with Metformin significantly decreased compared with that of mice without Metformin in kidney (0.90 ± 0.38 vs. 1.91 ± 0.49, p=0.037, n=4, respectively). Other organs showed tendency to decrease (liver: 0.35 ± 0.14 vs. 0.89 ± 0.14, p=0.09; lung: 1.43 ± 1.19 vs. 1.74 ± 2.89, p=0.64, n=4, respec- tively). Conclusion: Findings from the current study demonstrate that a constitu- tional predisposition to tumor may be enhanced by diet, and that high inci- dence of tumor death may be suppressed by a small dose of orally adminis- tered Metformin acting on ameliorating genes, including Ddit4 gene, in mice fed a high calorie diet. We hope this Metformin research will act as a step- ping stone for making an effective drug for the anti-cancer effects especially among diabetic patients. Diabetologia (2014) 57:[Suppl1]S1–S564 S 322 1 C PS 060 SGLT2 inhibitors: safety 798 Long-term renal safety with dapagliflozin treatment A. Ptaszynska1, T. Mansfield1, E. Johnsson2, S. Parikh3, Y. Yavin1, J.F. List1; 1Bristol-Myers Squibb, Princeton, USA, 2AstraZeneca, Mölndal, Sweden, 3AstraZeneca, Wilmington, USA. Background and aims: Dapagliflozin (DAPA) reduces hyperglycaemia by inhibiting renal glucose reabsorption. Because of this mechanism of action renal safety has been monitored closely throughout the DAPA clinical devel- opment programme. Materials and methods: Renal safety was assessed in data pooled from 13 placebo (PBO)-controlled trials lasting up to 24 weeks (short-term [ST] data- set), 9 PBO-controlled trials lasting up to 102 weeks (ST + long-term [LT] dataset), and from all 21 phase 2b/3 trials, including an active-controlled study up to 208 weeks (all Ph2b/3 dataset) in type 2 diabetes mellitus (T2DM) patients. Frequency of adverse events (AEs), serious AEs, and abnormal labo- ratory tests were summarised. Results: Among a total of 9339 patients in the all Ph2b/3 dataset, estimated glomerular filtration rate (eGFR) at baseline (BL) was ≥ 60 mL/min/1.73 m2 in 8268 (88.5%) patients and ≥ 30 to < 60 mL/min/1.73 m2 in 1070 (11.5%) patients. With DAPA, in the ST+LT dataset, after a small initial decrease at week 1 (mean change from BL: -4.2 vs 0.5 mL/min/1.73 m2 [DAPA 10 mg vs PBO]) mean eGFR returned toward baseline by week 24 (mean change from BL: -1.4 vs -0.7 mL/min/1.73 m2 [DAPA 10 mg vs PBO]) and was stable to week 102. In the ST and ST+LT datasets, AEs related to renal function were more frequent with DAPA (Table); most were small, reversible changes in serum creatinine. In both PBO- and DAPA-treated subjects, these AEs oc- curred > 10 times more frequently over 24 weeks, and 6-8 times more fre- quently over 102 weeks, in subjects with eGFR ≥ 30 to < 60 mL/min/1.73 m2 at BL (moderate renal impairment) compared with those with eGFR ≥ 60 mL/min/1.73 m2 at BL (Table). Serious AEs (SAEs) related to renal function were balanced across treatments and were recorded in 0.15% of patients in both DAPA and control groups in the all Ph2b/3 dataset (DAPA, 9 of 5936; control, 5 of 3403) and there were no instances of acute tubular necrosis or acute nephritis. Marked laboratory abnormalities were uncommon and simi- lar between groups; for example, serum creatinine ≥ 2.5 mg/dL was observed in 3 (0.2%) of both DAPA- and PBO-treated patients at week 102, but in no patient with eGFR ≥ 60 mL/min/1.73 m2 at BL. Shifts between categories of albuminuria (normal, micro, and macro) at week 24 in the ST dataset were similar for DAPA and PBO; overall, albuminuria did not worsen in either group. Renal safety data with DAPA were similar in a 4-year active-controlled trial (N=816). Conclusion: Pooled data from 21 phase 2b/3 trials in T2DM patients, in- cluding 13 ST and 9 LT studies, show that DAPA 10 mg is associated with a slight excess of renal AEs, mostly transient creatinine changes, but not of renal SAEs, compared with PBO. The difference is driven primarily by the subgroup of patients with eGFR ≥ 30-< 60 mL/min/1.73 m2 at BL. There is no indication of renal toxicity with DAPA. These data provide further evidence of a good renal safety profile for DAPA in the treatment of T2DM. Clinical Trial Registration Number: NCT00660907 Supported by: AstraZeneca/BMS 799 Efficacy and safety of canagliflozin (CANA) in patients with type 2 diabetes mellitus (T2DM) who progressed to stage 3A chronic kidney disease during treatment R. Roussel1, D. de Zeeuw2, G. Law3, G. Meininger3; 1Hôpital Bichat, Paris, France, 2University of Groningen, Netherlands, 3Janssen Research & Development, LLC, Raritan, USA. Background and aims: CANA, an SGLT2 inhibitor, has provided reductions in HbA1c, body weight (BW), and systolic blood pressure (SBP) across Phase 3 studies of patients with T2DM. CANA was also associated with transient reductions in estimated glomerular filtration rate (eGFR) that stabilized or at- tenuated over the treatment period. This analysis evaluated the impact of de- creases in eGFR on the efficacy and safety of CANA in patients with T2DM. Materials and methods: Changes from baseline in HbA1c, BW, and SBP were assessed in the subset of patients with eGFR ≥60 mL/min/1.73 m2 at baseline who had a reduction in eGFR to ≥45 and S 323 1 C 8.1%, and 6.4% and serious AEs were 12.4%, 10.0%, and 7.9%. The incidence of volume depletion-related AEs was 4.5%, 4.6%, and 2.0% with CANA 100 and 300 mg and non-CANA; discontinuations and serious AEs related to vol- ume depletion were infrequent across groups. Conclusion: Among patients with T2DM with reductions from eGFR ≥60 mL/min/1.73 m2 at baseline to eGFR ≥45 and S 324 1 C GTT was performed at day 80 and 120 of treatment. Immunofluorescence analysis of nephrin in the renal cortex and H&E in the pancreas was per- formed in the different experimental groups. Results: Empa treatment enhanced significant (P2000 vs. Veh S 325 1 C Conclusion: Consistent with the lack of effect on bone mineral density previ- ously reported, in these pooled clinical data analyses dapagliflozin therapy was not associated with an increase in fractures overall or in subgroups at risk (i.e. older patients, postmenopausal women and patients with renal im- pairment). Clinical Trial Registration Number: 00263276, 00357370, 00528372, 00528879, 00643851, 00663260, 00683878, 00736879, 00859898, 00976495, 00831779, 01095653, 00660907, 00680745, 00972244, 00673231, 01294423, 00984867, 00855166, 01031680, 01042977 Supported by: BMS/AstraZeneca 805 Two-year efficacy and safety of dapagliflozin for patients with type 2 diabetes mellitus and a history of cardiovascular disease I. Gause-Nilsson1, T.W. DeBruin2, J. Sugg2, S.J. Parikh2, E. Johnsson1, L.A. Leiter3; 1AstraZeneca, Mölndal, Sweden, 2AstraZeneca, Wilmington, USA, 3University of Toronto, Canada. Background and aims: Dapagliflozin (DAPA) is a selective sodium-glu- cose co-transporter 2 inhibitor with glucuretic and osmotic diuretic effects. The benefit/risk profile of DAPA was evaluated in a double-blind, placebo (PBO)-controlled phase 3 study of patients (pts) with type 2 diabetes mellitus (T2DM) inadequately controlled on usual care and preexisting cardiovascu- lar disease (CVD). Primary outcome results for 24 weeks (wks) plus a 28-wk extension were previously reported; here we report results for 104 wks. Materials and methods: Pts (HbA1c 7%-10%) were randomised 1:1 to re- ceive 10 mg/d DAPA or PBO for 24 weeks, followed by consecutive extension periods of 28 and 52 wks. Pts were stratified by age (< or ≥ 65 years [y]) and insulin (INS) use; INS doses were reduced by 25% at randomisation. Efficacy data for 104 wks (observed cases) are descriptive (mean [95% CI]); glycaemic end points exclude pts who received rescue therapy (DAPA, n=130 [27.5%]; PBO, n=245 [50.5%]). Results: In the efficacy population (N=962), mean age was 64 y; 47% were ≥ 65 y. Mean T2DM duration was 13 y. CVD history was mainly CHD (76.5%); 93% of pts had a history of hypertension. Most pts re- ported using 1 (45.4%) or 2 (34.7%) oral antihyperglycemic agents; 60.5% of patients used INS, 20% as monotherapy. Of the 964 randomised pts, 157 (33.6%) and 162 (32.5%) in the DAPA and PBO groups, re- spectively, were eligible to continue into the 52-week extension period. Over 104 wks DAPA maintained reductions vs PBO for HbA1c, body weight (BW), systolic blood pressure (SBP), and fasting plasma glucose (FPG) (Ta- ble). Pts in the DAPA group showed a mean decrease of -8.7 IU/d in daily INS dose compared with PBO. A larger proportion of pts achieved a com- bined reduction of HbA1c ≥ 0.5%, BW ≥ 3% and SBP ≥ 3 mm Hg with DAPA compared with PBO (4.2% vs 1.1%); proportions were lower in both groups compared with earlier time points. In the safety population (N=965), for DAPA vs (vs) PBO, ≥ 1 adverse event was observed in 77.0% vs 72.5% of pts, hypoglycaemic events in 30.7% (none major) vs 29.4% (2 major) of pts, and cardiac disorders in 10.0% vs 13.5% of pts, respectively. Events of genital infection were reported in 7.7% of DAPA vs 0.4% of PBO-treated pts; events of urinary tract infection were reported in 11.8% of DAPA vs 6.8% of PBO- treated pts. For DAPA vs PBO, mean percent changes were 3.6 vs -2.0 mg/ dL in total cholesterol, 4.3 vs -1.3 mg/dL in HDL-C, and 5.5 vs -1.6 mg/dL in LDL-C, respectively. Conclusion: When added to standard of care over 104 wks in older pts with advanced T2DM and preexisting CVD, DAPA treatment resulted in sus- tained reductions in HbA1c, FPG, SBP, and BW. Safety and tolerability were similar to PBO, except for a small excess of genital and urinary infections. Clinical Trial Registration Number: NCT01042977 806 Safety and tolerability of empagliflozin (EMPA) in phase III trials and their extensions in patients with type 2 diabetes (T2DM) M. Roden1, L. Merker2, A.V. Christiansen3, F. Roux4, S. Hantel5, A. Salsali6, G. Kim5, T. Meinicke5, S.S. Lund5, T. Hach5, H.J. Woerle5, U.C. Broedl5; 1Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research, Düsseldorf, 2Diabetes- und Nierenzentrum, Dormagen, Germany, 3Boehringer Ingelheim Danmark A/S, Copenhagen, Denmark, 4Boehringer Ingelheim France, Reims, France, 5Boehringer Ingelheim, Ingelheim, Germany, 6Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, USA. Background and aims: Four 24-week randomized, double-blind, Phase III trials assessed the efficacy and safety of EMPA vs placebo (PBO) as mono- therapy, add-on to metformin, add-on to metformin plus sulphonylurea, and add-on to pioglitazone ± metformin in patients with T2DM. Patients had the option to continue in 52-week double-blind extension trials. We assessed the safety and tolerability of EMPA using pooled data from these trials. Materials and methods: In every trial, patients received EMPA 10 mg, EMPA 25 mg, or PBO until the last patient who entered the extensions had been treated for 76 weeks. Data were pooled from 2477 patients (N=825 on PBO, N=830 on EMPA 10 mg, N=822 on EMPA 25 mg) of whom 68.6% were treated in the extensions. Adverse events (AEs) were assessed descriptively in patients who took ≥1 dose of study drug. AEs included confirmed hypo- glycaemic AEs (plasma glucose ≤70mg/dl and/or requiring assistance), AEs consistent with urinary tract infection (UTI), genital infection and volume depletion, identified using prospectively defined searches of investigator-re- ported AEs based on 77, 89 and 8 MedDRA preferred terms, respectively, and hepatic events identified from 4 Standardised MedDRA Queries. Results: Total exposure was 1029, 1202 and 1162 patient-years in the PBO, EMPA 10 mg and EMPA 25 mg groups, respectively. The percentages of pa- tients with any AE(s), with serious AE(s), and with AE(s) leading to treat- ment discontinuation were similar in the PBO, EMPA 10 mg and EMPA 25 mg groups (Table). Confirmed hypoglycaemic AEs were reported in slightly higher percentages of patients on EMPA 10 mg or EMPA 25 mg than PBO, but very few required assistance. Events consistent with UTI were reported by similar percentages of patients on PBO, EMPA 10 mg and EMPA 25 mg, and by more female than male patients. Events consistent with genital infection Diabetologia (2014) 57:[Suppl1]S1–S564 S 326 1 C were reported by a higher percentage of patients on EMPA than PBO, and by more female than male patients. AEs consistent with volume depletion were reported by 2 patients (0.2%) on PBO, 10 (1.2%) on EMPA 10 mg and 6 (0.7%) on EMPA 25 mg. Bone fractures were reported by 21 patients (2.5%) on PBO, 16 (1.9%) on EMPA 10 mg and 10 (1.2%) on EMPA 25 mg. Hepatic events were reported by 28 patients (3.4%) on PBO, 15 (1.8%) on EMPA 10 mg and 25 (3.0%) on EMPA 25 mg. Conclusion: Based on assessment of AEs, EMPA 10 mg and EMPA 25 mg as monotherapy or add-on therapy for ≥76 weeks were well tolerated in patients with T2DM. Clinical Trial Registration Number: NCT01289990 Supported by: Boehringer Ingelheim and Eli Lilly 807 Reduced risk of hypoglycaemic events with dapagliflozin vs glipizide as add-on therapy in type 2 diabetes mellitus: 4-year data from a phase 3 study K. Rohwedder1, E. Johnsson2, S. Parikh3; 1AstraZeneca, Wedel, Germany, 2AstraZeneca, Mölndal, Sweden, 3AstraZeneca, Wilmington, USA. Background and aims: Dapagliflozin (DAPA), a selective inhibitor of the sodium-glucose co-transporter 2, reduces plasma glucose by increasing renal glucose excretion. Its insulin-independent mechanism results in a low risk of hypoglycaemia in contrast to sulphonylureas. In a phase 3, randomised, double-blind trial of DAPA versus glipizide in combination with metformin (MET) in adults with type 2 diabetes mellitus inadequately controlled by MET alone (N=814), DAPA was associated with reduced hypoglycaemic events versus glipizide over 52 weeks. We now report the 4-year hypoglycae- mia- and key safety-related data. Materials and methods: The phase 3, randomised, double-blind trial of DAPA (≤ 10 mg/d) versus glipizide (≤ 20 mg/d) in combination with MET (median 2000 mg/d) was extended from 52 weeks to 4 years. DAPA and glip- izide were down-titrated if medically indicated. Results: At the start of the study, 406 patients were randomised to the DAPA arm and 408 patients to the glipizide arm. The 4-year study was completed by 161 patients in the DAPA arm and by 141 patients in the glipizide arm. Including data after rescue, the rate of hypoglycaemia was ~10-fold less with DAPA (Table). Most events were minor (capillary or plasma glucose < 3.5 mmol/L [< 63 mg/dL]). The majority of hypoglycaemic events occurred dur- ing the first year of treatment: in year 1, 3.4% versus 39.7%, year 2, 1.6% ver- sus 23.6%, year 3, 1.5% versus 37.2%, and in year 4, 2.2% versus 28.4% in the DAPA and glipizide arms, respectively. Study medication was down-titrated in the DAPA arm only during year 1, and in only 2.7% of patients. In the glipizide arm, 15.8% of patients had a down-titration in year 1, 1.9% in year 2, and 4.8% in years 3 and 4. Key adverse events are also shown in the Table. The effect of therapy on HbA1c attenuated over time in both arms, but DAPA showed more persistent benefits versus glipizide up to year 4 (Table). Conclusion: Fewer patients experienced hypoglycaemic events over 4 years with DAPA than with glipizide as add-on therapy to MET. DAPA treatment in combination with MET was efficacious and well tolerated over a 4-year period. Clinical Trial Registration Number: NCT00660907 Supported by: AstraZeneca/BMS Diabetologia (2014) 57:[Suppl1]S1–S564 S 327 1 C PS 061 Insulin secretagogues 808 Risk factors associated with treatment discontinuation and down-titration in type 2 diabetes patients treated with sulfonylureas K. Iglay1, Y. Qiu1, C.-P. Fan2, Z. Li2, J. Tang2, P. Laires1; 1Merck & Co., Inc., Whitehouse Station, 2Asclepius JT, New York, USA. Background and aims: Sulfonylurea (SU) therapy can be discontinued or down-titrated for reasons such as hypoglycemia or weight gain. A retrospec- tive cohort study using the MarketScan database was conducted to identify risk factors associated with therapy changes (ie, discontinuation or down- titration). Materials and methods: Patients (pts) were included when the following criteria were satisfied: first SU prescription (Rx) (index date) between 2009 and 2011, ≥18 years of age on the index date, and ≥ 1 year continuous enroll- ment pre- and post-index. Pts with type 1, gestational or secondary diabetes, insulin use before the index date, or ≥ 2 SUs on the index date were excluded. Therapy changes were determined during the 1 year post-index period. Dis- continuation occurred when the gap between the end date of current SU fill and the start date of subsequent fill was ≥ 90 days apart. Down-titration oc- curred when an SU fill had a lower equivalent dose than that on the index date. The Kaplan-Meier method was used to estimate 3- and 6-month therapy change rates. Cox regression was used to identify risk factors associated with therapy changes in hazard ratios (HRs). Results: 104,082 pts were included, of which 55,233 (53.1%) experienced therapy changes in the 1-year post-index period. 3- and 6-month therapy change rates were 23.2% and 38.9%, respectively. Major risk factors associ- ated with therapy changes were post-index hypoglycemic events (discon- tinuation HR= 1.78 [1.68, 1.90], p S 328 1 C 811 Impact of hypoglycaemic events and HbA1c level on SU discontinuation and down-titration Y. Qiu1, P. Laires1, C.-P. Fan2, Z. Li2, J. Tang2, K. Iglay1; 1Merck & Co., Inc., Whitehouse Station, 2Asclepius JT, New York, USA. Background and aims: A retrospective cohort study using GE Centricity electronic medical records (EMR) was conducted to assess the association of hypoglycemic events and HbA1c level with discontinuation and down- titration of sulfonylureas (SUs) among patients with type 2 diabetes mellitus (DM). Materials and methods: Patients with ≥12 months EMR activity before and after initial SU prescription (index date) in 2010 - 2012 were included. Pa- tients were excluded if they had type 1, gestational or secondary DM, re- ceived insulin before the index date, used ≥2 SUs on the index date, or had incomplete prescription data in 1 year post-index. Discontinuation occurred when consecutive SU prescriptions (Rx) were ≥90 days apart. Down-titration occurred when a subsequent SU Rx had a lower equivalent dose than the index dose. Post-index hypoglycemic events (identified using ICD-9 codes) occurred after the index date and before therapy changes or the end of the 1 year post-index. Multivariate Cox regression was used to assess the asso- ciation of post-treatment hypoglycemic events and HbA1c with discontinu- ation/down-titration. Results: 28,371 patients were included in the study, of which 13,459 (47.4%) were discontinuers, 717 (2.5%) down-titraters, and 14,195 (50.0%) continu- ers. 0.6% of the continuers had hypoglycemic events in the 1 year post-in- dex period, compared to 3.1% of down-titraters and 0.8% of discontinuers (p S 329 1 C 814 Predictors of glycaemic response to add-on therapy with a DPP-IV inhibitor: a retrospective cohort study using the primary care THIN database J. Mamza1, R. Mehta2, R. Donnelly1, I. Idris1; 1School of Medicine, 2Trent Research Design Services, University of Nottingham, UK. Background and aims: Apart from baseline HbA1c, little is known about clinical or biochemical factors that affect the magnitude of the glycaemic re- sponse to a DPP-IV inhibitor when used in routine practice as part of com- bination therapy in patients with type 2 diabetes. Similarly, there is limited information about the effectiveness of a DPP-IV inhibitor in different drug combinations. These subjects were investigated using a large UK Primary Care database. Materials and methods: A cohort of 25,183 patients with type 2 diabetes, newly treated with a DPP-IV inhibitor between June 2007 and May 2013, was sourced from UK General Practices via the Health Initiative Network (THIN) database. The index date was defined as the date of initiation of DPP- IV inhibitor therapy. Baseline clinical parameters of patients with suboptimal glucose control (HbA1c>6.5% or 48mmol/mol) 6 months or more after start- ing a DPP-IV inhibitor in combination with other glucose-lowering therapy (n=17,697) were compared against 12-months follow-up data using multi- variate logistic regression. Response to DPP-IV inhibitor therapy was defined as HbA1c 1% (13mmol/mol) reduction at 12 months. A com- parative analysis of on-treatment effects of various combination regimens were examined using a propensity score matching technique. Associations were examined using t test and Chi squared test. Results: Among patients whose glucose-lowering therapy was intensified by co-administration of a DPP-IV inhibitor, independent predictors of response were baseline HbA1c [1.57, 95%CI (1.53-1.61)], and use of metformin [1.31, 95%CI (1.20-1.44)] and aspirin [1.14, 95%CI (1.06-1.23)]. The independ- ent predictors of non-response included obesity [0.82, 95%CI (0.72-0.93)], previous hypoglycaemia events [0.84, 95%CI (0.77-0.93)], total cholesterol level [0.94, 95%CI (0.90-0.97)], concurrent use of lipid lowering drugs [0.88, 95%CI (0.81-0.97)], and combination therapy involving metformin, sulpho- nylurea and thiazolidinediones (TZDs) [0.80, 95%CI (0.70-0.91)]. Overall, intensification of therapy with the addition of a DPP-IV inhibitor resulted in significant reductions in HbA1c (0.5%, 6.5mmol/mol), body weight (1Kg) and total cholesterol 0.1mmol/L (p S 330 1 C 816 Vildagliptin as add-on to insulin improves glycaemic control without increased risk of hypoglycaemia in Asian (predominantly Chinese) patients with type 2 diabetes mellitus G. Ning1, W. Wang1, L. Li2, J. Ma3, X. Lv4, M. Yang5, W. Wang5, M. Woloschak6, V. Lukashevich6, W. Kothny7; 1Ruijin Hospital, Shanghai Jiao-Tong University School of Medicine, 2Shengjing Hospital of China Medical University, Shenyang, 3Nanjing First Hospital, 4General Hospital of Beijing Military Region, 5Beijing Novartis Pharma Co., Ltd., Shanghai, China, 6Novartis Pharmaceuticals Corporation, East Hanover, USA, 7Novartis Pharma AG, Basel, Switzerland. Background and aims: There are limited data from randomized controlled trials evaluating the efficacy and safety of oral antidiabetic drugs in combina- tion with insulin in Asian patients with T2DM, who may be at higher risk of hypoglycaemia due to lower weight and increased insulin sensitivity. Here we present the efficacy and safety of vildagliptin in combination with insulin with or without metformin from a dedicated Asian study in a predominantly Chinese population with T2DM. Materials and methods: In this 24-week, multicentre, double-blind, place- bo-controlled trial, patients with T2DM inadequately controlled (HbA1c 7.5-11.0%) on stable therapy with long-acting, intermediate-acting or pre- mixed insulin, with or without concomitant metformin were randomized to receive vildagliptin 50 mg bid (n=146) or placebo (n=147). The primary effi- cacy endpoint was change in HbA1c from baseline to week 24. Key secondary endpoints included change in FPG and the proportion of patients achieving HbA1c S 331 1 C and sustained reductions in body weight compared with an increase in body weight with glimepiride (GLIM). Patients had the option to participate in a dedicated body composition sub-study in which we aimed to determine the effects of EMPA vs GLIM on the amount and distribution of body fat. Materials and methods: Body composition was evaluated in 91 randomised patients in the sub-study. Using whole body dual energy X-ray absorptiom- etry (DXA), changes from baseline to week 52 and week 104 in trunk fat, limb fat, total fat mass and fat-free mass were assessed in 62 patients (36 re- ceiving EMPA and 26 receiving GLIM). Using magnetic resonance imaging (MRI), changes from baseline to week 52 and week 104 in abdominal vis- ceral adipose tissue (VAT) and abdominal subcutaneous adipose tissue (SAT) were assessed in 51 patients (29 receiving EMPA and 22 receiving GLIM). Changes from baseline were evaluated using mixed model repeated measures (MMRM) analyses. Results: In the 91 patients evaluated in the sub-study, baseline mean [SD] age was 55.7 [10.4] years, weight was 85.1 [14.7] kg, BMI was 31.9 [4.9] kg/m2 and 60% had BMI ≥30 kg/m2. EMPA significantly reduced trunk fat, limb fat and total fat mass vs GLIM at week 52 and week 104 (table). EMPA also signifi- cantly reduced fat-free mass vs GLIM at week 52 but not at week 104 (table). Adjusted mean (SE) changes from baseline in abdominal VAT at week 52 were -15.5 (5.2) cm2 with EMPA compared with +10.0 (6.1) cm2 with GLIM (p S 332 1 C Conclusion: CANA treatment reduces BW due to UGE-associated caloric loss. The sustained UGE increase leads to a period of weight loss followed by a plateau in which a lower equilibrium BW is maintained. The plateau in BW is predicted to be due to sustained increases in EI that match the caloric loss to UGE, with only minimal changes in EE predicted. Clinical Trial Registration Number: NCT01081834 Supported by: Janssen Research & Development, LLC 821 The SGLT2 inhibitor empagliflozin ameliorates early features of diabetic nephropathy in BTBR ob/ob type 2 diabetic mice with and without hypertension F. Gembardt1, C. Bartaun1, N. Jarzebska1, E. Mayoux2, V.T. Todorov1, B. Hohenstein1, C. Hugo1; 1Department of Internal Medicine III - Nephrology, University Hospital Carl Gustav Carus, Dresden, Germany, 2Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach a. d. Riss, Germany. Background and aims: Diabetic nephropathy is the leading cause of end- stage renal disease in man in the western world. Recent development of so- dium-glucose co-transporter 2 (SGLT2) inhibitors offers a new anti-diabetic therapy via enhanced glucose excretion. Whether this strategy exerts benefi- cial effects on developing type 2 diabetic nephropathy is unclear. We investi- gated the effects of the SGLT2 inhibitor empagliflozin in BTBR ob/ob mice, which spontaneously develop type 2 diabetic nephropathy. Materials and methods: In a first experiment, BTBR ob/ob mice either re- ceived a diet containing 300ppm empagliflozin or equicaloric placebo chow for 12 weeks. In a second experimental setting, BTBR ob/ob mice received an- giotensin (Ang) II and were separated in the same 2 diet groups for 6 weeks. Urine, blood, and tissues were harvested and blood pressure was monitored by tail cuff measurements. Results: In both experiments, empagliflozin treatment enhanced glucosu- ria, (6,987±987µmol/mg Crea vs 2,182±972µmol/mg Crea; P S 333 1 C Conclusion: DAPA 10 mg induces a modest BP reduction across BP cat- egories and may thereby contribute to a reduced cardiovascular risk in hy- pertensive as well as nonhypertensive T2DM pts. The potential of DAPA to reduce cardiovascular events is currently under investigation in a large-scale outcomes study, DECLARE. Clinical Trial Registration Number: NCT00263276, NCT00357370, NCT00528372, NCT00528879, NCT00683878, NCT00859898, NCT00680745, NCT00972244, NCT00673231, NCT00984867, NCT00855166, NCT01031680, NCT01042977 Supported by: AstraZeneca and BMS 823 Sodium glucose cotransporter 2 inhibition with empagliflozin reduces microalbuminuria in patients with type 2 diabetes D. Cherney1, M. von Eynatten2, S.S. Lund3, S. Kaspers3, S. Crowe3, H.J. Woerle3, T. Hach3; 1Toronto General Hospital, University of Toronto, Canada, 2Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, USA, 3Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany. Background and aims: Empagliflozin (EMPA) is a sodium glucose cotrans- porter 2 (SGLT2) inhibitor that lowers HbA1c, blood pressure (BP) and weight in patients with type 2 diabetes (T2D) and has similar potential in type 1 diabetes (T1D). It has also been shown that EMPA reduces renal hy- perfiltration associated with T1D, suggesting a renoprotective reduction in intraglomerular pressure. Based on preclinical data, we hypothesised that the addition of EMPA to standard care, including stable doses of renin an- giotensin system inhibitors (RASi), would reduce microalbuminuria versus placebo, consistent with attenuated renal injury. Materials and methods: This post-hoc analysis assessed the effect of EMPA on urine albumin to creatinine ratio (UACR). Data were pooled from 458 patients with T2D and pre-existing microalbuminuria (UACR 30-300 mg/g; mean [SD] estimated glomerular filtration rate [eGFR; modification of diet in renal disease] 87.9 [23.3] mL/min/1.73m2; mean [SD] age 56.7 [10.1] years) who had participated in 1 of 4 Phase III randomised placebo-controlled tri- als. Changes in UACR were assessed in 438 patients who received EMPA 10 mg (n=141), EMPA 25 mg (n=150) or placebo (n=147) for 24 weeks. Results: After controlling for baseline log-transformed UACR, HbA1c, sys- tolic BP, eGFR, region, trial and treatment, EMPA 10 mg and EMPA 25 mg significantly reduced UACR by 30% and 25%, respectively, compared with placebo at week 24 (Table; p S 334 1 C on the long term use of empagliflozin, it is premature to speculate about the clinical outcomes with empagliflozin. Studies are currently underway to ad- dress this question. 825 Dapagliflozin lowered ambulatory BP in patients with type 2 diabetes mellitus and hypertension inadequately controlled by a renin- angiotensin system blocker ± another agent N. Iqbal1, M.A. Weber2, T.A. Mansfield1, J.F. List1, A. Ptaszynska3; 1Global Clinical Research, Bristol-Myers Squibb, Princeton, 2Department of Medicine, SUNY Downstate College of Medicine, New York, 3Bristol-Myers Squibb, Princeton, USA. Background and aims: Hypertension is common in patients with type 2 dia- betes mellitus (T2DM) and treatment usually starts with an ACE inhibitor (ACEi) or an angiotensin receptor blocker (ARB), with other antihyperten- sive (AHT) therapies added if required. Dapagliflozin reduces renal glucose reabsorption accompanied by weight reduction and a diuretic effect, both contributing to BP reduction. Here we assess the effects of dapagliflozin in two studies of patients with inadequately controlled T2DM and hyperten- sion. Materials and methods: Two randomized, double-blind studies assessed the effects of 12 weeks of dapagliflozin 10 mg or placebo on BP in patients with inadequately controlled T2DM (HbA1C 7.0-10.5%) and hypertension (seated systolic BP [SBP] / diastolic BP: 140-164 / 85-104 mmHg) despite receiving glucose-lowering drugs and an ACEi or ARB (Study 1) plus a 2nd AHT agent (Study 2). Preliminary results have been presented previously; here we describe ambulatory BP data. Results: In Study 1, 613 patients on an ACEi or ARB were randomized to dapagliflozin or placebo. At Week 12, mean 24-hour ambulatory SBP was 2.9 (SEM: 1.01) mmHg lower with dapagliflozin versus placebo (p=0.0043, Fig- ure) and night-time SBP was reduced versus placebo (-2.34 [1.16] mmHg). In Study 2, 449 patients on an ACEi or ARB plus another AHT drug (thiazide/ thiazide-like diuretics in ~44% of patients, calcium channel blockers ~27%, beta blockers ~27%) were randomized to dapagliflozin or placebo. At Week 12, mean 24-hour ambulatory SBP was lower with dapagliflozin versus pla- cebo (-4.5 [1.37] mmHg, p=0.0012, Figure) and a reduction in night time SBP was also evident versus placebo (-3.9 [1.51] mmHg). Dapagliflozin was well tolerated and heart rate showed no significant change from baseline. Conclusion: Dapagliflozin significantly reduced mean 24-hour ambulatory SBP versus placebo in patients with T2DM and hypertension in combination with an ACEi or ARB ± 1 AHT over 12 weeks. Clinical Trial Registration Number: NCT01137474, NCT01195662 Supported by: BMS and AstraZeneca Diabetologia (2014) 57:[Suppl1]S1–S564 S 335 1 C 826 Impact of weight loss on weight-related quality of life (WRQL) and health satisfaction (HS): evidence from a head-to-head study of canagliflozin (CANA) vs sitagliptin (SITA) S. Traina1, A. Slee2, C. Neslusan1; 1Janssen Global Services, LLC, Raritan, 2Axio, Seattle, USA. Background and aims: The positive health impacts of performing diabetes self-care (e.g., healthy food choices, increased physical activity, and medica- tion taking) are not always apparent to a person with type 2 diabetes mellitus (T2DM); therefore, adherence can be difficult. People with T2DM are more likely to adhere to treatments that offer noticeable benefits, such as conveni- ence, avoidance of hypoglycemic episodes, and weight loss vs. those that do not. Previous studies have demonstrated associations between WRQL and disease outcomes and between HS and performance of self-care behaviors. In this study, we explore the impact of the amount of weight loss demonstrated with CANA treatment on WRQL as measured by the Impact of Weight on Quality of Life (IWQOL-Lite) and physical HS (which includes satisfaction with weight, energy, appetite, sleep, physical activity, and general health) as measured by the Current Health Satisfaction Questionnaire (CHES-Q). In a previously reported study of CANA vs. SITA in dual therapy with metformin and background diet and exercise, CANA was associated with a mean weight loss of -2.1 kg to -2.5 kg vs. SITA after 52 weeks. Materials and methods: In that study, the IWQOL-Lite and CHES-Q were administered at baseline and over time. Post hoc analyses were used to ex- plore the strength of the relationship between weight change and improve- ment in IWQOL-Lite total and physical HS scores. Empirical distributions of scores were examined to investigate floor and ceiling effects. The final analyti- cal sample was limited to subjects with baseline scores that were not at the ceiling (e.g., allowing for improvement in score). Differences in least squares mean (LSM) change in weight by changes in IWQOL-lite and CHES-Q after 52 weeks were compared using ANCOVA models adjusting for selected base- line covariates (i.e., weight, age, and gender). Logistic regression, adjusted for the same covariates, was used to generate odds ratios (ORs) for score changes associated with weight change. Results: Empirical distributions of change scores indicated that 29.1% of IWQOL-Lite and 5.4% of CHES-Q scores were at the ceiling at baseline; these scores did not decline over time and were thus excluded from the analyti- cal sample. Subjects with improved IWQOL-Lite total and CHES-Q physical scores lost more weight than those with no score improvement (IWQOL- Lite: -3.54 kg; 95% CI -4.07, -3.01 and -2.33 kg; 95% CI -2.78, -1.87, respec- tively; LSM difference -1.21 kg; 95% CI -1.92, -0.51; CHES-Q: -3.50 kg; 95% CI -3.89, -3.12 and -1.86 kg; 95% CI -2.25, -1.47, respectively; LSM difference -1.65 kg; 95% CI -2.17, -1.12). Weight loss of 2 kg was associated with a 68% greater chance of IWQOL-Lite total score improvement (OR 1.68; 95% CI 1.16, 2.42) and a 72% greater chance of CHES-Q physical domain score im- provement (OR 1.72; 95% CI 1.28, 2.31). Conclusion: Results suggest that the magnitude of weight loss shown in clini- cal trials to be associated with CANA vs. SITA is noticeable by and important to people living with T2DM. Treatments providing adequate glycemic con- trol that also offer a weight loss benefit are likely to improve WRQL and HS, and in turn, may support persistent and consistent performance of healthy behaviors. Clinical Trial Registration Number: NCT01106677 Supported by: Janssen Research & Development, LLC PS 063 GLP-1 based therapies: efficacy I 827 The 2 week fasting glucose as a predictor of response to once weekly dulaglutide 1.5 mg T. Forst1, S. Gough2, G. Grunberger3, V. Pechtner4, R. Shaginian5, H. Wang6, L. Fernandez6; 1Profil Mainz, Germany, 2University of Oxford and Oxford University Hospitals NHS Trust, UK, 3Grunberger Diabetes Institute, Bloomfield Hills, USA, 4Lilly Diabetes, Eli Lilly and Company, Neuilly-Sur-Seine, France, 5Lilly Diabetes, Eli Lilly and Company, Houten, Netherlands, 6Lilly Diabetes, Indianapolis, USA. Background and aims: The identification of patients most likely to respond to a specific treatment is important when attempting to individualize patient care. Predictors of treatment success for weekly GLP-1 receptor agonists are, to date, largely unknown. This analysis was therefore conducted to as- sess whether laboratory fasting blood glucose (FBG) in patients with type 2 diabetes mellitus (T2DM) measured early in treatment with the once weekly GLP-1 receptor agonist dulaglutide (DU) 1.5 mg predicts treatment response. Materials and methods: Post hoc analyses were conducted separately for 2 double-blind, randomized Phase 3 studies (AWARD-5, in combination with metformin, and AWARD-1, in combination with metformin and pioglita- zone) in patients with T2DM assigned to once weekly DU 1.5 mg. Baseline tertile values from AWARD 5 were used to categorize FBG at baseline and weeks 2 and 8 as follows: Low (L, S 336 1 C to evaluate the 6-month change in HbA1c for patients initiating EQW or LIRA. Materials and methods: Patients with T2DM prescribed EQW or LIRA be- tween 01 February 2012 and 31 May 2013 were identified. Baseline HbA1c measures were from 75 days prior to 15 days after initiating EQW or LIRA, with follow-up measures documented at 6 months (± 45 days). Adjusted linear regression models compared the difference in mean HbA1c change. A priori defined sensitivity analyses were performed in the subgroup of pa- tients with baseline HbA1c ≥7.0% and no prescription for insulin during the 12-month pre-index period. Results: For EQW and LIRA respecitvely, mean (SD) age of the main study cohort was 58.0 (11.0) and 58.1 (11.0) years, mean (SD) baseline HbA1c was 8.3% (1.7) and 8.4% (1.6), and 48.1% and 54.1% of patients were women. In adjusted models, change in HbA1c did not differ between EQW and LIRA during 6 months of follow-up (Table). Results were consistent in the sub- group analyses. Conclusion: In a real-world setting, HbA1c similarly improved in patients initiating EQW or LIRA. Supported by: Bristol-Myers Squibb/AstraZeneca 829 Effectiveness of lixisenatide before breakfast or the main meal using CGM with AGP analysis E. Strock1, R. Mazze1, M. Powers1, A. Monk1, S. Richter2, R. Bergenstal1, E. Souhami3, N. Demil3, B. Ahrén4; 1International Diabetes Center, Minneapolis, 2Park Nicollet Institute, Minneapolis, USA, 3Sanofi, Paris, France, 4Lund University, Sweden. Background and aims: Lixisenatide is a once-daily glucagon-like peptide-1 receptor agonist that has been shown by 7-point self-monitoring blood glu- cose to lower postprandial and fasting glycaemia. The aim of this study was to understand, in a subset of type 2 diabetes mellitus (T2DM) patients in a multicentre randomized trial, the change in diurnal glucose patterns using continuous glucose monitoring (CGM) prior to and during lixisenatide treat- ment based on the time of administration: before breakfast (BK) or the main meal (MM) as defined by the patient. Materials and methods: Blinded CGM was used 14 days before and during lixisenatide treatment (BK or MM). CGM data were analyzed using Ambu- latory Glucose Profile (AGP) analysis to detect changes in diurnal glucose patterns. Results: Sixty-nine patients completed (T2DM 7.8±4.9y; age 57.8±10.2y; 53.6% female). AGP analysis showed significant reduction in 24h glu- cose exposure (AUC) in both groups (BK: 4918.1±652.3 to 3681.2±699.6 mg/dL*24h, p S 337 1 C documented (≤3.88 mmol/L [70 mg/dL]) symptomatic hypoglycemia events were low: A30 2%, A50 1%, Pbo 3%; no severe events reported. Conclusion: Albi monotherapy resulted in robust, durable A1C reduction through Wk 156 and was generally well tolerated. Clinical Trial Registration Number: NCT00849017 Supported by: GSK 831 Harmony 3 year 3 Results: albiglutide vs sitagliptin and glimepiride in patients with type 2 diabetes mellitus on metformin J.E. Matthews1, B. Ahren2, J. Ye1, M.C. Carr3, M.W. Stewart3; 1GlaxoSmithKline, Research Triangle Park, USA, 2Department of Clinical Sciences, Division of Medicine, Lund Universtiy, Sweden, 3GlaxoSmithKline, King of Prussia, USA. Background and aims: This 3-year (y), randomized (rzd), double-blind, pla- cebo (Pbo) controlled, study examined albiglutide (Albi) 30mg vs placebo (Pbo), sitagliptin (Sita), or glimepiride (SU) in patients (pts) with T2DM (A1C 53-85.8 mmol/mol [7-10%]) on Metformin (Met). Materials and methods: Pts meeting predefined hyperglycemia criteria qualified for blinded dose titration (SU 2-4mg, Albi 30-50mg). Pts received hyperglycemic rescue if prespecified rescue criteria were met & continued receiving rzd medication. Pts (%) who completed the 3y study (including res- cue): Pbo 53, Sita 61, SU 60 & Albi 61. Results: Primary endpoint at Wk 104 (baseline A1c 65 ± 8.7 mmol/mol [8.1 ± .8%]), Albi treatment difference was superior to Pbo (-9.9 mmol/mol [-.91%]; P S 338 1 C 833 Safety and tolerability of liraglutide 3.0 mg in overweight and obese adults: the SCALE obesity and prediabetes randomised trial C. le Roux1, D.C.W. Lau2, A. Astrup3, K. Fujioka4, F. Greenway5, A. Halpern6, M. Krempf7, R. Violante Ortiz8, J.P.H. Wilding9, C.B. Jensen10, C.B. Svendsen10, X. Pi-Sunyer11; 1University College Dublin, Ireland, 2University of Calgary, Canada, 3University of Copenhagen, Denmark, 4Scripps Clinic, La Jolla, 5Louisiana State University, Baton Rouge, USA, 6University of São Paulo Medical School, Brazil, 7Université de Nantes, France, 8Instituto Mexicano del Seguro Social, Ciudad Madero, Mexico, 9University of Liverpool, UK, 10Novo Nordisk A/S, Søborg, Denmark, 11Columbia University, New York, USA. Background and aims: The primary objective of this trial was to investigate the efficacy of the GLP-1 receptor agonist (RA) liraglutide 3.0 mg, as adjunct to diet and exercise, for weight management. Key safety and tolerability data are presented. Materials and methods: Individuals (BMI ≥27 kg/m2 with ≥1 comorbidity or ≥30 kg/m2) were advised on a 500 kcal/day deficit diet and exercise pro- gramme and randomised 2:1 to once daily sc liraglutide 3.0 mg (n=2487) or placebo (n=1244). Certain event types (deaths, select CV disorders, pan- creatitis, neoplasms, thyroid disorders requiring thyroidectomy) underwent blind independent assessment. Results: Baseline characteristics: age 45.1 years, 78.5% female, body weight 106.2 kg, BMI 38.3 kg/m2. Liraglutide 3.0 mg induced greater weight loss (primary endpoint; 8.0%) vs placebo (2.6%) after 56 weeks (estimated treat- ment difference [ETD] -5.4%; p S 339 1 C reduction (2.1 vs 1.7%, ETD -0.44, 95% CI -0.71 to -0.16, p S 340 1 C PS 064 GLP-1 based therapies: efficacy II 837 Harmony 4: 3 year efficacy of albiglutide (albi) vs insulin glargine (glar) in patients with type 2 diabetes mellitus P.N. Weissman1, M.W. Stewart2, D.T. Cirkel3, J. Ye4, P.D. Ambery5; 1Endocrinology Associates, Miami, 2GlaxoSmithKline, King of Prussia, USA, 3GlaxoSmithKline, Stevenage, UK, 4GlaxoSmithKline, Research Triangle Park, USA, 5MedImmune, Cambridge, UK. Background and aims: This 3 year (y) randomized, open label, Ph 3 study assessed efficacy and safety of Albi 30 mg once weekly vs Glar in T2DM pts uncontrolled (A1c 53.0-85.8 mmol/mol [7-10%]) on metformin (met) ± sul- fonylurea (SU). Materials and methods: Pts were treated to a target A1c ≤53.0 mmol/mol (7.0%) and FPG ≤5.6 mmol/L (100 mg/dL). If needed, pts could uptitrate Albi to 50 mg QW and Glar per prespecified criteria. Patients could continue if hyperglycemic rescue was required. Primary endpoint (PE) was change from baseline in A1c at wk 52 in Albi vs Glar. Results: Baseline demographics were similar between groups; mean age 56 y, A1c 67.2 mmol/mol (8.3%), 82% on met + SU. Treated N=504/241 (Albi/ Glar). A1c decreased in both groups and was maintained for 3 y. The PE showed non-inferiority between Albi and Glar at wk 52 with a treatment difference of 1.20 mmol/mol (95% CI: 0.44, 2.95) (0.11% [95% CI: -0.04%, 0.27%]). At wk 156, a similar proportion of patients in each group required rescue: Albi 56.2%, Glar 48.0%, p=0.1515. In pts completing 3 y with and without rescue, A1c reduction was durable in both groups. Change in FPG favored Glar, whereas change in weight favored Albi. AEs through wk 156 for Albi/Glar were (% pts): nausea 13.3%/7.5%; diarrhea 10.9%/7.9%; vomiting 5.4%/5.4%. Prerescue documented (≤3.9 mmol/L [70 mg/dL]) symptomatic and severe hypoglycemia events were less with Albi vs Glar (18.5%/0.6% vs 30.3%/0.8%). Injection site reactions occurred in 17.5% of Albi and 10.0% of Glar pts. Conclusion: Albi showed similar A1c improvement at wk 156 to Glar with modest weight loss and acceptable tolerability. Clinical Trial Registration Number: NCT00838916 Supported by: GSK 838 Harmony 5 - year 3 Results: albiglutide vs placebo and vs pioglitazone in triple therapy (background metformin and glimepiride) in people with type 2 diabetes P. Shamanna1, P. Home2, M.W. Stewart3, J.M. Mallory3, D.M. Miller3, C.R. Perry3, M.C. Carr3; 1Bangalore Diabetes Centre, India, 2Newcastle University, Newcastle upon Tyne, UK, 3GlaxoSmithKline, King of Prussia, USA. Background and aims: A 3-year, randomized, double-blind, multicenter study evaluated once weekly (QW) GLP-1 receptor agonist albiglutide (Albi) vs placebo (Pbo) and vs pioglitazone (Pio) in people with A1C 53-85.8 mmol/ mol [7.0-10.0 %] on background dual therapy of metformin and glimepiride. Materials and methods: Uptitration of Albi 30 to 50 mg QW and Pio 30 to 45 mg QD was allowed if needed. If hyperglycemic rescue was required, people remained in the study. Results: Baseline characteristics were similar between groups. At Week 52 (primary endpoint), Albi significantly decreased A1C vs Pbo but did not meet criteria for noninferiority vs Pio. Among people remaining in the study for 3-years (including those who received rescue medication), glycemic con- trol was enhanced on Albi and Pio and maintained on Pbo (figure). Weight loss (kg) was observed on Albi and Pbo (-0.6 and -0.6), but weight increased on Pio (5.9 kg). Adverse events were higher on Albi than Pbo and Pio for injection site reactions (14.4%, 5.2%, 5.1% of participants, respectively), diar- rhea (13.7%, 10.4%, 8.3%), and nausea (11.1%, 6.1%, 7.2%), but not vomiting (2.6%, 3.5%, 5.1%). The incidence of SAEs (14.4%, 18.3%, 17.3%) and AEs leading to withdrawal (8.1%, 10.4%, 10.5%) was lower on Albi than Pbo and Pio. Conclusion: Albi in triple therapy gives effective glucose lowering to 3-years, was well tolerated, and was associated with less weight gain than Pio. Clinical Trial Registration Number: NCT00839527 Supported by: Research funded by GSK 839 A network meta-analysis to compare once weekly dulaglutide versus other GLP-1 receptor agonists in patients with type 2 diabetes A. Padhiar1, J.C. Thompson1, J.N. Eaton1, N.S. Hawkins1, K. Norrbacka2, M. Reaney2, R. Shaginian3, K.S. Boye4, N. Varol2; 1ICON Health Economics, Oxford, 2Eli Lilly and Company, Windlesham, UK, 3Eli Lilly and Company, Houten, Netherlands, 4Eli Lilly and Company, Indianapolis, USA. Background and aims: A network meta-analysis (NMA) was performed to compare the efficacy of once weekly dulaglutide 1.5mg to other GLP-1 recep- tor agonists (GLP1 RAs) licensed for use in the USA/Europe for the treat- ment of hyperglycaemia in type 2 diabetes (T2D). Materials and methods: A systematic review was conducted in CENTRAL, EMBASE, MEDLINE and recent conference proceedings from ADA, EASD, IDF-WDC, IDF-WPR and AASD (1990-Jan 2014) to identify randomised controlled trials of adults with T2D treated with liraglutide (1.2mg QD, 1.8mg QD), exenatide (5μg BID, 10μg BID, 2mg QW), lixisenatide (20µg QD) and/or albiglutide (30mg QW, 50mg QW). The AWARD 1, 2, 5 and 6 trials provided data for once weekly dulaglutide 1.5mg. The primary end- point of the NMA was relative reduction in HbA1c. To adjust for potential heterogeneity, networks were stratified by background therapy: metformin monotherapy (add-on to MET) and metformin in combination with sulfony- lurea (SU) and/or thiazolidinediones (TZD) (add-on to MET ± SU ± TZD). Networks were further stratified by study duration of 16-36 weeks and 37-56 weeks to generate four networks. Baseline HbA1c was adjusted for using me- ta-regression. The NMA was performed within a Bayesian framework; both fixed and random effects models were assessed. Results: Baseline HbA1c adjusted results are presented as mean differences in HbA1c (%) change from baseline and 95% credible intervals (CrI) for once weekly dulaglutide 1.5mg versus other GLP1 RAs in Table 1. Random effects Diabetologia (2014) 57:[Suppl1]S1–S564 S 341 1 C models are presented where convergence was achieved; otherwise, fixed ef- fects models are presented. The 37-56 weeks add-on to MET network was disconnected; therefore a NMA could not be conducted. In the 16-36 week network meta-analyses, once weekly dulaglutide 1.5mg showed improved control of HbA1c (%) compared to exenatide 5µg BID and lixisenatide 20µg QD as add-on to both MET and MET ± SU ± TZD, and exenatide 10µg BID as add-on to MET ± SU ± TZD only with statistically significant reductions from baseline. In the 37-56 week add-on to MET ± SU ± TZD NMA, once weekly dulaglutide 1.5mg showed a significant reduction in HbA1c (%) com- pared to albiglutide 30mg QW and exenatide 10µg BID. A higher baseline HbA1c (%) was associated with greater reductions in HbA1c (%) but the effect was non-significant. Conclusion: In the AWARD clinical trial program, once weekly dulaglutide 1.5mg demonstrated significant improvements in HbA1c (%) reduction from baseline. In these NMAs, we have shown statistically significant improve- ments in HbA1c (%) control with once weekly dulaglutide 1.5mg in seven comparisons. The results in the remaining comparisons were not statistically significant. Once weekly dulaglutide 1.5mg was not statistically significantly inferior to other GLP1 RAs in any of the comparisons. Supported by: Eli Lilly and Company 840 EVIDENCE 1 P. Gourdy1, A. Penfornis2, G. Charpentier3, S. Madani4, L. Martinez5, E. Eschwège6, J.-F. Gautier7; 1Institut de Recherche sur les Maladies Métaboliques et Cardiovasculaires, Toulouse, France, 2University of Franche-Comté, Besançon, 3Centre Hospitalier Sud Francilien, Corbeil, 4Novo Nordisk, Paris, 5Université Pierre & Marie Curie, Paris, 6INSERM, Villejuif, 7Lariboisière Hospital, Paris, France. Background and aims: We report 2-year data on effectiveness and toler- ability of the GLP-1 analogue liraglutide in the EVIDENCE observational, postmarketing study. Materials and methods: EVIDENCE is a 2-year multicentre, observational, postmarketing outpatient study requested by the French National Health Au- thority in order to evaluate the efficacy and safety of liraglutide in clinical practice. Diabetologists and general practitioners in France recruited patients starting treatment with liraglutide. Patients and physicians completed ques- tionnaires at study entry, 3 months and 6 months, then at 6-month inter- vals for a further 18 months. The primary objective was to determine the percentage of patients still taking liraglutide and at HbA1c target ( S 342 1 C Conclusion: This study shows that a difference in baseline renal status did not have an effect on efficacy outcomes in lixisenatide- vs placebo-treated patients, and a uniform placebo-adjusted effect of lixisenatide was observed across renal categories. Supported by: Sanofi 842 Predicting treatment success with exenatide twice daily (EBID) at 6 mo: Are baseline postprandial glucose excursions important? J. Shaw1, B. Gallwitz2, G. Schernthaner3, J. Han4, J. Malloy5; 1Baker IDI Heart and Diabetes Institute, Melbourne, Australia, 2Eberhand-Karls-University, Tubergen, Germany, 3Rudolfstiftung Hospital, Vienna, Austria, 4AstraZeneca, 5Bristol-Myers Squibb, San Diego, USA. Background and aims: Information on predicting glycaemic treatment suc- cess in patients (pts) before initiation of specific therapies, particularly GLP- 1 receptor agonists, is limited. As EBID effectively suppresses postprandial blood glucose (BG) through mechanisms including decreased gastric empty- ing, we assessed whether a greater postprandial BG defect was predictive of treatment success for EBID and, conversely, whether a pt with a greater fast- ing BG defect was less likely to achieve treatment success with EBID. Materials and methods: A pooled dataset (24-30 wks) from 8 EBID stud- ies that collected self-monitored BG data was used. Treatment success at 6 mo was defined as HbA1c S 343 1 C 844 Sustained glycaemic control with exenatide once weekly versus insulin glargine: associations with baseline factors and early treatment response M.E. Trautmann1, J. Malloy2, J. Han3; 1Diabetes Research, Hamburg, Germany, 2Bristol-Myers Squibb, 3AstraZeneca, San Diego, USA. Background and aims: To assist decision making in type 2 diabetes mel- litus therapy, this post hoc analysis explored the association between base- line characteristics or early (≤12 week) treatment responses with sustained glycaemic control, comparing the glucagon-like peptide-1 receptor agonist exenatide once weekly (QW) with titrated insulin glargine (glargine). Materials and methods: Three-year data were analysed from the open-label DURATION-3 study, in which patients received exenatide QW (N=233) or glargine (N=223) in addition to metformin alone or with a sulphonylurea. Loss of, or inability to achieve, HbA1c control was defined as HbA1c >7% at 2 consecutive visits or HbA1c >9% at 1 visit after 26 weeks of treatment. Base- line characteristics considered were: age, gender, body mass index, weight, background glucose-lowering medication, disease duration, creatinine clear- ance, HbA1c, fasting serum glucose (FSG), Homeostasis Model Assessment (HOMA) -S (insulin sensitivity) and HOMA-B (β-cell function). Early treat- ment responses considered were HbA1c, FSG and weight. Cox proportional hazard models were used to analyse the intention to treat (ITT) population and logistic regression was used to analyse the 3-year Completer population (exenatide QW, N=140; glargine, N=147). Results: Of the ITT patients, 50% on exenatide QW versus 43% on glargine achieved the sustained glycaemic goal; of 3-year Completers, 43 versus 33% achieved the glycaemic goal. The median duration of HbA1c control was 25.0 versus 16.7 months for exenatide QW versus glargine (ITT; Kaplan-Meier; P=0.03). The hazard ratio (glargine/exenatide; Cox model) of inability to achieve the glycaemic goal was 1.4 (P=0.006). In 3-year Completers (both treatments), stepwise logistic regression showed that only baseline HbA1c and HOMA-B were significantly associated with sustained glycaemic con- trol. Regardless of therapy, every 1% higher baseline HbA1c (>7%) increased the risk of losing glycaemic control by 90% (P=0.0002); every 10% higher baseline HOMA-B decreased the risk of losing glycaemic control by 11% (P=0.0288). The rate of weight change in weeks 1-8 was the most relevant ear- ly treatment response predictor of HbA1c control (P=0.0640). The medium slope of weight change was -0.0308 kg/day for exenatide QW (-0.86 kg in 4 weeks) versus 0.0018 kg/day for glargine (0.05 kg in 4 weeks). The odds (glar- gine/exenatide QW) of losing glycaemic control if the same weight change was seen in both groups was 1.2 (weight loss) and 1.1 (weight gain), indicat- ing that glargine-treated patients were more likely to lose glycaemic control over time regardless of weight gain or loss. For a 1 kg decrease in weight over weeks 1-4 with exenatide QW, the odds of losing glycaemic control decreased by 30%; for a 1 kg increase in weight over weeks 1-4 with glargine, the odds increased by 29%. Conclusion: Despite continued glargine titration, more patients achieved sustained glycaemic control with exenatide QW. Baseline HbA1c and HO- MA-B significantly impacted the odds of sustained control. The analysis of early treatment responses from this study suggest that the rate of weight change following treatment initiation (weeks 1-8) may potentially predict the chance of patients achieving longer term sustained glycaemic control. Clinical Trial Registration Number: NCT00641056 Supported by: Bristol-Myers Squibb/AstraZeneca 845 Daily blood glucose variability with exenatide once weekly versus basal insulin in 3 randomised controlled trials J. Vora1, J. Malloy2, M. Zhou3, E. Hardy4, N. Iqbal5, M. Trautmann6; 1Royal Liverpool University Hospital, UK, 2Bristol-Myers Squibb, San Diego, 3Bristol-Myers Squibb, Hopewell, 4Astra Zeneca, Wilmington, 5Bristol- Myers Squibb, Princeton, USA, 6Diabetes Research, Hamburg, Germany. Background and aims: Glucose lowering therapy aims to improve overall glycaemic control in patients with type 2 diabetes mellitus. However, data on blood glucose variability with different treatments are scarce. Because ba- sal insulin mainly affects fasting plasma glucose and exenatide once weekly (QW) controls both fasting plasma glucose and postprandial plasma glucose, we hypothesized that daily blood glucose variability would be improved more by exenatide QW than by titrated basal insulin in randomised controlled tri- als. Materials and methods: After 26 weeks, 7-point self-monitored blood glu- cose profiles were assessed in 3 randomised controlled trials of exenatide QW: 2 versus insulin glargine (once daily at night) in global and Japanese populations; 1 versus insulin detemir (once daily at night [90%] or twice daily [10%]). All patients had failed ≥1 oral drug; Japanese patients discontinued sulfonylureas at randomisation. Outcomes assessed were mean amplitude of glucose excursion (MAGE; average daily mealtime glucose excursion), blood glucose range (average difference between maximum and minimum daily blood glucose) and blood glucose variability (average SD of daily blood glucose). Results: In all 3 studies, HbA1c reductions were greater with exenatide QW than basal insulin (−1.5 vs −1.3% P S 344 1 C 846 Relationship between changes in postprandial glucagon, patient characteristics, and response to lixisenatide as add-on to oral antidiabetics C.J. Bailey1, S.T. Azar2, L. Blonde3, D. Dicker4, M. Puig-Domingo5, F.G. Eliaschewitz6, E. Nikonova7, R. Roussel8, K. Sakaguchi9, L. Sauque-Reyna10, M. Nauck11; 1School of Life and Health Sciences, Aston University, Birmingham, UK, 2American University of Beirut Medical Center, Lebanon, 3Ochsner Medical Center, New Orleans, USA, 4Hasharon Hospital-Rabin Medical Center, Petach-Tikva, Israel, 5Universitat Autònoma de Barcelona, Badalona, Spain, 6CPClin – Clinical Research Center, São Paulo, Brazil, 7Global Medical Affairs, Diabetes Division, Sanofi, Bridgewater, USA, 8Université Paris- Diderot, France, 9Kobe University Graduate School of Medicine, Japan, 10Instituto de Diabetes Obesidad y Nutricion S.C., Cuernavaca, Mexico, 11Diabeteszentrum Bad Lauterberg, Germany. Background and aims: Lixisenatide, a once-daily prandial glucagon-like peptide-1 receptor agonist, reduces postprandial (PP) glycaemic excursions and HbA1c. We report an exploratory analysis of the GetGoal-M and S trials in patients with type 2 diabetes mellitus (T2DM) with different changes in PP glucagon levels in response to lixisenatide treatment. Materials and methods: Patients (n=423) were stratified by their change in 2 hour PP glucagon level between baseline evaluation and Week 24 of treat- ment with lixisenatide as add-on to oral antidiabetics (OADs) into groups of Greater Change (GC; n=213) or Smaller Change (SC; n=210) in plasma glucagon levels (median change -23.57 ng/L). ANOVA and Chi-squared tests were used for the comparison of continuous and categorical variables, respec- tively. Baseline and endpoint continuous measurements in each group were compared using paired t-tests. Results: Mean change from baseline in 2 hour PP glucagon levels for the GC vs SC groups was -47.19 vs -0.59 ng/L (p S 345 1 C 848 Durability of glycaemic response with dapagliflozin as add-on therapy in type 2 diabetes inadequately controlled with metformin; 4-year data versus glipizide S. Del Prato1, M. Nauck2, K. Rohwedder3, E. Johnsson4, S. Parikh5; 1University of Pisa, Italy, 2Diabetes Center, Bad Lauterberg, 3AstraZeneca, Wedel, Germany, 4AstraZeneca, Mölndal, Sweden, 5AstraZeneca, Wilmington, USA. Background and aims: Dapagliflozin (DAPA), a selective inhibitor of the sodium-glucose co-transporter 2, reduces plasma glucose by increasing renal glucose excretion. Its insulin-independent mechanism results in a low risk of hypoglycaemia in contrast to sulphonylureas. In a phase 3 trial of DAPA (≤ 10 mg/d) versus glipizide (≤ 20 mg/d) in patients (N=814) with type 2 dia- betes mellitus inadequately controlled by metformin (MET) alone (median 2000 mg/d), the change in HbA1c with DAPA was statistically noninferior to glipizide over 52 weeks. Further to these previously presented data on the primary end point we report the durability of glycaemic control over 4 years. Materials and methods: The durability of glycaemic control was evaluated by calculating the coefficient of failure (CoF; the slope of the regression line vs time) for HbA1c and fasting plasma glucose (FPG) from 18 weeks (end of titration period) to 208 weeks. DAPA and glipizide were down-titrated if medically indicated. The CoF was determined for 3 groups: patients who had at least 3 postbaseline values from after week 18 to week 208 (full analysis set), all patients who completed 208 weeks of treatment (all completers), and patients who completed 208 weeks of treatment but who were not down- titrated (completers not down-titrated). Patients who required rescue medi- cation were excluded from the analyses. Results: The CoF for DAPA was significantly lower compared with glipizide, irrespective of which glycaemic variable (HbA1c or FPG) or analysis set was employed (Table). For the full analysis set, the CoF for HbA1c showed a rise of 0.2%/year (95% CI, 0.1 to 0.3) with DAPA and a rise of 0.6%/year (95% CI, 0.5 to 0.7) with glipizide. Corresponding values for FPG were 0.4 mmol/L/ year (95% CI, 0.2 to 0.5) for DAPA and 0.8 mmol/L/year (95% CI, 0.5 to 1.1) for glipizide. More patients treated with DAPA completed the 208-week study than patients treated with glipizide (161 vs 141, respectively). Fewer DAPA- treated patients had their dose down-titrated (3 vs 42, respectively; Table). For all completers and for completers who were not down-titrated, the CoFs for HbA1c and FPG were significantly lower with DAPA compared with glipizide (Table). Thus, the higher rates of discontinuation or down-titration with glipizide did not account for the lower CoF with DAPA. Treatment with DAPA in combination with MET was well tolerated over the 208-week pe- riod. Conclusion: Glycaemic durability over a 4-year period was significantly bet- ter with DAPA compared with glipizide as demonstrated by the CoF. DAPA in combination with MET was well tolerated. Clinical Trial Registration Number: NCT00660907 Supported by: AstraZeneca / BMS 849 Empagliflozin twice daily versus once daily as add-on to metformin in patients with type 2 diabetes S. Ross1, C. Thamer2, J. Cescutti3, T. Meinicke4, H.J. Woerle4, U.C. Broedl4; 1University of Calgary, LMC Endocrinology Centres, Canada, 2Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany, 3Boehringer Ingelheim France S.A.S., Reims, France, 4Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany. Background and aims: A randomized, double-blind, placebo-controlled, parallel-group study compared the efficacy and safety of empagliflozin (EMPA) twice daily (bid) versus once daily (qd) regimens as add-on to met- formin (MET) in patients with type 2 diabetes (T2DM). Materials and methods: Patients (n=983) were randomised to EMPA 12.5 mg bid (n=219), EMPA 25 mg qd (n=218), EMPA 5 mg bid (n=219), EMPA 10 mg qd (n=220) or placebo (PBO; n=107) added on to stable dose MET bid (≥1500 mg/day) for 16 weeks. The primary endpoint was change from base- line in HbA1c at week 16. The secondary endpoint was change from baseline in fasting plasma glucose (FPG) at week 16. Exploratory endpoints included changes from baseline in body weight, systolic and diastolic blood pressure (SBP and DBP) at week 16. The primary analysis was a test of non-inferiority of EMPA 12.5 mg bid versus EMPA 25 mg qd and of EMPA 5 mg bid versus EMPA 10 mg qd. The superiority of EMPA doses versus placebo was also test- ed. Efficacy was evaluated in 965 patients (mean [SD] age 58.2 [10.3] years; weight 89.0 [18.5] kg; BMI 31.8 [5.2] kg/m2; HbA1c 7.77 [0.80] %). Results: Reductions from baseline in HbA1c were non-inferior for EMPA 12.5 mg bid versus 25 mg qd and for EMPA 5 mg bid versus 10 mg qd. Reductions from baseline in HbA1c were statistically significant with all EMPA doses compared with PBO. FPG changes with EMPA were consistent with HbA1c changes. EMPA qd and bid doses reduced body weight, SBP and DBP to a similar extent. Adverse events (AEs) were reported by 45.7%, 41.7%, 43.8%, 50.0% and 47.7% of patients on EMPA 12.5 mg bid, EMPA 25 mg qd, EMPA 5 mg bid, EMPA 10 mg qd and PBO, respectively. Confirmed hypoglycaemic AEs ( S 346 1 C 850 Systematic review and network meta-analysis to compare dapagliflozin with other diabetes medications in combination with metformin for adults with type 2 diabetes G. Wygant1, A.H. Barnett2, M.E. Orme3, P. Fenici4, R. Townsend5, M. Roudaut6; 1Bristol-Myers Squibb, Princeton, USA, 2Diabetes Centre, Birmingham, 3ICERA Consulting Ltd, Swindon, UK, 4Bristol-Myers Squibb (at the time of research), Paris, France, 5AstraZeneca (at the time of research), Brussels, Belgium, 6Bristol-Myers Squibb, Paris, France. Background and aims: Effective management of type 2 diabetes mellitus (T2DM) and its complications is required in order to address the growing burden of the disease on healthcare resources. However, there are many fac- tors to consider when assessing the suitability of available treatments. A net- work meta-analysis (NMA) update was undertaken to evaluate the sodium glucose co-transporter-2 (SGLT-2) inhibitor, dapagliflozin, versus other dia- betes medications as add-on to metformin. This update allowed inclusion of a new drug class (glucagon-like peptide-1 [GLP-1] analogues), a new time- point (24-weeks) and covariate analysis. Materials and methods: The systematic review identified recent randomised controlled trials (2011-July 2013) involving T2DM patients inadequately con- trolled on metformin monotherapy. Comparators licensed in the EU includ- ed dipeptidyl peptidase-4 inhibitors (DPP-4i), thiazolidinediones (TZDs), glucagon-like peptide-1 (GLP-1s) analogues, sulfonylureas (SUs) and dapa- gliflozin. Bayesian NMA was conducted for outcomes at 24- and 52-weeks, and both fixed and random-effect models were explored. Covariate analyses were performed to assess confounding baseline parameters where feasible. Results: Of 2247 articles retrieved, 16 were included in the review. Combined with 19 from the pre-2011 analysis, a total of 19 and 8 studies were included in the 24-week and 52-week basecase NMA, respectively. There were suffi- cient data to analyse mean change in HbA1c, systolic blood pressure (SBP), change in weight, as well as the proportion of patients experiencing hypo- glycaemia at both 24 and 52-weeks. There were no significant differences in change in HbA1c or SBP between dapagliflozin and the other classes of diabetes drugs as add-on to metformin at 24 or 52-weeks. Significant weight loss results were seen by 24-weeks for dapagliflozin compared to DPP-4i and TZD and at 52-weeks for dapagliflozin compared to SU, DPP-4i and TZD: -4.66 kg (-6.43, -2.90); -2.59 kg (-4.53, -0.66) and -4.76 kg (-7.28, -2.24) at 52 weeks, respectively. Dapagliflozin resulted in a significantly lower risk of hypoglycaemia compared to SU (OR: 0.05 [0.01, 0.19]) over 52-weeks. Co- variate analysis indicated that a higher HbA1c at baseline was predictive of a larger treatment effect. Conclusion: This NMA update supports previous findings that effects on HbA1c are largely similar between drug classes. However, dapagliflozin compared with DPP-4i, TZDs and SUs offered superior weight control when added to metformin and was associated with a significantly reduced risk of hypoglycaemia in comparison to SUs. The wider evidence base compared to previous analysis increases the confidence in the results. Supported by: BMS and AstraZeneca 851 Fixed dose combinations of empagliflozin and linagliptin for 52 weeks in drug-naïve subjects with type 2 diabetes A. Lewin1, R. DeFronzo2, S. Patel3, D. Liu4, R. Kaste4, H.J. Woerle5, U.C. Broedl5; 1National Research Institute, Los Angeles, 2University of Texas Health Sciences, San Antonio, USA, 3Boehringer Ingelheim Ltd., Bracknell, Berkshire, UK, 4Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, USA, 5Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany. Background and aims: A randomized, double-blind, parallel group Phase III study evaluated the efficacy and safety of fixed dose combinations (FDCs) of empagliflozin/linagliptin (EMPA/LINA) in drug-naïve subjects with type 2 diabetes (T2DM). Materials and methods: Subjects were randomized to EMPA 25 mg/LINA 5 mg (n=137), EMPA 10 mg/LINA 5 mg (n=136), EMPA 25 mg (n=135), EMPA 10 mg (n=134), or LINA 5 mg (n=135) for 52 weeks. Primary analysis was at week 24. Exploratory endpoints at week 52 were changes from baseline in HbA1c, body weight, systolic and diastolic blood pressure (SBP and DBP), and percentage of subjects with baseline HbA1c ≥7% who reached HbA1c S 347 1 C Conclusion: In drug-naïve subjects with T2DM, FDCs of EMPA 25 mg/ LINA 5 mg and EMPA 10 mg/LINA 5 mg for 52 weeks significantly reduced HbA1c vs LINA 5 mg. HbA1c reductions were significantly greater with EMPA 10 mg/LINA 5 mg vs EMPA 10 mg, but not with EMPA 25 mg/LINA 5 mg vs EMPA 25 mg. FDCs were well tolerated, with overall safety profiles similar to those known for the individual components. Clinical Trial Registration Number: NCT01422876 Supported by: Boehringer Ingelheim and Eli Lilly 852 Improvement in glycaemic control and reduction in body weight over 52 weeks with dapagliflozin as add-on therapy to metformin plus sulphonylurea S. Matthaei1, K. Bowering2, K. Rohwedder3, A. Grohl4, E. Johnsson5; 1Diabetes-Zentrum Quakenbrück, Germany, 2University of Alberta, Edmonton, Canada, 3AstraZeneca, Wedel, 4Aptiv Solutions, Köln, Germany, 5AstraZeneca, Mölndal, Sweden. Background and aims: Dapagliflozin (DAPA), a highly selective sodium- glucose co-transporter 2 inhibitor, improves glycaemic control and promotes body weight loss in patients with type 2 diabetes mellitus (T2DM) as mono- therapy or in combination with other antihyperglycaemic drugs. This phase 3 study evaluated DAPA as an add-on therapy to metformin (MET) plus sul- phonylurea (SU). Materials and methods: Patients (men and women aged ≥ 18 years) with T2DM (HbA1c ≥ 7.0%-≤ 10.5% at randomisation) who were receiving a stable (≥ 8 weeks prior to enrolment) dose combination of MET (≥ 1500 mg/d) and SU (maximum tolerated dose of at least half maximum dose) were eligible to participate. Patients were randomised to receive DAPA 10 mg or placebo (PBO) once daily for 52 weeks (24-week double-blind period and 28-week extension). Results: At week 52, HbA1c and fasting plasma glucose (FPG) improved with DAPA versus PBO (HbA1c -0.8% vs -0.1%; FPG -1.5 mmol/L [-27.6 mg/dL] vs 0.6 mmol/L [11.5 mg/dL]; Table). Over 52 weeks more patients achieved the American Diabetes Association recommended glycaemic goal of HbA1c < 7.0% with DAPA (27.3%) versus PBO (11.3%). Both body weight and sys- tolic blood pressure were reduced with DAPA versus PBO. Total, low-density lipoprotein, and high-density lipoprotein cholesterol increased and triglycer- ides decreased from baseline to week 52 with DAPA versus PBO. No patient discontinued as a result of a lack of glycaemic control and fewer patients on DAPA (10.1%) than on PBO (42.7%) were rescued for failing to reach glycae- mic targets (week 4-16, FPG >13.2 mmol/L (240 mg/dL); week 16-24, FPG >11.1 mmol/L (200 mg/dL); week 24-52, HbA1c > 8.0%). In DAPA versus PBO groups the frequency of adverse events (AEs) was 69.7% versus 73.4%, respectively; serious AEs were 6.4% versus 7.3%, and hypoglycaemic events were 15.6% versus 8.3% (1 event of hypoglycaemia led to discontinuation in the DAPA treatment arm). Genital infections were reported by 10.1% versus 0.9% of patients with DAPA versus PBO (women 14.3% vs 2.0%; men 4.3% vs 0%). Urinary tract infections were reported by 10.1% versus 11.0% of patients with DAPA versus PBO (women 12.7% vs 22.4%; men 6.5% vs 1.7%). Conclusion: DAPA 10 mg resulted in sustained glycaemic benefits and weight loss over the duration of this 52-week study in patients with T2DM and inadequate glycaemic control on a background combination of MET and SU. Clinical Trial Registration Number: NCT01392677 Supported by: AZ/BMS 853 Efficacy and safety of canagliflozin in patients with type 2 diabetes mellitus who were, or were not, on antihyperglycaemic agents at screening G. González-Gálvez1, E. Jodar2, K. Kim3, M. Alba4, C. Tong4, G. Meininger4; 1Instituto Jalisciense de Investigacion en Diabetes y Obesidad, Guadalajara, Mexico, 2Department of Endocrinology, Hospital Quiron, Madrid, Spain, 3Department of Internal Medicine, Dongguk University Ilsan Hospital, Dongguk University School of Medicine, Goyang, Republic of Korea, 4Janssen Research & Development, LLC, Raritan, USA. Background and aims: Canagliflozin (CANA), a sodium glucose co-trans- porter 2 (SGLT2) inhibitor, demonstrated glycaemic improvement and was generally well tolerated in a Phase 3 study of patients with type 2 diabetes mellitus (T2DM) inadequately controlled on diet and exercise. In that study, the efficacy and safety of CANA monotherapy compared with placebo (PBO) were evaluated in patients who were, or were not, on antihyperglycaemic agents (AHAs) at screening. Materials and methods: In this randomised, double-blind, PBO-controlled study, patients with T2DM inadequately controlled with diet and exercise (N = 584; age, 55.4 y; HbA1c, 8.0%; fasting plasma glucose [FPG], 9.5 mmol/L; body mass index, 31.6 kg/m2) received CANA 100 or 300 mg or PBO once daily. Change in HbA1c from baseline to Week 26 was assessed in the overall modified intent-to-treat population and subsets of patients who were not on AHAs at screening (n = 303), or on AHAs (oral AHA monotherapy [except PPAR-γ agonists] or low-dose combination metformin + sulphonylurea) and entered the AHA washout period (n = 281). Results: Patients on AHAs at screening had a higher baseline mean age and T2DM duration, lower mean estimated glomerular filtration rate (eGFR), and higher HbA1c and FPG. Relative to PBO, CANA 100 and 300 mg low- ered HbA1c in patients not on AHAs (differences of -0.72% and -1.05%) and on AHAs at screening (differences of -1.11% and -1.29%; Figure) at Week 26; changes were similar to those in the overall population. The overall incidence of adverse events (AEs) was higher with CANA 100 and 300 mg than PBO in patients not on AHAs (59% and 52% vs 46%) and on AHAs at screening (62% and 68% vs 52%), but the incidence of serious AEs and AEs leading to discontinuation were low across groups in both subsets. Diabetologia (2014) 57:[Suppl1]S1–S564 S 348 1 C Conclusion: CANA treatment lowered HbA1c and was generally well toler- ated in patients with T2DM who were, or were not, on prior AHA therapy at screening. Clinical Trial Registration Number: NCT01081834 Supported by: Janssen Research & Development, LLC 854 Canagliflozin reduces both HbA1c and body weight in patients with type 2 diabetes mellitus on background metformin G. Langslet1, J.A. Davidson2, V. Valentine3, U. Vijapurkar4, W. Canovatchel4, G. Meininger4; 1Lipid Clinic, Oslo University Hospital, Norway, 2Touchstone Diabetes Center, The University of Texas Southwestern Medical Center, Dallas, 3Northside Family Medicine, Albuquerque, 4Janssen Research & Development, LLC, Raritan, USA. Background and aims: Canagliflozin (CANA), a sodium glucose co-trans- porter 2 (SGLT2) inhibitor, is approved for the treatment of adults with type 2 diabetes mellitus (T2DM). Reductions in HbA1c and body weight (BW) have been observed with CANA in 2 studies as add-on to metformin over 52 weeks. This post hoc analysis describes the distribution of change from baseline in HbA1c and BW in these studies at Week 52. Materials and methods: This analysis used data from 2 separate randomised, double-blind, active-controlled Phase 3 studies of CANA 100 and 300 mg versus either sitagliptin (SITA) 100 mg (Study 1: N = 1,284; HbA1c, 7.9%; BW, 87.2 kg) or glimepiride (GLIM; Study 2: N = 1,450; HbA1c, 7.8%; BW, 86.6 kg). Study 1 included a treatment arm where patients received placebo during the 26-week core period and switched to SITA during the 26-week extension period (n = 183); this arm was excluded from the current analysis. Results: In Study 1, least squares (LS) mean HbA1c changes from baseline were −0.73%, −0.88%, and −0.73%, and LS mean BW changes from baseline were −3.8%, −4.2%, and −1.3% with CANA 100 and 300 mg and SITA 100 mg, respectively. In Study 2, LS mean HbA1c changes from baseline were −0.82%, −0.93%, and −0.81%, and LS mean BW changes from baseline were −4.2%, −4.7%, and 1.0% with CANA 100 and 300 mg and GLIM, respec- tively. A greater proportion of patients had reductions in both HbA1c and BW with CANA 100 and 300 mg compared with SITA 100 mg (67.7%, 74.7%, and 50.6%, respectively) or GLIM (72.4%, 78.5%, and 26.8%, respectively; Figure). More patients had a decrease in HbA1c with either no change or an increase in BW with SITA or GLIM versus CANA 100 and 300 mg in both studies. Conclusion: CANA provided reductions in both HbA1c and BW versus SITA 100 mg or GLIM in most patients with T2DM as add-on to metformin at 52 weeks. Clinical Trial Registration Number: NCT01106677, NCT00968812 Supported by: Janssen Research & Development, LLC 855 Glycaemic efficacy of canagliflozin is largely independent of baseline beta cell function or insulin sensitivity D. Matthews1, B. Zinman2, C. Tong3, G. Meininger3, D. Polidori4; 1The Oxford Centre for Diabetes, Endocrinology and Metabolism, UK, 2Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, University of Toronto, Canada, 3Janssen Research & Development, LLC, Raritan, USA, 4Janssen Research & Development, LLC, San Diego, USA. Background and aims: Canagliflozin (CANA), a sodium glucose co-trans- porter 2 (SGLT2) inhibitor, lowers plasma glucose and HbA1c in patients with type 2 diabetes mellitus (T2DM) by increasing urinary glucose excre- tion. Because CANA’s mechanism of action is insulin-independent, it was hypothesised that the glycaemic efficacy of CANA would be independent of baseline beta-cell function (BCF) or insulin sensitivity (IS). This hypothesis was tested using data from 4 placebo (PBO)- and 2 active-controlled studies to measure BCF (HOMA-%B) and IS (HOMA2-%S). Materials and methods: The first analysis pooled data from 26-week PBO- controlled studies (N = 2,313; baseline HbA1c = 8.0%, HOMA2-%B = 49, HOMA2-%S = 60). Patients were divided into tertiles of HOMA2-%B or HO- MA2-%S and, given the potential impact of baseline HbA1c, further divided based on baseline HbA1c (HbA1c S 349 1 C with the biggest differences seen in patients with higher baseline HbA1c and in patients with lower HOMA2-%S (Figure B). Conclusion: Consistent with expectations based on the insulin-independent mechanism of action of CANA, the glycaemic efficacy of CANA is largely independent of baseline measures of BCF or IS. Supported by: Janssen Research & Development, LLC 856 Effects of sodium glucose co-transporter 2 inhibitors in patients with type 2 diabetes: a systematic review with meta-analysis of randomised clinical trials H. Storgaard1, L.L. Gluud2, M. Christensen1,3, F. Knop1, T. Vilsbøll1; 1Diabetes Research Division, Department of Medicine, Gentofte Hospital, University of Copenhagen, Hellerup, 2Gastrounit, Hvidovre Hospital, University of Copenhagen, 3Department of clinical pharmacology, Bispebjerg Hospital, University of Copenhagen, Denmark. Background and aims: Several randomised clinical trials (RCTs) have as- sessed the effects of sodium glucose co-transporter 2 inhibitors (SGLT-2i) for patients with type 2 diabetes. We performed a systematic review with meta- analyses of RCTs on SGLT-2i for ≥12 weeks. The primary outcomes were changes in HbA1c, body weight and adverse events. Materials and methods: Random effects meta-analysis was performed. Bias and heterogeneity were assessed in subgroup, sensitivity, regression and se- quential analyses. We included 24 placebo-controlled RCTs on 12-102-week treatment with canagliflozin 300 mg (7 trials), dapagliflozin 10 mg (13 trials) or empagliflozin 25 mg (4 trials) and 7 trials with active controls. Results: Random effects meta-analysis of 3,425 patients randomised to SGLT-2i and 3,234 patients randomised to placebo found that SGLT-2i re- duced HbA1c (mean difference -0.71%; CI -0.80 to -0.63) and body weight (-2.0 kg; -2.2, -1.9). The analysis was stable to assessments of bias and the results were confirmed in sequential analyses. Beneficial effects in favour of SGLT2i on fasting plasma glucose (-1.6 mM; CI -1.9 to -1.4), systolic blood pressure: (-4.4 mmHg; -5.2 to -3.6) and diastolic blood pressure: (-1.6 mmHg; CI -2.1 to -1.1) were detected. Analyses on adverse events showed that 6% of patients randomised to SGLT-2i developed genital or urinary tract infections (relative risk 1.7; CI 1.4 to 2.1); number needed to harm 50). Conclusion: We conclude that SGLT-2i have clinically relevant effects on glycaemic control and body weight and should be considered as second line treatment after metformin in obese patients with type 2 diabetes. The risk benefit balance may not be favourable in patients prone to genital or urinary tract infections. PS 066 Novel therapies 857 Pharmacokinetics of once weekly dulaglutide in patients with type 2 diabetes mellitus A. de la Peña1, C. Loghin2, X. Cui3, X. Zhang1, J.S. Geiser1, J. Martin1, J.Y. Chien1; 1Eli Lilly and Company, 2Clinical Pharmacology, Eli Lilly and Company, Indianapolis, 3inVentiv Health Clinical, Burlington, USA. Background and aims: The aim of these analyses was to characterize the pharmacokinetics (PK) of once weekly dulaglutide in patients with type 2 diabetes mellitus (T2DM). Materials and methods: PK meta-analysis of Phase 1 data and population PK modeling of Phase 2 and Phase 3 studies were used in order to perform this characterization. Results: The PK of dulaglutide was well described by a 2-compartment model with first-order absorption. After subcutaneous (SC) administration, dula- glutide was slowly absorbed. Time to reach maximum concentration (Cmax) at steady state ranged from 24 to 72 hours (median = 48 hours); absolute bioavailability was 47% and half-life was 4.7 days. Steady state was achieved between 2 and 4 weeks of dosing. At steady state for dulaglutide 1.5 mg, the mean peak (Cmax) and total area under the concentration-time curve (AUC) exposures of dulaglutide were 114 ng/mL (range 56 to 231 ng/mL) and 14000 ng·h/mL (range 6940 to 26000 ng·h/mL), respectively; the accumulation ratio was approximately 1.56. Intra-patient variability was S 350 1 C participants treated with carnitine orotate complex showed increase in mean liver attenuation values and liver attenuation index (LAI) after 12 weeks treatment(p S 351 1 C diet-induced and age-associated hepatic insulin resistance. Here, we exam- ined the impact of selective hepatic knockdown of mammalian Indy protein (mINDY) expression using anti-sense oligonucleotides (ASOs). Materials and methods: We studied the effect of mINDY knockdown on hepatic glucose metabolism in 4 week high fat fed rats (n=15 per group) as- sessed by hyperinsulinemic-euglycemic (HEC) clamp studies. Results: After 4 weeks of ASO treatment, mINDY mRNA expression was re- duced by 91% (P S 352 1 C Materials and methods: We studied 6 patients with lypodystrophy, 4 gener- alized (2 congenital generalized lypodystrophy CGL, and 2 acquired general- ized lypodystrophy, AGL) and the other 2 with familial partial lypodystro- phy (FPL). We determined basal, and after 3 and 6 month of leptin therapy: Creatinine, urea, microalbuminuria, proteinuria and % HbA1c. We evaluate which patients were on Angiotensin-converting enzyme (ACE) inhibitors therapy. In 2 of the 6 patients renal biopsy was done. We performed percuta- neous renal biopsies in 2 patients. Results: HbA1c and proteinuria improved after leptin therapy. Patient 2 didn’t develop proteinuria. Four patients have proteinuria at baseline and af- ter leptin improved significantly (patients 3, 5 and 6) or even turned negative (patient 1). Patients 1,3 and 5 were on enalapril. Renal Biopsy: patient 1 has diagnosis of mesangial proliferative glomeruloesclerosis and Patient 3 devel- oped focal segmental glomeruloesclerosis (FSGS) Conclusion: Renal function was evaluated in a group of patients with lipod- ystrophy during the course of a therapeutic trial of recombinant human lep- tin. It was describe that in some cases treated with recombinant human leptin there was an exacerbation of their underlying renal disease. In our patients we demonstrated efficacy of leptin to improved renal disease. Clinical Trial Registration Number: NCT00025883 865 Ranolazine induces skeletal muscle hypertrophy activating Ca2+/ calmodulin pathway I. Terruzzi1, P. Senesi2,3, A. Montesano2, L. Luzi2,3; 1Div. Metabolic and Cardiovascular Sciences. Metabolism, Nutrigenomics and Cellular Differentiation, San Raffaele Scientific Institute, 2Department of Biomedical Sciences for Health, University of Milan, 3Metabolism Research Centre and Department of Endocrinology and Metabolic Diseases, San Donato Hospital and Scientific Institute, Milan, Italy. Background and aims: The underlying causes of type 2 diabetes are a com- bination of a defect in insulin secretion from pancreatic β cells and an im- pairment in insulin-mediated glucose disposal (insulin resistance) in insulin target tissues, such as skeletal muscle. Ranolazine (RAN) is a novel anti-is- chemic and antianginal drug that reduces angina frequency in patients with chronic stable angina, a manifestation of coronary artery disease. RAN de- creases myocardial ischemia by improving sodium-calcium homeostasis via inhibition of the late phase of the inward sodium current (late INa). Recently, RAN has been shown to lower haemoglobin A1c (HbA1c) in patients with diabetes mellitus. The mechanism by which RAN improves glycaemic control is unknown. Previous studies in isolated human and murine β cell showed that RAN preserves pancreatic cells mass and promotes glucose-stimulated insulin secretion. However, RAN hypoglycaemic mechanisms in skeletal muscle is jet not studied. Objective of the present study was to determine the effect of RAN on skeletal muscle metabolism and differentiation. Materials and methods: We examined RAN action on skeletal muscle using C2C12 murine myoblastic cell. After treatment of C2C12 with 1, 10 or 25µM RAN for 24 hours to study the RAN dose-response relationship: 10µM RAN was considered the dose able to stimulate morphological changes and hyper- trophic process in neo-formed myotubes. 10µM RAN was added to C2C12 during proliferation, differentiation and neo formed myotubes to assess RAN possible interference in insulin pathway. Results: RAN enhanced myoblasts proliferative capacity decreasing the ex- pression of cell cycle inhibitors (p21 protein/cyclins). Interestingly, RAN did not modify p70S6 kinase activation showing it did not modulate the classic signaling pathways involved in cell growth. Western and Immunofluores- cence studies revealed that during differentiation, RAN improved neo myo- tubes formation, but did not stimulate kinases involved in skeletal muscle differentiation and glucose uptake (ERKs and AKT pathways). Neo formed myotubes analysis showed that RAN activated Ca2+/calmodulin dependent protein kinase (CaMKII), which plays an important role in GLUT4 upregula- tion and in myotube differentiation, confirming that RAN improved skeletal muscle metabolism a without affecting p70S6 and ERKs pathway. Conclusion: Taken together, our results demonstrate that RAN have a posi- tive action on skeletal muscle cells differentiation and metabolism, activating not AKT insulin mediated pathway, but Ca2+/calmodulin signaling pathway. This finding provides interesting evidence on the innovative use of RAN in diabetic condition, not only to improve insulin secretion of β cell but also insulin sensitivity in skeletal muscle. 866 Lipolytic and insulinotropic effects of HM12525A, a novel long-acting GLP-1/glucagon dual agonist S. Jung1, Y. Park1, J. Kim1, J. Lee1, Y.-M. Lee1, Y. Kim1, J. Kang1, M. Trautmann2, M. Hompesch2, S. Kwon1; 1Hanmi Parmaceutical, Seoul, Republic of Korea, 2Profil Institute, Chula Vista, USA. Background and aims: Oxyntomodulin, an alternative cleavage product of proglucagon, is a gut hormone which shows enhanced body weight loss and improved glycemic control by activating GLP-1 (GLP-1R) and glucagon re- ceptor (GCGR), respectively. But its clinical application is limited due to a short half-life. We have developed the long acting GLP-1/glucagon dual ago- nist, HM12525A, by conjugating a novel GLP-1/glucagon dual agonist with constant region of human immunoglobulin via non-peptidyl linker. In a pre- vious study, we demonstrated that once weekly administration of HM12525A exerted potent body weight loss and improved glycemic control in obese and diabetic animal models. However, the underlying modes of action for HM12525A are still poorly defined. The aim of this study was to investigate the molecular basis for the beneficial effects of HM12525A in vitro and in vivo system. Materials and methods: To evaluate the effect of HM12525A on lipid droplet formation, 3T3-L1 cells were incubated with HM12525A during adipogenic differentiation. After 10~14days, the cells were stained with Oil-red O to determine triglyceride content. Fully differentiated 3T3-L1 adipocytes were treated with HM12525A, and the phosphorylation of hormone-sensitive li- pase (HSL) was evaluated through western blot analysis. To measure lipolytic activity, conditioned medium was collected following HM12525A treatment for 4hr, and the amount of glycerol was quantified. Serum concentration of insulin in normal mice during ipGTT was quantified to evaluate the insulino- tropic and insulin sensitizing potency of HM12525A. Results: Consistent with in vivo results in which HM12525A administra- tion significantly reduced the fat mass of diet-induced obesity (DIO) mice, HM12525A inhibited the lipid droplet formation in 3T3-L1 adipcoytes in a dose-dependent manner. Of note, these inhibitory effects were partially re- served by either GLP-1R or GCGR antagonist, suggesting that the dual ago- nism of HM12525A synergistically exerts its lipolytic action. At the molecu- lar level, phosphorylation of HSL, a key marker for the activation of lipolysis, and the following glycerol release were increased upon HM12525A treatment in 3T3-L1 adipocytes. As to the insulinotropic effects in pancreatic β-cells, HM12525A increased insulin secretion in RINm5F cells. In line with in vitro results, HM12525A administration significantly increased insulin secretion as well as insulin sensitivity, thereby attenuating the glucose excursion during ipGTT in normal mice. Conclusion: Our results indicate that a well balanced dual agonism of HM12525A mediates synergistic effects on lipolysis of adipocytes. Moreover, HM12525A improves glucose tolerance by enhancing both insulin secretion and insulin sensitivity in β-cells and normal mice. Therefore, our results col- lectively demonstrate that a novel GLP-1/glucagon dual agonist HM12525A mediated lypolytic and insulinotropic effects through which anti-obesity and anti-diabetic potentials were conferred. Diabetologia (2014) 57:[Suppl1]S1–S564 S 353 1 C PS 067 Non-glycaemic effects of DPP4 inhibitors 867 Renoprotective effect of gemigliptin, dipeptidyl peptidase 4 inhibitor, on streptozotocin-induced type 1 diabetic model M.-K. Kim1, Y.-A. Jung1, G.-S. Jung2, Y.-K. Choi2, K.-H. Bae2, I.-K. Lee2, K.-G. Park2, J.-E. Lee3, E. Jung4, M.-J. Kim5, E. Kim6; 1Department of Internal Medicine, Keimyung University School of Medicine, 2Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, 3Department of Internal Medicine, CHA Gumi Medical Center, Gumi, 4Department of Internal Medicine, Catholic University of Daegu School of Medicine, 5Pooren Meerae Internal Medicine Clinic, 6Department of Internal Medicine, Daegu Fatima Hospital, Republic of Korea. Background and aims: Gemigliptin is a dipeptidyl peptidase 4 (DPP4) in- hibitor, which is currently used in the treatment of patients with type 2 dia- betes through augmentation of GLP-1 activity. In addition to GLP-1 activity, growing body of evidence shows that DPP4 play various roles in metabolism and inflammation through enzymatic activity and non-enzymatic activity. Some studies reported that DPP 4 activity increased in the urine and kidney of the diabetic patients and inhibition of DPP 4 activity attenuated diabetic neuropathy, retinopathy and renal ischemia-reperfusion injury. Therefore, the aim of this study was to determine whether gemigliptin has renoprotec- tive effects on the kidney of streptozotocin (STZ)-induced type 1 diabetic mice model. Materials and methods: Diabetes was induced by single intraperioneal in- jectin of streptozotocin (150 mg/kg/body weight). Diabetic mice was treated without or with an oral dose of gemigliptin 300mg/kg/ day for 8 weeks. Re- nal injury was observed by electron microscopy and light microscopy. We also measured serum glucose and urinary albumin excretion and evaluated fibrotic markers using immunohistochemical staining,qRT-PCR and western blot analysis. Results: Blood glucose was significantly higher in diabetic mice than control mice, and gemigliptin did not reduce blood glucose levels of STZ-induced diabetic mice. Diabetic mice exhibited marked increased kidney/body weight ratio and urinary albumin excretion, but gemigliptin treatment significantly reduced kidney/body weight ratio and albuminuria. Moreover, gemigliptin treatment significantly reduced glomerular basement membrane thickness of kidney compared with STZ-induced diabetic mice. Histological examina- tion showed renal fibrosis was induced by STZ, but gemigliptin treatment significantly attenuated STZ-induced renal fibrosis. In addition, immunhis- tochemical staining showed decrease of type 1 collagen and fibronectin in the kidney of STZ-induced diabetic mice treated with gemigliptin. To test whether attenuation of renal fibrosis by gemigliptine is GLP-1 dependent, we examined the effects of gemigliptin on TGF-β-stimulated fibrotic gene expression in cultured renal cells. The results showed that gemigliptin-treated RMCs and NRK-52E cells showed markedly decrease in type 1 collagen and fibronectin expression through inhibition of smad3 activity. Conclusion: In conclusion, our data showed that gemigliptin has renoprotec- tive effect on diabetic nephropathy regardless of glucose lowering effect. The present study raises the possibility that gemigliptin could be used to prevent the progress of diabetic nephropathy including patients with type 1 diabetes. 868 Pituitary adenylate cyclase-activating polypeptide, a substrate of DPP-4, protects glomerular podocytes from inflammatory injuries K. Sakamoto, M. Takemoto, K. Kuno, T. Ishikawa, P. He, R. Ishibashi, K. Kobayashi, H. Kawamura, H. Tokuyama, Y. Maezawa, K. Yokote; Department of Clinical Cell Biology and Medicine, Chiba University Graduate School of Medicine, Japan. Background and aims: Diabetic nephropathy (DN) is a leading cause of end-stage kidney disease; however, to date, there are few available treatment options. Although the mechanisms of DN are not well understood, inflam- mation plays a crucial role in the initiation and/or progression of DN. Di- peptidyl peptidase-4 inhibitors (DPP4i) have recently been introduced for use as oral hypoglycemic agents. In addition to reducing blood glucose lev- els, DPP4i have been found to possess pleiotropic actions, e.g. protecting the kidney from injuries. Although the mechanisms underlying the pleiotropic effects of DPP4i are not clear, it has been speculated that substrates, including incretins, which are stabilized by DPP4i may play a role in these effects. Pi- tuitary adenylate cyclase-activating polypeptide (PACAP) is a neuropeptide, which was originally isolated from the ovine hypothalamus and has diverse biological functions. PACAP is known to be one of the substrates of DPP4. Furthermore, in studies of kidney pathology, PACAP has been found to have renoprotective effects. However, the specific cell types within the kidney that are protected by PACAP have not yet been identified. In the present study, we investigated the effects of PACAP on kidney cells and the mechanisms by which PACAP decreases the expression of inflammatory cytokines. Materials and methods: We used immunohistochemistry (IHC), western blotting (WB), and real-time polymerase chain reaction (RT-PCR) to evalu- ate the expressions of PACAP receptors in kidney tissue, isolated glomeruli and cultured podocytes. Next, we evaluated if PACAP induces the phospho- lylation of cAMP response element-binding protein (CREB) in cultured po- docytes using WB. In addition, podocytes were stimulated with lipopolysac- charide (LPS) and PACAP, and the effects of PACAP on the expression of inflammatory cytokines were examined using WB or RT-PCR. Activation of NF-κB and ERK were evaluated by nuclear localization of NF-κB and phos- pholylaiton of ERK, respectively. Results: We found that VPAC1, one of the PACAP receptors, is expressed in podocytes, a key player of glomerular filtration barrier. PACAP (10 nM) significantly increased the cellular contents of cAMP in cultured podocytes. In the presence of PACAP, CREB was increased by 1.9 ± 0.5 fold (mean ± SEM). In cultured podocytes, LPS increased the expression of interleukin-6 (IL-6) and monocyte chemo-attractant protein-1 (MCP-1) by 4.2 ± 0.6 fold and 32.8 ± 3.3 fold, respectively. In the presence of PACAP, the increased expression of IL-6 and MCP-1 was significantly attenuated by 20% and 25%, respectively, through the protein kinase A signaling pathway. In the presence of LPS, NF-κB transnuclear localization and phosphorylation of ERK were significantly increased, and PACAP significantly reduced these increases by 51% and 81%, respectively. Conclusion: In the present study we demonstrated that PACAP has anti- inflammatory effects on glomerular podocytes. To date, treatment options for DN are limited and PACAP may be useful in the prevention/attenuation of DN. 869 Effect of two years of sitagliptin treatment on renal function in elderly patients with type 2 diabetes mellitus S. Kurioka; Komatsu Hospital, Neyagawa, Japan. Background and aims: Many elderly patients with type 2 diabetes suffer from complications of chronic kidney disease and diabetic nephropathy. In this study, we examined the effect of a DPP-4 inhibitor, sitagliptin, on renal function in elderly patients with type 2 diabetes. Materials and methods: A total of 70 patients aged ≥ 65 years with type 2 diabetes mellitus, which included 42 males (60%), were examined. Patients had the following characteristics: age (mean ± SD) was 74 ± 5 years, BMI 24.5 ± 3.5 kg/m2, duration of diabetes 14 ± 10 years, HbA1c 8.3 ± 1.0%, serum creatinine 0.82 ± 0.27 mg/dl, eGFR 69.3 ± 23.9 ml/min/1.73m2, and an urine albumin-to-creatinine ratio (UACR) of 94.1 ± 172.0 mg/gCr. Patients had been treated with either 25 mg (n = 6 [9%]) or 50 mg (n = 64 [91%]) of sitag- liptin over a two-year period. HbA1c, eGFR and UACR were assessed yearly. eGFR data had been collected 1 year before commencement of sitagliptin. Of the 70 patients, 22 (31%) had eGFR < 60ml/min/m2 and 28 (40%) had UACR > 30 mg/gCr. All endpoints were evaluated by ANOVA. Results: HbA1c values before treatment, 1 year, and 2 years after sitagliptin treatment were 8.3, 7.5, and 7.3, respectively (before vs. after treatment, p < 0.0001). Gradually decreasing eGFR values were noted, as evidenced by values of 71.0, 69.3 (p < 0.05 when compared with eGFR 1 year before treat- ment), 65.6 (p < 0.05 when compared with eGFR before treatment), and 63.0 (p < 0.005 when compared with eGFR before treatment) at 1 year before treatment, before treatment, 1 year, and 2 years after treatment, respectively, while eGFR values were not decreased in patients with reduced renal func- tion (eGFR < 60ml/min/m2), as evidenced by values of 52.5, 47.6 (p < 0.05 when compared with eGFR 1 year before treatment), 47.8 (p = 0.88 when compared with eGFR before treatment), and 45.5 (p = 0.54 when compared with eGFR before treatment), respectively. UACR values were significantly decreased in patients with micro- or macroalbuminuria (UACR> 30 mg/ gCr), as evidenced by values of 207.3, 136.5 (p = 0.27 when compared with Diabetologia (2014) 57:[Suppl1]S1–S564 S 354 1 C UACR before treatment), and 106.6 (p < 0.05 when compared with UACR before treatment) before treatment, 1 year, and 2 years after treatment, re- spectively. Conclusion: Despite the gradual decline in eGFR values in patients with nor- mal renal function, eGFR did not decrease in patients with reduced renal function. These findings suggest that sitagliptin exhibits a renal protective effect, and is effective in maintaining long-term glycemic control in elderly patients with type 2 diabetes. 870 A DPP-4 inhibitor suppresses atherosclerotic lesions in the aorta and coronary arteries with decrease of macrophage infiltration in cholesterol- fed rabbits T. Hirano1,2, S. Yamashita2, M. Takahashi2, H. Hashimoto2, M. Goto2; 1Diabetes, Metabolism, and Endocrinology, Showa University School of Medicine, Tokyo, Japan, 2Sanwa Kagaku Lab., Mie, Japan. Background and aims: Several studies have demonstrated suppression of atherosclerosis by dipeptidyl peptidase-4 (DPP-4) inhibitors in hypercho- lesterolemic mice. However, it remains unknown whether DPP-4 inhibitors might also exert anti-atherogenic effects in bigger animals. We examined the effect of anagliptin, a DPP-4 inhibitor, on the development of atherosclerosis in the aorta and coronary arteries in cholesterol-fed rabbits. Materials and methods: Japanese white rabbits were fed either a normal diet or a diet containing 0.5% cholesterol. The cholesterol-fed rabbits were given drinking water not mixed (n = 18) or mixed with 3 mg/mL of ana- gliptin (n = 16) for 12 weeks. The lipoprotein fractions were measured by high-performance liquid chromatography. The serum oxidative stress mark- ers 8-hydroxy-2’-deoxyguanosine (OHdG) and malondealdehyde (MDA) were measured by ELISA and TBARS, respectively. Inflammatory cytokine gene expressions in the carotid artery were quantified by real time-qPCR. We measured the lesion area in the aorta and the coronary arteries in seven cross- sections of the heart. Values were expressed as mean ± SE. Results: Dietary cholesterol intake led to a marked increase of the serum total-cholesterol (TC) level (37.9 ± 3.9 vs. 0.59 ± 0.04 mmol/L), with the most striking increase seen in the VLDL fraction (30−80 nm) among the major lipoproteins. No significant changes of the body weight, water intake, HbA1c, serum lipids and lipoproteins, or glucose response to intravenous glucose loading were observed following the administration of anagliptin. The plas- ma DPP-4 activity was suppressed by 86%, and the plasma active GLP-1 lev- els doubled. Dietary cholesterol intake resulted in the development of severe atherosclerosis in the aorta, with a ratio of the lesion area to the total aortic surface areas of 22.0 ± 2.3%, and anagliptin treatment induced marked sup- pression of the percent lesion area to 8.6 ± 2.1% (p < 0.001). Atherosclerotic lesions were also observed in the coronary arteries. The four major coronary arteries increased in area with cholesterol feeding (1.14 ± 0.12 vs. 0.81 ± 0.06 mm2), while anagliptin treatment attenuated this change (0.87 ± 0.07 mm2). The intimal formation in the coronary arteries in the cholesterol-fed mice was attenuated by anagliptin treatment (0.04 ± 0.02 vs. 0.13 ± 0.05 mm2). The alpha-SMA-positive and macrophage-positive areas in the coronary ar- teries were suppressed 54 and 78% respectively, by anagliptin treatment (p < 0.05). Notably, the ratio of the macrophage area to the plaque area was substantially decreased by 83%. The serum 8-OHdG and MDA showed no significant changes following the anagliptin treatment. However, the arterial gene expressions of the inflammatory cytokines TNF-alpha and interleukin-6 were markedly reduced by approximately 90% (p < 0.001−0.05), and the ex- pressions of macrophage chemoattractant protein-1and CD26 in the carotid arteries were partially reduced (by about 30%) following the anagliptin ad- ministration. Conclusion: Our study demonstrated for the first time that a DPP-4 inhibi- tor suppressed the development of atherosclerosis in the aorta and coronary arteries in bigger animals than rodents, by inhibiting the inflammatory re- sponses in the vessels. Supported by: Sanwa Kagaku 871 A DPP-4 inhibitor remarkably suppresses foam cell formation in peritoneal macrophages obtained from db/db diabetic mice, comparison with a SGLT2 inhibitor and pioglitazone M. Terasaki, M. Nagashima, M. Hiromura, K. Notomi, T. Hirano; Department of Medicine, Showa University School of Medicine, Tokyo, Japan. Background and aims: We reported that a dipeptidyl peptidase-4 inhibi- tor (DPP-4i) confer an anti-atherosclerotic effect in diabetic apolipopro- tein-E null mice with significant suppression of foam cell formation in macrophages(Mφ). However, it remains to be elucidated whether a DPP-4i exerts anti-atherogenic properties beyond glycaemic control. We evaluated the suppressive effect of DPP-4i on foam cell formation in Mφ obtained from db/db diabetic mice, and compared with other glucose-lowering agents, a sodium-glucose co-transporter (SGLT)2-inhibitor and pioglitazone. Materials and methods: Male db/db mice at 9-week-old were fed normal diet containing none (n=23), alogliptin (0.02% w/w, n=12), pioglitazone (0.02%, n=12), or ipragliflozin (0.0014%, n=13) for 4 weeks. In subset of animals at age of 13 weeks, glucose (0.5g/kg body weight) was administrated orally, and bled 0, 15, 30, 60, and 120 min thereafter (OGTT). Peritoneal Mφ were obtained from mice at age of 13 weeks after the injection of thioglycollate. Foam cell formation was determined by the incorporation of [3H]-oleate into cholesteryl-oleate stimulated by oxidized-LDL in Mφ. CD36 and Acetyl-CoA acetyltransferase(ACAT)-1 gene expression was measured by RT-PCR. Results: Food and water intakes were comparable among treated groups, whereas final body weight was increased in piogitazone group by 6% and decreased in ipragliflozin group by 5% compared with non-treated con- trol group. Pioglitazone and ipragliflozin decreased fasting blood glu- cose (FBG), HbA1c and glucose-AUC in OGTT compared with control group(FBG:6.1±0.4, 10.0±1.6, and 19.4±1.8mmol/l, HbA1c: 5.2±0.2, 5.7±0.3, and 8.4±0.2%, AUC: 790±280, 2123±260, 4183±293mmol/l×min, respec- tively). Alogliptin group exhibited mild reductions of FBG, HbA1c, and AUC (16.0±1.3mmol/l, 7.0±0.4%, and 3413±463mmol/l×min). Despite mild ame- lioration of glycemic control with alogliptin, the Mφ foam cell formation was 36% decreased, which was comparable to those with pioglitazone or ipragli- flozin (35 and 30%, respectively). Foam cell formation was critically regulated by glycemic control (HbA1c) in each group (r=0.70~0.94, p S 355 1 C 872 Increase in beta cell function and improvement in HOMA-2beta index induced by saxagliptin in patients with latent autoimmune diabetes in adults R. Buzzetti1, P. Pozzilli2, R. Frederich3, N. Iqbal3, B. Hirshberg4; 1Department of Experimental Medicine, Sapienza University of Rome, 2Department of Endocrinology and Diabetology, Policlinico Universitario Campus Bio-Medico, Rome, Italy, 3Bristol-Myers Squibb, Princeton, 4AstraZeneca, Wilmington, USA. Background and aims: We used the presence of glutamic acid decarboxy- lase antibodies (GADA) to identify patients with latent autoimmune diabetes of adults, a population which typically responds poorly to oral antidiabetic medications. We previously reported that saxagliptin (SAXA; 2.5, 5, and 10 mg/d) was generally well tolerated and produced greater reductions vs placebo (PBO) in HbA1c and fasting and postprandial plasma glucose, with consistent treatment effects in GADA-positive and GADA-negative patients. Materials and methods: In this analysis, we assessed the effects of SAXA on β-cell function (postprandial C-peptide AUC and HOMA-2β) in GADA- positive (n=133) and GADA-negative (n=2576) patients from 5 placebo- controlled clinical trials. Results: There were little or no changes from baseline to week 24 in fast- ing C-peptide concentrations across patient and treatment groups (Table). SAXA produced similar increases relative to PBO in C-peptide AUC and HOMA2-β% in GADA-positive and GADA-negative patients, with no evi- dence of an interaction of GADA status on treatment effects, although the number of GADA-positive patients limits this interpretation. Conclusion: These results suggest that SAXA improves β-cell function over 24 weeks in patients with and without GADA at baseline. Clinical Trial Registration Number: NCT00121641, NCT00316082, NCT00121667, NCT00295633, NCT00313313 Supported by: BMS and AZ 873 Efficacy of DPP-4 inhibitors and their effect on blood cyclosporine levels in patients with post-transplantation diabetes E. Kang1, J. Bae1, Y. Lee1, C. Ahn1, B. Cha1, H. Lee1, K. Huh2, M. Kim2, Y. Kim2; 1Internal Medicine, 2Transplantation Surgery, Yonsei University Health System, Seoul, Republic of Korea. Background and aims: Dipeptidyl peptidase-4 (DPP-4) inhibitors have been widely used in kidney-transplant patients with diabetes. However the efficacy and interactions with immunosuppressants according to individual DPP-4 inhibitors are not widely studied. Therefore we tried to compare the glucose- lowering efficacy of DPP-4 inhibitors in patients with post-transplantation diabetes and evaluate the drug interaction with immunosuppressant, cyclo- sporine. Materials and methods: A total of 91 renal allograft recipients with diabetes who began to take DPP-4 inhibitors after transplantation were enrolled. The glucose-lowering efficacy of three DPP-4 inhibitors, vildagliptin, sitagliptin, and linagliptin were compared using glycosylated hemoglobin (HbA1c) after treatment for three months. Changes in blood cyclosporine levels were also assessed in 48 patients treated with cyclosporine. Results: There were no significant differences in HbA1c levels among vilda- gliptin, sitagliptin, and linagliptin treatment (-0.29±1.71% vs. -0.40±1.87% vs. -0.82±1.32%, P=0.559). However, blood cyclosporine level was signifi- cantly increased in the sitagliptin group, compared with vildagliptin and linagliptin group after three months of treatment (P=0.005). Difference in cyclosporine blood level was significant between vildagliptin group and sit- agliptin group (P=0.004). Conclusion: There were no significant difference in glucose-lowering efficacy among vildagliptin, sitagliptin, and linagliptin in kidney-transplant recipi- ents with diabetes. However, drug interaction between cyclosporine was sig- nificant in patients with sitagliptin treated group. Physicians should consider drug interaction when prescribing sitagliptin in kidney-transplant patients receiving cyclosporine as an immunosuppressant. 874 A comparison of the effects of the DPP-4 inhibitor sitagliptin and the sulfonylurea glimepiride on metabolic parameters and endothelial function H. Nomoto1, H. Miyoshi1, A. Nakamura1, T. Kondo1, N. Manda2, Y. Kurihara3, T. Atsumi1, S. Aoki4; 1Medicine II (Immunology and Metabolism), Hokkaido University, 2Manda Memorial Hospital, 3Kurihara Clinic, Sapporo, 4Aoki Clinic, Sapporo, Japan. Background and aims: DPP-4 inhibitors improve hyperglycemia in a glu- cose-dependent manner, and have been reported to possess favorable effects on atherosclerosis in animal experiments. However, it has not been eluci- dated whether DPP-4 inhibitors are able to improve endothelial function in patients with type 2 diabetes. Therefore, we investigated the efficacy of the DPP-4 inhibitor, Sitagliptin (Sita) on endothelial function and glycemic me- tabolism compared with Glimepiride (Gli) therapy. Materials and methods: This study was a multicenter, prospective, rand- omized parallel-group comparison. Study inclusion criteria were current metformin treatment and inadequate glycemic control (HbA1c levels of 6.9% to 8.4%) with sufficient control of blood pressure and lipid profile. Patients who suffered severe atherosclerosis, liver damage and renal dysfunction were excluded. Flow mediated dilation (FMD), Endo PAT, a comprehensive panel of hemodynamic parameters (Task Force® Monitor), and serum metabolic markers were assessed before and after the 26 week treatment period. All FMD assessments were performed by the same individual in a quiet, tem- perature-controlled setting. Patients were randomly assigned to once daily Sita (50mg) or Gli (0.5 to 2mg) therapy according to age, body mass index and baseline FMD values. Statistical analysis was performed using a Mann- Whitney U test and a p value < 0.05 was considered significant. Results: Forty-six men and thirty-three women aged 59.0 ± 10.4 years with HbA1c levels of 7.1 ± 0.4 % were enrolled. After 26 weeks, improvements in HbA1c levels were similar between the Sita and the Gli group (p = 0.48). Contrary to expectations, the observed improvements in %FMD and Endo PAT were not statistically different between groups (%FMD; Sita 5.9 to 6.0 %, Gli 5.7 to 6.1 %, RHI; Sita 1.9 to 2.3, Gli 2.1 to 2.3). Plasma HDL-cholesterol, adiponectin and TNF-α levels were significantly improved in the Sita group (p < 0.05) and LDL-cholesterol also tended to be improved (p = 0.07). Blood Diabetologia (2014) 57:[Suppl1]S1–S564 S 356 1 C pressure, cardiac index, total peripheral resistance index and most other met- abolic parameters were not different. Conclusion: In type 2 diabetic patients without advanced atherosclerosis, changes in endothelial function were similar between Sita and Gli groups af- ter 26 weeks. However, early Sita therapy was associated with more favorable effects on adipokine and lipid profiles compared with Gli therapy. Therefore, it is probable that over the long term these beneficial changes in glycemic and lipid control will aid in the prevention of atherosclerosis. Clinical Trial Registration Number: UMIN000004955 875 Co-administration of the DDP-4 inhibitor linagliptin and native GLP-1 induce synergistic body weight loss and appetite suppression in DIO rats M. Feigh1, H.B. Hansen1, G. Hansen1, S. Paulsen1, M. Mark2, N. Vrang1, T. Klein2; 1Gubra, Hørsholm, Denmark, 2Boehringer Ingelheim, Biberach, Germany. Background and aims: Linagliptin is a dipeptidyl peptidase (DPP)-4 in- hibitor approved for the treatment of type 2 diabetes. DPP-4 inhibitors are weight-neutral, suggesting that elevation of endogenous incretin (GLP-1) levels is not sufficient to promote weight loss per se. However, it may be pos- sible that long-term exogenous GLP-1 treatment in the context of concurrent DPP-IV inhibition may yield synergistic effects on appetite and body weight regulation. Hence, we evaluated the chronic metabolic effects of linagliptin and native GLP-1(7-36) co-administration in DIO rats. Materials and methods: Male diet-induced obese (DIO) rats (n= 8-10 per group) were treated with either linagliptin (1.5 mg/kg, PO, BID; 0.5 mg/kg, SC, BID) or native GLP-1 (0.4 mg/kg, SC, BID) as monotherapy, and com- pared to co-administration of linagliptin and GLP-1 for a total of 28 days. Body weight, food intake and body composition was measured. Also, fore- brain preprodynorphin gene expression levels were analyzed to assess for po- tential central effects on central endogenous opioidergic neurotransmission. Results: In DIO rats, monotherapy with linagliptin did not significantly in- fluence Food intake, body weight and adiposity during the study period. In contrast, combined linagliptin and GLP-1 treatment induced a reduction in food intake in conjunction with a pronounced body-weight (~8% compared with baseline) and whole-body fat mass (~20% vehicle-corrected) lowering effect at study end, which was superior to GLP-1 administration per se. No- tably, the anorexigenic effect of linagliptin and GLP-1 co-administration was associated with a marked increase in chow preference at the expense of pal- atable high-fat carbohydrate diet intake. Interestingly, combined linagliptin and GLP-1 treatment increased preprodynorphin mRNA levels in the cau- date-putamen, an effect not obtained with administration of the compounds individually. Conclusion: These data demonstrate that combined treatment with linaglip- tin and GLP-1 synergistically reduces body weight in obese rats. This anti- obesity effect is caused by appetite suppression and change in diet preference, presumably associated with increased dynorphin activity in dopaminergic forebrain regions involved in reward anticipation and habit learning. In con- clusion, linagliptin and GLP-1 co-administration may therefore hold promise as a novel therapeutic principle for combined weight and diabetes manage- ment in obese patients. Supported by: Boehringer Ingelheim 876 Effects of sitagliptin on body fat and intrahepatic lipid content in Japanese overweight patients with type 2 diabetes Y. Nagai, H. Kato, Y. Sada, S. Asai, A. Ohta, Y. Tanaka; Metabolism and Endocrinology, Internal Medicine, St. Marianna University School of Medicine, Kawasaki, Japan. Background and aims: Previous studies showed that human hepatocyte and adipocyte have the receptor of glucagon-like peptide-1 (GLP-1). Actually, GLP-1 analog, exendin-4 decreases intrahepatic lipid (IHL) and body fat in obese patients with type 2 diabetes. However, concrete changes of body fat and IHL by DPP-IV inhibitors were not fully evaluated. Thus, the aim of this study was to evaluate the effect of DPP-IV inhibitor, sitagliptin on body fat and IHL in overweight Japanese patients with type 2 diabetes. Materials and methods: This was a prospective, 24-week, single center, open labeled comparison study. The subjects were 20 Japanese type 2 diabetic pa- tients (male: 11, female: 9) with BMI>25 kg/m2. Nine of them were drug naïve and the remaining 11 received metformin. After a 4 to 8 weeks period of in- structing lifestyle modification (28 kcal/kg of ideal body weight and 150 - 200 kcal daily exercise), subjects were randomly assigned to receive sitagliptin 25mg titrated up to 50 mg (S) or glimepiride 0.5 mg titrated up to 1 mg (G). At the baseline, week 12 and week 24 after starting each treatment, body fat and IHL at segment 6 of the liver were evaluated by dual energy X-ray ab- sorptiometry (DEXA) and 1H-magnetic resonance spectroscopy (1H-MRS), respectively. Results: Though HbA1c and GA levels were significantly decreased in both groups after 24 weeks (HbA1c; 7.3 ± 0.1 to 6.5 ± 0.1% (S) and 7.1 ± 0.2 to 6.6 ± 0.1%(G), GA; 18.2 ± 1.0 to 14.8 ± 1.0% (S) and 17.0 ± 0.5 to 15.7 ± 0.4% (G), all p S 357 1 C PS 068 Incretin based therapies in special 877 Efficacy and safety of saxagliptin in older participants in the SAVOR- TIMI 53 trial L.A. Leiter1, H. Teoh2, D.L. Bhatt3, O. Mosenzon4, B.M. Scirica3, P. Kumar5, A. Smahelova6, B. Hirshberg7, C. Stahre8, R. Frederich9, F. Bonnici10, I. Raz11; 1Li Ka Shing Knowledge Institute and Keenan Research Centre for Biomedical Science, St. Michael‘s Hospital, University of Toronto, 2St. Michael‘s Hospital, University of Toronto, Canada, 3Brigham and Women‘s Hospital, Harvard Medical School, Boston, USA, 4The Diabetes Unit, Hadassah Hebrew University - Medical Center, Jerusalem, Israel, 5Bangalore Diabetes Hospital & Centre for Diabetes & Endocrine Care, Bangalore Diabetes Hospital, India, 6Department of Internal Gerontometabolic Clinic, Charles University in Prague, Hradec Kralove, Czech Republic, 7AstraZeneca, Wilmington, USA, 8AstraZeneca, MÖLNDAL, Sweden, 9Bristol-Myers Squibb, Princeton, USA, 10UCT Private Academic Hospital, Cape Town, South Africa, 11Hadassah Hebrew University - Medical Center, Jerusalem, Israel. Background and aims: There are limited data from randomised clinical trials (RCT) on the efficacy and safety of antihyperglycaemic treatments in the el- derly and very elderly despite the high prevalence of diabetes in these popula- tions. The aim of this analysis was to examine the cardiovascular (CV) effects, glycaemic efficacy, and safety of saxagliptin (SAXA) in the elderly (≥65 years, N=8,561) and very elderly (≥75 years, N=2,330) participants in the SAVOR- TIMI 53 trial. Materials and methods: Individuals ≥40 years old (N=16,492) with HbA1c ≥6.5% and ≤12.0% were randomised (1:1) to double-blind saxagliptin (5mg or 2.5mg OD) or placebo treatment for a median follow-up of 2.1 years. Results: The exposure to study medication was similar in all age groups. The hazard ratio (HR) for SAXA vs. placebo for the primary CV endpoint (myo- cardial infarction, ischemic stroke, CV death) was 0.92 for those ≥65 years vs. 1.15 for those S 358 1 C symptoms suggestive of hypoglycaemia confirmed by a self-monitored blood glucose measurement of S 359 1 C Secondary endpoints included other pharmacokinetic parameters (Cmax, tmax and t1/2), safety and tolerability of semaglutide. Results: Exposure of semaglutide in subjects with mild or moderate renal impairment or ESRD was similar to that in subjects with normal renal func- tion (Table). Subjects with severe renal impairment had a 22% higher ex- posure of semaglutide than those with normal renal function and the 95% CI (1.02, 1.47) exceeded the ‘no effect’ limits (0.70, 1.43). One subject with severe renal impairment reported two major hypoglycaemic events. There were no appreciable changes in laboratory safety parameters or vital signs; no serious adverse events (AEs) were noted. One subject withdrew due to gastrointestinal AEs. Conclusion: In three of four groups with renal impairment, there was no effect on semaglutide exposure. In the severe renal impairment group, sema- glutide exposure was increased. Semaglutide was well tolerated. The long- term effect of semaglutide on renal impairment will be investigated in future trials. Clinical Trial Registration Number: NCT00833716 Supported by: Novo Nordisk 883 VIRTUE study: effects of vildagliptin versus sulphonylureas as monotherapy or combined with metformin in Muslim patients with type 2 diabetes fasting during Ramadan I. Shaltout1, F. Pathan2, R. Medlej3, M. Al-Arouj4, A.A.K. Hassoun5, M.S. Chawla6, D. Knap7, J.A. Vaz8; 1Cairo University, Cairo, Egypt, 2BIRDEM Hospital, Dhaka, Bangladesh, 3Chronic Care Center, Saint Joseph University, Beirut, Lebanon, 4Dasman Diabetes Institute, Dasman, Kuwait, 5Dubai Diabetes Center, United Arab Emirates, 6Lina Diabetes Care Center, Mumbai, India, 7Novartis Pharma AG, Basel, Switzerland, 8Novartis Healthcare Pvt. Ltd., Hyderabad, India. Background and aims: For Muslim patients with type 2 diabetes mellitus (T2DM), fasting during Ramadan can increase the risk of serious complica- tions such as hypoglycaemia. This post-hoc analysis of the VIRTUE study assessed, in routine clinical practice, the effectiveness and safety of vildaglip- tin relative to sulphonylureas (SUs) as monotherapy or when combined with metformin in Muslim patients with T2DM fasting during Ramadan. Materials and methods: VIRTUE was a multicentre, prospective, 16-week, observational study conducted in the Middle East and Asia, and included adult Muslim patients with T2DM who were receiving vildagliptin or SUs (either as monotherapy or combined with metformin as per routine care) for ≥4 weeks and S 360 1 C % (p S 361 1 C range while the risk with glimepiride 2 mg/day increased with lower HbA1c (Figure). The increase for lower levels of HbA1c was more pronounced in the glimepiride 2 mg/day subgroup than in the full set of patients treated with glimepiride (adjusted HR = 0.08 [95% CI 0.0, 0.15]). Conclusion: Taken together, the data show a substantially lower risk of con- firmed hypoglycaemia with vildagliptin compared to low dose (2 mg/day) glimepiride, and indicate that the previously reported results are not driven by high doses of glimepiride. Supported by: Novartis 887 Incidence of fractures in patients with type 2 diabetes in the SAVOR- TIMI 53 trial K. Strojek1, O. Mosenzon2, J. Davidson3, B.M. Scirica4, G. Leibowitz2, R.G. Bretzel5, J.E. Villena6, M. Alvarsson7, B. Hirshberg8, C. Stahre9, A. Parker10, D.L. Bhatt11, I. Raz12; 1Department of Internal Diseases Diabetology and Cardiometabolic Diseases, Silesian Medical University Zabrze, Poland, 2The Diabetes Unit, Hadassah Hebrew University - Medical Center, Jerusalem, Israel, 3The University of Texas Southwestern Medical Center, Dallas, 4Brigham and Women‘s Hospital Harvard Medical School, Boston, USA, 5St. Josefs Hospital Balserische Stiftung, Giessen, Germany, 6Hospital Nacional Cayetano Heredia, Lima, Peru, 7Dept. of Endocrinology, Metabolism and Diabetes, Karolinska University Hospital, Stockholm, Sweden, 8AstraZeneca, Wilmington, USA, 9AstraZeneca, Mölndal, Sweden, 10The Diabetes Unit, AstraZeneca LP, Wilmington, USA, 11The Diabetes Unit, Brigham and Women‘s Hospital, Harvard Medical School, Boston, USA, 12Diabetes Unit, Hadassah Hebrew University - Medical Center, Jerusalem, Israel. Background and aims: Patients with type 2 diabetes have an increased risk of bone fractures, the predisposing factors for which in diabetes are unknown. In the SAVOR-TIMI 53 trial, fractures were considered an adverse event of special interest (AESI) and information regarding fractures was actively col- lected by the investigators. Materials and methods: We compared the incidence of fractures among the 8,280 patients who were assigned to treatment with saxagliptin to the inci- dence in the 8,212 patients who were assigned to placebo. Results: During the median follow-up of 2.1 years, 235 (2.8%) and 236 (2.9%) patients in the saxagliptin and placebo groups experienced a fracture; HR (95% CI) = 0.99 (0.83-1.19). The treatment exposure-adjusted rate of bone fracture (first event only) was 14 per 1000 patient-years of follow-up in both groups. Fracture risk was similar in patients treated with saxagliptin or pla- cebo across different subgroups defined by race, cardiovascular risk, renal function and duration of diabetes. A multivariable analysis of the entire study population showed the risk of fracture was associated with older age (p=0.001), female gender (p S 362 1 C 889 Safety profile of the DPP-4 inhibitors vildagliptin and sitagliptin in patients with type 2 diabetes and severe renal impairment V. Lukashevich1, J. Foley1, A. Schweizer2, M.S. Rendell3, W. Kothny2; 1Novartis Pharmaceuticals Corporation, East Hanover, USA, 2Novartis Pharma AG, Basel, Switzerland, 3Creighton University School of Medicine, Omaha, USA. Background and aims: Therapeutic management of type 2 diabetes mellitus (T2DM) with declined kidney function is challenging especially in patients with severe renal impairment (RI). Since the actions of the incretin hormones GLP-1 and GIP are glucose-sensitive, hypoglycaemia secondary to increased drug exposure is unlikely to occur with DPP-4 inhibitors (DPP-4i), making this drug class very attractive in treating patients with RI. It has been sug- gested that safety profiles of DPP-4i may differ in patients with RI, driven by whether or not unchanged drug or metabolites are renally excreted. Here we present a large pooled analysis of safety data for two different DPP-4i elimi- nated via the kidney, vildagliptin (Vilda), which is primarily hydrolyzed, and sitagliptin (Sita), excreted as unchanged drug, from two studies in 368 pa- tients with T2DM and severe RI. Materials and methods: Data were pooled across two double-blind studies of similar design (a new study comparing Vilda and Sita at the doses rec- ommended by label in severe RI and a previously reported study comparing Vilda and placebo [Pbo]). Patients with severe RI [eGFR (MDRD) S 363 1 C Materials and methods: Patient-level data were pooled from 40 double- blind, controlled studies of vildagliptin, used either as monotherapy or com- bination therapy, ranging in duration from 12 to ≥104 weeks. A retrospective meta-analysis of prospectively adjudicated CV events by a blinded independ- ent expert committee was performed using Mantel-Haenszel risk ratios (RR) to compare the vildagliptin treatment group (approved doses of 50 mg bid and qd) to the comparators groups (placebo and active comparators).The pri- mary endpoint was a major adverse CV events (MACE) composite of myo- cardial infarction, stroke and CV death. Additional assessments included the individual components of the composite endpoint. Results: Patients included in the pooled dataset [treated with vildagliptin (N=9599; 9250 subject year exposure [SYE]) or comparators (N=7847; 7317 SYE)] had a mean age of ~57 years, slight male predominance (~55/45%), mean BMI of 30.5 kg/m2 (nearly 50% obese), mean HbA1c of 8.1%, and a mean T2DM duration of 5.6 years]; 46% of patients had dyslipidemia, 58% hypertension and 27% were 65 years or older. The incidences and risk ratios for the MACE composite endpoint as well as the individual components are shown in the Figure below. A MACE occurred in 111 (1.16%) patients receiv- ing vildagliptin and 108 (1.52%) patients receiving comparators, the resulting RR being 0.86 (95% CI 0.66, 1.11), indicating no increased risk of MACE with vildagliptin. Similar risk ratios were also seen for each individual component of the composite endpoint (Figure). Conclusion: This large meta-analysis demonstrates that vildagliptin is not associated with an increased risk for adjudicated MACE events relative to comparators, confirming the CV safety profile of the drug in a largely ex- panded pooled population. Supported by: Novartis 892 Lower respiratory tract infections and the use of dipeptidyl peptidase-4 inhibitors: a population-based matched cohort study J.-L. Faillie1, L. Azoulay2, K. Filion2, V. Patenaude2, S. Suissa2, P. Ernst2; 1Department of Medical Pharmacology, Montpellier University Hospital, France, 2Centre for Clinical Epidemiology, Lady Davis Institute, Montreal, QC, Canada. Background and aims: It is hypothesized that dipeptidyl peptidase-4 (DPP- 4) inhibitors, also known as gliptins, may alter the immune response and play a role in the occurrence of infections. The aim of this study was to determine whether the use of DPP-4 inhibitors is associated with an increased risk of community-acquired lower respiratory tract infection, when compared to sulfonylureas. Materials and methods: A population-based matched cohort study was con- ducted among linkable patients with records in both the United Kingdom Clinical Practice Research Datalink (CPRD) and the Hospital Episodes Sta- tistics (HES) database. Between January 1, 2007 and March 31, 2012, each new user of DPP-4 inhibitors was matched to two patients prescribed sulfo- nylureas either in monotherapy or in combination, on three variables: age, duration of treated diabetes (defined as the time between a first ever non- insulin prescription and cohort entry) and use of sulfonylurea in the prior year. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of community-acquired lower res- piratory tract infection, comparing DPP-4 inhibitors users to sulfonylurea users. In secondary analyses, types of DPP-4 inhibitors, and the outcome of hospitalization for community-acquired pneumonia were assessed. All mod- els were adjusted for comorbidities and diabetes-related variables (such as hemoglobin A1c and previous use of anti-diabetic drugs). Results: Overall, 6933 new users of DPP-4 inhibitors were compared to 17,026 users of sulfonylureas. The crude incidence rates of community-ac- quired lower respiratory tract infection was 8.44 per 100 person-years (95% CI: 7.71 - 9.23) for DPP-4 inhibitors users and 7.54 (95% CI: 7.21 - 7.89) for sulfonylurea users. Compared to sulfonylureas, the use of DPP-4 inhibitors was not associated with an increased risk of community-acquired lower res- piratory tract infection and hospitalization for community-acquired pneu- monia (0.90, 95% CI: 0.78 - 1.04 and 0.76, 95% CI: 0.44 - 1.31, respectively) (table 1). In the subgroup analysis by type of DPP-4 inhibitors, the risk of community-acquired lower respiratory tract infection ranged from a 17% de- creased risk (HR: 0.83, 95% CI: 0.71 - 0.97) for sitagliptin, to a 31% increased risk for vildagliptin (HR: 1.31, 95% CI: 1.00 - 1.72). Conclusion: Overall, the use of DPP-4 inhibitors as a class does not appear to be associated with an increased risk of community-acquired lower respira- tory tract infections or hospitalization for community-acquired pneumonia. Additional research is needed to determine whether there is a within-class effect. 893 Pregnancy outcomes after unintentional exposure to vildagliptin C. McNeill1, C. Chu1, P.M. Paldánius2, W. Kothny2; 1Novartis Pharmaceuticals Corporation, East Hanover, USA, 2Novartis Pharma AG, Basel, Switzerland. Background and aims: Type 2 diabetes mellitus (T2DM) is increasingly di- agnosed in younger adults and as the population of women with diabetes becomes younger, T2DM and pregnancy increasingly co-exist. Pregnancies complicated by diabetes are associated with an increased risk of adverse outcomes such as preterm delivery, abortions, congenital abnormalities and neonatal mortality. However, data on outcomes in women exposed to newer oral anti-diabetes drugs (OADs), including vildagliptin, are limited. In the absence of adequate studies, most OADs carry warnings against use during pregnancy. Nevertheless, some use of these drugs occurs during pregnancy, either accidentally or intentionally. This analysis was performed to evaluate outcomes in pregnant women exposed to vildagliptin treatment. Materials and methods: The Novartis Safety Database was searched to iden- tify all known records (cut-off date: 30 September 2013) of exposure to vilda- gliptin (± metformin) during pregnancy. The database includes clinical study and post-marketing surveillance reports, as well as spontaneous reports. Maternal outcomes (number of live births, pre-term deliveries, abortions or caesarean sections) and neonatal outcomes (number of congenital anomalies or deaths) were analysed and presented as descriptive data when available. Results: A total of 32 pregnancies in 31 women with T2DM were identified from the database: 14 cases from controlled clinical trials, 15 spontaneous reports and 3 cases from post-marketing studies. The availability of demo- graphic and baseline characteristics data was limited. The mean age was 33.9 (range 19 - 44) years and 29.0% of women had a previous history of an abor- tion (elective or spontaneous). Limited information was available on cumula- tive vildagliptin dose or trimester(s) of exposure. 23 out of 32 pregnancies (71.9%) resulted in a live birth and more than half of all the pregnancies (n=18, 56.3%) had an unremarkable outcome of normal newborns without complications. Three (9.4%) pregnancies resulted in normal newborns with pregnancy complications (one case of gestational hypertension, haemorrhage during pregnancy and placenta previa) and two (6.3%) resulted in preterm deliveries. Four caesarean sections were reported (all live births), one in the preterm and three in term pregnancies. There were eight (24.9%) spontane- ous or missed abortions, but their relationship to vildagliptin was unclear. Additionally one case of therapeutic abortion was reported. No congenital abnormalities or neonatal deaths were reported. Conclusion: Although the numbers are low, this analysis provides pre- liminary information on pregnancy outcomes in T2DM after exposure to vildagliptin. No prospective studies of vildagliptin have been performed in pregnant women with T2DM. As with most OADs, vildagliptin is also not recommended for use during pregnancy. Supported by: Novartis Pharma AG Diabetologia (2014) 57:[Suppl1]S1–S564 S 364 1 C PS 070 Clinical studies with DPP4 inhibitors 894 Oral glucose lowering with linagliptin plus metformin is a viable initial treatment strategy in patients with newly diagnosed type 2 diabetes and marked hyperglycaemia B. Gallwitz1, S.A. Ross2, A.E. Caballero3, S. Del Prato4, D. Lewis-D’Agostino5, Z. Bailes6, S. Thiemann7, S. Patel6, H.-J. Woerle7, M. von Eynatten5; 1Dept. Medicine IV, Universitätsklinikum Tübingen, Germany, 2LMC Endocrinology Centres, University of Calgary, Canada, 3Joslin Diabetes Center and Harvard Medical School, Boston, USA, 4Department of Endocrinology and Metabolism, Section of Diabetes, University of Pisa, Italy, 5Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, USA, 6Boehringer Ingelheim Ltd, Bracknell, UK, 7Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany. Background and aims: Newly diagnosed type 2 diabetes (T2D) patients commonly present with marked hyperglycaemia. This condition has rarely been studied for novel oral diabetes drugs and insulin is often proposed as the preferred starting therapy. Materials and methods: We explored oral glucose-lowering combina- tion therapy in newly diagnosed (≤12 months) T2D patients with marked hyperglycaemia (n=316) utilising prespecified exploratory subgroup analyses from a randomised double-blind study of initial combination of linagliptin+metformin versus linagliptin. Baseline mean±SD age and HbA1c was 48.8±11.0 years and 9.8±1.1%, respectively. The primary endpoint was HbA1c change from baseline to week 24. Results: Mean±SE HbA1c reduction was -3.4±0.2% versus -2.5±0.2% with linagliptin+metformin and linagliptin, respectively, in patients with baseline HbA1c ≥9.5%, and -2.1±0.2% versus -1.4±0.2% in patients with baseline HbA1c S 365 1 C with medication, assessed as TSQM-9 scores for “convenience”, showed an increase from baseline to week 24 for the vildagliptin group and a decrease in the insulin group (9.6 ± 19.1 score points vs. -5.6 ± 22.0 score points; p3%. That more patients achieved the HbA1c target of S 366 1 C Conclusion: In this matched cohort study, pts with T2DM who started with MET+SITA dual therapy progressed to insulin therapy at a slower rate than those who started with MET+SU dual therapy. Supported by: MSD 899 Combination with linagliptin improves the tolerability of metformin in type 2 diabetic patients previously labelled as intolerant to metformin G. Gusmão, M. Moureira, F. Sosa; Endocrinologia, Hospital Santa Helena, Jequié - Bahia, Brazil. Background and aims: Metformin is the first-line drug for the treatment of type 2 diabetes but gastrointestinal intolerance to it is frequent. We sought to determine if patients withdrawn from metformin treatment due to intoler- ance would be able to tolerate it when rechallenged, and if the addition of linagliptin would improve the probability of tolerating metformin. Materials and methods: From the clinical records of our Diabetes Clinic, we identified type 2 diabetic patients aged 30 to 80 with HbA1c > 7% who had withdrawn from metformin due to gastrointestinal intolerance but had presented no serious adverse effects and had no significant renal insufficiency or other contraindications for metformin use, were not taking incretin-based therapy and were possible candidates for treatment with metformin and/or linagliptin. After at least 6 months from withdrawal, they were randomised to treatment with metformin alone (500 mg twice daily for 2 weeks, then 1000 mg twice daily for 10 additional weeks if the previous dose was tolerated), or with metformin (same pattern) plus linagliptin 5 mg daily, in an open fash- ion. Tolerance was assessed at weeks 2 and 12 by questionnaire performed by personnel blinded to the treatment status. Fasting glucose and HbA1c were measured at baseline and 12 weeks Results: We found 84 consecutive patients who were offered to participate. 25 declined, 30 received metformin plus linagliptin (ML) and 29 metformin alone (M). Their age was 63.2 ± 7.6 years, 56% were female; at baseline their fasting plasma glucose (FPG) was 9.2 ± 1.6 mmol/l and their HbA1c was 8.2 ± 0.9%. There were no significant differences between the groups for any of these variables. In the M group, 8 patients (28%) tolerated the low-dose and 5 (17%) the full-dose metformin (odds/ratio 3.94; 95% CI 1.32-11.76, p=0.014). In the ML group, 18 patients (60%) tolerated the low-dose and 14 (47%) the full-dose metformin (odds/ratio 4.20; 95% CI 1.26-13.96, p=0.019). By intention to treat, at week 12 the FPG in the M group was reduced by 1.1 ± 0.5 mmol/l and HbA1c was reduced by 0.2 ± 0.1%. In the ML group FPG was reduced by 2.7 ± 1.2 mmol/l and HbA1c was reduced by 0.9 ± 0.3% (both p S 367 1 C 901 Duration of maintenance of dual therapy with metformin and sitagliptin in type 2 diabetes: the Odyssée observational study P. Valensi1, G. De Pouvourville2, N. Benard3, C. Chanut-Vogel3, C. Kempf4, C. Moisan3, J. Dallongeville5; 1Department of Endocrinology Diabetology Nutrition-Jean Verdier Hospital, CRNH-IdF, CINFO, Bondy, 2ESSEC, Paris, 3MSD, Courbevoie, 4CSD, PARIS, 5Institut Pasteur, Lille, France. Background and aims: Sulfonylurea and DPP-4 inhibitors are usually pre- scribed for T2DM patients in combination with metformin. Odyssée, a pro- spective, real-world, observational study conducted in France in primary care practices, compared the duration of maintenance of treatment without modification (withdrawal, substitution, or add-on therapy) in T2DM patients in whom dual therapy with metformin + sitagliptin (MetSita) or metformin + sulfonylurea (MetSu) was initiated. Materials and methods: Odyssée was a multicenter, longitudinal study. It was conducted in a randomly selected sample of general practitioners in France. Patients were not randomized and were followed for a period of up to three years. Treatments were prescribed at the discretion of the physician. Physicians were expected to include all patients who met the eligibility crite- ria: Adult (aged ≥18 years) patients with type 2 diabetes who had initiated a de novo treatment with metformin and sitagliptin dual therapy or metformin and sulfonylurea dual therapy within the previous eight weeks were eligible for the study. Patient follow-up and changes to treatment were performed according to the physician’s clinical practice and were not specified in the study protocol. Results: At baseline, differences between the two arms (MetSita [n = 1874] and MetSu [n = 733]) were modest (mean age: 62.4 vs 64.2 years, BMI: 30.3 vs 29.6 kg/m2, diabetes duration: 6.4 vs 7 years, respectively). Mean HbA1c lev- els were similar (7.5 vs 7.6%). The median treatment duration for patients in the MetSita group was longer than the MetSu group (median treatment dura- tion 43.2 vs 20.2 months, respectively; between-group difference 23 months, log-rank p S 368 1 C PS 071 Non-glycaemic endpoints 903 Liraglutide 3.0 mg reduces body weight and improves cardiometabolic risk factors in overweight/obese adults: the SCALE obesity and prediabetes randomised trial M. Krempf1, A. Astrup2, C. le Roux3, K. Fujioka4, F. Greenway5, A. Halpern6, D.C.W. Lau7, R. Violante Ortiz8, J.P.H. Wilding9, C.B. Jensen10, S.K. Lillleøre10, X. Pi-Sunyer11; 1Clinique d‘endocrinologie, CHRU Nantes, France, 2University of Copenhagen, Denmark, 3University College Dublin, Ireland, 4Scripps Clinic, La Jolla, 5Louisiana State University, Baton Rouge, USA, 6University of São Paulo Medical School, Brazil, 7University of Calgary, Canada, 8Instituto Mexicano del Seguro Social, Ciudad Madero, Mexico, 9University of Liverpool, UK, 10Novo Nordisk A/S, Søborg, Denmark, 11Columbia University, New York, USA. Background and aims: A 5-10% weight loss has been shown to improve the multiple cardiometabolic comorbidities associated with chronic obesity. This phase 3 trial investigated the effects of liraglutide 3.0 mg, as adjunct to diet and exercise, on weight loss (primary endpoint) and cardiometabolic risk fac- tors (waist circumference, BP, lipids and other biomarkers). Materials and methods: Individuals (BMI ≥27 kg/m2 with ≥1 comorbidity or ≥30 kg/m2) were advised on a 500 kcal/day deficit diet and exercise pro- gramme and randomised 2:1 to once-daily s.c. liraglutide 3.0 mg (n=2487) or placebo (n=1244). Randomisation was stratified by prediabetes status (ADA 2010 criteria) and BMI. Data are shown for the full analysis set (exposed indi- viduals with ≥1 post-baseline assessment) with LOCF. The trial has an ongo- ing 2-year extension for participants with prediabetes. Results: Baseline characteristics were: age 45.1 years, 78.5% female, body weight 106.2 kg, BMI 38.3 kg/m2, 61.2% with prediabetes. At week 56, in- dividuals on liraglutide 3.0 mg achieved more weight loss (8.0%) than those on placebo (2.6%) (estimated treatment difference [ETD] -5.4% [95% CI 5.8; -5.0]; p S 369 1 C Conclusion: Liraglutide 3.0 mg, as adjunct to diet and exercise, was superior to placebo on all co-primary endpoints and generally well tolerated. Clinical Trial Registration Number: NCT01272219 Supported by: Novo Nordisk 905 Liraglutide improves carotid intima-media thickness in patients with type 2 diabetes and non-alcoholic fatty liver disease: an 8-month prospective pilot study R.V. Giglio, M. Rizzo, A.M. Patti, D. Nikolic, V. Di Bartolo, G. Castellino, G. Pecoraro, C. Rinaudo, F. Emanuele, A. Bonfiglio, M. Soresi, G. Montalto; Biomedical Department of Internal Medicine and Medical Specialties, University of Palermo, Italy. Background and aims: It has been shown in the last years that GLP-1 ana- logues, such as liraglutide, have several anti-atherogenic properties beyond their effects on glucose metabolism, including that on subclinical atheroscle- rosis. Yet, the effects of liraglutide in subjects with non-alcoholic fatty liver disease (NAFLD) are largely unknown. Materials and methods: We included in a 8-month prospective study 29 sub- jects with type-2 diabetes and NAFLD (16 men and 13 women, mean age: 61±10 years), who were matched for age and gender with another group of 29 subjects with type-2 diabetes (T2DM) but without NAFLD (16 men and 13 women, mean age: 61±8 years). The diagnosis of NAFLD was based on ultrasonographic and biochemical data. All subjects were naïve to incretin- based therapies and treated with metformin only. Liraglutide was given, on top of meformin, at a dosage of 0.6 mg/day for two weeks, followed by a dose of 1.2 mg/day for the rest of the study. At baseline and every 4 months fasting plasma samples were taken for laboratory analyses and carotid-intima me- dia thickness (IMT) was assessed by B-mode real-time ultrasound. Statistical analysis was performed by ANOVA and the Spearman correlation method. Results: From baseline to 4 months and 8 months of liraglutide therapy we found significant reductions in HbA1c in both groups of patients (from 8.9±1.5 to 6.6±1.2 to 6.5±1.1% in subjects with T2DM and NAFLD, and from 8.7±0.6 to 7.1±1.1 to 6.9±0.9% in subjects with T2DM only, p S 370 1 C cal activities of daily living questionnaire (APPADL) were administered at baseline and 26 weeks. Analysis of covariance models were fitted with factors of pooled country, baseline score, HbA1c strata (≤8.5, >8.5) and treatment. Results: DU 1.5 mg was noninferior to LIRA 1.8 mg in HbA1c change from baseline. Significant reductions in weight were observed with both DU 1.5 mg and LIRA 1.8 mg (LS mean [SE]: -2.90 [0.22] and -3.61 [0.22] kg, re- spectively; p=0.01 DU 1.5 mg vs LIRA 1.8 mg). Both groups had significant improvements from baseline in IW-SP, with no between-group differences. Only DU-treated patients had significant improvements from baseline in APPADL (p=0.014); between-group differences were not significant. Both groups had significant improvements from baseline in EQ-5D VAS with no between-group differences. In the EQ-5D UK population index score, pa- tients treated with DU 1.5 mg had significant improvements from baseline (p=0.031), while there was no significant change with LIRA 1.8 mg and no significant between-group difference. Conclusion: Both once weekly DU 1.5 mg- and once daily LIRA 1.8 mg- treated patients experienced glycaemic and weight benefits. Significant im- provements from baseline were observed in all PRO measures for DU 1.5 mg; significant improvements were only observed in the EQ-5D VAS and IW-SP for LIRA 1.8 mg. There were no significant between-group differences observed in any PRO measures. Clinical Trial Registration Number: NCT01624259 Supported by: Eli Lilly and Company 908 Results of the albiglutide HARMONY program prospective major adverse cardiovascular event (MI, stroke, cardiovascular death, and unstable angina) meta-analysis P.D. Ambery1, J. Donaldson2, J. Ye3, J. McMurray4; 1MedImmune, Cambridge, 2GlaxoSmithKline, Uxbridge, UK, 3GlaxoSmithKline, Research Triangle Park, USA, 4University of Glasgow, UK. Background and aims: In keeping with guidance issued by the FDA in 2008 requiring demonstration of the cardiovascular (CV) safety of new glucose- lowering agents for type 2 diabetes, GSK conducted a meta-analysis of CV events for albiglutide. Materials and methods: Possible CV events were collected prospectively across 1 Phase II and 8 Phase III albiglutide studies for blinded adjudication by an independent Clinical Endpoint Committee. Subjects were randomized to albiglutide or to placebo or active comparators (sitagliptin, insulin lispro, insulin glargine, pioglitazone, liraglutide, and glimepiride). Five of the Phase III studies were up to 3 years in duration. There were 5107 subjects in the safety population of whom 2524 were exposed to albiglutide (4870 person- yrs) and 2583 to comparators (5213 person-yrs). Results: The primary endpoint was the first occurrence of MACE+ (MACE + unstable angina) for albiglutide versus all comparators. MACE (CV death, non-fatal myocardial infarction, or non-fatal stroke) was a pre-defined sec- ondary endpoint. Results are shown in the table. Conclusion: Although the upper bound of the 95% CI for MACE+ events in this prospective adjudicated metaanalysis was below 1.8, because it is above 1.3, a cardiovascular endpoint study is required for albiglutide. No difference was seen in heart failure events or all-cause mortality for albiglutide versus comparators in this meta-analysis. Supported by: GSK 909 Safety and efficacy of liraglutide in patients with type 2 diabetes and end-stage renal disease: an investigator-initiated, randomised, placebo- controlled trial T. Idorn1, F.K. Knop2,3, M.B. Jørgensen1, T. Jensen4, M. Resuli5, P.M. Hansen5, K.B. Christensen6, J.J. Holst3, M. Hornum1, B. Feldt-Rasmussen1; 1Department of Nephrology, Copenhagen University Hospital, 2Diabetes Research Division, Department of Internal Medicine, Gentofte Hospital, Hellerup, 3Department of Biomedical Sciences, The NNF Center for Basic Metabolic Research, Faculty of Health Sciences, University of Copenhagen, 4Department of Endocrinology, Copenhagen University Hospital, 5Department of Internal Medicine, Hillerød Hospital, 6Department of Biostatistics, Institute of Public Health, University of Copenhagen, Denmark. Background and aims: Diabetes is the leading cause of end-stage renal dis- ease (ESRD), however, few antidiabetic agents are available to these patients. We evaluated safety and efficacy of liraglutide treatment in patients with type 2 diabetes (T2D) and dialysis-dependent ESRD. Materials and methods: 24 patients with T2D and ESRD and 23 patients with T2D and normal kidney function (71% and 74% treated with insulin, re- spectively) were randomly allocated to 12 weeks of double-blinded liraglutide (titrated dose of 0.6 mg, 1.2 mg or 1.8 mg) or placebo treatment (1:1) injected subcutaneously once-daily. 20 participants (1:1) in each group completed the study period. Dose-corrected plasma trough liraglutide concentration was evaluated at the final trial visit as the primary outcome measure. Additional safety and efficacy parameters were assessed. Results: Liraglutide-treated ESRD patients were titrated slower than liraglu- tide-treated controls, but ended at comparable doses at 12 weeks (1.33±0.13 and 1.26±0.06 mg/day, respectively; p=0.61). Dose-corrected plasma trough liraglutide concentrations at the final trial visit were increased by 49% (con- fidence interval: 6-109%, p=0.02) in liraglutide-treated ESRD patients com- pared with liraglutide-treated controls (plasma concentrations 11,975 [9,379- 15,289] and 8,057 [6,306-10,293] pmol/l/mg, respectively) (see Figure). Nau- sea and vomiting were more frequent in liraglutide-treated ESRD patients compared with liraglutide-treated controls. Severe adverse events were more frequent in the ESRD group (n=7, including n=6 during liraglutide treat- ment) compared with controls (n=1 during placebo treatment), although none were assessed to be related to the trial medication. Glycaemic control was improved in all treatment arms assessed by HbA1c and blood glucose measurements, and baseline insulin doses were significantly reduced in both liraglutide-treated groups (p S 371 1 C Conclusion: Liraglutide treatment appears to be applicable in patients with ESRD and T2D. Slower dose escalation and reduced treatment doses may be advisable. Clinical Trial Registration Number: NCT01394341 Supported by: Novo Nordisk A/S 910 Influence of liraglutide treatment in obstructive sleep apnoea in obese type 2 diabetes patients C. Abreu1, F. Gómez-Peralta1, J. Castro1, E. Alcarria1, M. Cruz-Bravo1, M. Llorente1, C. Albornos1, C. Moreno1, M. Cepeda2, F. Almodóvar3; 1Unidad de Endocrinología y Nutrición, Hospital General de Segovia, 2Servicio de Medicina Interna, Hospital General de Segovia, 3Unidad de Endocrinología y Nutrición, Hospital Universitario Fundación Alcorcón, Madrid, Spain. Background and aims: Liraglutide, a long-acting human glucagon-like pep- tide-1 analogue, is a novel class of type 2 diabetes mellitus (T2DM) treatment that exerts its glucose-lowering effect through different mechanisms that lead to clinically significant reductions in HbA1c and fasting and postprandial glucose levels with moderate weight loss. Concerns about the health impact of obstructive sleep apnoea (OSA) among obese patients with T2DM have been increasing. We aimed to evaluate the impact of liraglutide in the OSA symptoms-somnolence in obese T2DM patients. Materials and methods: This was a retrospective observational, unicentric study conducted in obese (BMI ≥ 30kg/m2) T2DM subjects receiving treat- ment with liraglutide for at least 3 months before study inclusion. Following routine clinical practice, data of the HbA1c, anthropometric measures, blood pressure, and the Epworth Sleepiness Scale (ESS) were collect at the study initiation (V1) and month 3 (V2). Results: A total of 109 subjects were evaluated (mean age, 59.7±10.9 years; female, 51.4%). Significant differences between V1 and V2 were found in mean values of body weight (106.0±16.8Kg vs 102.3±15.8Kg; p S 372 1 C also reduced glycaemic variability (SD - standard deviation: 1.4 ± 0.5 vs. 2.0 ± 0.6, p S 373 1 C 914 Effect of dipeptidyl peptidase-4 inhibitor sitagliptin on CD4+ T lymphocytes and monocytes in patients with type 2 diabetes A. Hattori1, H. Tokuyama1, F. Shimada2, M. Nieda3, Y. Maezawa1, K. Kobayashi1, H. Kawamura1, M. Takemoto1, K. Yokote1; 1Clinical Cell Biology and Medicine, Chiba University Graduate School of Medicine, 2Diabetes and Metabolism, National Hospital Organization Chiba Medical Center, 3Japan Biotherapy Institute, Tokyo, Japan. Background and aims: Dipeptidyl Peptidase-4 inhibitors (DPP-4i) are cur- rently used widely because of their efficacy and relative low risk of side effects. However several adverse effects of DPP-4i have been reported, e.g. raising the risk of pancreatitis and autoimmune diseases. DPP-4 is identical to CD26, a molecule expressed on immune cells particularly activated helper T cell (CD4+T;Th), but the effect of DPP-4i on immune cells is still unclear. The aim of this study is to evaluate the effect of DPP-4i on human T lymphocytes, especially Th, and monocytes in vivo and ex vivo. Materials and methods: Thirty-two Japanese patients with type 2 diabetes (19 male and 13 female, mean age 62.5 years, mean HbA1c 7.6%) taking Sitag- liptin 50mg/day as the first or the additional medication were enrolled to the study. Blood sampling was performed before and 8 weeks after the treatment. Leukocyte counts were performed by automatic blood cell counter. Surface markers including CD26 and CD163 on lymphocytes and monocytes were measured by flow cytometer after purification of peripheral blood mononu- clear cells. Ex vivo intracellular cytokine assay was performed to detect sub- sets of Th: Th0, Th1 and Th2 . The concentration of soluble CD26 (=DPP-4), soluble CD163 (macrophage/monocyte activation marker), TNF-α (Th1 cy- tokine) and IL-4 (Th2 cytokine) in patients’ serum were measured by ELISA. Results: Cell numbers (T cell numbers/μL) and percentages of circulating T lymphocytes were significantly decreased (1252±77/μL vs 1131±78/μL, p=0.02 and 65.4±8.9% vs 62.2±11.3%, p=0.02) after the administration of DPP4i, although those of total lymphocytes did not changed. In the T lym- phocyte subset, cell numbers and percentages of Th decreased significantly (853±316/μL vs 764±286/μL, P=0.01 and 44.8±8.3% vs 43.0±9.5%, p=0.04), but not Tc (CD8+T). Ex vivo assay showed the significant increase of Th1/2 ratio (8.7±1.0 vs 10.7±1.5, p=0.03) but TNF-α and IL-4 concentrations did not change. The change of Th1/2 ratio in 8 weeks were more correlated to the change of Th2 (r=-0.90) than a change of Th1(r= 0.58). The number of CD26+Th cells and serum level of soluble CD26 did not show any significant change. The percentage of circulating CD163+ monocytes (as it is called M2 phenotype monocyte) were significantly decreased (10.0±2.3% vs 2.7±0.5%, p=0.006) though total monocyte number did not. The concentration of solu- ble CD163 was significantly decreased (623±57ng/mL vs 584±60ng/mL, p=0.008) Conclusion: These findings suggest that DPP-4i possibly effect the balance of circulating T lymphocyte subset, particularly Th, and monocyte, but not effect Th1 and Th2 cytokines. These observation may indicate the direct effect of DPP-4 (CD26) inhibition on the differentiation from Th0 to Th2 rather than to Th1. On the other hand the decrease of soluble CD163 possibly re- flects the suppression of macrophage/monocyte activation that could lead to improvement of chronic inflammation. Further detailed study is needed to clarify the effect of DPP-4i to human T lymphocytes and monocytes. Supported by: Diabetes Masters Conference 915 A dipeptidyl peptidase-4 inhibitor directly suppresses inflammation and foam cell formation in monocytes/macrophages beyond incretins K. Nohtomi, M. Terasaki, K. Kohashi, M. Hiromura, M. Nagashima, T. Hirano; Showa University School, Department of Medicine, Tokyo, Japan. Background and aims: We reported that incretins (Glucagon-like peptide (GLP)-1 and glucose-dependent insulinotropic polypeptide (GIP)) and a dipeptidyl peptidase-4 inhibitor (DPP-4i) confer an anti-atherosclerotic effect in apolipoprotein-E null mice. It remains unclear whether the anti- atherogenesis of DPP-4i is wholly dependent on incretins or partly initiated via direct mechanisms independent of incretins. In this study we sought to elucidate the direct action of DPP-4i on anti-atherogenesis by examining the in vitro effect of incretins and DPP-4i on the expression of inflammatory cy- tokines and foam cell formation in monocytes/macrophages. Materials and methods: Mouse peritoneal macrophages were obtained from the ascites by injection of thioglycolate. Monocytes/macrophages were ob- tained from human peripheral blood mononuclear cells through a CD14 column. Foam cell formation was determined by the incorporation of [3H]- oleate into cholesteryl-oleate stimulated by oxidized-LDL. Results: Exendin (Ex)-4 (5nM) and GIP (1nM) elicited cyclic (c)AMP gen- eration and suppressed the LPS-induced gene expression of inflammatory molecules such as Interleukin (IL)-1beta, IL-6, TNF-alpha, and nuclear fac- tor-kappa B by 10~50% in U937 human monocytes. This suppressive effect, however, was attenuated by an inhibitor of adenylate cyclase and mimicked by 8-bromo-cAMP or forskolin. Teneligliptin (1-10 nM) substantially sup- pressed the LPS-induced expression of inflammatory cytokines by 20~75% without affecting the cell viability or DNA synthesis.Similarly, treatments with antibody or small interfering RNA of CD26 suppressed the expression of inflammatory cytokines.Teneligliptin with DPP4-resistant incretins (Ex- 4 and DAla(2)GIP) additively suppressed the expression of these cytokines. Ex-4, GIP, and teneligliptin all suppressed LPS-induced extracellular signal- regulated kinase (ERK) phosphorylation to the same extent. The suppres- sion of p-ERK by incretins was cAMP-dependent, while the suppression by teneligliptin was not. Teneligliptin suppressed cell surface toll-like receptor (TLR)-4-and TLR-5-mediated cytokine production, but had no effect on the intracellular TRLs-induced cytokines. Ex-4 and GIP slightly suppressed foam cell formation in mouse peritoneal macrophages and human macrophages whereas teneligliptin did not. Foam cell formation was enhanced by 2-fold in peritoneal macrophages of diabetic mice (db/db) and in the macrophages of patients with poorly controlled type 2 diabetes. Ex-4, GIP, and teneliglip- tin suppressed foam cell formation in the macrophages of diabetic mice and patients by 25-50%. The gene expressions of acyl-CoA cholesterol acyltrans- ferase (ACAT)-1 and CD36 were up-regulated in diabetes, and incretins and teneligliptin attenuated the upregulation. CD26 mRNA and transcripts cord- ed by exons 3-13 of the GLP-1 receptor and exons 1-14 of the GIP receptor were detectable in the monocytes/macrophages of mouse and human. Conclusion: The present in vitro study suggests that DPP-4i suppresses in- flammation partly via ERK inactivation and foam cell formation by down- regulating CD36 and ACAT-1 in monocytes/macrophages. This suppressive effect appears to be Independent of incretin’s action via cAMP. Supported by: Tanabe-Mitsubishi phama.co. 916 Liraglutide associated reversal of obesity and age dependent cardiac fibrosis correlates with improved vascular reactivity and modulation of macrophage gene expression A.E. Dear1, T. Gaspari2, I. Welungoda2, Y. Hu1, R.E. Widdop2, R.W. Simpson1; 1Medicine, Monash University, 2Pharmacology, Monash University, Melbourne, Australia. Background and aims: Cardiac fibrosis observed in obesity and myocardial aging results from excessive deposition of extracellular matrix (ECM) com- ponents, contributes to a detrimental increase in myocardial stiffness and is a major factor in the development of cardiac hypertrophy and the transition to heart failure. Excessive deposition of cardiac ECM is characterised by in- creased macrophage infiltration, differentiation of fibroblasts to the secretory myofibroblast phenotype and enhanced deposition of type I and III collagens. Expression of the glucagon like peptide-1 receptor (GLP-1R) in cardiac cells together with cardioprotective effects of GLP-1 and GLP-1R agonists in in vivo studies suggests a possible role for GLP-1 and GLP-1R agonists in the modulation of cardiac fibrosis. Our current study aimed to determine the effect of the GLP-1R agonist liraglutide in obesity and age induced models of cardiac fibrosis and identify associated molecular mechanisms. Materials and methods: In vivo experiments utilised 5 month old C57Bl/6J mice maintained on a high fat diet (HFD) for 16 weeks to evoke obesity in- duced hypertension and cardiac fibrosis. 20 month old C57Bl/6J mice main- tained on a normal chow diet were utilised as a model of aged induced car- diac fibrosis. Treatment regimens for both obesity and age induced cardiac fibrosis consisted of liraglutide 300μg/kg s.c., twice daily for 4 weeks. In vitro studies utilised the murine macrophage cell line RAW264.7 stimulated with lipopolysaccharide (LPS) (100ng/ml) and treated with liraglutide (100nM), with or without the specific GLP-1R antagonist exendin-9 (100nM) for 48 hours. Results: Treatment of 5 month old C57Bl/6J mice with liraglutide prevented HFD-induced increase in blood pressure, reduced fat mass and importantly protected against increased interstitial cardiac fibrosis over vehicle treated animals (n=4, p S 374 1 C liraglutide treatment significantly inhibited age associated cardiac fibrosis (n=4-6, p S 375 1 C 920 Effect of liraglutide on beta cell mass and function in alloxan diabetic mice K. Tamura, K. Minami, H. Takahashi, M. Kudo, S. Seino; Kobe University Graduate School of Medicine, Japan. Background and aims: Although GLP-1 receptor agonists have been shown to have beneficial effects on both beta-cell mass and function, the mechanism of the effect of the drugs on beta-cells is still unclear. The objective of this study is to characterize the effect of liraglutide, a long-acting GLP-1 analog, on beta-cell mass and function in alloxan-induced diabetic mice. Materials and methods: We crossbred Ins2-CreER knock-in mice with R26R-YFP mice to generate Ins2-CreER/R26R-YFP double knock-in mice, in which the pancreatic beta-cells can be labeled specifically and perma- nently upon injection of tamoxifen. These mice enabled us to discriminate beta-cell neogenesis from self-replication. Pancreatic beta-cells were labeled by 5 consecutive injections of tamoxifen (4 mg/head/injection) to male mice at 6 weeks of age. Ten days after the last injection, diabetes was induced by a single dose of alloxan (60 mg/kg, i.p.). On the following day, mice with blood glucose concentration above 300 mg/dl were used for the study. Liraglutide (200 µg/kg s.c.) or vehicle (saline) was administered once daily from the day following alloxan injection for 30 days. Food intake, body weight, blood glu- cose levels, serum insulin levels and serum glucagon levels were measured. Beta-cell mass was estimated by immunofluorescent staining of pancreas sec- tions. Proliferation rate was evaluated by Ki67 immunostaining and apopto- sis rate was evaluated by TUNEL-method. To investigate beta-cell function, oral glucose tolerance test was performed. Results: About 80% of beta-cells were destroyed by alloxan injection, lead- ing to hypoinsulinemia and hyperglycemia in the mice. Liraglutide treat- ment ameliorated hyperglycemia in alloxan-diabetic mice. We found that beta-cell mass in liraglutide-treated group was two-fold higher than that in vehicle group, which correlates with the increment of serum insulin levels. In liraglutide-treated group, proliferation rate was significantly higher and apoptosis rate was significantly lower than in vehicle-treated mice. To exam- ine the effect of liraglutide on beta-cell neogenesis, we injected tamoxifen to the alloxan-treated Ins2-CreER/R26R-YFP mice. As a result, the percentage of YFP-labeled beta-cells was not changed significantly before and after lira- glutide treatment, suggesting that the contribution of neogenesis is, if any, limited to the increment of beta cell mass by liraglutide treatment. Further- more, oral glucose tolerance test showed that the insulin secretory response to glucose was recovered following 30 days liraglutide treatment, and that this effect was sustained two weeks after drug withdrawal. Conclusion: Chronic treatment of alloxan-induced diabetic mice with lira- glutide preserved beta-cell mass by increasing cell proliferation and decreas- ing cell death. In addition, liraglutide treatment achieved sustained improve- ment of insulin secretory function of pancreatic beta-cells. Supported by: MEXT, JST, Novo Nordisk 921 Peripheral glucagon-like peptide-1 receptor agonism decreases VLDL production and fasting dyslipidaemia in insulin resistance J. Taher, C. Baker, C. Cuizon, M. Naples, S. Farr, K. Adeli; The Hospital for Sick Children, Toronto, Canada. Background and aims: Glucagon-like peptide-1(GLP-1) is an incretin hor- mone that stimulates pancreatic insulin release to reduce plasma glucose levels and thus acts as an important drug target in the treatment of type 2 diabetes (T2D). While GLP-1 receptor (GLP-1R) agonism is known to en- hance satiety and induce weight loss, our laboratory has also shown that GLP- 1 regulates lipoprotein metabolism by decreasing intestinal chylomicron production and may also play a similar role in the liver. Hepatic lipids are packaged with apolipoprotein B-100 (apoB100) into VLDL and secreted into the plasma. Dysregulation of VLDL production results in the fasting dyslipi- demia that is observed in patients with T2D. In a similar way to intestinal me- tabolism, we postulate that GLP-1R agonism regulates fasting dyslipidemia by decreasing VLDL production in insulin resistance. We further postulate that this occurs through either a peripheral or central pathway independent of changes in food consumption. Materials and methods: Fructose-fed insulin resistant Syrian golden ham- sters were given twice daily intraperitoneal (i.p) injections of the GLP-1R agonist exendin-4 (ex4; 20ug/kg) for 7 days and placed into metabolic cag- es. Alternatively, fructose-fed hamsters received an intracerebroventricular (ICV) administration of ex4 (250ng) into the third ventricle of the brain for 4 days. Plasma was subsequently collected over a 6 hour time course follow- ing i.p poloxamer administration to prevent lipoprotein uptake. Livers were excised for PCR analysis. Results: Peripheral ex4 prevented fructose-induced fasting hypertriglyc- eridemia (vehicle(veh)=3.4 fold triglyceride (TG)increase; ex4=1.4 fold TG increase) and hypercholesterolemia (veh=1.3 fold increase; ex4=0.8 fold de- crease). Ex4 also decreased VLDL-triglyceride( 1.28mmol/L vs 0.55mmol/L; p S 376 1 C PS 073 Incretin based therapies: mechanistic / miscellaneous 922 Near-normalisation of glycaemic control in patients with type 2 diabetes with a glucagon-like peptide-1 receptor agonist in combination with exercise P.F. Mensberg1,2, S. Nyby1,2, P.G. Jørgensen3, H. Storgaard1, J. Sivertsen3, M.T. Jensen3, J.J. Holst4, B. Kiens2, E.A. Richter2, F.K. Knop1, T. Vilsbøll1; 1Diabetes Research Center, Department of Medicine, Gentofte Hospital, Hellerup, 2Department of Nutrition, Exercise and Sports, Section of Molecular Physiology, Copenhagen, 3Department of Cardiology, Gentofte Hospital, Hellerup, 4Department of Biomedical Sciences, NNF center for Basic Metabolic Research, Copenhagen, Denmark. Background and aims: Exercise and glucagon-like peptide-1 receptor ago- nist (GLP-1RA) treatment, respectively, improve glycaemic control in pa- tients with type 2 diabetes (T2D). We investigated the effects of exercise in combination with GLP-1RA treatment in patients with T2D. Materials and methods: Thirty-three untrained obese patients with T2D were included in a randomised double-blinded placebo-controlled clinical trial (23 men, age: 56±10 years (mean±SD); weight 99±16 kg; BMI: 32±4 kg/ m2; fasting plasma glucose (FPG): 10.1±2.8 mmol/l; HbA1c: 65±14 mmol/ mol (8.1±1.3%)), treated with diet only (n=3) or metformin (n=30). Patients were randomised to liraglutide (titrated to 1.8 mg once-daily) and exercise or placebo and exercise during 16 weeks. Both groups had 3 60-minute su- pervised training sessions per week; 2 spinning sessions and 1 session with resistance training. Results: HbA1c, FPG and weight reductions were greater with liraglutide and exercise vs. placebo and exercise (Table 1). Both groups exhibited similar re- duction in body fat mass and increase in maximal oxygen uptake. The most common adverse event was transient nausea (liraglutide: 24%, placebo: 6%). Conclusion: Exercise alone for 16 weeks has no or little effect on glycaemic control and body weight, but combined with GLP-1RA treatment it can near- normalise glycaemic control and cause robust weight loss in overweight, dys- regulated patients with type 2 diabetes. Clinical Trial Registration Number: NCT01455441 Supported by: Novo Nordisk A/S 923 Transfer of liraglutide from blood to cerebrospinal fluid is limited and an unlikely contributor to body weight loss in patients with type 2 diabetes M. Christensen1,2, A.H. Sparre-Ulrich1,3, B. Hartmann4, U. Grevstad5, M.M. Rosenkilde3, J.J. Holst4, T. Vilsbøll1, F.K. Knop1; 1Diabetes Research Division, Department of Medicine, Gentofte Hospital, Hellerup, 2Department of Clinical Pharmacology, Bispebjerg Hospital, 3Department of Neuroscience and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, 4NNF Center for Basic Metabolic Research, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, 5Department of Anesthesiology, Gentofte Hospital, Hellerup, Denmark. Background and aims: The mechanisms by which subcutaneously adminis- tered glucagon-like peptide-1 (GLP-1) receptor agonists confer weight loss in humans are incompletely understood. A centrally mediated effect by passage of the blood-brain barrier has been suggested from animal studies. There are no human data on blood-brain transfer of GLP-1 receptor agonists. Materials and methods: We measured liraglutide in plasma and cerebro- spinal fluid (CSF) obtained by lumbar puncture in 8 patients with type 2 diabetes (male: 4/8, age: 63±2 years [mean±SEM], BMI: 30±1 kg/m2, Hba1c 56±4 mmol/mol) treated with 1.8 mg liraglutide subcutaneously for at least 1 month and with weight loss 8.4±0.4 kg in response to treatment. Liraglutide was measured with a radioimmunoassay (antibody no. 98302) specific for the intact N-terminus of the GLP-1 moiety of liraglutide. For validation of the CSF-assay, we obtained control CSF from subjects (n=9) not treated with liraglutide. Results: Plasma-liraglutide was in the nanomolar range (31±4.4 nmol/L). When applying the assay to the CSF control samples, we measured a picomo- lar baseline level of 17±1.2 pmol/L. Recovery of liraglutide added to CSF was 100 % and sensitivity < 2 pmol/L. The non-corrected CSF-liraglutide concen- trations in the liraglutide-treated patients were 23±1.8 pmol/L (comparison to untreated controls, Mann-Whitney test: P0.6). Conclusion: Blood-brain barrier transport of liraglutide appears very limited and CSF-liraglutide does not seem to correlate with the plasma liraglutide concentrations or with clinical weight loss. Therefore, our results do not sup- port that blood-to-CSF transfer is a significant contributor to the anorectic effects of GLP-1 receptor agonists. 924 Exenatide promotes hyperbolic-like specular changes in insulin sensitivity and beta cell function in non-human primates F. Casiraghi1, A. Caumo1,2, S. Benedini1,2, L. Luzi1,2, F. Folli3; 1I.R.C.C.S Policlinico San Donato, Milan, 2University of Milan, Italy, 3UTHSCSA, San Antonio, USA. Background and aims: The GLP-1 receptor agonist Exenatide (EXE) im- proves glycemic control in type 2 diabetes mellitus (T2DM) by enhancing glucose-dependent insulin secretion, suppressing inappropriate glucagon secretion, slowing gastric emptying, and regulating appetite. However, the effects of continuous chronic EXE infusion on insulin sensitivity and β-cell function are not completely understood. We investigated the effect of con- tinuous EXE infusion on a well characterized non-human primate (baboon) model of insulin resistance and T2DM. Materials and methods: We examined the effect of a continuous IV Infu- sion (via tether system) of: i) EXE (0.014 ug/kg.h, n=10) or ii) saline (SAL, n=10) on insulin sensitivity and β-cell function in non-diabetic baboons (NDB). At baseline, baboons received a two-step Hyperglycemic Clamp (+125 and +225 mg/dl) followed by an IV Arginine bolus (0.15 g/kg) (HC+A), with a total duration of 210 minutes. Immediately after HC+A, ba- boons underwent a partial pancreatectomy (~30% total mass) and started EXE or SAL infusion for 13 weeks. At the end of treatment, EXE infusion was stopped for 72 hours before HC+A was repeated and remnant pan- creas was collected after humane euthanasia, at necropsy. Insulin sensitiv- ity was calculated by linear regression from the two-step dose-response relationship between the incremental (above baseline) glucose clearance and the incremental insulin concentration measured during the final 30 min of each step. Beta-cell function was calculated by linear regression from the two-step dose-response relationship between the incremental C-peptide con- centration and the incremental glucose concentration measured during the final 30 min of each step. Diabetologia (2014) 57:[Suppl1]S1–S564 S 377 1 C Results: We found a significant increase in insulin sensitivity (6.6±3.4 to 11.2±5.5 104 ml/kg per μU/ml, p S 378 1 C Conclusion: Lixisenatide and liraglutide + insulin glargine improved glycae- mic control in T2DM ± metformin, albeit with differing gastric emptying mechanisms of action and safety/tolerability profiles. Clinical Trial Registration Number: NCT01596504 Supported by: Sanofi 927 The influence of DPP-4 inhibitors on fat metabolism in type 2 diabetes patients A.S. Ametov, D.G. Gusenbekova; Department Endocriniligy and Diaetology RMAPO, Ministry of Health, Moscow, Russian Federation. Background and aims: To evaluate the effect of sitagliptin in combination with metformin on glucose toxicity and lipotoxicity in patients with type 2 diabetes and obesity. Materials and methods: The study involved 82 patients (55 women, 27 men, mean age 56,1 ± 5,47 years) with obesity, lipid metabolism disorders, who have not reached target levels HbA1s (average HbA1c 8,3 ± 1,6%) after met- formin and diet therapy. The first group of patients (42 patients) received co-formulated drug, consisting of sitagliptin 100mg and metformin 2g once a day; the second comparison group (40 patients) received metformin 1.5-2.0 g/day. Dynamics of fasting glycemia, postprandial glycemia, glycated hemo- globin, weight, BMI, WC, WHR, lipid profile (total cholesterol, triglycerides, LDL, HDL, apoβ protein), insulin, proinsulin, leptin, adiponectin, insulin resistance using the index HOMA IR and functional activity of β-cells (by HOMA-β index) was evaluated at baseline and at 6 months of therapy. In ad- dition, MRI was performed to assess visceral fat in all the patients. Results: At 6 months patients in both groups demonstrated significant posi- tive changes in the levels of fasting glucose, postprandial glycemia and gly- cosylated hemoglobin. In group I, HbA1 decreased from 8,3 ± 1,6% to 6,6 ± 1,24% (p S 379 1 C 929 Effects of standard dose of vildagliptin vs sitagliptin on blood glucose fluctuations evaluated by long-term self-monitoring blood glucose (SMBG) K. Kimachi, H. Miyoshi, N. Aikawa, H. Nomoto, H. Kameda, K. Cho, T. Kondo, T. Atsumi; Department of Medicine II (Immunology and Metabolism), Hokkaido University School of Medicine, Sapporo, Japan. Background and aims: No previous studies have compared the DPP-4 inhib- itors, vildagliptin and sitagliptin, in terms of fluctuations of daily blood glu- cose profile by using long-term SMBG. The aim of this study was to evaluate the effects of switching from sitagliptin to vildagliptin, or from vildagliptin to sitagliptin on 12-week glucose fluctuations using SMBG in Japanese patients with type 2 diabetes. Materials and methods: We conducted a multicenter, prospective, open- label controlled trial with blinded endpoint analysis. Thirty-two patients with type2 diabetes mellitus had been treated by vildagliptin or sitagliptin in each standard use (vildagliptin 50mg twice daily or sitagliptin 50mg once daily), and were allocated to two groups who received vildagliptin during 12 weeks then switched to sitagliptin during 12 weeks, or vice versa. SMBG systems were distributed to all patients at the beginning and measurements of seven- point SMBG (daily profile of blood glucose level of before and 2 hours af- ter each meal, and at bedtime) during the trial were used to determine: 1) the mean amplitude of glycemic excursions (MAGE), 2) M-value, 3) mean (± standard deviation) blood glucose level, and 4) fasting and postprandial blood glucose level. Plasma glycosylated hemoglobin (HbA1c), glycoalbumin (GA), 1,5-anhydroglucitol (1,5AG), immunoreactive insulin (IRI), Proinsu- lin/IRI ratio, C-peptide immunoreactivity (CPR) and glucagon levels were measured. Results: The MAGE was significantly lower in patients taking vildagliptin than sitagliptin (61.9±27.3 vs. 72.9±38.4 mg/dL; p=0.04). In patients taking vildagliptin, the mean blood glucose level was significantly lower (151.0±31.6 vs. 157.5±29.2 mg/dL; p=0.02), M-value was significantly lower (11.7±2.4 vs. 14.5±2.5; p=0.01) and the blood glucose level after supper was signifi- cantly lower (172.2±40.8 vs. 190.6±42.2 mg/dL; p=0.04). 1,5-anhydroglu- citol (1,5AG) was significantly higher(1.40±2.67 vs. -0.90±1.59; p=0.003), and proinsulin/IRI ratio was significantly lower(-0.039±0.13 vs. 0.51±0.11; p=0.01). There were no significant differences in plasma HbA1c, GA, fasting CPR, or glucagon levels between patients taking vildagliptin and sitagliptin. Conclusion: Vildagliptin was superior to reduce the fluctuations of daily blood glucose profile compared with sitagliptin in Japanese patients with type 2 diabetes. Clinical Trial Registration Number: UMIN000005627 930 Factors associated with progression of type 2 diabetes and impact of treatment with saxagliptin in the SAVOR-TIMI 53 trial G. Leibowitz1, O. Mosenzon1, D.L. Bhatt2, B. Hirshberg3, C. Wei3, A. Cahn1, G. Jermendy4, W.H.H. Sheu5, J. Lopez-Sendon6, A. Avogaro7, B.M. Scirica2, I. Raz1; 1The Diabetes Unit, Hadassah Hebrew University - Medical Center, Jerusalem, Israel, 2Brigham and Women‘s Hospital, Harvard Medical School, Boston, 3AstraZeneca, Wilmington, USA, 4Bajcsy-Zsilinszky Hospital, Budapest, Hungary, 5Department of Internal Medicine, Division of Endocrinology and Metabolism, Taichung Veterans General Hospital, Taiwan, 6Cardiology Department, Hospital Universitario La Paz, Madrid, Spain, 7Dept. of Medicine, University of Padova, Italy. Background and aims: In type 2 diabetes (T2D), glycemic control often deteriorates over time, requiring intensification of treatment. We aimed to identify factors associated with progression of diabetes and studied the im- pact of saxagliptin, a DPP-4 inhibitor, on diabetes progression. In addition, we evaluated the effect of saxagliptin on beta cell function as reflected by a decline in HOMA2-β. Materials and methods: We studied the association of clinical and biochemi- cal parameters with diabetes progression in the SAVOR-TIMI 53 study, a ran- domized clinical trial of 16,492 patients with T2D treated with saxagliptin vs. placebo added to current anti-diabetic medications for a median of 2.1 years. Diabetes progression was defined by 1) HbA1C increase ≥0.5%, 2) initiation of new anti-diabetic medications, 3) increase in oral medication dose or 4) ≥25% increase in insulin dose for ≥3 months. HOMA2-β was measured at baseline and at year 2 in 4134 patients (25.1% of trial). Results: Progression of diabetes during the study occurred in 54.7% of all subjects. Compared with placebo, treatment with saxagliptin decreased the risk of diabetes progression (OR 0.60; 95% CI 0.57-0.65; p S 380 1 C PS 074 The potential future of insulin therapy 931 IDegAsp provides superior FPG control and reduced hypoglycaemia vs BIAsp 30 in insulin-naive adults with type 2 diabetes: a randomised phase 3 trial E. Franek1, M. Haluzík2, S. Canecki Varžić3, M. Sargin4, S. Macura5, J. Zacho5, J. Christiansen6; 1Central Clinical Hospital MSWiA, Warsaw, Poland, 2General University Hospital and 1 Faculty of Medicine, Praha, Czech Republic, 3Clinical Hospital Center Osijek, Croatia, 4Kartal Training and Research Hospital, Istanbul, Turkey, 5Novo Nordisk A/S, Copenhagen, Denmark, 6Aarhus University Hospital, Denmark. Background and aims: Insulin degludec/insulin aspart (IDegAsp) is a 70%/30% combination of insulin degludec (IDeg) and insulin aspart (IAsp), providing both a basal insulin with an ultra-long duration of action and a rapid-acting bolus insulin in a single injection.The aim of this trial was to compare the efficacy and safety of IDegAsp and biphasic insulin aspart 30 (BIAsp 30), each administered twice daily (BID) ± metformin in insulin- naive adults with type 2 diabetes (T2DM). Materials and methods: This was a 26-week randomised, open-label, mul- tinational, treat-to-target trial. Subjects (mean: age 58.9 years, duration of diabetes 9.5 years, HbA1c 8.4%, FPG 10.2 mmol/L, BMI 31.2 kg/m 2) were randomised 1:1 to IDegAsp BID (n=197) or BIAsp 30 BID (n=197). Insulin was administered before breakfast and the main evening meal and titrated to a pre-breakfast and pre-main evening meal self-measured plasma glucose (SMPG) target of ≤5 mmol/L. Patient characteristics were similar in both treatment arms. Results: IDegAsp was non-inferior to BIAsp 30 based on mean change in HbA1c from baseline (primary endpoint; estimated treatment difference [ETD] +0.02% [95% CI -0.12; 0.17]), end of treatment mean HbA1c was 49.0 (6.6) and 48.0 (6.5) (mmol/mol [%]), respectively. IDegAsp was superior to BIAsp 30 in lowering FPG (ETD -1.0 mmol/L [95% CI -1.42; -0.59]; p S 381 1 C safety issues were identified. Both formulations were well tolerated and no injection site reactions occurred. Conclusion: The PK and PD results of BC Combo showed a stronger pran- dial effect with a faster onset and also a longer basal action for BC Combo indicating the potential of an improved metabolic control vs. MIX when used as once daily treatment. Clinical Trial Registration Number: NCT01981031 933 Long-acting basal insulin (HM12470) offers once-weekly dosing potential with a favorable PK, PD and mitogenic profile I. Choi1, S. Hwang1, J. Kim1, S. Jung1, D. Kim1, Y.-M. Lee1, Y. Kim1, M. Trautmann2, M. Hompesch2, J. Son1, S. Kwon1; 1Hanmi Pharmaceutical Co., Ltd., Gyeonggi-do, Republic of Korea, 2Profil Institute, Chula Vista, USA. Background and aims: The novel long-acting basal insulin (HM12470) has been developed for once-weekly injection by conjugating an insulin analog with the constant region of a human immunoglobulin fragment via non-pep- tidyl linker. This study investigated the in vitro properties, pharmacokinetics and pharmacodynamics of HM12470 in normal and diabetic animal models to evaluate the once-weekly dosing potential and the mitogenic potency. Materials and methods: Pharmacokinetics of HM12470 was evaluated after subcutaneously administration to mice, rats, dogs, pigs, and monkeys. Based on these preclinical PK data, human plasma profile of HM12470 was predict- ed for once-weekly dosing by the Wajima Css-MRT method. Blood glucose lowering was monitored in db/db mice after HM12470 was subcutaneously injected. In vitro analysis, the lipogenic potency in fully differentiated 3T3- L1 adipocyte cells, and cell proliferative potency in MCF-7 and SaOs-2 cells were evaluated. Results: In a pharmacokinetic study, subcutaneously injected HM12470 ex- hibited a half life of ~ 43 hr in normal rats, while insulin degludec showed 2.9 hr of half-life. The extended half-life was also confirmed in other species such as mice, dogs, pigs, and monkeys. The improved pharmacokinetic profile had contributed to prolonged glucose lowering efficacy in db/db mice. Based on the results from three different species (mice, rats, and dogs), human phar- macokinetics was predicted that half-life time of HM12470 was expected to be ~132 hr and the peak-to-though ratio was to be 1.6 on once weekly dosing. In vitro mitogenic potency of HM12470 was assessed by using cell prolifera- tion in MCF-7 and SaOs-2 cells. Compared to its lipogenic efficacy assessed in fully differentiated 3T3-L1 adipocyte cells, the mitogenic to lipogenic po- tency ratio was significantly lower than that of human insulin. Conclusion: HM12470 has a prolonged pharmacokinetic and pharmaco- dynamic profile in animal studies, and the predicted half-life and peak-to- trough ratio of HM12470 in human are promising once-weekly dosing po- tential with flat exposure profile. Moreover, HM12470 is expected to have low mitogenic potency based on in vitro results. 934 The ultra-rapid biochaperone insulin lispro shows a faster onset of action and stronger early metabolic effect than insulin lispro alone G. Andersen1, B. Alluis2, G. Meiffren2, A. Ranson2, O. Soula2, G. Soula2, R. Soula2, A. Fischer1, L. Nosek1, F. Schliess1, T. Heise1; 1Profil Institut für Stoffwechselforschung GmbH, Neuss, Germany, 2Adocia, Lyon, France. Background and aims: We investigated the pharmacodynamic characteris- tics of BioChaperone insulin lispro (BC LIS), a novel insulin lispro formula- tion with BioChaperone aimed at accelerating the absorption from the sub- cutaneous tissue. Materials and methods: In this double-blind, crossover study people with type 1 diabetes received 0.2 U/kg of BC LIS or insulin lispro (LIS) alone under automated glycaemic clamp conditions (ClampArt®, target blood glucose 100 mg/dL, clamp duration 6h post-dosing). Results: Thirty-seven patients were enrolled and 36 completed this study (age 42.7±13.4 years (mean±SD), BMI 24.8±1.7 kg/m², HbA1c 7.6±0.6%). Mean glucose infusion rates (GIR) are given in the figure. Compared with LIS, BC LIS showed ultra-rapid properties with a faster onset of action (23.1±7.0 vs. 34.4±15.3 min, p S 382 1 C 935 Comparison of duration of action of 2 insulin glargine products, LY2963016 and insulin glargine, in subjects with type 1 diabetes mellitus H. Linnebjerg1, T. Heise2, X. Zhang1, M.E. Seger1, D. Coutant1, L. Chua3, E. Lam3; 1Eli Lilly and Company, Indianapolis, USA, 2Profil, Neuss, Germany, 3Lilly-NUS Centre for Clinical Pharmacology, Singapore. Background and aims: LY2963016 (LY IGlar) and insulin glargine (Sanofi- Aventis; IGlar) are both insulin glargine products, with identical amino acid sequences. Even with identical amino acid sequences, proteinbased therapeu- tics manufactured by distinct processes must be shown to be similar. This study compared the duration of action of LY IGlar and IGlar. Materials and methods: This was a Phase 1, randomised, doubleblind, 2peri- od, crossover, glucose clamp study. Fasted male subjects with type 1 diabetes mellitus received a single subcutaneous dose of 0.3 U/kg LY IGlar and IGlar on 1 occasion each. A minimum 7day washout separated doses. Pharmaco- dynamic effects were assessed with a euglycaemic clamp (Biostator®, glucose target level 100 mg/dL [5.6 mmol/L]) lasting up to 42 hours postdose. Du- ration of action was defined as the time post dosing at which the subject›s blood glucose level was consistently >150 mg/dL (8.3 mmol/L) without any glucose infusion. Results: Twenty male subjects aged 2354 years with type 1 diabetes mellitus participated in this study. Based on timetoevent (survival) analysis, median duration of action was estimated to be 37.1 and 40.0 hours for LY IGlar and IGlar, respectively, while mean duration of action (standard error) was 23.8 (3.8) and 25.5 (3.9) hours for LY IGlar and IGlar, respectively. In 14 of the 40 clamps (7 LY IGlar, 7 IGlar), subjects’ duration of action exceeded the 42hour clamp period. The mean profiles of the pharmacodynamic variables (glucose infusion rate and blood glucose) were similar between treatments over the 42hour clamps (see figure below). LY IGlar was well tolerated, with similar adverse event profiles observed following LY IGlar and IGlar dosing. No safety concerns from clinical laboratory, vital signs or ECG data were noted. Conclusion: Similar pharmacodynamic effects and duration of action were demonstrated for LY IGlar and IGlar in this 42hour clamp study in subjects with type 1 diabetes mellitus. Clinical Trial Registration Number: NCT01600950 Supported by: Eli Lilly and Company 936 Higher early insulin exposure and greater early glucose-lowering effect with faster-acting insulin aspart in patients with type 1 diabetes mellitus H. Haahr1, U. Hövelmann2, L. Brøndsted1, C.L. Adrian1, L. Nosek2, T. Heise2; 1Novo Nordisk A/S, Søborg, Denmark, 2Profil, Neuss, Germany. Background and aims: Faster-acting insulin aspart (faster aspart) is insulin aspart (IAsp) in a new formulation containing two new excipients (nicoti- namide and arginine), which result in a faster initial absorption after s. c. injection. Materials and methods: In total, 52 patients with type 1 diabetes mellitus (mean ± standard deviation age: 40.3 ± 12.0 years; haemoglobin A1c: 7.3 ± 0.7%) were randomised to a single dose (0.2 U/kg s. c.) of faster aspart or IAsp under glucose clamp conditions (Biostator; blood glucose target 100 mg/dl; clamp duration 12 hours post-dose) in a crossover design. Results: Compared with IAsp, faster aspart had a faster onset of appearance, i.e., time from trial drug administration until the first time serum insulin as- part concentrations reached the lower limit of quantification (mean difference [95% confidence interval (CI)]: -6.33 minutes [-7.30; -5.36]), earlier time to 50% Cmax (median difference [95% CI]: -11.0 minutes [-13.5; -9.0]), earlier tmax (median difference [95% CI]: -7.5 minutes [-17.5; 0.0]), and greater early exposure up to 2 hours post-dose (e.g., 4.5-fold more insulin aspart in the cir- culation in the first 15 minutes post-dose); total exposure was similar (Table). Faster aspart had an earlier and higher glucose-lowering effect (indicated by higher glucose infusion rates, GIR) in the first 2 hours post-injection versus IAsp (Table) and an earlier time to 50% GIRmax (mean difference [95% CI]: -7.81 minutes [-13.19; -2.44]). The maximum GIR and total glucose-lowering effect (Table) were similar between faster aspart and IAsp. No safety or toler- ability issues were identified; in particular, no injection site reactions occurred. Diabetologia (2014) 57:[Suppl1]S1–S564 S 383 1 C Conclusion: Faster onset and higher early exposure with faster-acting insulin aspart led to a greater early glucose-lowering effect, indicating that faster as- part has the potential to improve postprandial glucose versus IAsp. Clinical Trial Registration Number: NCT01618188 Supported by: Novo Nordisk 937 Biphasic pharmacokinetic and pharmacodynamic profiles associated with concentrated insulin BIOD-531 show rapid onset and basal duration of action L. Morrow1, M. Hompesch1, L. Canney2, P. Pichotta2, A. Krasner2, E. De Souza2; 1Profil Institute for Clinical Research, Chula Vista, 2Biodel Inc., Danbury, USA. Background and aims: Formulations of human insulin and insulin analogs containing citrate and EDTA have been shown to be more rapidly absorbed than commercially available insulin products. The aim of this study was to compare the pharmacokinetic (PK) and pharmacodynamic (PD) profiles of BIOD-531, a concentrated (400 U/ml or U-400) formulation of recombinant human insulin containing EDTA, citrate and MgSO4, to products marketed as combined prandial/basal insulins. Materials and methods: A single-center, randomized, double-blind four- period crossover study employing 24-hour euglycemic clamps, evaluating the PK and PD profiles of BIOD-531 at two doses (1 U/kg and 0.5 U/kg) vs. Humulin® R U-500 (1 U/kg, U-500R) and Humalog® Mix75/25™ (0.5 U/kg, HU 75/25) in obese non-diabetic subjects. Results: Thirteen subjects (age 39.8 ± 13.5 yrs; weight 99.5 ± 12.6 kg) were randomized. BIOD-531 was associated with an increased rate of absorption vs. U-500R as measured by multiple PK parameters including a 92% shorter time to early ½ Tmax (11.0 ± 1.9 min) vs. U-500R(135.3 ± 34.9 min), a 43% shorter time to Tmax (223.8 ± 62.3 min) vs. U-500R (393.3 ± 58.3 min) and a 765% increase in early insulin exposure as measured by AUCins0-30min (2966 ± 383 mU*min/L) vs. U-500R (343 ± 74 mU*min/L). BIOD-531 was associ- ated with a more rapid glucose lowering effect as measured by multiple PD parameters including a 66% faster time to onset of action (7.2 ± 4.1 min) vs. U-500R (21.4 ± 6.7 min) and 169% increase in AUCGIR0-60min (108.5 ± 22.0 mg/kg vs. U-500R (40.4 ± 10.0 mg/kg). These differences were all statistically significant. BIOD-531 was associated with an increased rate of absorption vs. HU 75/25 as measured by multiple PK parameters including a 66% shorter time to early ½ Tmax (16.4 ± 4.9 min) vs. HU 75/25 (47.9 ± 2.6 min), an 18% shorter time to Tmax (131.3 ± 43.4 min) vs. HU 75/25 (160.0 ± 11.9 min) and a 917% increase in early insulin exposure as measured by AUCins0-30min (1200 ± 141 min) vs. HU 75/25 (118 ± 22 mU*min/L). BIOD-531 was associated with a more rapid glucose lowering effect as measured by multiple PD parameters including a 59% faster time to onset of action (14.6 ± 6.0 min) vs. HU 75/25 (35.9 ± 7.9 min) and 375% increase in AUCGIR0-60min (68.9 ± 13.4 mg/kg) vs. HU 75/25 (14.5 ± 4.7 mg/kg). With the exception of Tmax, these differences were all statistically significant. The duration of action of BIOD-531 is ap- proximately 18 hours, shorter than that of either comparator. All study drugs were well tolerated. Conclusion: BIOD-531 is associated with more rapid onset of action than either marketed prandial/basal insulin product suggesting that it may provide improved mealtime glucose control with low injection volumes. Further- more, the duration of action of BIOD-531 is consistent with a basal insulin. Supported by: NIDDK Award R43DK096604 938 Pharmacokinetics and elimination of basal insulin peglispro (BIL) versus PEG-only in rats following a single IV or SC dose M. Knadler, J.A. Martin, B.B. Ellis, P.L. Brown-Augsburger, A.T. Murphy, V.J. Wroblewski; Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, USA. Background and aims: BIL (LY2605541) is a novel, 20-kDa PEGylated in- sulin lispro that has a large hydrodynamic size. It has a prolonged duration of action which is related to a delay in insulin absorption and a reduction in clearance. The aim of this study was to examine the pharmacokinetics and elimination of BIL compared with 20-kDa PEG only in rats. Materials and methods: Male Sprague Dawley rats were given a 0.5-mg/kg IV or 2-mg/kg SC dose of BIL with 14C-labeled 20-kDa PEG portions of BIL or 125I-labeled lispro portions of BIL. PEG only ( 14C- 20-kDa PEG) was dosed at 10 mg/kg IV or SC. Blood for serum, urine, and feces were collected up to 336 h postdose. Samples were analyzed for 14C using liquid scintillation counting, for 125I using gamma counting, for BIL using ELISA, and for PEG- related products using liquid chromatography/mass spectrometry (MS) or size-exclusion chromatography. Results: AUCs for BIL were 41% and 16% of circulating 14C after IV and SC dosing, respectively. AUCs for BIL were 63% and 31% of circulating 125I for IV and SC dosing, respectively. After SC dosing, t½ of 14C was 4.4-fold longer than that of 125I. After IV dosing, BIL Cl was 2.4-fold faster than that of 14C radioequivalents and 1.6-fold faster than that of 125I radioequivalents indicating catabolism of BIL into products with longer half-lives. The t½ and Cl of 14C after PEG-only dosing were similar to those after BIL dosing. SC bioavailability of radiolabeled BIL was 23%−29%; absorption was ≥75%. The bioavailability of 20-kDa PEG-only was 78%. Excretion of radioactivity after IV or SC dosing of 14C-BILwas generally similar. In the BIL study, 14C was eliminated in urine and feces (urine:feces ratio, 0.9 for IV and 1.35 for SC). In the PEG-only study, most of the 14C was recovered in urine after IV or SC ad- ministration. In the BIL study, ≥86% of 125I was eliminated in urine mainly as small peptides, amino acids or free 125I. After SC 14C-BIL dosing, a mean ±SD of 2.5% ± 0.66% of the dose was excreted intact in urine. Based on qualitative MS analyses, the main PEG-related product found in serum and urine after BIL dosing was 20-kDa PEG with lysine attached. After PEG-only dosing, the main urinary product appeared to be intact PEG based on size-exclusion chromatography. Conclusion: These data suggest that elimination of BIL is via catabolism to smaller peptides and/or amino acids and PEG with lysine attached. BIL is cleared faster than the products from catabolism. After BIL dosing, the PEG portion is eliminated via biliary and renal routes; after PEG alone, most of the 14C dose is eliminated renally. The PEG in BIL would only partially depend on renal function for elimination, unlike PEG-only. The protein attached to PEG may direct the elimination of PEG. Supported by: Eli Lilly and Company Diabetologia (2014) 57:[Suppl1]S1–S564 S 384 1 C PS 075 Insulin treatment in the „real world“ 939 Pump use is less frequent in minority youth: transatlantic analysis in three large registries representing Austria, Germany, England, Wales and the United States B. Rami-Merhar1, D. Maahs2, J. Warner3, S. Hofer4, W. Tamborlane5, J. Hermann6, A. Icks7, D. Schatz8, N. Foster9, R.W. Holl6; 1Medical University of Vienna, Austria, 2University of Colorado, Denver, USA, 3on behalf of the National Paediatric Diabetes Audit and the Royal College of Paediatrics and Child Health, Cardiff, UK, 4Medical University of Innsbruck, Austria, 5Yale School of Medicine, New Haven, USA, 6University of Ulm, 7Department of Public Health, Heinrich Heine University Düsseldorf, Germany, 8University of Florida College of Medicine, Gainesville, 9JAEB Center, Tampa, USA. Background and aims: Use of insulin pumps has recently increased among children and adolescents with type-1 diabetes in many parts of the world. Pa- tient and family preferences, beliefs and policies of healthcare professionals, expectations on metabolic goals, but also financial aspects / reimbursement by health insurance interact on the decision for or against CSII in individual patients. Each of these components depends on society and may be different for minority youth. Ideally, important treatment choices should be independ- ent from the socio-economic background of a patient. The aim of this study was to evaluate this claim based on data from 3 large, multicenter registries. Materials and methods: In total, 54,767 children and adolescents ( S 385 1 C 941 Adherence to insulin treatment in insulin naive type 2 diabetic patients: results of telephonic intervention D. Gogas Yavuz1, H. Bilen2, S. Sancak3, T. Galip4, Z. Hekimsoy5, I. Şahin6, M. Yılmaz7, H. Aydın8, A. Atmaca9, M. Sert10, P. Karakaya11, D. Arpacı12, A. Oğuz13, N. Guvener14, Dropout Study Group; 1Endocrinology and Metabolism, Marmara University School of Medicine, Istanbul, 2Endocrinology and Metabolism, Ataturk University School of Medicine, Erzurum, 3Endocrinology and Metabolism, Fatih Sultan Mehmet EAH, Istanbul, 4Endocrinology and Metabolism, Sakarya Devlet Hastanesi, Sakarya, 5Endocrinology and Metabolism, Celal Bayar University Medical Faculty, Manisa, 6Endocrinology and Metabolism, Inonu University School of Medicine, Malatya, 7Endocrinology and Metabolism, Namık Kemal University Medical Faculty, Tekirdag, 8Endocrinology and Metabolism, Yeditepe University Medical School, Istanbul, 9Endocrinology and Metabolism, Ondokuzmayıs University Medical Faculty, Samsun, 10Endocrinology and Metabolism, Cukurova University Medical School, Adana, 11Endocrinology and Metabolism, Bakırkoy EAH, 12Endocrinology and Metabolism, Sakarya Devlet Hastanesi, 13Endocrinology and Metabolism, Medeniyet University Medical Faculty, 14Endocrinology and Metabolism, Baskent University Medical Faculty, Istanbul, Turkey. Background and aims: The aim of the present study was to assess the efficacy of telephone support on insulin treatment adherence in insulin naive type 2 diabetic patients using different insulin treatment regimens (basal, basal- bolus and premix ) in third care medical centers in Turkey Materials and methods: In this 12 week-open label randomized multi- center study A total of 1456 insulin naive type2 diabetic patients (mean age 56±12yrs, diabetes duration 6,0±6 yrs, 49.1% female) were recruited in 14 third care endocrine centers from different city’s in Turkey. Patients randomized to telephonic intervention (TI) (50.3%) and control group (C ) (%49.7). Primary outcome was medication adherence. Results: At the end of the 12th week 83.2 % of TI group and %70.3 of the Control group were adherent to treatment (p S 386 1 C relatively unchanged. Increased hypoglycaemia on insulin does not appear to be a necessary consequence of increased need for exogenous insulin with time in routine care. Supported by: Sanofi 944 Insulin treatment across Europe: results from the GUIDANCE study R.C. Vos, G.E.H. Rutten, GUIDANCE; Julius Center for Health Science and Primary Care, University Medical Center Utrecht, Netherlands. Background and aims: Given the increasing prevalence of type 2 diabetes (T2DM), insulin treatment will be needed for a growing number of patients. Traditionally the conversion to insulin therapy has been managed in the sec- ondary care setting, but more and more primary care physicians are involved in insulin therapy. We compared insulin treated patients across eight Euro- pean countries, stratified to treatment setting. Materials and methods: Data were available from the medical records of 1960 insulin treated T2DM patients of eight European countries, from mainly primary care (Belgium, France, Germany, Sweden, the Netherlands) and mainly secondary care (Italy, Ireland, the United Kingdom) countries. Patients from ‘mainly primary care’ and ‘mainly secondary care’ countries were compared for patients variables: age, sex, duration of diabetes, HbA1c, BMI, micro- and macro vascular complications, type of insulin treatment; and service variables: access of a nurse or a dietician, physician’s number of T2DM patients treated, and physician’s feelings about the diabetes guidelines. MLwiN for binomial multilevel analysis was used to analyse the relation of the patient and service variables with insulin treated patients from mainly primary and secondary care countries. Results: Type of insulin treatment differed across European countries; insu- lin treated patients from mainly secondary care countries used more often short acting analogues (OR 2.54, 95%CI 1.89-3.43) and less often either NPH (OR 0.45, 95%CI 0.28-0.73), long acting analogues (OR 0.50, 95%CI 0.36- 0.68) or short acting human insulin (OR 0.59, 95%CI 0.37-0.92) compared to those in mainly primary care countries. Further, insulin treated patients from the mainly secondary care countries showed higher HbA1c levels (OR 1.49, 95%CI 1.34-1.66) and had more often access to a dietician (OR 2.54, 95%CI 1.89-3.43), but were less likely to have macro vascular complications (OR 0.65, 95%CI 0.49-0.86). No clinical relevant differences were found for the other variables. Conclusion: The results of this large European cross-sectional study show that T2DM patients on insulin and mostly treated in a primary care setting in five European countries did not differ from the same category of patients mainly treated in outpatient clinics or hospitals in three other countries, with the exception of type of insulin, HbA1c and the presence of macro vascular complications. The decision to treat a T2DM patient in primary or secondary care is likely not based on clinical arguments. 945 Metabolic control in type 2 diabetic patients with convential or intensified insulin therapy: a retrospective single centre survey over a period of 22 years G. Kramer, N. Kuniss, C. Kloos, N. Mueller, A. Saemann, G. Wolf, U.A. Mueller; Dept. Internal Medicine III, University Hospital Jena, Germany. Background and aims: The 2012 EASD/ADA position statement on man- agement of hyperglycaemia in type 2 diabetes recommends mealtime insulin, consisting of one to three injections, when basal insulin has been titrated to an acceptable fasting glucose but HbA1c remains above target. The progres- sion from basal insulin to a twice daily pre-mixed insulin (CT) is regarded as a “less studied alternative”. Multiple insulin injection therapy (MIT) is regard- ed to be more easily adaptable to the patients needs. Therefore it is supposed that MIT is associated with better metabolic control and less hypoglycaemia compared to CT in patients with diabetes type 2 (DM2). Materials and methods: HbA1c, insulin dose, body weight and incidence of non severe and severe hypoglycaemia were analysed in respect to the strategy of insulin therapy applied (MIT or CT). 20943 visits of 1447 insulin treated people with DM2 in an university outpatient department for endocrinology and metabolic diseases from a period of 22 years (9.1.1990 to 20.9.2012) were included. Data were drawn from the electronic patient record EMIL©. Be- cause the strategy of insulin therapy changed over the long period, we com- pared visits with CT and visits with MIT. All visits with one or two insulin injections were classified as CT, visits with more than two injections as MIT. Visits with basal insulin only were not included. Non severe hypoglycaemia was defined as a condition with typical symptoms (e.g., sweating, feeling shaky, impaired concentration) which disappeared quickly after carbohydrate ingestion or a plasma glucose < 2.7 mmol/l with or without typical symp- toms. Severe hypoglycaemia was defined as necessity of intravenous glucose injection. All patients participated in a structured education programme at least once. The normal range and mean value of HbA1c changed from 1999 to 2012, therefore we adjusted HbA1c according the mean normal value of healthy people to the DCC trial (5.05 %). Results: MIT was used in 13901 (66.4 %) of all 20943 visits. Compared to CT these patients were younger (62.0 ± 10.1 vs. 68.7 ± 9.9 years; p < 0.0001), time since diabetes diagnosis was longer (16.5 ± 8.7 vs. 15.8 ± 8.1 years; p < 0.0001), had higher BMI (32.8 ± 6.1 vs. 30.9 ± 5.3 kg/m2; p < 0.0001), a higher daily insulin dose (76.4 ± 52.0 vs. 46.5 ± 26.9 IU/day; p < 0.0001), and lower HbA1c (7.7 ± 1.4 vs. 7.9 ± 1.4 %; p < 0.0001). In MIT more non severe hy- poglycaemic incidences per week (0.3 ± 0.7 vs. 0.2 ± 0.8; p = 0.01) occurred, episodes of severe hypoglycaemia/last 12 months were rare and comparable (0.02/Patient/year) with both insulin therapy strategies. Multiple, linear re- gression analysis showed that the only parameter associated with HbA1c was the insulin dose (β = 0.222, p < 0.001). Higher HbA1c is weakly associated with higher daily insulin dose, but not age (β = -0.019, p = 0.471), time since diabetes diagnosis (β = 0.060, p = 0.023) as well as incidence of non severe hypoglycaemia/week (β = 0.006, p = 0.797). Conclusion: MIT and CT insulin therapy result in comparable metabolic control with HbA1c below 8 %. MIT is associated with higher BMI and Diabetologia (2014) 57:[Suppl1]S1–S564 S 387 1 C higher incidence of non severe hypoglycemia. People with DM2 should be actively involved in the decision which insulin therapy to choose considering the life style and convenience. Therapy strategies can be swopped if necessary. 946 Less nocturnal hypoglycaemia and weight gain with new insulin glargine 300 U/ml vs 100 U/ml: 1-year results in people with type 2 diabetes using basal insulin and OADs (EDITION 2) H. Yki-Järvinen1, R.M. Bergenstal2, G.B. Bolli3, M. Ziemen4, M. Wardecki5, I. Muehlen-Bartmer4, M. Maroccia6, M.C. Riddle7; 1University of Helsinki, Finland, 2International Diabetes Center at Park Nicollet, Minneapolis, USA, 3University of Perugia, Italy, 4Sanofi-Aventis Deutschland GmbH, Frankfurt am Main, Germany, 5Sanofi, Warsaw, Poland, 6Umanis, Levallois-Perret, France, 7Oregon Health & Science University, Portland, USA. Background and aims: EDITION 2 investigated glycaemic control and hy- poglycaemia using new insulin glargine 300 U/ml (Gla-300) versus glargine 100 U/ml (Gla-100) in people with type 2 diabetes mellitus on basal insulin plus OAD(s). Materials and methods: In EDITION 2, 811 adults with type 2 diabetes mellitus and inadequate control of HbA1c were randomised to receive either Gla-300 or Gla-100 for 6 months. In this 6-month open-label extension, par- ticipants continued to receive Gla-300 or Gla-100 once daily plus OADs; 315 (78%) participants in the Gla-300 group and 314 (77%) participants in the Gla-100 group completed 12 months of treatment. Results: Improved control of HbA1c was maintained at 12 months with each regimen. Over 12 months, per participant-year event rates of confirmed (≤3.9 mmol/l [≤70 mg/dl]) or severe nocturnal hypoglycaemia were 37% lower with Gla-300 than Gla-100 (1.74 vs 2.77, RR: 0.63 [95% CI: 0.42 to 0.96]). Fewer participants experienced ≥1 confirmed (≤3.9 mmol/l [≤70 mg/dl]) or severe nocturnal hypoglycaemic event with Gla-300 than Gla-100 (RR: 0.84 [95% CI: 0.71 to 0.99]). Severe hypoglycaemia at any time of day was infre- quent, experienced by 7 participants in the Gla-300 group and 6 participants in the Gla-100 group. Body weight increase was observed in both groups, and was significantly less with Gla-300 than Gla-100 (LS mean [95% CI]: 0.42 [0.04 to 0.80] versus 1.14 [0.76 to 1.52] kg, p=0.0091). No between-treatment differences in adverse events were seen. Conclusion: Over 12 months of treatment, people with type 2 diabetes mel- litus using Gla-300 and OADs had comparable glycaemic control, and experi- enced fewer nocturnal hypoglycaemic events and less weight gain compared with those using Gla-100. Clinical Trial Registration Number: NCT01499095 Supported by: Sanofi PS 076 Glucose variability in insulin treatment 947 New insulin glargine 300 U/ml: glycaemic control and hypoglycaemia in insulin-naïve people with type 2 diabetes mellitus (EDITION 3) G.B. Bolli1, M.C. Riddle2, R.M. Bergenstal3, M. Ziemen4, K. Sestakauskas5, H. Goyeau6, P.D. Home7; 1University of Perugia, Italy, 2Oregon Health & Science University, Portland, USA, 3International Diabetes Center at Park Nicollet, Minneapolis, USA, 4Sanofi-Aventis Deutschland GmbH, Frankfurt am Main, Germany, 5Sanofi, Chilly Mazarin, 6EXPERIS IT, Nanterre, France, 7Newcastle University, Newcastle upon Tyne, UK. Background and aims: EDITION 3 studied the efficacy and safety of new in- sulin glargine 300 U/ml (Gla-300) vs glargine 100 U/ml (Gla-100) in people with type 2 diabetes mellitus uncontrolled on non-insulin therapy. Materials and methods: In this 6-month, multicentre, open-label study, participants were randomised to once-daily Gla-300 or Gla-100 in the even- ing while stopping sulfonylureas. Insulin was titrated seeking FPG 4.4-5.6 mmol/l. The primary endpoint was HbA1c change to month 6. The main secondary endpoint was percentage of participants with ≥1 confirmed (≤3.9 mmol/l [≤70 mg/dl]) or severe nocturnal (00:00-05:59 h) hypoglycaemic event from week 9 to month 6. Results: In total, 878 participants were randomised (baseline characteristics: BMI 33.0 kg/m2; diabetes duration 9.8 years; HbA1c 8.5 % [70 mmol/mol]). HbA1c decreased similarly with both treatments. LS mean changes (SE) were −1.42 [0.05] % (−15.5 [0.5] mmol/mol) with Gla-300 and −1.46 (0.05) % (−16.0 [0.5] mmol/mol) with Gla-100, and LS mean difference (95% CI) was 0.04 (−0.09 to 0.17) % (0.4 [−1.0 to 1.9] mmol/mol). The relative risk of expe- riencing ≥1 confirmed (≤3.9 mmol/l [≤70 mg/dl]) or severe nocturnal hypo- glycaemic event with Gla-300 vs Gla-100 was 0.76 (95% CI: 0.59 to 0.99) for the 6-month treatment period, 0.74 (0.48 to 1.13) for baseline to week 8, and 0.89 (0.66 to 1.20) (p=0.45) for week 9 to month 6. Using a stricter thresh- old of S 388 1 C Conclusion: In insulin-naïve people with type 2 diabetes mellitus, new insu- lin glargine 300 U/ml provides comparable effective glycaemic control, with less hypoglycaemia, compared with glargine 100 U/ml. Clinical Trial Registration Number: NCT01676220 Supported by: Sanofi. 948 Similar efficacy and safety with LY2963016 insulin glargine compared with insulin glargine in patients with type 2 diabetes mellitus: the ELEMENT 2 study P. Hollander1, J. Rosenstock2, A. Bhargava3, L.L. Ilag4, R.K. Pollom4, W.J. Huster4, M.J. Prince4; 1Baylor Endocrine Center, Dallas, 2Dallas Diabetes and Endocrine Center, 3Iowa Diabetes and Endocrinology Research Center, Des Moines, 4Eli Lilly and Company, Indianapolis, USA. Background and aims: LY2963016 (LY IGlar) and insulin glargine (Sanofi- Aventis; IGlar) are both insulin glargine products, with identical amino acid sequences. Even with identical primary structure, protein-based therapeutics manufactured by distinct processes must be shown to be clinically similar. Materials and methods: This was a 24-week, Phase 3, randomized, double- blind, parallel study to compare the efficacy and safety of LY IGlar and IGlar. The primary objective was to test the non-inferiority (0.3% margin) of LY IGlar to IGlar as measured by change in HbA1c from baseline to 24 weeks in patients with T2DM on ≥2 oral antihyperglycaemic medications (OAMs). Testing for non-inferiority of IGlar to LY IGlar was also performed and pre-specified as a complementary hypothesis, which if met along with the primary aim, would demonstrate equivalent efficacy between LY IGlar and IGlar. Patients were insulin-naïve (HbA1c ≥7.0% to ≤11.0%) or previously on IGlar (HbA1c ≤11.0%). Insulin-naïve patients started on 10 U/day while those previously on IGlar started at an equivalent dose to their prestudy IGlar dose. Both groups followed a patient-driven titration algorithm until fasting blood glucose reached ≤5.6 mmol/L. For blinding purposes, patients randomized to treatment were provided syringes and insulin vials contained in a partial cover that concealed distinguishing differences in vial appearance. Results: Both treatment groups had within-group similarly significant (p S 389 1 C in target SG range (4.4-7.8 mmol/l) vs Gla-100, with less hypoglycaemia, a mean glucose profile for all subjects with less glucose excursion, and lower within- and between-day glucose variability. Clinical Trial Registration Number: NCT01658579 Supported by: Sanofi 950 Getting the time right: is there any relationship between insulin administration-to-meal ingestion time interval and blood glucose control in type 2 diabetes? U. Bujacz-Jedrzejczak, J. Loba, L. Czupryniak; Medical University of Lodz, Poland. Background and aims: Patients treated with short-acting or premixed analog insulins are advised to take them immediately before a meal while those treated with classic human insulin are instructed to take insulin 20-30 minutes prior to a meal. The patients find the latter piece of advice difficult to comply with, and whether the compliance with recommended insulin dose- to-meal time interval (I-to-M, in minutes) has any relevance to metabolic control is unknown. We conducted a prospective observational study aiming at assessing the relationship between glucose control and mean duration of I-to-M. Materials and methods: The study group consisted of 90 type 2 diabetes patients (mean age 64±12 years, duration of diabetes 9.4±4.7 years, HbA1c 7.5±1.5%, BMI 30.8±4.4 kg/m2), treated with insulin twice daily (before breakfast and evening meal), who were asked to record the actual hours of insulin injections and meal initiation for 30 consecutive days for I-to-M to be calculated. HbA1c was measured within 10 days after recordings were completed. Results: There was no correlation between HbA1c and mean I-to-M in the whole study group. However, in the patients treated with analog insulins (n=51) there was a trend to positive correlation between HbA1c and mean I-to-M, while in those treated with human insulins (n=39) the trend was op- posite (Fig. 1). Conclusion: We conclude that compliance with the recommended pre-meal insulin dose timing might have some, albeit minor, impact on overall glucose control in type 2 diabetes. 951 The impact of pure protein load on glucose levels in type 1 diabetes patients treated with insulin pumps T. Klupa1,2, T. Benbenek-Klupa3, B. Matejko1, S. Mrozińska1, M. Małecki1,2; 1Department of Metabolic Diseases, Jagiellonian University, 2University Hospital, Krakow, Poland, 3DiabMed Enterprise, Kraków, Poland. Background and aims: It is well recognized that the mealtime insulin re- quirement in type 1 diabetes (T1DM) patients is driven mostly by carbohy- drate content and controlling it may improve glucose levels. However, it is uncertain whether the calculation of other food components may optimize glycemic control. The aim of this study was to estimate the impact of pure protein load on glucose levels in T1DM patients treated with personal insulin pumps. Materials and methods: We examined 10 T1DM patients (6 females, 4 males, mean age - 32.3 years, mean T1DM duration - 11.7 years, mean HbA1c - 6.85%) treated with insulin pump (Medtronic Paradigm 722 or Veo) equipped with a continuous glucose monitoring system (CGMS) option. In phase I of the study, baseline insulin infusion was optimized to minimize dif- ferences in fasting glucose levels to less than 30 mg/dL between any two time points between 9 am and 3 pm. The procedure of optimization was based on a retrospective analysis of CGMS records and included increasing or decreas- ing the rate of basal insulin infusion two hours before the observed rise or fall in glucose level. Then, the new settings of the basal infusion rate were re- checked with CGMS to meet the study criteria. In phase II, the patients were exposed to single, pure protein load (Protifar, Nutricia) at the dose of 0.3 g/kg (0.34 mL of Protifar/kg dissolved in 200 mL of water) of body weight admin- istered at 9 a.m. Such a dose of protein is an equivalent of the usual protein portion in a medium-size meal, based on the dietary recommendations for patients with diabetes (15-20% of energy). The rate of basal insulin infusion during protein load was the same as defined initially in phase I, no modifica- tions or extra insulin boluses were permitted. CGMS record was performed in both phases, during which the patients avoided physical activity. Glucose patterns were defined during 6 hours of phase I (fasting) and phase II (protein load). Results: Mean baseline glucose level was 119.8 and 117.6 mg/dL for Phase I and Phase II, respectively (p=0.68). Mean maximal glucose level was 146.4 and 145.2 mg/dL for Phase I and Phase II, respectively (p=0.85). Mean maxi- mal glucose level increment was similar for the entire 6-hour lasting fasting and protein load test (26.6 mg/dL vs. 27.6 mg/dL, respectively, p=0.78). There was no difference between the change in baseline vs. 6th hour glucose levels for the fasting state vs. protein load test (12.5 mg/dL and 19 mg/dL, respec- tively, p=0.04). The glucose variability assessed by CGMS-based standard de- viation of mean glucose levels was 36.4 and 37.9 mg/dL for Phase I and Phase II, respectively (p=0.01). Diabetologia (2014) 57:[Suppl1]S1–S564 S 390 1 C Conclusion: The administration of a pure protein load does not seem to have a clinically significant impact on glucose levels in T1DM patients treated with insulin pumps. Thus, we do not recommend protein content counting for glycemic control in this group of patients. 952 Effects of insulin degludec and insulin glargine on day-to-day fasting blood glucose variability in patients with type 1 diabetes T. Nakamura1, K. Sakaguchi1, A. So1, S. Nakajima2, M. Takabe2, H. Komada1, Y. Okuno1, Y. Hirota1, T. Nakamura2, K. Iida2, M. Kajikawa2, M. Nagata2, W. Ogawa1, S. Seino3; 1Diabetes, Metabolism and Endocrinology, Kobe University Graduate School of Medicine, 2Best Basal Insulin Study Group, Kobe, 3Molecular and Metabolic Medicine, Kobe University Graduate School of Medicine, Japan. Background and aims: Insulin degludec (IDeg), a novel ultra-long acting insulin analogue, has been suggested to have a longer and flatter glucose- lowering effect than insulin glargine (IGlar). It is thus possible that IDeg yields lower glycemic variability, in particular at fasting state. We have com- pared the effects of IDeg and IGlar on day-to-day variability of fasting blood glucose (FBG) in type 1 diabetes subjects treated with basal-bolus insulin injections. Materials and methods: This open-label, multicenter, randomized crossover trial investigated the effects of 4-week basal-bolus insulin treatment with ei- ther IDeg or IGlar as a basal insulin. Patients who were found to satisfy the inclusion criteria were randomly allocated IGlar (first period) / IDeg (sec- ond period) or IDeg (first period) / IGlar (second period) groups. The last week of each treatment period was the data-collecting phase during which the patients were directed to examine blood glucose level 7 times per day. The level of glycoalbumin was measured at the end of each treatment period. The primary end point was the standard deviation (SD) and the coefficient of variation (CV) of FBG during the last week of the treatment period. Second- ary end points included the level of glycoalbumin, the dose of daily insulin, the intra-day variability of blood glucose level, and the frequency of severe hypoglycemia. Data are shown as means ± SD, and were statistically analyzed with the paired t-test. Results: Thirty-six randomized subjects (17 in the IDeg/IGlar and 19 in the IGlar/IDeg groups) were recruited, and the data of 32 subjects who com- pleted the trial were analyzed (age: 57.4 ± 13.9 years, BMI: 22.6 ± 3.2 kg/m2, HbA1c: 7.4 ± 0.8 %). The mean (139.5 ± 31.8 vs. 154.2 ± 37.2 mg/dL, P = 0.03) and the SD (46.9 ± 17.5 vs. 57.5 ± 24.6 mg/dL, P = 0.02) of FBG were smaller during the IDeg treatment period than those during the IGlar treatment pe- riod whereas the CV was not different between the two periods (IDeg 34.3 ± 13.3 vs. IGlar 37.1 ± 13.0 mg/dL, P = 0.29). The level of glycoalbumin was similar in the two treatment periods (IDeg 21.5 ± 3.0 vs. IGlar 21.8 ± 3.6 mg/ dL, P = 0.39). The daily dose of basal insulin in the IDeg period was slightly but significantly smaller than that in the IGlar period (IDeg 11.0 ± 5.2 vs. IGlar 11.8 ± 5.6 units/day, P S 391 1 C Conclusion: Similar to carbohydrate counting errors, the direction and ROC at the time of the insulin dose can have major impacts on subsequent rates of hypoglycemia and hyperglycemia. Accordingly, in order to optimize glyce- mic control, the direction and ROC should be considered when determining a meal insulin dose. This model suggests that using CGM data to determine a dose has glycemic advantages relative to using a static SMBG measurement. 955 Least glucose variability is observed with the combination of a GLP-1 agonist and basal insulin among four commonly used insulin regimens in type 2 diabetes (VARIATION study) H.S. Bajaj1, R. Aronson2, C. Ye3, K. Venn2, A.L. Patrick2; 1LMC Diabetes & Endocrinology, Brampton, 2LMC Diabetes & Endocrinology, Toronto, 3McMaster University, Hamilton, Canada. Background and aims: The majority of insulin management trials to date have focused on HbA1c as the primary measure of glycemic control. Limited data pertaining to glucose variability with various insulin regimen exists in the literature. This cohort study compares glucose variability in patients with well-controlled type 2 diabetes (DM) among four commonly-used insulin regimen (A= basal insulin + oral drugs, B= basal insulin + GLP-1 agonist, C= pre-mix insulin and D= basal-bolus insulin). Materials and methods: Eighty patients were recruited from five of our dia- betes clinics in Ontario, Canada between November 2013 and March 2014. Study inclusion criteria were 1. type 2 DM; 2. age between 18 to 80 years; 3. BMI ≤ 45 kg/m2; 4. stable insulin regimen for a minimum of 6 months; 5.clin- ically stable HbA1c value of ≤7.5% within 3 months before study enrolment. Participants were instructed to follow usual diet and exercise regimen dur- ing the 6-day, masked Continuous Glucose Monitoring (CGM) period. Daily glucose standard deviation (SD) was chosen as the primary outcome for the study. Six patients were excluded from analysis (three for clinical instability; three for missed medication doses during 6-day CGM period). Results: Key baseline clinical characteristics as well as CGM results are pre- sented in Table 1. Each of the four study cohorts had an average HbA1c of 7.0%. Statistically significant differences were observed on inter-cohort com- parisons on Analysis of Covariance (ANCOVA) modeling - with the lowest SD observed in cohort B compared to cohorts C and D (1.8 mmol/L vs. 2.1 and 2.2 mmol/L; p=0.05 and 0.01, respectively). SD of cohort B was also nu- merically less compared to that for cohort A (2.0 mmol/L), however this did not reach statistical significance (p=0.15). Self- reported hypoglycemia epi- sodes during the 6- day study period were highest among cohort D compared to the other three cohorts (p=0.01 vs. cohorts A & B; p=0.03 vs. cohort C). No other statistically significant differences were observed on inter-cohort comparisons for CGM measurements (frequency, duration and degree of hy- poglycemia or hyperglycemia). Conclusion: Among patients with Type 2 DM who have well-controlled HbA1c on insulin, least glucose variability is observed in the combination regimen using basal insulin with a GLP-1 agonist. This reduced glucose vari- ability with GLP-1 agonists suggests their complementary glycemic action with basal insulin and is explained by known mechanisms of action for this incretin-based injectable medication class - with multi-pronged effects on both alpha and beta cells in the pancreas as well as appetite suppression. 956 Sodium glucose co-transporter-2 (SGLT2) inhibitor empagliflozin (EMPA) in type 1 diabetes (T1D): impact on diurnal glycaemic patterns B.A. Perkins1, D.Z. Cherney2, H. Partridge1, N. Soleymanlou3, H. Tschirhart1, B. Zinman1, R. Mazze4, N. Fagan5, S. Kaspers6, H.J. Woerle6, U.C. Broedl6, O. Johansen7; 1Department of Medicine, Division of Endocrinology, University of Toronto, 2Department of Medicine, Division of Nephrology, Toronto General Hospital, University of Toronto, 3Boehringer Ingelheim Canada Ltd./Ltée, Burlington, Canada, 4International Diabetes Center, Park Nicollet Institute, Minneapolis, 5Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, USA, 6Boehringer Ingelheim, Ingelheim, Germany, 7Boehringer Ingelheim Norway KS, Asker, Norway. Background and aims: We recently reported improved glycemic control, de- creased hypoglycemia incidence, and reduced insulin dose in T1D patients treated with open-label EMPA. To further characterize the effects, we ana- lysed glucose patterns by continuous glucose monitoring (CGM). Materials and methods: In this 8-week single-arm open-label pilot study of EMPA (NCT01392560), we compared ambulatory glucose profiles (AGP) during 2-week intervals in a placebo run-in, end-of-treatment, and post- treatment period. Glucose exposure was measured by total area under the curve (AUCTOTAL in mg/dL∙h), AUCNIGHT (23:00 to 6:59h) and AUCDAY (7:00 to 22:59h). Glucose Variability was represented by inter-quartile range and Glucose Stability by the average hourly absolute change (in mg/dL/hr) of all CGM readings in the 2-week periods. Results: 40 patients aged 24±5 years, and A1c 8.0±0.9% completed the trial. 39 patients had data available for the analysis of the CGM endpoints. Though A1c decreased (8.0±0.9% to 7.6±0.9%, p S 392 1 C PS 077 Old and new insulin formulations 957 A single dose pharmacokinetic study of basal insulin peglispro in subjects with hepatic impairment S.L. Choi1, C. Loghin2, P. Garhyan3, M.P. Knadler3, E. Lam1, V. Sinha4, N. Porksen3, H. Linnebjerg3; 1Eli Lilly-NUS Centre for Clinical Pharmacology, Singapore, 2Clinical Pharmacology, Eli Lilly and Company, Indianapolis, 3US Endocrinology/Diabetes Care, Eli Lilly and Company, Indianapolis, 4Office of Clinical Pharmacology, Center for Drug Evaluation and Research, Silver Spring, USA. Background and aims: Basal insulin peglispro (BIL; LY2605541) is being developed for the treatment of type 1 and type 2 diabetes mellitus (T2DM) as a once daily subcutaneous (SC) administration. BIL is a novel, PEGylated insulin lispro that has a large hydrodynamic size. It has prolonged duration of action which is related to a delay in insulin absorption and a reduction in clearance. The patient population may include those with hepatic impair- ment, so the pharmacokinetics (PK) of BIL was assessed in subjects with varying degrees of hepatic impairment compared to healthy subjects with normal hepatic function. Materials and methods: In this parallel-group, open-label study, 35 subjects received a single SC dose of 0.33 U/kg (0.075 mg/kg) BIL and blood samples for PK were taken up to 9 days postdose. BIL PK was determined using a spe- cific, validated enzyme-linked immunosorbent assay method. PK parameters were calculated using noncompartmental analyses. The effect in subjects with hepatic impairment was compared to subjects with normal hepatic function using the ratio of least squares (LS) geometric means of primary PK param- eters: area under the concentration time curve from zero to infinity (AUC[0- ∞]) and maximum observed drug concentration (Cmax). AUC(0-∞) and Cmax were log transformed and analyzed using an Analysis of Variance model with group as a factor. Results: In total, 35 subjects were assessed (12 with normal hepatic func- tion; 8 mild, 8 moderate, 7 severe hepatic impairment, based on the Child- Pugh criteria). Subjects were nondiabetic, apart from one subject with T2DM (moderate hepatic impairment). Subjects were male (57%) and female (43%), 41-69 years old, with mean body mass index (SD) 27.7 kg/m2 (4.52 kg/m2). Healthy subjects were matched to hepatically impaired subjects for age (±10 years), weight (±10 kg), and gender. Ratios of the LS geometric means and 90% confidence intervals (CIs) for AUC(0-∞) for mild, moderate, and severe hepatically impaired subjects to the healthy subjects were 0.789 (0.556, 1.12), 0.744 (0.525, 1.06), and 0.782 (0.543, 1.12), respectively. Ratios (90% CIs) for Cmax were 0.728 (0.398, 1.33), 0.905 (0.495, 1.65), and 0.840 (0.448, 1.58), re- spectively. Median time of Cmax and mean half-life were similar between the hepatically impaired subjects and healthy subjects. Healthy subjects reported 7 treatment emergent adverse events (TEAEs) (5 of 12 subjects; 41.7%) and hepatically impaired subjects reported 7 TEAEs (1 TEAE reported by 1 of 8 [12.5%] mild; 3 TEAEs reported by 2 of 8 [25.0%] moderate, and 3 TEAEs reported by 2 of 7 [28.6%] severe). The majority of TEAEs were mild in sever- ity. TEAEs experienced by >1 subject included hypoglycemia, nausea, fatigue, and headache. Conclusion: Single doses of BIL were well tolerated by hepatically impaired and healthy subjects in this study. A substantial overlap of AUC(0-∞) and Cmax was observed across all groups. Although the mean AUC(0-∞) and Cmax were lower in hepatically impaired subjects, the ratios between hepatically impaired and healthy subjects were not statistically different. Based on data from this study, requirement for dose adjustment of BIL in subjects with he- patic impairment would not be anticipated. Clinical Trial Registration Number: NCT01751399 Supported by: Eli Lilly and Company 958 Conventional insulin therapy (CT) vs multiple insulin injection (MIT) in type 2 diabetes: analysis of a quality assurance project (DPV) C.S. Kloos1, G. Kramer1, U.A. Müller1, R. Schiel2, S. Hofer3, S. Merger4, W. Kerner5, R.W. Holl6; 1Internal Medicine, Klinikum für Innere Medizin III, Jena, 2Medigreif Inselklinik Heringsdorf, Germany, 3Medical University of Innsbruck, Austria, 4Pediatric Department, University Ulm, 5Karlsburg Hospital, 6Institute for Epidemiology and Medical Biometry, Ulm, Germany. Background and aims: People with type 2 diabetes and insulin therapy re- ceive predominantly an intensive form of insulin treatment with multiple in- sulin injections (MIT). Evidence from randomized studies for the superiority of MIT over CT in type 2 diabetes is lacking, but MIT produced increased weight gain, higher costs of therapy and more hypoglycaemic episodes. This survey provides data from a project of quality assurance in Germany and Austria (DPV) with documentation of insulin treatment in type 2 diabetes. Materials and methods: Patients older than 18 years with type 2 diabetes and insulin treatment were extracted from the DPV database. CT was defined as 2 insulin injection time-points per day, MIT as more than 2 insulin injection time-points per day. Patients with less than 2 daily insulin injection time- points and patients with LADA (positive ß-cell antibodies) were excluded from the analysis. Patients with insulin therapy and Metformin as concomi- tant therapy were included, but not evaluated separately. Results: A total of 88 519 (CT 23 118 (26%); MIT 65 401 (74%)) patients were analysed. Patients with CT (44% male) compared to patients with MIT (53% male) were older (74.6 (10.2) vs. 66.7(12.0) years), had comparable time from diabetes diagnosis (13.0 (9.5) vs. 13.4 (9.8) years), lower BMI (29.1 (5.7) vs. 31.7 (6.7)kg/m2), lower daily insulin dose (39.2 (20.2) vs. 65.3 (41.1) IE/day) and better HbA1c (7.81 (1.9) vs. 8.14 (2.0)%; p S 393 1 C effects of its mono-components throughout the dose/exposure range (see Figure). Conclusion: The PK properties of IDeg and Lira are preserved in the IDegLi- ra co-formulation; both mono-components contribute to glycaemic control across the recommended dose range. Supported by: Novo Nordisk A/S 960 Aspart pharmacokinetics in children with type 1 diabetes during closed loop insulin delivery: comparison between diluted (U20) and standard (U100) insulin strengths Y. Ruan1, D. Elleri1, M. Tauschmann1, J.M. Allen1, M.E. Wilinska1, D.B. Dunger2, R. Hovorka1; 1Wellcome Trust-MRC Institute of Metabolic Science, University of Cambridge, 2Department of Paediatrics, University of Cambridge, UK. Background and aims: To compare pharmacokinetics of two different con- centrations of insulin aspart in children aged 3 to 6 years with type 1 diabetes (T1D) during closed loop insulin delivery. Materials and methods: Young children with T1D underwent an open-label, randomized, two-period crossover study involving two occasions in a clinical research facility, 2 to 6 weeks apart, from 17:00 day 1 to 8:00 day 2 with an evening meal consumed at 17:00. In random order, diluted (1:5 dilution with normal saline, 20 IU/ml) or standard strength (100 IU/ml) insulin aspart was administered via an insulin pump as a meal bolus and then overnight dur- ing closed loop advised by a model predictive algorithm. Plasma insulin was measured every 30 to 60 min. A compartment model adopting Bayesian pa- rameter estimation was used to measure time-to-peak of insulin concentra- tion (tmax), insulin metabolic clearance rate (MCRI), and background insulin concentration (insc). Results: Eleven children [6 male; age 5.07(range 3.75-6.96yrs); A1C 60(14) mmol/mol; BMI sds 1.0(0.8); duration of diabetes 2.2(1.0)yrs; mean (SD); total daily dose 12.9(10.6,16.5)IU; fasting C-peptide 5(5,17.1)pmol/l; median(IQR)] participated in the study. Pharmacokinetic parameters (Table) were not different between the two aspart concentrations but tmax showed less inter-subject variability following administration of diluted aspart (SD 8.7 vs. 14.4min; diluted vs. standard strength; P =0.047). insc but not the other parameters was positively correlated with total daily insulin dose (Spearman’s rs = 0.7; P =0.014) but not with fasting C-peptide (rs = -0.19; P = 0.39) sug- gesting that it does not reflect residual insulin secretion. Conclusion: Diluting insulin aspart does not change its pharmacokinetics, but may result in less variable absorption and could be used in young people with T1D undergoing closed loop insulin therapy. Clinical Trial Registration Number: NCT01557634 Supported by: JDRF, Diabetes UK and EU FP7 SPIDIMAN project 961 New insulin glargine 300 U/ml: efficacy and safety of flexible vs fixed dosing intervals in people with type 2 diabetes mellitus N. Jeandidier1, M.C. Riddle2, G.B. Bolli3, P.D. Home4, R.M. Bergenstal5, M. Ziemen6, I. Muehlen-Bartmer6, M. Wardęcki7, L. Vinet8, H. Yki-Järvinen9; 1University of Strasbourg, France, 2Oregon Health & Science University, Portland, USA, 3University of Perugia, Italy, 4Newcastle University, Newcastle Upon Tyne, UK, 5International Diabetes Center at Park Nicollet, Minneapolis, USA, 6Sanofi-Aventis Deutschland GmbH, Frankfurt am Main, Germany, 7Sanofi, Warsaw, Poland, 8EXPERIS IT, Nanterre, France, 9University of Helsinki, Finland. Background and aims: The smoother, more prolonged PK and PD profiles of new insulin glargine (300 U/ml; Gla-300) vs glargine 100 U/ml (Gla-100), with glycaemic control extending beyond 24 h, provide a rationale for flexible timing of daily injections depending on individual lifestyle. These 3-month sub-studies, of Gla-300 injected each evening, compared the efficacy and safety of using a fixed 24-h dosing interval with a flexible dosing regimen allowing between-injection intervals of 24 ± 3 h on at least 2 days each week. Materials and methods: The background multicentre, 6-month, open-label studies compared Gla-300 vs Gla-100 in people with type 2 diabetes mellitus using basal + meal-time insulin (EDITION 1) or basal insulin + OADs (EDI- TION 2). Participants (EDITION 1, N=109; EDITION 2, N=89) using Gla- 300 were randomised at month 6 to continue the fixed regimen or move to flexible intervals. Efficacy and safety were evaluated 3 months later (month 9). Results: The frequency of maintaining a 24 ± 1 h interval between injections was ~88% with the fixed regimen and ~60% with the flexible regimen. HbA1c change (primary endpoint) was comparable in fixed vs flexible regimens (Ta- ble). Percentages of participants with ≥1 hypoglycaemia at any time of day (24 h), or ≥1 nocturnal (00:00-05:59 h) hypoglycaemia, were also compara- ble. Severe hypoglycaemia was experienced by only 1 participant, in the EDI- TION 1 sub-study. There were no differences in adverse events. Diabetologia (2014) 57:[Suppl1]S1–S564 S 394 1 C Conclusion: Occasional flexibility in timing of daily Gla-300 injections by ± 3 h resulted in similar efficacy and safety compared with advising injections at a fixed time each day. Clinical Trial Registration Number: NCT01499082, NCT01499095 Supported by: Sanofi. 962 Unmet needs in HbA1c goal achievement (< 7.0%) by patients with type 2 diabetes mellitus treated with basal insulin: pooled randomised controlled trial data and real-world clinical practice data P. Chava1, L. Blonde1, S. Brunton2, R. Zhou3, J. Meyers4, K. Davis4, M. Dalal5, A. DiGenio6; 1Ochsner Medical Center, New Orleans, 2University of North Carolina, Chapel Hill, 3Medpace Inc., Cincinnati, 4RTI Health Solutions, Research Triangle Park, 5Sanofi US, Inc., Bridgewater, 6Isis Pharmaceuticals, Inc., Carlsbad, USA. Background and aims: Despite the efficacy of basal insulin therapy in pa- tients with type 2 diabetes mellitus (T2DM), many patients do not reach glycaemic goals. Here we use both randomized clinical trial (RCT) data and “real-world” data from a retrospective study of electronic medical records (EMR) to assess the characteristics of T2DM patients who do not achieve target glycaemic control while on basal insulin (HbA1c ≥ 7.0%) in order to better understand the clinical characteristics of these patients. Materials and methods: Patients with T2DM aged ≥ 18 years treated with basal insulin, but with HbA1c ≥ 7.0% were identified by analysing results from 11 RCTs at 6 months after the index date, defined as date of first prescrip- tion for basal insulin, and results from the GE Centricity EMR database at 6-months and at 12-months follow-up after the index date. Patients were stratified based on fasting plasma glucose (FPG) levels (< 130 or ≥ 130 mg/ dL). Information on patient demographic and clinical characteristics were available from both sources of data. Results: The table shows that in the RCTs around 52% achieved HbA1c < 7.0% (recommended by the ADA for most diabetic patients). HbA1c < 7.0% was achieved by fewer patients (~27%) in the EMR database at both 6 and 12 months. Among those with HbA1c ≥ 7.0%, 54.9% of RCT patients and 17.6% and 27.7% of EMR patients at 6 and 12 months, respectively, had FPG < 130 mg/dL. About half of the RCT patients not achieving goal had a FPG ≥ 130 mg/dL. This suggested that these patients needed further basal insulin titra- tion, whereas patients at goal likely required postprandial glucose control. In the EMR cohort, > 70% of patients were likely to need additional basal insulin titration. Conclusion: Failure to adequately titrate basal insulin is an unmet need in many T2DM patients, which is evident even in RCTs. When basal insulin is adequately titrated and FPG is controlled, additional postprandial treatment may be needed. Another important unmet need is understanding the causes of failure to achieve control of FPG with basal insulin. Supported by: Sanofi US, Inc. 963 New insulin glargine 300 U/ml: glycaemic control and hypoglycaemia in a meta-analysis of phase 3a EDITION clinical trials in people with type 2 diabetes mellitus R.A. Ritzel1, R. Roussel2, G.B. Bolli3, L. Vinet4, H. Yki-Järvinen5; 1Klinikum Schwabing, Städtisches Klinikum München GmbH, Germany, 2Assistance Publique Hôpitaux de Paris, Bichat Hospital, France, 3University of Perugia, Italy, 4EXPERIS IT, Nanterre, France, 5University of Helsinki, Finland. Background and aims: The EDITION 1, 2 and 3 studies compared the effi- cacy and safety of new insulin glargine 300 U/ml (Gla-300) with insulin glar- gine 100 U/ml (Gla-100) in people with type 2 diabetes mellitus on basal and mealtime insulin, basal insulin and OADs, and no prior insulin, respectively. A meta-analysis of these studies was conducted to study efficacy and safety. Materials and methods: Meta-analysis of these three studies enabled glycae- mic control and hypoglycaemia to be examined over 6 months in a large, heterogeneous type 2 diabetes mellitus population (Gla-300, N=1247; Gla- 100, N=1249). Results: Mean change in HbA1c was comparable for Gla-300 and Gla-100 (LS mean change [SE]: −1.02 [0.03] % for both groups). Gla-300 was associated with reduced risk of experiencing hypoglycaemia compared with Gla-100 (nocturnal and at any time of day; Table). Nocturnal hypoglycaemic event rates were consistently lower with Gla-300 than Gla-100. Severe hypogly- caemia was rare in both treatment groups (2.3% with Gla-300 vs 2.6% with Gla-100). Weight gain with Gla-300 and Gla-100 was slight (LS mean change [SE]: 0.49 [0.10] kg and 0.75 [0.10] kg, respectively), with a trend for less weight gain with Gla-300 (LS mean difference: −0.26 [95% CI: −0.52 to 0.01] kg, p=0.058). Conclusion: Gla-300 provides comparable glycaemic control to Gla-100 in type 2 diabetes mellitus, with consistently less hypoglycaemia at any time of the day and less nocturnal hypoglycaemia. Diabetologia (2014) 57:[Suppl1]S1–S564 S 395 1 C Clinical Trial Registration Number: NCT01499082/ NCT01499095/ NCT01676220 Supported by: Sanofi 964 Long-term efficacy and safety of insulin degludec in combination with bolus insulin aspart in children and adolescents with type 1 diabetes N. Thalange1, L.C. Deeb2, V. Iotova3, T. Kawamura4, G.J. Klingensmith5, A. Philotheou6, J. Silverstein7, S. Tumini8, A. Ocampo Francisco9, O. Kinduryte10, T. Danne11; 1Norfolk & Norwich University Hospital, Norwich, UK, 2Larry Deeb PA, Tallahassee, USA, 3University Hospital Sveta Marina, Varna, Bulgaria, 4Osaka City University, Japan, 5Barbara Davis Center for Childhood Diabetes, Aurora, USA, 6UCT Private Academic Hospital, Cape Town, South Africa, 7University of Florida, Gainesville, USA, 8Ospedale „SS. Annunziata“, Chieti, Italy, 9Novo Nordisk A/S, Søborg, 10Novo Nordisk A/S, Bagsværd, Denmark, 11Children‘s Hospital auf der Bult, Hannover, Germany. Background and aims: This 1:1 randomised controlled, open-label, treat-to- target trial investigated the efficacy and safety of insulin degludec (IDeg) and insulin detemir (IDet), both in combination with bolus insulin aspart (IAsp) in children and adolescents with type 1 diabetes for 26 weeks (n=174 IDeg; 176 IDet) followed by a 26-week extension (n=152 IDeg; 128 IDet). Materials and methods: The trial included 85 children aged 1-5 years (IDeg: 43), 138 children 6-11 years (IDeg: 70) and 127 adolescents 12-17 years (IDeg: 61). Randomised subjects [baseline mean (SD) diabetes duration 4.0 (3.5) years, HbA1c 8.1 (1.1)%, fasting plasma glucose (FPG) 8.7 (5.1) mmol/L], re- ceived IDeg once daily + IAsp or IDet once/twice daily +IAsp. Results: Non-inferiority of IDeg vs. IDet at 26 weeks was confirmed for HbA1c (primary endpoint): estimated treatment difference (ETD) 0.15% [-0.03; 0.32]95% CI (non-inferiority margin: 0.4%, p S 396 1 C PS 078 Insulin immunogenicity 965 Injection depth does not affect the pharmacokinetics or pharmacodynamics of insulin lispro in healthy obese or normal weight subjects K. Yeo1, A. de la Peña2, E. Catton2, D.A. Ignaut2; 1Lilly-NUS Centre for Clinical Pharmacology, Singapore, 2Eli Lilly and Company, Indianapolis, USA. Background and aims: An 8-mm needle length is commonly used world- wide, and recent recommendations suggest that using shorter, 4-, 5-, or 6-mm needles may be optimal for most patients to avoid intramuscular injection and still maintain adequate glucose control. The goal of this analysis was to compare the pharmacokinetics (PK) and pharmacodynamics (PD) of insulin lispro after administration at a 5-mm injection depth or an 8-mm injection depth in 2 separate populations: normal weight (Study 1) or obese (Study 2). Materials and methods: In both open-label, randomized, 2-period crosso- ver euglycemic clamp studies, subjects received single doses of insulin lispro (0.25 U/kg) on 2 different occasions; PK and PD measures were collected under euglycemic conditions for up to approximately 6 hours post-dose. Results: In Study 1, 16 healthy normal weight subjects (13 male; age 31±9 years; BMI=23±2 kg/m2) were dosed at a 5-mm and an 8-mm injection depth on 2 separate occasions. In Study 2, 16 healthy obese subjects (11 male; mean age 41±11 years; BMI=34±3 kg/m2) were also dosed at a 5-mm and an 8-mm injection depth. In both studies, PK parameters [AUC(0-tlast), AUC(0-∞), and Cmax] and PD parameters (Gtot, Rmax) were log-transformed prior to analysis using a mixed effects model. The table shows PK and PD parameters for both studies. Conclusion: There were no statistical differences in PK or PD of insulin lis- pro following administration at the 5-mm injection depth or 8-mm injection depth in either study. Injection depths in the 5-8 mm range did not affect the PK or PD of insulin lispro in either normal weight or obese subjects. These observations are consistent with clinical data in pediatric, adult, and obese patients with type 1 and type 2 diabetes mellitus, where the 5-mm needle was found to be as efficacious and safe as an 8-mm needle. 966 Development of a clinical predictive score for requirement of hypoglycaemic agents for optimal control of blood glucose during glucocorticoid treatment H. Morita1, M. Yamauchi1, I. Mori1, T. Ikeda1, Y. Kitada2, K. Taguchi3, K. Fujioka3, H. Okada4, T. Usui5, M. Kawashima1, K. Kajita1, T. Ishizuka3; 1Department of General Internal Medicine, Gifu University Graduate School of Medicine, 2Department of General Internal Medicine, Hashima City Hospital, 3Department of General Internal Medicine, Gifu Municipal Hospital, 4Department of General Medicine, Gifu Prefectural Central Medical Center, 5Department of General Medicine, Japanese Red Cross Gifu Hospital, Japan. Background and aims: Long-term glucocorticoid (GC) treatment may be necessary for patients with various autoimmune diseases. The highest dose of GC is usually administered at the Initiation of the therapy, which often causes hyperglycemia and needs hypoglycemic agents in some patients. The aim of the present study is to elucidate risk factors related to GC-induced hypergly- cemia and to develop a simple method to predict in advance the requirement of hypoglycemic agents to achieve optimum control of blood glucose during GC treatment. Materials and methods: We retrospectively evaluated 230 patients (57±19 years, 57% of whom were women) with autoimmune diseases admitted be- tween 2004 and 2013 in our department to develop a predictive scoring sys- tem. Inclusion criteria indicated that the dosage of prednisolone treatment should be more than 5 mg/day for 4 weeks at least. Exclusion criteria included use of hypoglycemic agents before GC administration or concomitant use of anti-cancer drugs. Results: Forty-three percent (Ninety-nine persons) of GC-administered pa- tients required hypoglycemic agents (19% for oral hypoglycemic agents, 5% for a GLP-1 analog or 20% for insulin). A logistic regression analysis revealed that risk factors of requirement of hypoglycemic agents during GC adminis- tration should be raised in accordance with age and fasting plasma glucose (FPG). A ROC analysis showed that cutoff values of age and FPG were 58 years old and 101 mg/dl, respectively. We scored patient profiles as follows: 0 point for FPG less than 101 mg/dl and age less than 58 years old, 1 point for FPG more than 101 mg/d and age more than 58 years old, respectively. Mean of the sum of the points was 1.27 for patients requiring hypoglycemic agents and 0.64 for others, respectively. The percentages of patients requiring hypo- glycemic agents in the sum of each point were 17% (12 / 72) at 0 point , 45% (48 /106) at 1 point and 75% (39 /52) at 2 points, respectively. The sensitivity and specificity were 87.9% and 45.8%, respectively, when 1 point was defined as the cut-off value of the sum of the scores. Conclusion: Use of a scoring system composed in accordance with age and FPG is a simple and effective way to predict the requirement of hypoglycemic agents for hyperglycemia in patients treated with GC. 967 Insulin antibodies are associated with lipoatrophy in adults with type 2 diabetes mellitus K. Takahashi, A. Hakozaki, M. Narazaki, N. Takebe, Y. Ishigaki; Department of Diabetes and Metabolism, Iwate Medical University, Morioka, Japan. Background and aims: The clinical significance of insulin antibodies (IA) cannot be ignored, even after the introduction of either recombinant human or analog insulin. Previous studies suggest potential associations of IA with hypoglycemia and hyperglycemia, as well as skin reactions. Furthermore, in recent studies, elevated IA levels were independent risk factors for coro- nary heart disease in insulin-treated elder adults, or for retinopathy in type 2 diabetic subjects with insulin therapy. In the subjects under insulin therapy, lipodystrophy, i.e. lipohypertrophy and lipoatrophy, are common problems at frequently injected sites. We occasionally notice that some diabetic subjects developing lipodystrophy tend to show high levels of IA. Indeed, strong as- sociation of lipodystrophy with IA has been reported in children and adoles- cents with type 1 Diabetes.Herein, we endeavored to ascertain clinical and immunological factors that are associated with lipodystrophy in adults with type 2 diabetes. Materials and methods: Subjects were inpatients of our University Hospital. The study groups consisted of 40 females and 42 males, of which mean age was 63± 12 years, duration of diabetes 15± 8.0 years, and HbA1c 9.0± 2.2%. After obtaining approval from the ethics committee of the University and Diabetologia (2014) 57:[Suppl1]S1–S564 S 397 1 C informed consent from all subjects, blood samples were collected. IA were measured by RIA kits of IA with reduced non-specific bindings (Yamasa Co, Tokyo, Japan.) Grading of lipodystrophy was applied as follows. Grade 0: no lipoatrophy or lipohypertrophy at any injection site, grade 1: moderate lipohypertrophy without increased palpable density of subcutaneous tissue, grade 2: severe hypertrophy with increased density of the injection site, and grade 3: lipoatrophy at anyinjection site. Statistical significance was analysed by the Mann-Whitney U test, c2 test, or multiple linear regression analyses. Results: Thirty-eight out of 82 subjects (46.3%) receiving current or previ- ous therapy with recombinant human or analogue insulin were IA positive (percent binding 9.6± 17%). We found lipoatrophy (grade 3) of at least one injection site in 3 subjects, severe lipohypertrophy (grade 2) in 8 and mod- erate lipohypertrophy (grade 1) in 23 adults. No alteration (grade 0) was found in the remaining 48 subjects. Subjects with lipoatrophy displayed sig- nificantly higher levels of percent binding with IA (66± 21%) than those with severe lipohypertrophy (8.0± 8.0%, P=0.03) or those without lipodystrophy (3.3± 6.3%, P=0.03). Multiple regression analyses revealed that lipoatrophy (β=0.78, t=7.99, P=0.000) and percent binding with IA (β=0.65, t=6.05, P=0.000, R=0.65) could significantly predict each other. Conclusion: To our knowledge, this is the first report that showed strong re- lationship between lipoatrophy and IA in subjects with type 2 diabetes under insulin therapy. However, it is the chicken or the egg with lipoatrophy and percent binding with IA. Further investigations should be needed to eluci- date the pathological relationship between them. In conclusion, we propose that the emergence of IA and its concomitant adverse effects should be taken into account, when starting insulin therapy in subjects with type 2 diabetes mellitus. 968 The immunogenicity of originator insulin glargine: incidence and absence of clinical sequelae A. Vlajnic1, L. Traylor2, J. Frimpter2, H. Khatami1, H. Schellekens3; 1Sanofi, Bridgewater, 2Sanofi US, Inc., Bridgewater, USA, 3Department of Pharmaceutical Sciences, Faculty of Sciences, Utrecht University, Netherlands. Background and aims: The immunogenicity of various basal insulins ap- pears to be variable, with no apparent clinical impact. This analysis assessed immunogenicity and its clinical relevance in trials of the originator insulin glargine (GLA) after > 10 years experience. Materials and methods: This was a retrospective analysis of supporting data from published registration studies of GLA vs NPH insulin (NPH) in pa- tients with type 1 diabetes mellitus (T1DM; n = 4) or type 2 diabetes mel- litus (T2DM; n = 3). Assessments included: change in antibody (Ab) levels (%bound/total [%B/T]) against GLA (IGAb) and human insulin (HIAb); proportion of patients with predefined change in Ab levels (≥ 20 units from baseline); and possible Ab-associated clinical effects. Results: In the T1DM studies, baseline IGAb and HIAb levels ranged from 21.72-29.35 and 20.50-29.91 (%B/T), respectively. Significantly larger reduc- tions in Ab levels from baseline were seen with GLA vs NPH in 2 studies for IGAb and HIAb (Table). The proportion of patients with predefined chang- es in Ab levels was low in all studies and treatment groups with available data: GLA, IGAb increase 2.0‒14.5%, decrease 2.0‒4.3% and HIAb increase 1.2‒8.3%, decrease 2.8‒4.3%; NPH, IGAb increase 0‒1.2%, decrease 0‒2.9% and HIAb increase 0‒0.8%, decrease 0‒2.7%. In most studies more GLA- treated patients showed an increase in IGAb; however, one study showed a decrease (p = 0.0029). In the T2DM studies, baseline HIAb and IGAb lev- els ranged from 9.86-17.97 and 6.83‒17.57 (%B/T), respectively. Ab level increases were less marked in GLA-treated patients (significant in 2 stud- ies) (Table). More T2DM patients had predefined changes in Ab levels vs T1DM patients: GLA, IGAb increase 0.01‒3.1%, decrease 0‒6.2% and HIAb increase 0.01‒2.5%, decrease 0‒9.2%; NPH, IGAb increase 0.01‒13.3%, de- crease 0‒2.8% and HIAb increase 0.01‒13.8%, decrease 0‒2.8%. Significantly more NPH-treated patients had predefined increases in IGAb and HIAb in Study 3002 (overall and insulin-naïve subgroup; all p < 0.0001). There was no evidence of insulin resistance, loss of glycaemic control, or increased hypo- glycaemia associated with Ab level change in T1DM or T2DM patients. There was no association between changes in Ab levels and in HbA1c in any treat- ment group; total insulin dose was unchanged or only slightly increased in a majority of T1DM patients. Two T2DM patients (1 GLA-treated and 1 NPH- treated) had isolated dose increases; HbA1c decreased from baseline in both. No correlation was seen between changes in Ab levels and hypersensitivity. Conclusion: Treatment with GLA is associated with generally low rates of Ab formation, without evidence of any clinical consequences. As some regula- tory agencies require assessment of immunogenicity for biologics, including biosimilars, these data may provide a potential benchmark for consideration of emerging data from GLA biosimilar studies. Supported by: Sanofi US, Inc. 969 Evaluation of immunogenicity of LY2963016 insulin glargine compared with insulin glargine in patients with type 2 diabetes mellitus L.L. Ilag, M.A. Deeg, W.J. Huster, T.M. Costigan, J.S. Zielonka, R.K. Pollom, M.J. Prince, R.J. Konrad; Eli Lilly and Company, Indianapolis, USA. Background and aims: LY2963016 (LY IGlar) is an insulin glargine product, with an identical amino acid sequence to insulin glargine (Sanofi-Aventis; IGlar), which demonstrates similar pharmacokinetic and pharmacodynamic profiles and no clinically meaningful differences from IGlar in safety or ef- ficacy in patients with type 2 diabetes mellitus (T2DM). When evaluating similar biological products, assessment of immunogenicity is important to examine potential differences between products in the incidence and severity of human immune responses and implications this might have for safety or efficacy. Materials and methods: This was a 24-week, Phase 3, randomised, double- blind, parallel study to compare the efficacy and safety of LY IGlar (N=376) and IGlar (N=380) in patients with T2DM on ≥2 oral antihyperglycaemic medications (OAMs). Patients were insulin-naïve (HbA1c ≥7.0% to ≤11.0%) or previously on IGlar (HbA1c ≤11.0%). Anti-insulin glargine antibodies (% binding) were measured for the overall 24-week treatment period. The rela- tionships between anti-insulin glargine antibody levels or treatment emer- gent antibody responses (TEAR) and selected efficacy and safety measures were evaluated using ANCOVA. Results: Approximately 40% of patients (LY IGlar, 41%; IGlar, 38%) entered the study on IGlar and the remainder (LY IGlar, 59%; IGlar, 62%) were insu- lin-naïve. Similar proportions of patients completed the 24-week treatment period (LY IGlar, 89%, IGlar 86%). As shown in Table 1, proportions (%) of patients with detectable antibodies at baseline and throughout the treatment period were similar for LY IGlar vs IGlar (p>.05). There were no differences in TEAR incidence between patients randomised to LY IGlar vs IGlar and clinical outcomes (HbA1c, basal insulin dose and total hypoglycaemia) were not affected by TEAR status (i.e., p>.05 for treatment-by-TEAR interaction for these outcomes) (Table 1). The incidence of treatment-emergent allergic events did not differ between LY IGlar and IGlar (p>.05) (Table 1). Clinical outcomes were also not affected by insulin antibody levels (p>.05). Diabetologia (2014) 57:[Suppl1]S1–S564 S 398 1 C Conclusion: These clinical safety data demonstrate a highly similar immu- nogenicity profile of LY IGlar to IGlar, with no effects of anti-insulin glargine antibodies on selected efficacy and safety outcomes in patients with T2DM. Clinical Trial Registration Number: NCT01421459 Supported by: Eli Lilly and Company 970 Demonstration of differential immune modulation by originator and biological copies of insulin glargine in an in vitro model of human immunity E. Luna1, P. Agrawal1, R. Mehta1, C. Vernhes1, C. Denys2, H. Khatami3, A. Schmidt4, N. Tennagels4, D.R. Drake III1; 1Sanofi Pasteur-Vaxdesign Campus, Orlando, USA, 2Sanofi, Paris, France, 3Sanofi, Bridgewater, USA, 4Sanofi, Frankfurt am Main, Germany. Background and aims: Because the manufacture of insulin analogs require sophisticated production and purification procedures, it is possible originator and biological copies of insulin might have varied immunogenicity profiles based on differences in the structure (protein aggregation, denaturation), pu- rity (contamination with endotoxins or others), and/or other characteristics of these products. To date, few studies have addressed the specific pathways of potential immune modulation triggered by biological copies of insulin glargine Materials and methods: Here, we used an in vitro model of immunity, termed the MIMIC® system, to study antigen presenting cell (APC) activation and cytokine secretion induced by a range of doses of originator and biologi- cal copies of insulin glargine that are available in some countries Results: By flow cytometric phenotype analysis, the originator and biological copy products triggered minimal changes in expression of activation markers on MIMIC® system-derived APCs. In contrast, some lots of the biological copies of insulin glargine, but not the originator product, were capable of inducing the secretion of some pro-inflammatory markers, such as IL-8, in the treated MIMIC® cultures. These immunogenicity differences appear to be due to the effect of the compounds on monocytic and/or lymphocytic popu- lations in the in vitro human cell system. Of note, all insulin analogs included in this analysis were comparable in their insulin receptor activity in vitro and typically generated similar analytical profiles. Conclusion: In contrast to the originator product, some of the biological copies of insulin glargine (variable by lot) triggered the induction of pro- inflammatory cytokines in the MIMIC® system, which might indicate a dif- ference in the biochemical composition of some of these insulin biological copies. While the mechanism and differences causing these effects is still un- der investigation, and the potential clinical impact is unknown, the data pre- sented here suggests there might be different immune modulation potentials between originator and copy insulin compounds. Further investigation of the potential clinical relevance and/or mechanisms regulating this response may be warranted. Supported by: Sanofi/ Sanofi Pasteur. 971 Binding kinetics, internalisation and localisation of long-acting basal insulin peglispro (BIL) and insulin receptor in cells V.V. Kiselyov, J.S. Moyers, C.B. Volk, J.X.C. Cao, G.Y. Wu, L. Ding, M.D. Michael; Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, USA. Background and aims: BIL (LY2605541) is a novel, PEGylated insulin lis- pro that has a large hydrodynamic size. It has a prolonged duration of action which is related to a delay in insulin absorption and a reduction in clear- ance. We compared the binding kinetics, intracellular localization, in vitro clearance, and effect on activation and localization of human insulin recep- tor (hIR) of BIL with those of insulin lispro and biosynthetic human insulin (BHI). Materials and methods: Cellular internalization and localization of BIL, BHI, insulin lispro, and tyrosine-phosphorylated hIR (hIR-pY) were ob- served by immunofluorescent confocal imaging in U2OS cells that overex- press hIR. Binding and kinetic properties of BIL and BHI were measured in IM9 cells using radio-ligand binding techniques. The in vitro clearances of BIL and BHI were studied in HEK cells overexpressing hIR. Results: Co-localization studies using antibodies to insulin or PEG, and the early endosomal marker, EEA1, showed that the internalization and subcel- lular localization pattern of BIL were similar to those of BHI and insulin lis- pro; all were rapidly internalized and co-localized with EEA1. During ligand washout, concomitant loss of insulin, PEG methoxy group, and PEG polymer backbone immunostaining was observed for BIL similar to the loss of in- sulin immunostaining observed for insulin lispro and BHI. Co-localization studies using an antibody to the lysosomal marker LAMP2 did not reveal evidence of lysosomal localization for insulin lispro, BHI, or BIL, using either insulin or PEG immunostaining reagents. In vitro clearance studies showed efficient insulin receptor-mediated-clearance of BHI and the insulin portion of BIL. In contrast, the concentration of PEG polymer backbone appeared to decrease significantly more slowly. This is consistent with a process whereby, upon internalization of BIL, the insulin moiety is degraded, whereas the PEG backbone is recycled out of cells, and is consistent with the above observa- tion that no PEG was detected in lysosomes. BIL and BHI both induced rapid phosphorylation and internalization of hIR. Co-staining of hIR-pY and EEA1 showed that hIR promptly entered early endosomes after BIL and BHI treatment. Minimal lysosomal localization of hIR-pY was detected with either ligand. Ligand washout studies suggested a more rapid dephosphoryla- tion of hIR-pY after treatment with BIL than with BHI. Whole cell binding analysis performed with IM9 cells also showed slightly faster hIR dissocia- tion kinetics for BIL than for BHI. The pH dependence of the dissociation rate and the magnitude of accelerated dissociation from the receptor at high concentrations of BIL were similar to those of BHI. Affinity of BIL for hIR was ~20-fold lower than that of BHI. Conclusion: The mechanism of insulin receptor activation by BIL is similar to that of BHI as evidenced by preservation of pH and ligand concentration dependences of the dissociation rate. The reduction of affinity to the insulin receptor is attributable mainly to a decrease in the association rate constant. The insulin and PEG moieties of BIL undergo a dynamic cellular process of rapid internalization and transport to early endosomes after stimulation of hIR-pY followed by loss of cellular immunostaining in a manner similar to those of insulin lispro and BHI suggestive of efficient transport of the PEG backbone out of cells. Supported by: Eli Lilly and Company 972 Pharmacological evaluation of once-weekly potentials by combination of long-acting insulin with long-acting exendin4 in an animal model S. Kwon1, C. Lim1, Y. Park1, J. Kim1, S. Jung1, I. Choi1, S. Lee1, J. Kang1, M. Trautmann2, M. Hompesch2, Y. Kim1; 1Hanmi Parmaceutical, Seoul, Republic of Korea, 2Profil Institute, Chula Vista, USA. Background and aims: Combination of basal insulin and GLP-1 receptor agonist could provide diabetic patients with improved efficacy and safety. However, the recently developed combination products require an once- daily administration, which is still frequent and thus negatively affects treat- ment adherence and efficacy. We have developed the novel long-acting basal insulin analog and exendin-4 analog, namely HM12470 (once-weekly regi- men) and HM11260C (once-weekly or monthly regimen), respectively, using Diabetologia (2014) 57:[Suppl1]S1–S564 S 399 1 C LAPSCOVERY technology. Because of their prolonged and flatness phar- macokinetic (PK) / pharmacodynamic (PD) properties, the combination of HM12470 and HM11260C could maximize the dosing convenience as well as the therapeutic advantages. The objective of this study was to evaluate the once-weekly potentials of HM12470 and HM11260C combination in animal model. Materials and methods: Human PK profile of HM12470 was predicted based on PK results of three different animal species using Wajima Css-MRT method and compared with that of HM11260C in human. db/db mice were chronically administrated with HM12470 and/or HM11260C. To evaluate the glucose lowering efficacy, 4h fasting blood glucose (FBG) levels were measured every 3-4days. Body weight was determined every 2days for proper drug administration. At the end of study, HbA1c levels were determined. To evaluate the PK profiles of combination treatment, serum concentration of HM12470 and/or HM11260C were quantified after single administration of HM12470 and/or HM11260C in SD rats. Results: The half-life of HM12470 in human was predicted as 132hr which was similar with HM11260C (140hr), suggesting their once-weekly com- bination potentials. In db/db mice, once-weekly mimetic administration (Q2D) of HM12470 and HM11260C showed significantly reduced FBG lev- els compared to that of individual compounds. Consistent with FBG con- trol, HM12470 and HM11260C combination group had lower HbA1c level (6.8±0.8%) than HM12470 (9.0±0.6%) or HM11260C (8.0±1.7%) groups, indicating improved glycemic control by combination treatment. Moreover, this improved glycemic control was maintained even by reducing HM12470 dose in combination treated group. Furthermore, combination treatment attenuated body weight gain observed in HM12470 only groups. As to the pharmacokinetic profiles, administration of HM12470 and HM11260C did not affect pharmacokinetic profiles of either individual compounds in SD rats, ruling out possible drug-drug interference. Conclusion: Based on PK prediction results, HM12470 could be a best partner of HM11260C for the once-weekly combination therapy. Of note, combination of HM12470 and HM11260C not only improved glycemic control, but also controlled unwanted body weight gain by insulin drug in db/db mice. Given that there were no PK interference between HM12470 and HM11260C, our results collectively demonstrate that combination of HM12470 and HM11260C provide optimal convenience as well as thera- peutic advantages, shedding light on ideal therapy for the management of diabetic patients. PS 079 Insulin treatment and hypoglycaemic reactions 973 Hypoglycaemia in people with type 2 diabetes achieving vs not achieving HbA1c target levels over 4 years in routine clinical practice: analysis of the CREDIT study L. Blonde1, N. Freemantle2, F. Calvi-Gries3, M. Vincent4, P. Home5; 1Ochsner Medical Center, New Orleans, USA, 2University College London, UK, 3AtlanStat, Rezé, France, 4Sanofi, Paris, France, 5Newcastle University, Newcastle upon Tyne, UK. Background and aims: To evaluate the incidence and rate of hypoglycaemia in people with type 2 diabetes achieving vs not achieving HbA1c < 7.0 % (53 mmol/mol) over 4 years after starting insulin therapy in routine clinical prac- tice in Europe, Canada, and Japan. Materials and methods: CREDIT (Cardiovascular Risk Evaluation in people with Type 2 Diabetes on Insulin Therapy) was a noninterventional study de- signed to examine the relationship between HbA1c and cardiovascular events and to provide insight into current, real-world practices of the use of insulin by 2999 people beginning insulin, with up to 54 months’ follow up (n = 2272 with 4-year data). For this analysis, records of symptomatic, severe, and noc- turnal symptomatic hypoglycaemia were collected for the 6 months prior to the 1, 2, 3, and 4-year follow-up intervals. Updated mean HbA1c was calcu- lated as the average of normalized HbA1c measurements from 1 month after insulin initiation up to follow-up year values. Hypoglycaemia events were categorised by whether or not the person experiencing the event had updated mean HbA1c < 7.0 % or not at the follow-up visit closest to the event. Results: Overall mean (SD) HbA1c declined from 80 (21) mmol/mol (9.5 [2.0] %) when starting insulin to 60 (14) mmol/mol (7.6 [1.3] %) at 4 years (median 56 mmol/mol [7.3 %]). During the last 6 months of year 1, 2, 3, and 4 follow-up, the proportion of people experiencing ≥ 1 hypoglycaemic event was higher for those who achieved target (range 17.7%-24.6%) than for those who did not (range: 15.8%-16.8%) (Table). This was also true for incidence rate of any documented symptomatic hypoglycaemia (1.18-1.40 vs 0.75- 0.96 events/person), and nocturnal symptomatic hypoglycaemia (0.14-0.24 vs 0.12-0.20 events/person). The percentage of people who experienced ≥ 1 severe hypoglycaemic event for the 4 half-yearly follow-up periods was also higher for those to target (range 3.0%-4.6%) than for those not (0.9%-1.6%). Mean rates of severe hypoglycaemia were 0.04-0.20 vs 0.02-0.04 events/per- son, while severe nocturnal hypoglycaemia was 0.02-0.06 vs 0.00-0.01 events/ person. Conclusion: People with type 2 diabetes starting insulin who achieve up- dated mean target HbA1c levels in routine clinical practice do experience higher rates of symptomatic and nocturnal hypoglycaemia than those failing to achieve target HbA1c. However, the difference between groups in the per- centage of individuals with hypoglycaemia seems to diminish over time, and rates for both groups are relatively modest. Moreover, the rate of severe hy- poglycaemia is low and could not be distinguished by whether target HbA1c is achieved. Initiation of insulin in the CREDIT population was associated Diabetologia (2014) 57:[Suppl1]S1–S564 S 400 1 C with a significant improvement in glycaemia. These hypoglycaemia data are consistent with a positive benefit risk ratio for this therapeutic strategy. Supported by: Sanofi 974 IDegAsp lowers the rate of hypoglycaemia vs biphasic insulin aspart 30 in adults with type 2 diabetes achieving glycaemic target (HbA1c S 401 1 C 976 Glycaemic control and hypoglycaemia in Japanese people with type 2 diabetes mellitus receiving new insulin glargine 300 U/mL in combination with OADs (EDITION JP 2) Y. Terauchi1, M. Koyama2, X. Cheng3, S. Shimizu4, T. Hirose5, on behalf of the EDITION JP 2 study group; 1Department of Endocrinology and Metabolism, Yokohama City University Graduate School of Medicine, 2Asia Pacific TSU, Sanofi Japan, Tokyo, Japan, 3Clinical Science & Operation, China R&D, Sanofi, Beijing, China, 4Biostatistics and Programming, Sanofi Japan, 5Division of Diabetes, Metabolism and Endocrinology, Department of Medicine, Toho University School of Medicine, Tokyo, Japan. Background and aims: EDITION JP 2 studied the efficacy and safety of new insulin glargine 300 U/mL (Gla-300) versus glargine 100 U/mL (Gla-100) in Japanese people with T2DM using basal insulin plus OAD(s). Materials and methods: In this 6-month, multicentre, open-label, phase 3 study, participants (n=241; mean age 60.8 years; mean BMI 25.3 kg/m2; mean duration of T2DM 14.0 years; mean HbA1c 8.0 %) were randomised to re- ceive once-daily Gla-300 or Gla-100 plus OAD(s). Insulin was titrated to tar- get FPG 4.4-5.6 mmol/L (80-100 mg/dL). The primary endpoint was HbA1c change from baseline to month 6. Results: HbA1c decreased similarly in both groups (LS mean [SE] −0.45 [0.06] % for Gla-300 and −0.55 [0.06] % for Gla-100; LS mean difference 0.10 [95% CI: −0.08 to 0.27] %). Fewer participants experienced any hypo- glycaemic events during 6 months with Gla-300 versus Gla-100. The number of participants with ≥1 confirmed (≤3.9 mmol/L) or severe hypoglycaemic event (24 h and nocturnal) was consistently lower with Gla-300 compared with Gla-100, and rate per participant-year was lower with Gla-300 versus Gla-100 (Table). Severe hypoglycaemia was infrequent in both groups. LS mean (SE) weight change was −0.62 (0.19) kg for Gla-300 and 0.37 (0.19) kg for Gla-100. Similar safety profiles were observed in both groups. Conclusion: In Japanese people with T2DM using basal insulin plus OAD(s), Gla-300 provides comparable effective glycaemic control with fewer hypogly- caemic events, particularly during the first 8 weeks. Clinical Trial Registration Number: NCT01689142 Supported by: Sanofi 977 The beneficial effect of insulin degludec on nocturnal hypoglycaemia and insulin dose in type 1 diabetic patients: a systematic review and meta- analysis of randomised trials K. Dzygalo1, D. Golicki2, A. Kowalska1, A. Szypowska1; 1Department of Paediatrics, 2Department of Experimental and Clinical Pharmacology, Medical University of Warsaw, Poland. Background and aims: Insulin degludec is a new-generation ultra-long act- ing basal insulin which offers a significantly more predictable glucose-lower- ing effect than other long-acting insulin analogues. To compare the effect of treatment with insulin degludec and long-acting insulin analogues glargine and detemir in type 1 diabetic (T1D) patients by means of a systematic review and meta-analysis. Materials and methods: The following electronic databases were searched up to January 2014: MEDLINE, EMBASE, The Cochrane Library. Additional references were obtained from the reviewed articles. There were included randomized controlled trials of at least 12 weeks‘ duration with basal-bolus regimen therapies in T1D patients. Results: Current analysis included 4 studies involving 1,846 T1D patients. The combined data from all trials showed a statistically significant reduction in the basal insulin dose (MD 0.042, 95% CI -0.067 to -0.018, p=0.001) and the total daily insulin dose (MD -0.072, 95% CI 0.016 to -0.027, p=0.002) in the degludec group compared to other long-acting analogues. There was also a significant reduction of nocturnal hypoglycemia in the degludec group compared to the controls (Rate Ratio 0.697, 95% CI 0.617 to 0.786, p=0.000). There were no differences between the groups in terms of HbA1c values, fast- ing plasma glucose (FPG), adverse events. Conclusion: Basal-bolus treatment with insulin degludec was superior to long-acting insulin analogues detemir and glargine in reducing the rate of nocturnal hypoglycemia. In comparison to other long-acting analogues, treatment with insulin degludec was safe, patients obtained similar metabolic control expressed by HbA1c and FPG levels with the added benefit of a re- duced basal and total insulin dose. 978 Hypo-/hyperglycaemia alert based on online glucose prediction using a global model for type 1 diabetes C. Yu1, C. Zhao1, Y. Fu2; 1Control Science and Engineering, Zhejiang University, Hangzhou, China, 2Control Science and Engineering, Becton Dickinson Technologies, Research Triangle Park, USA. Background and aims: Avoiding hypo/hyperglycemia is a major challenge for diabetes mellitus. A timely alert of hypo/hyperglycemia episodes before they occur can allow enough time for the patient to take necessary actions to avoid hypo/hyperglycemia. Materials and methods: A global model, which is developed based on one subject and then applied to new subjects without any customization, is used for early hypo/hyperglycemia alert based on online glucose prediction using clinical data provided by BD. Two autoregressive (AR) modeling methods are considered using standard least squares algebra: (i) global prediction model; (ii) subject-dependent model. Both are developed based on time-series glu- cose data with 2 min as sampling interval. Also, to evaluate the role of high- frequency glucose data, the glucose data are filtered by a first-order low-pass filter with 30min the as threshold. Online short-term glucose predictions are made for 15 subjects for alert of hyper/hypoglycemia based on filtered and non-filtered data. Results: The results are summarized in Table 1 as evaluated by time lag which is the time difference between the predicted hyper/hypoglycemia and the measured event. Conclusion: The results indicate that the accuracy of global model is compa- rable to or even better than that of subject-dependent model for alert of hy- per/hypoglycemia based on the index of time lag. However, the sensitivity for hypos using global model is lower than that using subject-dependent model. Also, the model developed based on filtered glucose data can get smaller time Diabetologia (2014) 57:[Suppl1]S1–S564 S 402 1 C lag of hyper/hypoglycemia alert than that based on raw glucose data since the high-frequency noises are removed after low-pass filtering. Supported by: National Natural Science Foundation of China (No. 61273166) 979 Assessment of glycaemic control and risk of hypoglycaemia for two basal-bolus algorithms in hospitalised patients with diabetes mellitus type 2 J.K. Mader1, K.M. Neubauer1, L. Schaupp1, F. Aberer1, K. Donsa2, T. Augustin2, S. Spat2, B. Hoell2, P. Beck2, J. Plank1, T.R. Pieber1; 1Internal Medicine / Endocrinology and Metabolism, Medical University of Graz, 2HEALTH, Joanneum Research GmbH, Graz, Austria. Background and aims: Current guidelines recommend pre-meal blood glu- cose (BG) levels of S 403 1 C PS 080 Gastrointestinal liners and e-health opportunities 981 Acute and sub-acute effects of the endobarrier on glucose homeostasis and appetite in obese uncontrolled type 2 diabetes mellitus patients G. Segal-Lieberman1, A. Lang2, M. Lahav2, N. Lieberman3, A. Paster1, N. Konvalina1, H. Kanety1, R. Hemi1, O. Cohen1; 1Institute of Endocrinology, Sheba Medical Center, Ramat-Gan, 2Institute of Gastroenterology, Sheba Medical Center, Ramat-Gan, 3Community Medical Devision, Clalit Medical Services, Tel-Aviv, Israel. Background and aims: Diabetes is a prevalent metabolic disease accompa- nied by an increased risk for cardiovascular disease as well as micro-vascular complications. To date, bariatric surgery has become one of the most effective means of providing durable, clinically-significant resolution of diabetes, with the majority of subjects with T2DM undergoing these procedures experienc- ing a remission of their disease. One of the theories behind the early resolu- tion of diabetes after the Roux-en-Y Gastric bypass operation claims that the mediating factor lies within the bypassing of the duodenum by the ingested food. The Endobarrier gastrointestinal liner system (by GI Dynamics®) is an endoscopically-delivered device that mimics gastric bypass surgery by shield- ing the duodenum and upper jejunum from contact with chyme. Materials and methods: An Endobarrier device was placed in the duode- num via an endoscopic procedure in 33 uncontrolled diabetic obese subjects. Acute and sub-acute effects of the device on glucose homeostasis were evalu- ated using a Continuous Glucose Monitoring (CGM) recording device for one week, as of two days prior to device implantation until four days after its removal. The effect of the device on appetite was evaluated using the 5-ques- tion visual analog scale, where each question was given a score between 1-10. Results: Our population included 17 females and 16 males with mean dia- betes duration of 10.4 (SD±5.8) years, most were insulin treated. Baseline HbA1c level was of 9.3% (SD±1.0) and baseline body mass index was 38.0 kg/m2 (SD±5.0). CGM data revealed a significant decrease in average daily glucose levels from 243.2 mg/dL (SD±63.4) one day before the procedure to 227 mg/dL (SD±53.6) (p value = 0.002) on the day of the procedure, while the subjects were fasted most of the day. As of the first day after the procedure, despite the fact that eating was initiated, prandial medications were termi- nated and basal insulin dose was lowered by 50%, the recorded glucose levels continued to decreased and reached a decrease of 30.3 mg/dL (SD±69.4) (p value= 0.048) vs. baseline on the last day of CGM recording. Whereas no significant change in the VAS score was recorded after one week, one month after the Endobarrier insertion, a significant change of 1.1 points (SD±1.7) (p value=0.003) vs. baseline was recorded suggesting a significant decrease in appetite one month after the procedure, despite a weight loss of 4.2 kg at this time-point. Conclusion: Endobarrier therapy in advanced uncontrolled obese diabetic patients significantly improves their glucose homeostasis, as early as the first day after the procedure. This improvement becomes even more prominent as days go by and eating becomes more regular. Despite a substantial weight loss, appetite is significantly suppressed one month after the insertion of the device. EndoBarrier seems to have an early lowering effect on blood glucose as well as a suppressive effect on appetite, rendering it an attractive option for the care of „diabesity“ in this challenging population. Clinical Trial Registration Number: NCT01718457 Supported by: IASD 982 Effects of duodeno-jejunal bypass liner on bile acid homeostasis in patients with severe obesity A.A. Young1, K.K. Brown1, J.D. Williams1, F. Pimentel2, A. Sharp2, P.P. Parra2, D. Turiel2, R. Munoz2, M. Arrese2, V. Irribarra2, J.F. Miguel2, L.M. Kaplan3, A. Levine4, D. Maggs5, A. Escalona2; 1GlaxoSmithKline, Research Triangle Park, USA, 2Pontificia Universidad Católica de Chile, Santiago, Chile, 3Obesity, Metabolism & Nutrition Institute, Massachusetts General Hospital, Boston, 4Wyss Institute, Harvard University, Cambridge, 5GI Dynamics, Lexington, USA. Background and aims: Endoscopic placement of a duodenojejunal bypass liner (DJBL) elicited a 16.1 ± 1.6% loss of body weight in 17 morbidly obese patients, and normalization of HbA1c ( S 404 1 C Results: In the overall cohort, there was prompt and robust weight loss ob- served that plateaued at 6-12 months (weight loss % at 52 weeks: overall -15.9%, non-DM -14.3%, DM -18.2%). A1C in the DM group decreased to 6.24±1.18% at 52 weeks. Lowering of both fasting and post-prandial glucose (see figure) in the DM group was observed 1 week post implant and was maintained to 52 weeks. GLP-1 responses were augmented and triglycerides and FFAs suppressed during meal challenges from 12 to 52 weeks (P S 405 1 C 986 Improvement of metabolic control in type 1 diabetic children using ELKa toolset: a randomised controlled trial A. Kowalska, K. Piechowiak, A. Szypowska; Department of Pediatrics, Medical University of Warsaw, Poland. Background and aims: In intensive insulin therapy, prandial in- sulin doses should be calculated before each meal with carbohy- drate (CHO) and fat/protein (FP) exchange counting. Therefore food weighing and complicated calculations need to be performed. ELKa system is an advanced toolset which helps performing calculation of food exchanges. It consists of ELKa software including database of vari- ous meals and nutrients and ELKaPlus digital kitchen scale. Among similar devices it is extraordinary due to opportunity of transmission weights in a real-time to a computer via USB port. After choosing the name of a particu- lar product from the list, the program gives precise information about the amount of CHO and FP exchanges in a serving. It simplifies every-day cal- culations and it increases the precision of results. A number of our patients use the system and it appears to be useful accessory, but no clinical trials were performed yet. The aim of this study was to investigate the benefit of using ELKa toolset in comparison with standard method of CHO and FP counting on metabolic control in type 1 diabetic children. Materials and methods: A randomized 26-week clinical trial was conducted on 106 patients (64 girls, 42 boys) aged 11 ± 4.2 years (range 1.8 - 17.2 years), with mean BMI 19.6 ± 3.4 (range 14.5- 33.2), diabetes duration 5 ± 3.1 years (range 1 - 14.3 years), and HbA1c 7.5 ± 1% (range 5.1%- 10%). All patients were treated with insulin pumps. Patients were randomly assigned into two groups: the group A (n=53) using ELKa system for food exchange counting and the group B (n= 53) using standard method of food calculation. There were no significant differences between groups in mean baseline HbA1c lev- els (group A: 7.6% ± 0.1; group B: 7.4% ± 0.1; p=0.395), BMI (A: 18.4; B: 19.7; p= 0.165), duration of diabetes (A: 4.3 years; B: 4.6 years; p= 0.265) and insulin/kg/24h (A: 0.8; B: 0.8; p=0.099). Control group was slightly older (6.9 years ± 0.5 vs. 5.3 years ± 0.5; p=0.042). At follow-up visits after 3 and 6 months, HbA1c, BMI, daily insulin dose and blood glucose values were as- sessed. The group A also was asked about the frequency of using the toolset. Results: 94 patients completed 26-week follow-up. There were no significant difference between groups with regard to HbA1c (group A - 7.4% ± 0.2, n=43 vs. group B - 7.6% ± 0.1, n=51; p= 0.222). Patients who declared using the toolset everyday (86-100% of meals) or for the most part of the week (51- 85% of meals) had significantly lower HbA1c (6.9 ± 0.2, n=21) in comparison with the control group (7.6 ± 0.1, n=51; p=0.0005). Moreover subgroup of patients who used ELKa only for 50-85% meals also achieved better result in a glycated hemoglobin than control group (HbA1c 6.8% ± 0.2, n=14; p= 0.0009). Conclusion: Our results confirm the beneficial effect of using ELKa system in day-to-day calculations of CHO and FP exchanges. Diabetic children who used the system for more than 50% meals achieved better metabolic control than children who performed food calculations without this tool. 987 Role of mobile technology to improve diabetes care in adults with type 1 diabetes: the REMOTE-T1D study S.K. Garg1, W.R. Hiatt2, P.A. Gottlieb1, C.R. Beatson1, F. Flacke3, V.N. Shah1, J.K. Snell-Bergeon1; 1Barbara Davis Center for Childhood Diabetes, University of Colorado Denver, 2CPC Clinical Research, University of Colorado Denver, Aurora, USA, 3Sanofi, Paris, France. Background and aims: This study evaluated the role of mobile technology to improve diabetes care in adults with type 1 diabetes (REMOTE-T1D). We hypothesized that the use of mobile technology (iPhone plus iBGStar®) will result in improvement in Patient Reported Outcomes (PRO). Materials and methods: This single-center, prospective, 6-month, open- label, investigator-initiated pilot study enrolled 100 adult patients with type 1 diabetes. Patients were randomized in a 1:1 fashion to an intervention group using self-monitoring of blood glucose (SMBG) with iBGStar vs. SMBG with Accu-Chek Nano® (control). All subjects wore a blinded, continuous glucose monitor (CGM) for four separate 7-day periods. Primary outcomes included PRO (hypoglycemia fear questionnaire) and the secondary outcomes includ- ed glucose control parameters. Insulin doses, hypoglycemic episodes and self-monitored blood glucose (SMBG) measures were obtained at each visit. At the conclusion of the study, satisfaction with the device was ascertained. Results: Baseline demographics and A1c values were similar between the two groups (mean SD ± A1c 7.7 ± 1.0 and 8.0 ± 0.9 in the control and iBGStar groups, respectively, p=0.15) . The Hypoglycemia Fear Scale improved com- pared to baseline in both groups to a similar extent at 3 months (-4.5 ± 9.5 in the iBGStar group vs. -3.0 ± 8.4 in the control group, p=0.65 between groups) and the changes remained similar between groups at 6 months (-3.9 ± 12.5 in the control group vs. -1.4 ± 10.0 in the iBGStar group, p=0.31). There was a significant improvement in change of A1c in the iBGStar group at 6 months when compared to baseline or the control group (Figure). The percentage of time spent in hypoglycemia (less than 70 mg/dL) as assessed by CGM read- ings was similar in the two groups throughout the study. There was a signifi- cant increase in insulin dose at 3 months in the iBGStar group (7.7 ± 20.9 units/day) but not in the control group (1.7 ± 18.3 units/day), with no differ- ence in hypoglycemic episodes (21.5 ± 15.5 in the iBGStar group vs. 25.5 ± 31 in the control group, p=0.48) throughout the study. The number of SMBG tests per day was similar at baseline (4.8 ± 1.1 in the iBGStar group and 5.2 ± 1.1 in the control group, p=0.07) and decreased significantly from baseline in both groups at 6 months (-0.3 ± 0.09 in the iBGStar group vs. -0.4 ± 0.9 in the control group, p=0.48). At the exit interview, when asked whether they would prefer the iBGStar vs. their past SMBG device, 87.5% preferred the iBGStar. Conclusion: We conclude that the use of iBGStar in this REMOTE-T1D study which allowed for easier communication of SMBG readings with the treating center by its email function, improved A1c at 6 months with no in- creased risk of hypoglycemia or change in PRO. This device was well-received by patients. Clinical Trial Registration Number: NCT1825382 Supported by: Sanofi 988 Long-term tele-monitoring of patients with type 2 diabetes mellitus: results of the Greek pilot of the renewing health multicentre randomised control trial G.E. Dafoulas1, A. Mavrodi2, P. Gkiata3, H. Giannakakos4, P. Stafylas5, M. Delizona6, V. Aletras2, P. Pechlivangolou7, K. Theodorou1, G. Koukoulis6, A. Bargiota6; 1Faculty of Medicine,University of Thessaly, Larisa, 2Department of Business Administration, University of Macedonia, Thessaloniki, 3Intermunicipal Digital Community of Central Greece, Trikala, 45th Regional Health Authority of Thessaly and Sterea, Larisa, Greece, 5Health Information Management SA, Brussels, Belgium, 6Department of Endocrinology and Metabolic Diseases, Regional University Hospital, Larisa, Greece, 7Toronto Health Economics and Technology Assessment Collaborative, University of Toronto, Canada. Background and aims: Evidence is required to assess the impact of long term telemedicine use in treatment of patients with type 2 diabetes mellitus (DMT2). The aim of the present study was to examine the impact of a long- term telemonitoring program for patients with DMT2 on glycemic control, Diabetologia (2014) 57:[Suppl1]S1–S564 S 406 1 C health-related quality of life (HRQOL), physical activity and compliance with the mediterranean diet compared to usual care. Materials and methods: In the Greek pilot of a prospective, randomized, single-blinded, multicenter, one year study 154 patients with DMT2 capable to use the telemonitoring device, with an HbA1c > 53 mmol/mol (7.0 % ac- cording to NGSP) were studied after they were randomly assigned in the te- lemonitoring group (IG), (N=74) and in the control group (CG), (N=80) and having signed the informed consent form. In the (IG) group patients‘ blood glucose profiles were collected weekly using a mobile phone health platform, for a period of one year. Allocated health professionals provided by phone the appropriate counseling on lifestyle and medication changes when required. Patients in (CG) group received usual care with face-to-face consultations. HRQOL was assessed using a generic (SF36v2) questionnaire and a disease- specific questionnaire, the Problem Areas in Diabetes (PAID) scale. Physi- cal activity was assessed using the self-administered short form instrument. International Physical Activity Questionnaire (IPAQ) and the compliance with the Mediterranean diet using the Mediterranean Diet Quality Index (KIDMED) adapted for Greek adults. Results: The table shows the outcome of the variables studied in both groups. A greater reduction in HBA1C was observed in the IG compared to the CG at the end of the study. There was a statistically significant improvement in the generic HRQOL in the MCS, in the disease specific HRQOL and the physical activity in the IG compared with the CG, but there was no improvement in KIDMED in neither or the two groups. Conclusion: Our preliminary results indicate that in patients with DMT2, home telemonitoring is more effective than usual care in improving glycemic control with concurrent improvement in patients quality of life and increase of their physical activity. However home telemonitoring does not seem capa- ble to empower patients with DMT2 with to follow a healthier diet. Clinical Trial Registration Number: NCT01498367 Supported by: ICT-PSP CIP EC, Greek Ministry of Health, Thessaly Prefecture PS 081 Improvements in continuous glucose monitoring 989 Accuracy of Continuous Glucose Monitoring (CGM) under free-living conditions during three home closed loop studies L. Leelarathna1, H. Thabit1, D. Elleri2, J.M. Allen1, A. Lubina-Solomon3, J.K. Mader4, C. Benesch5, M. Stadler6, T.R. Pieber4, S. Arnolds5, S.R. Heller3, S.A. Amiel6, D.B. Dunger2, M.L. Evans1, R. Hovorka1; 1Institute of Metabolic Science, University of Cambridge, 2Department of Paediatrics, University of Cambridge, 3Academic Unit of Diabetes, Endocrinology and Metabolism, Department of Human Metabolism, University of Sheffield, UK, 4Division of Endocrinology and Metabolism, Department of Internal Medicine, Medical University of Graz, Austria, 5Profil Institut für Stoffwechselforschung GmbH, Neuss, Germany, 6Diabetes Research Group, Kings College, London, UK. Background and aims: Closed loop systems modulate insulin delivery based on glucose levels measured by commercial CGM devices. We evaluated the accuracy of Freestyle Navigator II CGM (Abbott Diabetes Care, CA, USA) against the capillary glucose measured by the in-built glucometer of Freestyle Navigator II CGM using Freestyle Lite glucose strips. Data collected dur- ing three unsupervised randomised open-label crossover home closed-loop studies were pooled for the current analysis. Materials and methods: Paired CGM and capillary glucose values (10,616 pairs) were collected from 57 participants with type 1 diabetes [41 adults (age 39±12 years, HbA1c 7.9±0.8%; mean±SD) recruited at five centres; 16 adoles- cents (15.6±3.6 years, HbA1c 8.1±0.8%) recruited at one centre]. Participants were instructed to calibrate the sensor according to manufacturer instruc- tions. Numerical accuracy was assessed by relative absolute difference (RAD) (|sensor - capillary| / capillary) x 100%) when capillary > 4.2 mM and by ab- solute difference (AD) |sensor - capillary| when capillary ≤ 4.2 mM. Clinical accuracy was assessed by the Clarke error grid analysis. Results: Total duration of sensor use for whole cohort was 2,002 days (48,052 hours). The average number of capillary glucose values taken by adults and adolescents was 5.9 ± 2.0 and 4.3 ± 2.4 per day, respectively. Overall sensor accuracy for the entire capillary glucose range (1.1 to 27.8 mmol/L was good with a median (IQR) RAD of 10.2% (4.5 - 19.0). Sensor accuracy stratified according to capillary glucose levels is shown in Table. Lowest RAD was ob- served in the hyperglycaemic range. Over 95% of pairs were in combined Clarke error grid zones A and B. Calculated per subject, median (IQR) RAD was 10.8% (8.7 -11.5). Ninety percent of participants had a median ARD < 13.5%. Diabetologia (2014) 57:[Suppl1]S1–S564 S 407 1 C Conclusion: In a cohort comprising 57 subjects with T1D aged 12 years and older, we observed high and consistent accuracy of FreeStyle Navigator II. Particularly at glucose levels above 10 mmol/l, this facilitates safe operation of closed loop as otherwise insulin over-delivery due to overestimated glucose levels could lead to hypoglycaemia. Our results support the use of Freestyle Navigator II in home closed loop studies and may contribute towards estab- lishing performance criteria for CGM in home use. Clinical Trial Registration Number: NCT01221467, NCT01440140 and NCT01666028 Supported by: JDRF, Diabetes UK and Seventh Framework Programme of the European Union 990 Comparison of paired and sensor-based glucose testing for meal bolus adjustment G.A. Hayter, E.S. Budiman; Abbott Diabetes Care, Alameda, USA. Background and aims: Meal boluses, whether based on carb counting or fixed bolus dosing, can benefit from periodic review of the post prandial glu- cose response. For carb counting, this can mean adjusting the carb ratio. For fixed bolus, this can be used to directly increase or decrease the fixed amount. The post prandial glucose response can be characterized by the difference be- tween the peak glucose in response to the meal and the glucose at meal start. Two glucose testing methods for measuring this difference are considered. The first approach relies on a pair of discrete strip tests, one at the meal start and one at some fixed time after meal start. This method is confounded by the fact that peak meal times vary widely. The second approach utilizes sensor data which is sampled periodically. The benefit of this method is that the peak time can be estimated which provides a better estimate of the glucose differ- ence. The precision of both methods will be compared, taking into account measurement error. Materials and methods: Sensor data from a feasibility study of a 14-day sen- sor wear, comprising of 55 subjects with diabetes, is used. These data have a 1 minute sample rate and will represent “truth” for this analysis. True meal start and peaks have been previously identified. Paired testing is simulated using the truth data, with the first reading at the true meal start and the second reading at a predetermined fixed time interval (e.g., 60 ,120 and 180 minutes). Sensor-based testing is simulated by decimating the truth data to longer sam- pling intervals (e.g., 5, 10, 15 minutes) and a meal detection algorithm is ap- plied to these decimated data to estimate meal start and peak times. Results: The figure summarizes the median and inter-quartile-range (IQR) of the relative difference (RD) between the “true” glucose meal response and the results from the six methods. The three sensor-based testing methods are shown to have much better precision than the paired testing methods; IQR RD for the best paired testing (Paired 120) is about 46% compared to about 25% for all of the sensor-based testing methods, and paired testing has substantially more negative bias. The difference in precision between the two test methods is primarily attributed to the wide variation in glucose meal response from patient to patient; the sensor-based method estimates the peak time based on the data while the paired testing method assumes a fixed dura- tion for the peak time. The figure also shows that performance for sensor- based testing is similar for sample periods up to 15 minutes. Conclusion: The sensor-based testing method offers a more consistent char- acterization of meal response than the paired testing method, resulting in higher confidence that bolus adjustments can be made safely and effectively. The higher level of precision may allow dose titration that results in lower overall glucose levels with lower risk of hypoglycemia, but these results must be evaluated in prospective trials. Supported by: Abbott Diabetes Care 991 A cell-based hybrid biosensor for automatic real-time quality control of islets and sensing of insulin demand F. Lebreton1, M. Raoux1, Y. Bornat2, J.-B. Floderer2, D. Bosco3, J. Gaitan1, P. Meda4, T. Berney3, B. Catargi1,5, S. Renaud2, J. Lang1; 1Dépt. Sciences & Technologies, Université de Bordeaux, UMR CNRS 5248, Pessac, 2Institut Polytechnique de Bordeaux, UMR CNRS 5218, Talence, France, 3Dépt de Chirurgie, Hôpitaux Universitaires de Genève et Université de Genève, 4Dépt. Physiologie & Métabolisme, Université de Genève, Switzerland, 5Hôpital St André, CHU de Bordeaux, France. Background and aims: Glucose sensors have achieved considerable progress in the treatment of diabetes but require complex algorithms to match the complexity of everyday life. In contrast, pancreatic islets have been shaped during 0.5 billion years of evolution to secrete insulin not only in response to glucose, but also to other nutrients and to modulate their functioning to take physiological conditions into account via hormone receptors. Islets may provide the optimal sensor that uses changes in electrical activity as the first integrative signals of regulation. We have set out to develop recording and real-time analysis of these signals (i) for long-term analysis in drug testing, (ii) to provide a device for pre-transplantation control of donor islets and (iii) to advance in the development of a novel sensor for insulin demand. Materials and methods: Mouse and human islets were cultured on multi- electrode arrays. Electrical activities received on each electrode were continu- ously amplified and analogically filtered. Output signals were recorded for of- fline analysis or digitally converted and processed (for up to 60 electrodes in parallel) in real time using integrated circuits that were specifically designed on a custom electronics board. Results: An acquisition system of islet electrical activity has been designed to process 60 channels in real time. This portable and autonomous system measures islet signals of extremely low amplitude (microvolts). Digital filters embedded on integrated circuits perform signal frequency separation, as well as frequency and statistical computation on individual signals. Extracellular recordings of human and mouse islet cells reveal two distinct glucose-de- pendent signals. According to pharmacology and experiments in Cx36-null mice, rapid action potentials (AP) are due to depolarization of single cells, whereas slower changes (slow waves, SW) result from coupling of multiple cells via gap junctions. In mouse islets, GLP-1 enhances SW frequencies at 8.2 mM glucose with maximal effect at 50 pM. In islets perfused with glucose, first increasing the concentration in small 0.5 mM steps from 3 to 11 mM and then decreasing by the same steps, both APs and SWs displayed a maximal effect at 10 mM. An exquisite glucose-dependency was present with a hyster- etic shape, characterized by a shift of the EC50 from 7.5 mM for the increasing phase of the glucose ramp to 8.5 mM glucose for the decreasing phase. Simi- lar characteristics were observed in human islets. Moreover, our algorithms correctly blind ranked signals recorded at different glucose concentrations. Conclusion: Glucose and GLP-1 signatures can be automatically computed from AP and SW signals. The islets provide natural integrated hysteretic Diabetologia (2014) 57:[Suppl1]S1–S564 S 408 1 C algorithms with rapid reaction to hyperglycemia and hormones as well as exquisite sensitivity to decreasing glucose levels, a safety latch against hypo- glycemia. In the long-term, this system may also provide a unique biosensor for insulin demand. Supported by: Région d’Aquitaine, ANR Emergence, French Min. of Research PhD scholarship 992 Hurdles in bioelectronics sensors: how to guide cells toward electrodes? E. Pedraza1, A. Karajić2, F. Lebreton1, M. Raoux1, J. Gaitan1, V. Ravaine2, A. Kuhn2, J. Lang1; 1Dept. Sciences & Technologies, Université de Bordeaux, UMR CNRS 5248, 2Université de Bordeaux, UMR CNRS 5255, Pessac, France. Background and aims: We are exploiting the innate, multi-parametric sens- ing ability of islets in order to develop a hybrid bioelectronics glucose sensor. In our sensor, multielectrode arrays (MEA) record the electrophysiological signal produced by beta cells in response to multiple environmental factors (glucose, lipids, hormones, etc.). Signal quality, and consequently overall sen- sor performance, relies critically on close cell-electrode proximity. However, attempts to improve cell-electrode co-localization, via chemical/physical modification of electrodes or dielectrophoresis, lead to additional cell stress and convoluted setups. In contrast, we present here a non-invasive method of further exploiting the electrical properties of beta cells to guide them directly toward electrodes via simple electrophoresis. Materials and methods: The electrophoretic motion of clonal beta cells (INS-1 832/13) (1E6 cells/mL) and murine islets (60 islets) in an electric field was investigated using a low conductivity, isotonic buffer (5 mM HEPES, 280 mM mannitol, 5 mM glucose). The electric charge at the surface of the INS-1 cells was measured in terms of their zeta potential using a Zetasizer Nano. Computational models describing the distribution of the electric field and predicting the path of cell motion were created to optimize electrode configu- ration. INS-1 cells were subjected to an electric field for 10 min and subse- quently assessed via proliferation assay (MTT) after 24 h, glucose stimulated response test after 48 h, and analysis of electrophysiological activity after 48 h. Student‘s t-test was used in statistical analysis. Differences were considered significant (p0.1). Conclusion: Our findings demonstrate that INS-1 beta clonal cells and disso- ciated mouse islets can be driven directly to target electrodes by applying an electric field. The positive results of the viability tests, insulin secretion tests, and electrical recordings demonstrate that this technique is not detrimental to the health and function of beta cells. Furthermore, its application can be extended to other cell types and other bioelectronics sensors. We are cur- rently investigating how to adapt this technique to improve the manipulation of whole islets, a task which is complicated by their higher density compared to single cells. Supported by: Excellence cluster Labex AMADEus 993 Sensor and software use for improved glucose control in individuals with diabetes managed by multiple daily injections of insulin: the SIGN study R.A. Ajjan1, K. Abouglila2, S. Bellary3, A. Collier4, B. Franke5, E. Jude6, G. Rayman7, A. Robinson8, B. Singh9; 1Molecular Vascular Medicine, Leeds University, 2County Durham & Darlington NHS Foundation Trust, Durham, 3Heart of England NHS Trust, Heartlands Hospital, Birmingham, 4University Hospital Ayr, 5The Rotherham Hospital NHS Foundation Trust, Rotherham, 6Tameside Hospital NHS Foundation Trust, Ashton-under-Lyne, 7Ipswich Hospital, 8Royal United Hospital, Bath, 9Royal Wolverhampton Hospital NHS Trust, New Cross Hospital, UK. Background and aims: Studies investigating the role of continuous glucose monitoring (CGM) to improve glycaemic control in type 1 and type 2 dia- betes patients remain relatively scarce. Our study aims were to: 1) Assess whether subjects with type 1 (T1D) or type 2 diabetes (T2D) managed by multiple daily injection (MDI) of insulin could improve glycaemic control by understanding CGM data and reviewing ambulatory glucose profiles (AGP) with their clinician. 2) Identify device features and glucose profile character- istics that enable improved clinical outcomes. Materials and methods: A UK, multicentre (n=9), 100 day study recruit- ed 105 MDI treated diabetes patients aged 18-82 years with HbA1c of 58- 108mmol/mol. Subjects were randomised to control or intervention group; the control group used a standard self-monitoring blood glucose device, FreeStyle Freedom Lite. The intervention group employed a FreeStyle Navi- gator CGM and were asked to disable the glucose alarm function. At days 30 and 45, subjects reviewed AGP and summary statistics with a clinician and completed a questionnaire at the end of the study. Post-hoc exploratory anal- ysis evaluated the frequency of high glucose variability, defined to be, have at least one period of the day (by each patient’s typical meals and bedtimes) with the 50th - 10th percentiles > 4.72 mmol/L. Results: In T1D intervention group (n=29), within group analysis showed significant reduction in hypoglycaemia ( S 409 1 C every 15 minutes and capillary SMBG test every 30 minutes; glucose was ma- nipulated to provide sufficient data in low and high glucose ranges. Results: On average, the CGM provided glucose readings with a mean abso- lute relative difference (MARD) of 9% relative to temporally-matched refer- ence YSI compared to 5% of that for the SMBG meter. The CGM readings were highly correlated with YSI with correlation coefficient of 0.97 compar- ing 0.99 of that for SMBG in the regression analyses. Using YSI reference, the CGM performed similarly as the SMBG meter. In particularly, for hypo- glycemia (< 70 mg/dL , i.e.3.9mml/L), the mean absolute difference (MAD) was 6.4 mg/dL ( 0.36 mmol/L) for the CGM compared to 4.2 mg/dL i.e. 0.23 mmol/L) for the SMBG; for hyperglycemia (> 180 mg/dL i.e.10 mmol/L), the MARD was 8% for the CGM compared to 5% for the meter. The Clarke Error Grid (CEG) showed that 92% of CGM-YSI matched pairs fall in the clinically accurate A zone with 7% in the B zone, 0.0% in the C zone, and 0.5% in D zone for CGM; and 98.5% of SMBG-YSI matched pairs fall in the A zone with 1.1% in the B zone, 0.0% in the C zone, and 0.4% in the D zone. Conclusion: The overall point accuracy, clinical accuracy,accuracy over the duration of wear, accuracy across the glycemic ranges, and reliability (98% of sensors lasted 7 days) were comparable to the SMBG meter. Accordingly, the CGM accuracy should not limit AP system development. Figure 1. Compari- son of CGM-YSI vs. SMBG-YSI Clinical Trial Registration Number: NCT02087995 995 Glucose concentrations in blood and tissue: variable time lag R. Chlup1,2, J. Krejci3, M. O´Connell4, B. Sebestova3, R. Plicka3, L. Brozova3, B. Doubravova3, H. Zalesakova2, J. Vojtek3, J. Bartek5; 1Dept. of Physiology and IInd Dept. of Medicine, Palacký University and Teaching Hospital, Olomouc, Czech Republic, 2Institute of Gerontology and Geriatrics, Moravsky Beroun, 3BVT Technologies, a.s., Strazek, Czech Republic, 4Probe Scientific Ltd, Bedford, UK, 5Dept. of Medical Chemistry and Biochemistry, Palacký University Olomouc, Czech Republic. Background and aims: Time constants dealing with carbohydrate metabo- lism are hot topics of diabetes physiology. Current methods derive glucose concentrations from different tissues using subcutaneous and whole blood measures. The aim of this pilot study was to assess the time delay between glucose concetration in blood and tissue in the course of increase and de- crease. Materials and methods: Three healthy volunteers, 3 persons with type 1 dia- betes (DM 1) and 4 persons with type 2 diabetes (DM 2) underwent concur- rent glucose measurements using standard equipment (1) continuous glucose monitoring system (CGMS), namely, sensor Enlite, transmitter Minilink and monitor Guardian, Medtronic Minimed, CA, USA, calibrated by glucom- eter Calla, Wellion, Austria, and (2) Beckman analyser to estimate venous plasma lglucose concentrations. Samples were taken in fasting state (just be- fore drinking 300 ml tea with 50 g glucose) and then 4 times (hourly). This test was performed on 3 consecutive days. Carelink Personal and MS Excel were applied to plot the 4-hour evolution of glucose concentration obtained by CGMS and by Beckman, respectively. In each volunteer the time differ- ence between ascending and descendig plots was estimated and averaged for healthy persons, for DM 1 and for DM 2, respectively. Results: The results are summarized in the Table showing 1) the mean time delay of defined glucose concentrations between the methods, 2) the mean period of time from the start of consuption of glucose to the maximum dif- ference betwen the rising glucose concentrations, 3) the mean period of time from the start of consumption of glucose to the maximum glucose concen- tration, 4) the mean delay betwen the decreasing glucose oncentrations. The results are relevant to common practice. Conclusion: In persons with diabetes glucose tolerance testing resulted in slower transport of glucose into subcutaneous tissue than in healthy subjects. The rate of change of glucose between blood and tissue varies, introducing er- rors in CGMS readings when calibrations are made in non-steady state con- ditions. CGMS can provide reliable plasma glucose concentrations but only if the system is calibrated during steady state. Hence, before each calibration an appropriate algorhithm to confirm the presence of a steady state should be applied. Supported by: Palacky University Olomouc; Probe Scientific Ltd 996 Glucose trend prediction in diabetic patients using Random Forests algorithm for analysis of CGM data K. Tabakov1, N. Myakina2; 1Department of Information Technology, Novosibirsk State University, Russian Federation, 2Laboratory of Endocrinology, Institute of Clinical and Experimental Lymphology, Novosibirsk, Russian Federation. Background and aims: Avoiding both hypoglycemia and hyperglycemia is the critical difficulty in the management of diabetes because of the impact of multiple factors on glycemic trends. Despite this, the development of new approaches to analysis of data obtained in continuous glucose monitoring (CGM) provides an opportunity for prediction of glucose level without quan- tifying of these factors. The aim of the study was to evaluate the possibility of using the Random Forest algorithm for glucose level prediction based on CGM data. Materials and methods: Total 50 records of 48 h CGM obtained from pa- tients with type 1 and type 2 diabetes mellitus were analyzed. Based on this dataset using Random Forests algorithm and R-STUDIO software a 50 lin- ear regression models were created for each case individually. Programming the Random Forests module was performed using the author‘s modification of the algorithm of local search, which allows you to automatically find the optimal settings of the program for a particular purpose. Various predictive windows of 50-360 min were chosen to predict glucose trends in these diabe- tes patients. The algorithm was trained individually for each case on a set of randomly selected segments of the CGM record that total was 50% of the en- tire record. Subsequently, the algorithm was verified on the data not included in the training set. Model adequacy was evaluated by the difference between the predicted and actual value and also by the mean square error of all the predicted values at hypoglycemic and hyperglycemic ranges. Results: Overall, a valid prediction has been provided at normal, hypoglyce- mic and hyperglycemic ranges in 50-180 min predictive window: the mean squared prediction error was 8.1, 8.6 and 9.8, respectively (p=0.0001). Conclusion: Random Forests algorithm can be successfully utilized to pre- dict glucose levels in diabetic patients based the CGM data. At the predictive window of 50-120 min (an important period related to postprandial glyce- mia) the model has a good predictive accuracy. Models like these could be employed in a semiclosed-loop device that can provide better glycemic con- trol in diabetic patients. Diabetologia (2014) 57:[Suppl1]S1–S564 S 410 1 C PS 082 Improving insulin pump treatment and continuous glucose monitoring use 997 Time delay, reference measurement errors and number of reference measurements may strongly alter MARD values for CGM devices H. Kirchsteiger1, G. Freckmann2, L. Heinemann3, V. Lodwig4, G. Schmelzeisen-Redeker4, M. Schoemaker4, L. del Re1; 1Institute for Design and Control of Mechatronical Systems, Johannes Kepler University Linz, Austria, 2Institute for Diabetes-Technology GmbH, Ulm, 3Science & Co, Düsseldorf, 4Roche Diagnostics GmbH, Mannheim, Germany. Background and aims: As a parameter describing the measurement quality of CGM systems most often the median (or mean) average relative deviation (MARD) is stated. For calculation of MARD paired CGM readings and re- sults of reference blood glucose measurements are used. However, this com- parison is affected by: a) measurements are made in different compartments (blood, interstitial fluid) which leads to a time-delay between the results; b) reference values are also affected by measurement errors; c) the selection and number of time points used for computation has an effect as well. Materials and methods: In order to evaluate the effect of these factors we re-analyzed data from a clinical study in which the performance of 3 differ- ent CGM systems (two of each kind, 6 in total) was evaluated in 12 patients with type 1 diabetes. We systematically varied the time delay between the two measurement methods to find the delay which results in the minimal MARD. The result was verified using a correlation based time delay analysis. We simulated measurement errors of the reference method and formulated an optimization problem, which gives as a solution the reference measure- ment values inside the error bounds, which result in the minimal MARD values. We also systematically selected different time points of paired refer- ence and CGM measurements which were excluded from the MARD com- putation: a) repeatedly excluding the time-point which contributes the most to the MARD; b) repeatedly excluding the time-point which contributes the least to the MARD. This results in two functional dependencies of the MARD from the nr. of samples used with the actual MARD value being in between the extreme cases. Results: MARD values decrease from originally 16.5%, 12.6%, and 16.7% for the CGM systems by 1.5%, 1.7%, and 2.5% when compensating for the ef- fect of the time-delay. MARD values varied in the worst case in the range of 11.0_21.8%, 6.9_16.3%, 10.7_21.6% when compensating for a 5% measure- ment error of the reference method and the estimated time-delay. The MARD values varied even more with different selections of the time-points of paired measurements: Reducing the number of paired measurements by 50% can re- sult in an improvement (= lower value) by nearly 50% of the original MARD value. However, selecting different time points can also result in an MARD value twice as high as the original value. Conclusion: MARD is greatly affected by various factors. Following a rigor- ous protocol that controls for certain factors that have an impact on MARD is crucial to enable uniform comparability among different clinical studies. POCT05-A guidelines should consider the effects discussed. Supported by: Roche Diagnostics 998 Using of RT-CGMS during perinatal period with special focus on newborn hypoglycaemia occurence: our ongoing experience K. Stechova1, K. Chaloupek2, D. Bartaskova1, I. Spalova3, M. Cerny2; 1Dpt.of Internal Medicine, 2NICU, Dpt.of Obstetrics and Gynaecology, 3Dpt.of Obstetrics and Gynaecology, University Hospital Motol, Prague, Czech Republic. Background and aims: In babies of diabetic mothers perinatal complications including hypoglycaemia are still seen more frequently than in newborns of healthy mothers. In our study by using real time continuous Glucose moni- toring system (RT-CGMS) we monitor pregnant women prior the labour and we continue to monitor their newborns after the birth as well to be able to follow complex glycaemia changes which influence the child prior and after the birth and correlate them with the newborns´s clinical status including hypoglycaemia occurence. Materials and methods: Newborns: We analysed by RT-CGMS 30 newborns of mothers suffering from pregestational diabetes, all mothers were treated by insulin. 23/30 of these children were born to type 1 diabetes (T1D) mothers, in 5/30 mothers were treated for T2D (type 2 diabetes) and in 2/30 cases the mother has MODY2 (maturity onset diabetes of the young). Median of gesta- tion age in diabetes group was 37 weeks (range 31-38 weeks). Newborns were classified into four clinical status cathegories for further analysis. The first Enlite sensor was placed immediately after the birth (connected to Guardian real time device, Medtronic, Minneapolis, MN, USA). Eight term newborns born to healthy mothers were analysed as controls too. The median of moni- toring in diabetes group was 5 days (range 3-8 days) and 4 days in controls (range 4-6 days). Mothers: in 9 T1D mothers their RT-CGMS records at least 3 days prior the labour and during the labour were available for the analysis too. In mothers the same type of RT-CGMS device was used as in newborns. Data were analysed by SPSS sw v.17. Notice - hypoglycaemia in newborns is considered as glycaemia ≤ 2,5 mmol/l. Results: RT-CGMS revealed higher frequency of hypoglycaemia episodes in newborns of diabetic mothers (only 6/30 newborns of diabetic mothers were without any hypoglycaemia). These hypoglycamia episodes were present not only shortly after the birth but after 3rd monitoring day as well (p=0.01). In control newborns hypoglycaemia episodes were present too (4/8) but during the first 4 days of life only and were less frequent and shorter. In T1D group the occurrence and severity of hypoglycaemia were influenced by: the pres- ence of macrosomia (p=0.022), length of the newborn (p=0.001), maternal HbA1c in 3rd trimester (p S 411 1 C percentage of subjects achieving HbA1c S 412 1 C 1002 Factors associated with successful continuous subcutaneous insulin infusion therapy in type 2 diabetes patients - the OPT2MISE trial O. Cohen1, Y. Reznik2, I. Conget3, R. Aronson4, J. Castañeda5, S. Huang6, S. Lee6, J. Shin6, OpT2mise Study Group; 1Ch. Sheba Medical Center., Tel-Hashomer, Israel, 2Centre Hospitalier Régional Universitaire de Caen Côte de Nacre, France, 3Endocrinology and Nutrition Department, University Hospital Clinic, Barcelona, Spain, 4LMC Diabetes & Endocrinology, Toronto, Canada, 5Medtronic Bakken Research Center, Maastricht, Netherlands, 6Medtronic, Northridge, USA. Background and aims: The recent randomized study OpT2mise, demon- strated that insulin pump therapy (CSII) provided significant improvement in glycemic control, as compared with injection therapy (MDI) in type 2 dia- betes (T2D). We aimed to assess the association between baseline or treat- ment related factors with glycemic improvements seen in the CSII arm of the study. Materials and methods: Sub-optimally controlled (HbA1c ≥ 8%) T2D sub- jects, despite an MDI optimization period, were randomly assigned to CSII or continuing an MDI regimen. Simple and multiple linear regression analysis were performed to assess the association of A1c change at 6 months and fac- tors such as age, gender, duration of diabetes, baseline A1c, BMI, treatment satisfaction score (DTSQs), total daily insulin dose, cognitive state score (MoCA) and glucose self-measurements (SMBG). Analyses were adjusted by baseline A1c. Results: A total of 331 subjects were randomized (45.6% women, age 56.0±9.6 yr, BMI 33.4±7.3 kg/m2, diabetes duration 15.1±8.0 yr, HbA1c 9.0±0.8%). Sub- jects assigned to CSII achieved significantly greater A1c reduction (p S 413 1 C 1004 Glucose variability outcome for type 1 diabetic patients switching to CSII: improved complexity patterns beyond glucose dispersion reduction M. Lytrivi, L. Crenier; Endocrinology, Erasme Hospital, Brussels, Belgium. Background and aims: Standard Deviation (SD) and MAGE are commonly reported indices of glucose variability (GV). However, they represent meas- ures of glucose dispersion rather than true quantifications of instability. These indices are also directly proportional to the mean, making them inappropri- ate for comparisons of glucose profiles at different glucose levels. The aim of this work was therefore to study the evolution of GV in type 1 diabetic pa- tients switching to continuous subcutaneous insulin infusion (CSII) in terms of both dispersion and complexity of the displayed patterns. For this purpose, we used metrics independent of the mean, as the Coefficient of Variation (CV), Approximate Entropy (ApEn) and the Detrended Fluctuation Analysis (DFA) method, which are, contrary to CV, sensitive to intrinsic complexity and self-similarities of glucose patterns rather than to value scattering. Materials and methods: CV, ApEn (m=1,r=0.2xSD) and DFA-α were com- puted from blinded Continuous Glucose Monitorings (CGMs) of 48h dura- tion (iPro2 with Enlite® sensors, Medtronic MiniMed, Northridge, CA) ob- tained from 20 adults with type 1 diabetes on MDI (baseline) and 6 months after initiation of CSII. Mean and SD of glucose, HbA1c, MAGE, the number of hypoglycemic episodes and time in hypoglycemia (TH) were also recorded. Results: CV was reduced by 4.9 ± 9.7 (p=0.036) on CSII, while change in CV (ΔCV) was strongly correlated with baseline CV (Pearson’s r=-0.77, p S 414 1 C metabolic control. The goal of this research was to examine if sleep duration and late eating habit in CSII treated T1DM patients were associated with gly- cemic control as assessed by HbA1c level and mean self-monitoring blood glucose (SMBG) readings. Materials and methods: We included 148 adult T1DM patients on insulin pump (100 females and 48 males) who filled a questionnaire concerning sleep duration (≤6h and >6h) and late eating (snacking after 10 p.m. - rarely, sometimes, often). Other sources of information about the patients includ- ed available medical records together with the results of biochemical tests (HbA1c measured with high performance liquid chromatography method- HPLC), memory read-outs from insulin pumps, data from blood glucose meters. Data from personal insulin pumps and blood glucose meters were downloaded regularly to the computer and analysed via CareLink Profes- sional (Medtronic) and Accu Chek 360 (Roche) software. Tests for differences between the two groups, univariate correlations analysis, and multivariate re- gression analysis were used. Results: The patients‘ mean age was 26 yrs, mean T1DM duration 13.4 yrs, mean HbA1c level 7.2%, mean BMI 23.3 kg/m2, while mean nighttime sleep duration was 7.2 h. Short sleep duration (26.3% of patients) as compared to long sleep duration was associated with worse glycemic control - HbA1c: 7.6 vs. 7.1%, p=0.0293; mean SMBG readings: 159 vs. 149 mg/dl, p=0.0491, re- spectively. Additionally, in a univariate correlation analysis we found an as- sociation between frequent late snacking (rarely: 54.1%, sometimes: 38.5%, often: 7.4%) and worse HbA1c level (r=0.19, p=0.0287) and SMBG readings (r=0.21, p=0.0135). In the multivariate regression analysis (dependent varia- bles - age, T1DM duration, gender, BMI, sleep duration and frequency of late snacking), independent predictors for HbA1c were the patient‘s age (β=-0.34, p=0.0013), T1DM duration (β=0.21, p=0.0440) and sleep duration (β=-0.23, p=0.0061). The analyzed model was significant (p=0.0017). In the multivari- ate regression analysis for mean SMBG readings, the same variables except for T1DM duration were significant Conclusion: In summary, short sleep duration seems to be associated with poorer glycemic control in T1DM patients treated with CSII but late snack- ing is not. 1007 Factors predictive of overnight closed loop performance during free living in children and adults with type 1 diabetes M. Tauschmann1, H. Thabit1, L. Leelarathna1, D. Elleri1, J.M. Allen1, A. Lubina-Solomon2, M. Stadler3, E. Walkinshaw2, A. Iqbal2, P. Choudhary3, S.R. Heller2, S.A. Amiel3, M.L. Evans1, D.B. Dunger1, R. Hovorka1; 1Wellcome Trust-MRC Institute of Metabolic Science, University of Cambridge, 2Academic Unit of Diabetes, Endocrinology and Metabolism, Department of Human Metabolism, University of Sheffield, 3Diabetes Research Group, Kings College London, UK. Background and aims: Overnight closed-loop (CL) insulin delivery is feasi- ble, safe, and effective in the home setting but outcomes vary between indi- viduals. Understanding the factors influencing the performance of overnight CL under free-living conditions may help target those likely to benefit most while improving the control algorithm. Materials and methods: We combined data collected during two randomized home studies in 16 adolescents (age 15.6±3.6 years; mean±SD) and 24 adults (age 43.2±11.7) with type 1 diabetes on insulin pump therapy (pooled HbA1c 8.0±0.8%; duration of diabetes 20.0±14.0 years, BMI 24.3±3.7 kg/m2, total daily insulin dose 46.2±17.3 U/day). Participants underwent, in random or- der, two periods of sensor augmented insulin pump therapy (SAP) or SAP combined with overnight CL insulin delivery, each lasting three-weeks (ado- lescents) or four-weeks (adults). Associations between baseline character- istics (age, HbA1c, duration of diabetes, BMI, total daily insulin dose and gender) and the performance of overnight CL (time in target, mean glucose) were examined using correlation analysis (Spearman’s correlation coefficient, r). Results: We analysed data on 866 CL nights at home. Lower baseline HbA1c (r= -0.43, p S 415 1 C context of their lived experience to support their own diabetes self-manage- ment. This study is ongoing, and final data are pending. Supported by: Animas Corp. 1009 OpT2mise study: the impact of insulin pump therapy on treatment satisfaction and resource utilisation in patients with type 2 diabetes R. Aronson1, O. Cohen2, I. Conget3, S. de Portu4, A.-S. Racault4, J. Shin5, Y. Reznik6; 1LMC Diabetes & Endocrinology, Toronto, Canada, 2Institute of Endocrinology, Ch. Sheba Medical Center, Tel Hashomer, Tel Aviv, Israel, 3Endocrinology & Nutrition Dept, Hospital Clínic i Universitari, Barcelona, Spain, 4Medtronic Int, Tolochenaz, Switzerland, 5Medtronic Inc, Northridge, USA, 6Endocrinology & Diabetes Dept, CHU Côte de Nacre, Caen, France. Background and aims: Achieving glycemic outcomes in insulin-using type 2 diabetes (T2D) patients often requires multiple daily injections (MDI) with basal and bolus insulin types. Insulin pump therapy (CSII) has recently been evaluated in T2D patients in a large prospective randomized controlled trial. Patient outcomes have been found to be positively associated with treatment satisfaction and safety, contributing to improved health and reduced health- care costs. We assessed the effects of CSII on treatment satisfaction (TS) and on medical resource utilization, in suboptimally controlled T2D patients par- ticipating in the OpT2mise study. Materials and methods: T2D subjects, suboptimally controlled (HbA1c ≥ 8%) despite an MDI optimization period, were randomly assigned to CSII or continuing MDI. The primary endpoint was the between-group difference in mean HbA1c change from baseline to 6 months. TS was measured by means of the Diabetes Treatment Satisfaction Questionnaire, status (DTSQs) and change (DTSQc). DTSQs questionnaires were collected at baseline and at 6 months; DTSQc was completed at 6 months. Data on medical resource use including diabetes-related hospitalizations, insulin usage and number of glu- cose testing strips were collected at each visit. Results: 331 subjects were randomized (45.6% women, mean±SD age 56.0±9.6 yr, BMI 33.4±7.3 kg/m2, diabetes duration 15.1±8.0 yr, HbA1c 9.0±0.8%). Subjects assigned to CSII achieved significantly greater HbA1c reduction than MDI arm (-1.1±1.2% vs. -0.4±1.1%, p S 416 1 C and alarms) was higher with Paradigm®Veo pumps (33%) compared to Ani- mas® pumps (12%) (p=0.007). Conclusion: Pump malfunctions remain common with modern pumps. In this study, we reported less complete failures than in our previous study. This may reflect an improvement of the pumps. However, the current high rates of mechanical and minor defects suggest changes in our practice and may be explained by the greater attention paid by patients and healthcare profession- als to pump defects. A multi-centre evaluation of pump failures is needed. 1011 A randomised three-period trial to compare the pharmacodynamics and pharmacokinetics of different glucagon dosages at different blood glucose levels T. Jax1,2, H. Blauw3,4, J. DeVries4, T. Heise5, for the PCDIAB consortium; 1Cardiometabolic Research, Profil Institut für Stoffwechselforschung GmbH, Neuss, 2Klinik für Kardiologie, Herzzentrum Wuppertal, Universität Witten/ Herdecke, Germany, 3Inreda Diabetic BV, Goor, Netherlands, 4Academic Medical Centre at the University of Amsterdam, Netherlands, 5Profil Institut für Stoffwechselforschung GmbH, Neuss, Germany. Background and aims: Bihormonal closed-loop systems deploy the admin- istration of glucagon in case of imminent or actual hypoglycaemia. The ad- ministration of glucagon in such a closed-loop system will not necessarily be limited to hypoglycaemia, but can also be done preemptively, trying to avoid hypoglycemia. Thus more information is needed about the pharmacody- namic effects of small doses of glucagon at non-hypoglycaemic blood glucose (BG) levels. We evaluated the pharmacokinetics and pharmacodynamics of different dosages of s.c. glucagon at different BG levels. Materials and methods: The study was conducted as an open, randomized 3-period cross-over experiment in six otherwise healthy patients with T1DM. At each of the 3 periods, different BG levels were established in 4 consecutive steps (144, 108, 72, and 50.4 mg/dl) and glucagon was given s.c. at each BG level in a randomized sequence (A: 0.11 mg at first 3 BG levels and 1.00 mg at the fourth, B: 3 x 0.22 mg and 0.66 mg, C: 3 x 0.44 mg and 0.33 mg). Results: Glucagon raised BG in a dose-dependent fashion at all BG levels. Mean glucose excursion ranged from 48.7 mg/dl to 111.0 mg/dl (Figure). The response after repeated dosing tended to decrease only with the 0.44 mg glucagon dose (mean glucose excursion from 97.3 mg/dl to 71.0 mg/dl), pos- sibly due to hepatic glycogen depletion. Time till maximum glucose concen- tration showed a dose dependency, with mean values increasing from 41.3 min to 79.2 min. Maximum plasma glucagon concentration increased with increasing glucagon dosage and was reached after around 10-25 minutes. For the 0.11 mg glucagon dose, plasma glucagon concentration tended to in- crease with decreasing baseline BG levels and in 3 patients no clear increase in glucagon concentration was seen at BG level of 144 mg/dl. Conclusion: In conclusion, s.c. administered glucagon produces a predict- able pharmacodynamic response at lower doses than the usual rescue dose and across a range of hypo- to hyperglycaemic BG levels. This supports the use of small glucagon doses in closed-loop systems to prevent impending hypoglycaemia. The results of this study can be used to further optimize the control algorithms of bihormonal closed-loop systems. Supported by: PCDIAB, European Commission FP7 grant 305654 1012 The impact of insulin pump therapy on glycaemic profile of patients with type 2 diabetes: data from the OpT2mise study I. Conget1, O. Cohen2, R. Aronson3, J. Castaneda4, R. Re4, J. Shin4, Y. Reznik5; 1Diabetes Unit. Endocrinology Department, Hospital Clinic i Universitari, Barcelona, Spain, 2Institute of Endocrinology, Ch. Sheba Medical Center, Tel Hashomer, Israel, 3LMC Diabetes & Endocrinology, Toronto, Canada, 4Medtronic, Northridge, USA, 5Endocrinology and Diabetes Department, Centre Hospitalier Régional Universitaire de Caen Côte de Nacre, Caen, France. Background and aims: Previous studies did not conclusively demonstrated a benefit of continuous subcutaneous insulin infusion (CSII) in patients with Type 2 Diabetes (T2D) previously treated with basal-bolus multiple daily in- jections (MDI). The OpT2mise study is a multicenter, randomized trial com- paring CSII to MDI in a large cohort of T2D subjects suboptimally controlled despite MDI therapy. Aim: To further understand the mechanism underlying the differences between both types of therapies, a study on the glucose met- rics, as a secondary endpoint, was performed. Materials and methods: Subjects with poor glycemic control (n=495) on MDI were enrolled into a run-in period for insulin dose optimization (Total Daily Dose ≥ 0.7 and ≤ 1.8 U/kg/d). Those showing an HbA1c ≥8% and ≤12% were then randomly assigned to CSII or continuing on MDI for 6 months. Blinded Continuous Glucose Monitoring (CGM, iPro2, Medtronic) data was collected for a 6 day period before and 6 months after randomization and changes in glucose metrics were evaluated (Glucose exposure, Glucose Vari- ability and Glucose Ranges). Results: A total of 331 patients were randomized, aged 56.0±9.6, 45.6% women, BMI 33.4±7.3, diabetes duration 15.1±8.0 yr, and HbA1c 9.0±0.8%. CSII achieved significantly greater HbA1c reduction than MDI (- 1.1±1.2% vs. - 0.4±1.1%, p 250 mg/dl (-12.2%, p < 0.001 %, and -6.4%, p < 0.05), more time in target (70-180 mg/dl; 11.8 %, p < 0.001) and no dif- ference in time exposure to SG < 70 mg/dl. Concerning glucose variability, there were no difference in 24h Standard Deviation (SD) SG, coefficient of variation (CV) or in mean amplitude of glucose excursions (MAGE). Conclusion: In comparison with MDI, CSII treatment in suboptimally con- trolled patients with T2D provides a significant improvement in glucose pro- file with increased time spent in target, less exposure to hyperglycemia and without an increase in the risk of hypoglycemia. Clinical Trial Registration Number: NCT01182493 Supported by: Medtronic Europe Diabetologia (2014) 57:[Suppl1]S1–S564 S 417 1 C PS 084 Supporting tools for insulin treatment 1013 Impact of a built-in insulin calculator feature on diabetes control: pilot study T. Haak1, M. Diamant2, T. Danne3, G. Freckmann4; 1Research Institue of the Diabetes Academy Mergentheim (FIDAM), Bad Mergentheim, Germany, 2VU University Medical Centre, Amsterdam, Netherlands, 3Diabetes Zentrum fur Kinder and Jugendliche Kinderkrankenhaus auf der Bult, Hannover, 4Institut fuer Diabetes-Technologie Forschungs- und Entwicklungsgesellschaft mbH an der Universitaet Ulm, Germany. Background and aims: To evaluate the impact of the FreeStyle InsuLinx me- ter with a built-in insulin bolus calculator on glycaemic control by observing time spent in euglycaemia (3.9 to 10.0mmol/L) in the assessment phase vs. the baseline phase for the intervention group (primary endpoint), and meas- ures of glycaemic control from a continuous glucose monitor (e.g. time spent in hypoglycaemia, secondary endpoints). Additional secondary endpoints considered in the study were: total daily dose of insulin (TDD), Hypoglycae- mia Fear Survey (HFS), Diabetes Distress Scale (DDS) and Diabetes Treat- ment Satisfaction Questionnaire (DTSQ) results. Materials and methods: The pilot study was a multicentre, randomised, two arm (control and intervention), parallel study conducted at 4 European study centres, with Type 1 diabetes subjects on Multiple Daily Injections (MDI, 3 or more per day) using U100 insulin (n=49). All subjects wore a masked Free- Style Navigator Continuous Glucose Monitor (CGM) for 14 days (baseline phase) where upon randomisation, 33 entered the intervention group (one subject was excluded from analysis as their Navigator had alarms switched on for both CGM phases) and 16 the control group. The intervention group proceeded to use the InsuLinx meter with built-in insulin calculator in either Easy (n=7, fixed dose meal time insulin calculator) or Advanced mode (n=25, rapid acting Insulin calculator with carb counting) for 45 days, the control group used a FreeStyle Freedom Lite meter for 45 days. After this time both arms continued with their designated meter together with a masked CGM for 14 days (assessment phase). Results: When comparing baseline versus assessment phase in the inter- vention group, no significant difference in time spent in euglycaemia (3.9 to 10.0mmol/L) was found (primary endpoint). Study statistics found a de- crease in the number of excursions per day S 418 1 C bachs alpha for BOLUS r= 0.78 and the CARB r = 0.67). Better bolus calcula- tion was associated with a higher level of education (r = 0.24, p S 419 1 C the evaluation against the respective system’s manufacturer’s measurement method (GOD for 7 systems, HK for 3 systems), individual reagent system lots showed 90% to 100% of results within ±15% or ±15 mg/dL of the com- parison method results, with 23 of 30 lots having at least 95% within these limits. When evaluated against the alternate comparison method, between 83% and 100% of results were within these limits for individual lots and 25 of 30 lots showed at least 95% within the limits. Eight of 10 systems fulfilled the minimum accuracy criteria of ISO 15197:2003 (at least 95% of results within ±20% or ±15 mg/dL of the comparison method results) against the respective manufacturer’s measurement method. Individual lots showed between 94.5% and 100% of the results within these limits and 28 of 30 lots showed at least 95% within these limits. Conclusion: In this evaluation, 7 of 10 systems met the minimum system ac- curacy criteria of ISO 15197:2013. Analysis of individual reagent system lots showed that specific evaluation results differed depending on the comparison method. These data suggest that the comparison method used in an evalua- tion can have an influence on an SMBG system’s apparent system accuracy. Investigator-initiated study supported by Sanofi 1018 Is fingerstick calibration beneficial for stable and consistent sensors? E.S. Budiman; Abbott Diabetes Care, Alameda, USA. Background and aims: Subcutaneously inserted glucose monitoring systems typically require fingerstick calibrations. Calibrations allow for sensor spe- cific adjustment and can correct for any sensor drift over the wear duration. When sensors can be manufactured with a sufficiently low in-vitro lot sensi- tivity CV and can remain stable in-vivo, fingerstick calibration may no longer be beneficial. Universal calibration provides a shared, one-time calibration factor for all sensors from the same lot. This analysis computes the expected average and expected 90th percentile sensor performance in fingerstick and universal calibration studies. Per-sensor mean absolute relative difference (MARD) distribution is used for the comparison. Materials and methods: A total of 33 subjects with diabetes were enrolled to wear four sensors simultaneously. Sensors from a lot with low in-vitro sensi- tivity CV (=2.9%) were used in the study. To examine the effect of fingerstick error, an emulator that allows raw sensor data from the actual clinical study to be recalibrated with any meter was used. No difference in per-sensor MARD distribution was observed when the same meter type from a different clinical study data was used for calibration (z-test p-value = 0.80). The fingerstick calibration study was repeated 1000 times using a simulated 15/15 error me- ter (i.e. 95% of the values within 15 mg/dL below 100 mg/dL reference value, and within 15% at or above 100 mg/dL). The universal calibration study is also repeated 1000 times, each time randomly selecting 3 sensors and using the median of the per-sensor sensitivities to calibrate the rest. Results: Each study generated a distribution of per-sensor MARD values. After repeating the study 1000 times, the range of average per-sensor MARD from the fingerstick calibration studies has a mean and standard deviation of 13.7% and 0.24%, respectively (Table 1). The table also shows the average from the 1000 universal calibration studies, as well as the 90th percentile of both studies. The per-sensor MARD distribution of the universal calibration studies, both in terms of the average and 90th percentile, is statistically better than that of the fingerstick calibration studies (p S 420 1 C practices and assess how CGM users respond to their CGM information in a real-world setting. Materials and methods: A 70 question clinical scenario-based survey was developed that probed 5 topics:subject characteristics and general CGM use, treating and preventing incidental and mealtime hypo- and hyperglycaemia and nocturnal issues. The survey was extensively beta-tested to make sure questions were clearly written and unambiguous. Clinicians that actively prescribe CGM from across the US were contacted and asked to recruit and provide a survey web link to regular Dexcom CGM users (>6 days a week on average) from their practice. Results: 222 of 300 respondents had T1D and are analyzed in this abstract: mean age (46 ± 14 years), duration of T1D (22±14 years), self-reported HbA1c (6.9% ± 0.8%), 52% male, 75% used CSII and 25% MDI (75% had used CGM >1 year), education- 65% university graduates or had another advanced de- gree. Key findings: The majority of the participants viewed the RT informa- tion (51%) or hypo/hyper alarms (30%) as the most important information; only 3.6% of subjects thought the computer downloads analysis was most helpful. To prevent hypoglycemia, 70% of participants reported they would prophylactically consume carbohydrates in response to a displayed glucose of 120mg/dl with a decreasing trend (angled or downward rate of change arrow). 42% of respondents stated that at least one time in the last 6 months, their CGM device alerted somebody around them to respond to their hypo- glycemia alarm when they themselves were unable to. 70% and 66% of par- ticipants report waking up at night at least once per week in response to alerts for hypo- and hyperglycemia respectively. In a scenario-based question about response to a low alert waking them up at midnight, 50% of respondents stated they would treat the low glucose without confirming with their meter. 65% of participants reported that since starting CGM, the number of bo- luses they took per day had increased. Daytime and nighttime glucose targets commonly changed- 57% lowered their daytime glucose targets. Participants reported a mean increase in correction insulin dose (unrelated to a meal) for 1 arrow up was 111%, a mean decrease in the correction dose for 1 arrow down was 41%, relative to the correction dose with a flat arrow. For a 50 gram glucose meal at euglycemia, the meal dose increased by 81% for 2 arrows up and decreased by 53% for 2 arrows down, relative to the meal dose with a flat arrow. Most (59%) participants also reported they adjusted the insulin timing in each of the scenarios, based on the direction and rate of change. Conclusion: Patients use CGM data in many of their clinical decisions and for multiple aspects of their diabetes management. These changes contribute to safer diabetes management and improved outcomes. The adjustments re- ported by patients in this survey to their mealtime and correctional insulin doses are much larger than current recommendations. PS 085 Psychological distress 1021 Metabolic control in type 1 diabetic children with self report depression: 3-year follow-up B.M. Zdunczyk1, K. Dżygało2, K. Piechowiak3, A. Szypowska2; 1Clinical Child and Adolescent, Clinical Hospital, 2Department of Pediatrics, Clinical Hospital, 3Department of Pediatrics, Clinical Hospital, Warsaw, Poland. Background and aims: Many studies show there is a higher risk of depres- sion in patients with diabetes than in healthy population. Depression has been associated with worse disease outcomes and readmissions in adults with diabetes. Less is known about the effect of depression on metabolic control and hospital admissions in children and adolescents with type 1 diabetes. The aim of this study was to examine if depression is associated with worse meta- bolic control and repeat hospitalisations in children and adolescents with type 1 diabetes in 3-year follow-up. Materials and methods: In years 2010-2012 we screened 477 children for depression: during the routine visit in the outpatient clinic 252 girls and 225 boys, with diabetes duration > 1 year filled in Polish version of Children’s Depression Inventory by M. Kovac, a self-report questionnaire consisting of 27 items. At the same time other data was collected: sex, age, diabetes dura- tion, HbA1c, BMI, daily insulin dose. Medical parameters of 139 children (65 boys, 74 girls, age 11,38 ±1,85 diabetes duration 4,53±3,05, ins/kg 0,81±0,20, BMI 19,18±0,2 , HbA1c 7,5±1,3 at baseline ), collected for 3 years were ana- lysed. Results: 5% (22/137 participants scored ≥ 13, indicating elevated depressive symptoms. There was no difference between children with CDI ≥ 13 and CDI < 13 in diabetes duration (p = 0.483) and age (p = 0.329). We found sig- nificant relationship between scores in CDI and HbA1c in subsequent years of follow up: 2010 - r=0.1753 p=0,039, 2011 - r=0.1349 p=0,037 and 2013 - r=0.2629 p=0,0018 but not 2012 - r=0.1159 p=0,183. CDI scores were not as- sociated with daily insulin dose for all years but the baseline (2010 r = 0.181, p = 0.033, 2011 r = 0.139 p= 0.125, 2012 r = 0.099, p= 0.273). We didn’t find the difference between participants with CDI ≥ 13 and CDI < 13 in HbA1c (p=0.573), daily insulin dose (p=0.113) or BMI (p= 0.541). At the same time children with CDI ≥ 13 had higher insulin dose/kg/d (p = 0.024). Surprising- ly we didn’t find the difference between children with CDI ≥ 13 and CDI < 13 in the number of hospitalisations (p=0.797) or psychological care (p = 0.064). Conclusion: 15% participants show elevated depressive symptoms as as- sessed with Children’s Depression Inventory at the baseline. Children and adolescents with higher scores on the CDI don’t differ from patients without depressive symptoms in age, BMI and daily insulin dose. Although depressive symptoms are associated with worse metabolic control children with higher scores in CDI achieve the same HbA1c level as their peers with lower CDI scores. This trend is seen for the whole period of follow-up. Our results show that depressive symptoms are not associated with higher use of hospital facili- ties. Still it is necessary to pay attention to emotional wellbeing of children and adolescents with diabetes type 1 regardless of HbA1c. 1022 The assessment of factors related to depressive symptoms in adult patients with type 1 diabetes A. Duda-Sobczak, D. Zozulinska-Ziolkiewicz, D. Pisarczyk-Wiza, B. Wierusz-Wysocka; Department of Internal Medicine and Diabetology, Poznan University of Medical Sciences, Poland. Background and aims: Depressive disorders are more common among peo- ple with diabetes compared with general population. The presence of depres- sive symptoms may reduce motivation and the treatment adherence and in fact worsen the treatment outcomes. The aetiology of depression is complex and not only related to diagnosed chronic disease. The aim of this study was to assess factors related to depressive symptoms in adult patients with type 1 diabetes. Materials and methods: 304 subjects (164 women), aged 43 (33-52) years, type 1 diabetes duration 26 (33-31) years, HbA1C 7.8 (7.1-8.8)% were in- cluded. Subjects diagnosed with chronic complications causing disability (blindness, end stage renal disease, painful peripheral neuropathy, limb am- putation) were excluded to avoid the presumable impact of irreversible dis- Diabetologia (2014) 57:[Suppl1]S1–S564 S 421 1 C ability on questionnaire answers. Depressive symptoms were assessed using Beck Depression Inventory (BDI). Additionally, Problem Areas in Diabetes Questionnaire (PAID), Multidimensional Fatigue Symptom Inventory-Short Form (MFSI-SF) were completed. Data on social, economic and lifestyle fac- tors were collected. Metabolic control expressed by HbA1C and medical histo- ry was assessed via medical chart review. The studied group was divided into two subgroups according to the presence or absence of depressive symptoms based on the BDI standard cut-offs. Results: 41% of study participants presented any depressive symptoms. They were more frequent among women (p S 422 1 C study has found PAID score to be weakly related to Hba1c. In our study, PAID 1 score is also related to poor metabolic control, figure 1. Conclusion: We conclude that PAID 1 is suitable as a rapid tool to disclose emotional distress in a busy clinical setting. PAID 1 has been easily imple- mented by the nurses and it is found valuable as an opening question in the consultation in order to discuss emotional distress and worries related to diabetes treatment, complications and daily life. Next step will be to study change in PAID 1 over time and how we can manage further screening and treatment of emotional distress in a structured way. 1026 Associations between glycaemic control, depressed mood, clinical depression and diabetes distress H. Ascher-Svanum1, D. Schuster1, H. Schmitt2, L. Marangell1, E. Dennehy1, A. Zagar1, D. Jiang1, D. Kendall1, J. Raskin3, R. Heine1; 1Eli Lilly and Company, Indianapolis, USA, 2Lilly, Brussels, Belgium, 3Lilly, Toronto, Canada. Background and aims: Although depression has been associated with poor glycemic control in patients with type 2 diabetes (T2DM), the available data are inconsistent and both methodological approaches and data quality vary significantly across studies. Specifically, depression has been defined and measured in multiple ways, including identification of a clinical syndrome, use of a depression rating scale and as measured by distress linked to diabetes and its management. This exploratory, post hoc analysis assessed the associa- tion between depression, evaluated by these 3 methods, and glycemic control over 2 years in patients withT2DM initiating insulin in a single treatment trial. Materials and methods: We analyzed data from a 24-month, prospective, observational study that evaluated glycemic response in patients with T2DM who initiated insulin therapy (N=985) in 5 European countries. Secondary measures included patient-reported diagnosis of depression at baseline, se- verity of depressed/anxious mood (EuroQol (EQ)-5D item) and diabetes- related distress (Psychological Distress domain of the Diabetes Health Pro- file, DHP-18). The latter two measures were assessed at baseline and 5 time points throughout the study. Glycemic control was measured by A1c at these same time points. Analyses employed t-tests to assess the unadjusted baseline difference in A1c between patients with and without the respective depres- sion parameter. The potential effect of demographic and clinical confounding variables was controlled through a linear model structure. Results: At baseline, A1c of patients with depressed mood (moderately or extremely depressed/anxious) was higher, but not statistically significantly different from patients without depressed mood (9.6% vs. 9.5%, respective- ly, p=.303). However, patients with depressed mood had higher A1c values throughout the 24 months follow up (p S 423 1 C uric acid), and markers of oxidative damage (sialic acid, urinary-8-OH- deoxyguanosine [u-8-OH-dG]) were assessed at baseline, and after six-and twelve months. Repeated measures ANOVA’s were employed to determine within- and between-group differences in depressive symptoms and meas- ured biomarkers at follow-ups. Results: Participants in the psychoeducational, physical exercise and the control group improved equally in depressive symptoms from baseline to 12-month follow-up (time versus time x group effect; F= 12.51 p S 424 1 C PS 086 Psychology and quality of life 1030 Physicians’ challenges when discussing the type 2 diabetes diagnosis with patients: insights from a cross-national study (IntroDia™) W.H. Polonsky1, A. Belton2, S. Down3, M. Capehorn4, V. Gamerman5, F. Nagel6, J. Lee6, D. Clark6, S. Edelman7; 1Department of Psychiatry, University of California San Diego and Behavioral Diabetes Institute, San Diego, USA, 2International Diabetes Federation, Brussels, Belgium, and The Michener Institute for Applied Health Sciences, Toronto, Canada, 3Somerset Partnership NHS Foundation Trust, Bridgwater, 4National Obesity Forum and Clifton Medical Centre, Rotherham, UK, 5Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, USA, 6Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany, 7Division of Endocrinology and Metabolism, University of California San Diego and Veterans Affairs Medical Center, San Diego, USA. Background and aims: Physician communication at diagnosis of type 2 diabetes (T2D) may impact patient self-care, quality of life, and outcomes. IntroDia™ is a survey of ~17000 T2D patients and physicians in 26 countries to investigate early physician-patient communication and its potential con- sequences. Materials and methods: As part of IntroDia™, we surveyed 6753 physicians from Asia, Europe, America, Africa, and Australia using an online survey that included a novel questionnaire of 12 challenges that physicians may en- counter at T2D diagnosis, and the Jefferson Scale of Physician Empathy. Results: Across countries, 76-100% of physicians (88% overall) agreed that conversations with patients at T2D diagnosis impact on patients’ disease ac- ceptance and treatment adherence. 92% of physicians wanted tools to help patients sustain behavioural change. Using factor analysis, the 12 items re- sulted in loading onto 2 factors (Table): Discouraged with Patients at Diag- nosis (DPD; α = 0.87), comprising 8 related challenges, and Frustrated with Situation at Diagnosis (FSD; α = 0.72), comprising 4 challenges. Correlation between these 2 factors suggested related, but distinct, groups of challenges (r = 0.64, p < 0.0001). Factor scores varied globally (DPD highest in France; FSD in Japan). Upon adjusting for demographic/clinical practice variables, regression models showed a negative relationship between physician empa- thy and perceived challenges for total score (all 12 items) as well as DPD and FSD (all p < 0.0001). Conclusion: Many physicians, especially those scoring lower on empathy, report significant challenges and frustrations during conversations with patients at diagnosis of T2D. Most physicians wanted tools to help patients sustain behavioural change. Supporting use of empathy-related skills may contribute to better patient outcomes. Supported by: Boehringer Ingelheim/Eli Lilly 1031 Health-related quality of life and the relation to metabolic syndrome, type 2 diabetes, degree of obesity and low-grade inflammation in obese individuals S.N. Slagter, M.M. van der Klauw, J. van Vliet-Ostaptchouk, J.C. Keers, H.L. Lutgers, A.P. van Beek, B.H.R. Wolffenbuttel; Endocrinology, University Medical Center Groningen, Netherlands. Background and aims: We assessed the differences in health-related quality of life (HRQL) in obese participants with and without metabolic syndrome (MetS) and type 2 diabetes (T2D), by degree of obesity and level of inflamma- tion (high sensitive C-reactive protein (hsCRP)). HsCRP is related to Obesity, MetS and T2D. Materials and methods: Participants were derived from the random pop- ulation-based LifeLines study in the Netherlands. In total, 53,266 subjects (39,540 normal weight and 13,726 obese; BMI >30 kg/m2), aged 18-80 years, were included. Subjects with diabetes were those with known T2D or had a fasting blood glucose >7.0 mmol/L. HRQL was assessed with the Short Form- 36. Sex-corrected physical (PCS) and mental component score (MCS) were calculated. MetS was measured with the NCEP ATP III (2005). ANCOVA was used to compare the PCS and MCS score, with age as covariate. Clinically relevant differences in PCS and MCS were defined as a minimum difference score of respectively, 2 and 3 points between groups of patients. We defined a poor outcome on individual scales of the PCS and MCS as a score below 80 points. Results on individual scales were stratified by gender. Results: PCS and MCS: Compared to normal weight individuals, obese sub- jects had a significantly lower PCS (ranging from 3.1-7.5 points difference), pending on their MetS and T2D status, degree of obesity and level of inflam- mation. However, no clinically relevant changes were found for MCS. Obese subjects with MetS had a significant lower PCS (49.7 vs. 51.1) and MCS (50.4 vs. 51.2) than obese subjects without MetS, however these differences were small. Having T2D only affected the PCS (50.7 vs. 47.5; p S 425 1 C the original studies that we identified. Observational studies reporting risk estimates for type 2 diabetes were assessed. To ascertain the validity of the eligible studies, the quality of each report was appraised with reference to the STROBE statement. In addition, the Newcastle-Ottawa Scale for assessing the quality of prospective cohort studies in meta-analyses was used to quantify the validity of each study. The adjusted relative risk and 95% confidence in- terval were calculated with the random effect model. Dose-response relations were evaluated using data from different nap categories in each study. Two authors independently confirmed the eligibility of the studies and collated the data. Any discrepancies were resolved through discussion. Sex was employed as an independent category for comparison. Publication bias was evaluated by the funnel plot, Begg test, and Egger test. Results: Among 1,035 potential studies, 225,717 Asian and Western subjects stratified into 5 categories were identified. The analyses performed in each study were well adjusted for several confounders for diabetes. Pooled analysis revealed that a longer nap time (≥60 min/day) significantly increased the risk of type 2 diabetes (relative risk 1.46 (1.23-1.74, p S 426 1 C Better patient involvement and a reassessment of non-commercial funding strategies might help create more valuable research at least from patients’ point of view. 1035 Glycaemic perception and other factors influencing adherence to self- monitoring blood glucose in patients with type 1 or type 2 diabetes M. Jansà Morató1, M. Vidal Flor1, M. Galindo Rubio2, M. Penalba Martínez3, L. Corredoira González4; 1Servicio de Endocrinología y Nutrición, Unidad de Diabetes, Hospital Clínic i Provincial de Barcelona, 2Servicio de Endocrinología y Nutrición, Hospital Clínico de San Carlos de Madrid, 3Servicio de Endocrinología y Nutrición, Hospital Universitari i Politècnic La Fe de Valencia, 4Departamento de Estadística, BioClever 2005 S.L., Barcelona, Spain. Background and aims: Self-monitoring blood glucose is an important tool to achieve good glycemic control. However, many patients measure their glu- cose concentrations less often than recommended. This study was conducted to assess the factors associated with adherence to self-monitoring blood glu- cose in patients with diabetes mellitus treated with insulin. Materials and methods: In this observational, prospective and multicenter study in patients with type 1 or type 2 diabetes treated with insulin as per usual clinical practice, socio-demographic and clinical characteristics were collected at the baseline visit. Glycemic perception was assessed through 4 monthly telephone calls asking patients about their perceived glycemic levels and then comparing these perceptions with their real values from the glu- cometer. A patient was considered adherent if the minimum number of glu- cose measurements established by the Spanish Society of Diabetes was car- ried out. To assess factors associated with adherence to self-monitoring blood glucose, a multivariate logistic regression were performed. Results: A total of 2.700 patients were recruited, of whom 2.029 (75%) were considered valid for the analysis. With an age range of 18 to 80 years, the mean duration of the disease was 16.5±10.5 years and 30% of patients had type 1 diabetes. The majority of patients (81%) had an unsatisfactory glyce- mic control (baseline HbA1c ≥7%), which was significantly associated with an incorrect glycemic perception. Only 17% of patients had a correct glyce- mic perception. More than half of the patients (54%) were adherent to the self-monitoring according to the SED rules (95%CI: 52%, 56%), this being 51% in patients with type 1 diabetes. In patients with an age range of 18 to 44 years, the recommendation of >21 measurements/week (OR: 0.031; 95%CI: 0.015, 0.062) and alcohol consumption (OR: 0.452; 95%CI: 0.254, 0.805) were associated to low adherence. In patients with an age range of 45 to 64 years, an increase in the number of insulin injections/day was associated to a high- eradherence (OR: 1.296; 95%CI: 1.073, 1.566), while the recommendation of >21 measurements/week (OR: 0.070; 95%CI: 0.048, 0.104) and the belief that it was possible to know their glucose levels without the glucometer (OR: 0.925; 95%CI: 0.877, 0.976) were associated to low adherence. In patients over 65 years, the recommendation of >21 measurements/week (OR: 0.112; 95%CI: 0.076, 0.163), obtaining the test strips from the health centre vs from the pharmacy (OR: 0.554; 95%IC: 0.378, 0.812) and discouragement in cases of off-target HbA1c levels (OR: 0.488; 95%IC: 0.337, 0.706) were associated to low adherence. Conclusion: Adherence to self-monitoring blood glucose in patients with diabetes varies depending on age and on being recommended >21 measure- ments/week as a common factor associated with low adherence in all age groups. Other variables associated with poor adherence by age group are: al- cohol intake in young patients, confidence in glycemic perception in middle age and lack of motivation for poor control in older people. 1036 The relation between illness perception, diabetes distress and depression in type 2 diabetes patients: pilot study in a Romanian sample A.S. Mocan1,2, S.S. Iancu1, I.A. Vereşiu1, D.E. Dumitraş3, S. Puşcaşu2, C.M. Mureşanu2, A.G. Şerban2, L.A. Pop2, A.S. Băban2; 1Diabetes, Nutrition and Metabolic Diseases Department, Emergency Clinical County Hospital, 2Psychology and Educational Sciences Faculty, Babes -Bolyai University, 3Department of Economic Sciences, University of Agricultural Science and Veterinary Medicine, Cluj-Napoca, Romania. Background and aims: Depression has a major negative effect on diabetes outcome. Illness perception and diabetes distress may contribute to emotion- al and somatic depressive symptoms. Our aim was to investigate the relation- ship between illness perception, diabetes distress and depression. Materials and methods: 204 type 2 diabetes patients were included in the study. Revised Illness Perception Questionnaire (IPQ-R) and Diabetes Dis- tress Scale (DDS) were used to investigate illness burden. Beck Depression Inventory was used for depression. Linear regression analysis was used to evaluate the association between the five components of IPQ-R (illness identity, causes, timeline, consequences and cure-control), respectively four components of DDS (emotional burden, physician-related distress, regimen- related distress and diabetes-related interpersonal distress) and emotional and somatic symptoms of depression. Results: Mean age of the participants was 64.61 years (range 32-81 years) and the mean HbA1c 7.43% (SD=1.46%). Depressive symptoms were found in 43.1% of patients of them 22.1% had subclinical depression. In depressive group, HbA1c was associated with somatic depressive symptoms (r = .262, p < 0.05). Emotional diabetes distress was identified as a mediator between ill- ness perception identity and somatic depression: first and second mediation linear regression R2 = .174, p < .005, R2 = .239, p< .005, Sobel test z-score = 2.853, p= 0.0043. Emotional diabetes distress was also identified as a media- tor between perception of illness consequences and somatic depression: first and second mediation linear regression R2 =. 172, p < .005, R2 = .217, p S 427 1 C to non-depressed individuals. Depressed women after 18 months of interven- tion were less likely to have an improvement in their plasma glucose (OR: 0.32; 95%CI: 0.12; 0.87) and blood pressure (OR: 0.29; 95%CI: 0.10; 0.94) compared to non-depressed women. Depression had no effect on the impact of intervention on body adiposity and lipid profile. Inclusion of inflammatory markers into the models did not influence the association of depression and improvement in cardiometabolic profile. Comparing models including or not the interaction variable (depression*interdisciplinary intervention) no differ- ence in results was found. Conclusion: Depression at baseline of a lifestyle intervention may predict a lower chance of improving long-term cardiometabolic profile particular- ly among women. This raises the necessity of increase health professionals awareness and public health policy makers’ attention to depressed people at high cardiometabolic risk. We suggest a focus on screening and management of depression as part of the intervention programs for lifestyle changes in at-risk individuals. Clinical Trial Registration Number: RBR #65N292 Supported by: FAPESP 1038 The impact of hyperglycaemia and obesity on hospitalisation costs and clinical outcome in general surgery patients L. Buehler1, M. Fayfman1, A.-S. Alexopoulos1, L. Zhao2, F. Farrokhi1, J. Weaver3, D. Smiley1, F. Pasquel1, G. Umpierrez1; 1Medicine-Division of Endocrinology, Emory University, 2Biostatistics, Rollins School of Public Health, 3Information Technology, Emory University, Atlanta, USA. Background and aims: The association between inpatient hyperglycemia and poor outcome in surgical patients is well established. It is not known if obesity increases rates of complications in surgical patients with and without hyperglycemia and diabetes. Materials and methods: We analyzed the effects of obesity and hypergly- cemia on hospitalization costs and clinical outcome in patients, with and without diabetes (DM), who underwent gastrointestinal, hepato-biliary and pancreatic surgery. Inpatient hyperglycemia was defined as BG ≥7.8 mmol/L. Overweight was defined by body mass index (BMI) between 25-30 kg/m2 and obesity as a BMI >30 kg/m2. Hospital cost was calculated using cost-charge ratios from Centers for Medicare & Medicaid Services. Hospital complica- tions included a composite of major cardiovascular events, pneumonia, bac- teremia, acute kidney injury, respiratory failure, and death. Results: Among 2366 surgery patients, hyperglycemia was present in 1799 patients (76%) of whom 506 (28%) had history of DM and 1293 (72%) had no DM before surgery. Compared to patients with normoglycemia, those with DM and non-DM with hyperglycemia had higher number of hospital com- plications (5% vs. 21% vs. 20%, p S 428 1 C PS 087 Individual perception of diabetes 1040 Somatic symptoms in Chinese patients with type 2 diabetes: frequency, risk factors, and the relation to quality of life Y. Zhang1, A. Luk1, W. Yang2, W. Jia3, W. Li4, L. Ji5, X. Guo6, A. Kong1, R. Ozaki1, R. Yeung1, N. Sartorius7, Y.-K. Wing8, J. Weng9, J. Chan1, China Diabetes and Depression Study Group; 1Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Shatin, 2Department of Endocrinology, China-Japan Friendship Hospital, Beijing, 3Department of Endocrinology, Shanghai 6th Hospital, 4Department of Endocrinology, Peking Union Medical College Hospital, Beijing, 5Department of Endocrinology, Peking University People‘s Hospital, 6Department of Endocrinology, Peking University First Hospital, Beijing, China, 7Association for the Improvement of Mental Health Programmes, Geneva, Switzerland, 8Department of Psychiatry, The Chinese University of Hong Kong, 9Department of Endocrinology, The Third Affiliated Hospital, Sun Yat-sen University, Guangzhou, China. Background and aims: Somatic symptoms are common in patients with chronic diseases which may reflect underlying emotional distress. We in- vestigated the burden of somatic symptoms, risk factors and its associations with psychological health and quality of life in Chinese patients with type 2 diabetes. Materials and methods: 2401 patients aged 18-75 years attending hospital- based clinics in 4 cities in China underwent detailed clinical-psychological assessment during a 12-month period in 2011-2012. Somatic symptom burden was measured using the 28-item Somatic Symptom Inventory (SSI); emotional distress was measured using the 21-item Depression, Anxiety Stress Scale (DASS); quality of life was measured using the Euroqol-5D. Results: In this cross-sectional analysis, the mean (SD) age of the study co- hort was 56.4 (10.6) years (46.6% women). Somatic symptoms were highly prevalent with a mean SSI score of 41.4 (15.1), and 14 of the 28 symptoms were reported by more than 30% of patients. Overall, 71.5% of patients re- ported at least one pain symptom, and 51.9% reported at least one gastro- intestinal symptom. The most frequent reported symptom was trouble with vision (55.9%), followed by joint pain (44.9%), fatigue (44.8%), and muscles soreness (42.7%). In multivariate linear regression, after controlling for study sites, somatic symptom severity (adjusted R2=0.483; P S 429 1 C 1043 Perception of type 1 diabetes management by adolescents, parents and healthcare professionals: divergence of opinions influences metabolic and psychological outcomes S. Schmit1, J. Jacobs1, J.-P. Bourguignon2, M. Fichelle3, S. Jellimann4, O. Ziegler5, C. De Beaufort3; 1Centre for Health Studies, CRP-Sante, Luxembourg, 2Adolescent Health Centre, CHU de Liège, Belgium, 3Department of Paediatric Endocrinology and Diabetology, Centre Hospitalier de Luxembourg, Luxembourg, 4Clinical Infantile and Genetic Medecine, CHRU de Nancy - Hôpital d‘Enfants, 5Department of Diabetology, Metabolic Diseases and Nutrition, CHRU de Nancy - Hôpital Brabois Adultes, France. Background and aims: It is a major challenge for adolescents with type 1 diabetes to manage their disease, due to physical changes and psychosocial factors that characterise this period of transition. Although parents play an important role in supporting their children, responsibility is shifting. This may enhance conflicts between parents and adolescents, creating an impor- tant barrier to efficient metabolic control as well as to quality of life. Phy- sicians and nurses represent another party supporting and influencing the youth’s self-management. This study aims at comparing the perception of these different parties of several aspects of diabetes management and at iden- tifying the influence of the convergence or divergence of these opinions on metabolic outcomes and the youth’s well-being. Materials and methods: Adolescents between 13 and 17 years visiting the outpatient clinic in 3 regions, as well as their parent(s), paediatric diabetolo- gist and diabetes nurse filled out questionnaires containing questions on the adolescent’s metabolic targets, treatment, well-being, quality of life, self- management and perception of self-management. Clinical data concerning blood glucose measurements, insulin regimen, episodes of severe hypogly- caemia and diabetic ketoacidosis and Hb1Ac level were collected via a Clini- cal Record Form. Answers of the 4 parties were compared to see whether convergent answers are related to metabolic control, well-being and quality of life. The study was approved by the three IRB’s and Informed consent was obtained from all participants. Results: The sum of distances between answers of the 4 parties on HbA1c targets and on well-being was correlated to HbA1c outcomes (R = 0.448, p < 0.001; R = 0.232, p = 0.036). The sum of distances between opinions on HbA1c aims, on hyperglycaemia management and on fear from hypogly- caemia were inversely correlated to the youth’s self-management score (R = -0.378, p < 0.001; R = -0.349; p = 0.001; R = -0.295, p = 0.003). Conclusion: These results confirm the importance of harmony between the different protagonists of the youth’s type 1 diabetes management. The themes that seem to be most important are Hb1Ac targets, perception of the adoles- cent’s well-being and perceptions of hyper- and hypoglycaemia management. Communication between protagonists proves to be extremely important in adolescent diabetes management, since the difficulty of this management may be related to the level of harmony between adolescents, parents and healthcare teams. These findings are an important element to be taken into account in the conception of patient education programmes. Supported by: INTERREG IV A 1044 The impact of glycaemic variability on quality of life in adults with type 1 diabetes M. Reddy1, R. Agha-Jaffar1, P. Herrero2, P. Georgiou2, M. El-Sharkawy2, P. Pesl2, N. Jugnee1, H. Thomson1, C. Toumazou2, D. Johnston1, N. Oliver1; 1Division of Diabetes, Endocrinology and Metabolism, 2Centre for Bio-Inspired Technology, Institute of Biomedical Engineering, Imperial College London, UK. Background and aims: We aimed to assess whether glycaemic variability (GV) has an impact on quality of life (QOL) in adults with established type 1 diabetes mellitus (T1DM) on either multiple daily injections (MDI) of in- sulin or continuous subcutaneous insulin infusion (CSII). There is evidence suggesting that GV reduces QOL in people with type 2 diabetes mellitus (T2DM), but this has not been fully explored in T1DM. Materials and methods: Subjects wore a retrospective continuous glucose monitor (CGM) for 5-6 days and completed the diabetes quality of life ques- tionnaire (DQOL). GV measures (SD, CONGA, LI, JINDEX, LBGI, HBGI, GRADE, MODD, MAGE, ADRR, MVALUE, MAG) were calculated using the EasyGV version 9.0 software. A linear regression analysis was then per- formed to assess whether there was a correlation between GV and measures of QOL. Results: 57 adult subjects with T1DM (51% male, 65% on CSII, 35% on MDI , mean (SD) age 41 (13) years, duration of diabetes 21 (12) years, HbA1c 7.9 (1.1) %, body mass index 25.2 (4.0) kg/m2,) were included in the analysis. No significant associations between GV measures and DQOL scores (Table 1) were demonstrated. The GV was significantly higher for those subjects on MDI compared to the CSII group (p-value S 430 1 C WHOQOL-BREF. Among the remaining 293 patients (age [mean±SD]: 59.1±9.6 years, 52.2% male), 67% of patients had at least one negative emo- tion, and 22% had significant levels of all three negative emotions. 50% of patients had moderate/ severe depressive symptoms, 43% had moderate/ severe anxiety symptoms, and 22% had moderate/severe levels of stress. Compared with patients with no/mild depression, patients who had moder- ate/ severe depression were younger (57.7±9.4 vs 60.3±9.7 years, p=0.024), had shorter diabetes duration (10.8±7.6 vs 12.8±8.2 years, p=0.030), used less insulin (38.8% vs 50.3, p=0.049) and exercised less frequently (43.3% vs 56%, p=0.030). They also had worse health related quality of life (EQ-5D: median[IQR] 0.85[0.75,1] vs 1[0.8,1] , p=0.004), worse quality of life in all 4 domains of the WHOQOL-BREF: physical health (13.0±2.2 vs 15.3±2.1, p S 431 1 C Materials and methods: The Danish National Diabetes Register was linked with socioeconomic and population registers covering the entire population. Twelve year diabetes incidence was analysed using multivariate Cox regres- sion for 2,161,700 people aged 30-59 years at baseline. Results: 93.2 % were of Danish origin, 3.6 % of non-Danish Western origin and 3.3 % of Non-western origin. 12 year incidence of diabetes was, respec- tively, 5.8%, 7.9 % and 11.4 %. Occupational activity was stratified by five main socioeconomic categories which are derived from the International Standard for the Classification of Occupations: professionals, managers, tech- nicians, workers skilled at basic level and unskilled workers. Professionals were used as reference. The results of the survival analysis are shown in the table with 99 % confidence intervals. Conclusion: Diabetes incidence increases with age, male gender, occupations with low socioeconomic status and among people of Non-western origin. The results indicate that undifferentiated upward adjustment of the retirement age has the potential to cause a marked increase in employees with diabetes, especially among already vulnerable groups. The social and economic impli- cations are unknown and need to be subject of future research. 1049 Relative rates of disability retirement among employees in Denmark with and without diabetes: a prospective analysis with 10 year follow-up B. Cleal1, K. Poulsen1, H. Hannerz2, L. Andersen2; 1Steno Health Promotion Center, Steno Diabetes Center, 2National Research Center for the Working Environment, Copenhagen, Denmark. Background and aims: Work and the workplace have been allocated a piv- otal role in the future of health promotion, with economic and social stability predicated, in many accounts, on the assumption that more people will be fit, able and willing to work for longer. In this light, it is important to determine what socio-demographic trends might challenge this vision of future pros- perity. Our ambition in this study was to estimate the proportion of disability retirements, among employees with diabetes, which could be attributed to socioeconomic status. This perspective was considered important since both incident diabetes and work-disability are marked by a social gradient. Ob- taining more detailed and nuanced information on the relationship between diabetes and disability retirement is, therefore, undertaken in order to inform efforts to reduce disability retirement among people with diabetes. Materials and methods: Using four national population registries, includ- ing the Danish National Diabetes Registry, we identified all employees in Denmark aged 20-59 years old at baseline (January 2001 ). These individuals were stratified according to their presence/absence in the diabetes registry and their socioeconomic status, and subsequently followed in our national registries until 31 December 2010. Results: 29,210 people with diabetes and 1,950,962 people without diabetes were included in the analysis. The follow-up yielded 3,163 cases of disabil- ity retirement among people with diabetes and 64,262 cases among people without diabetes. With Professionals without diabetes as reference, the results indicate heightened levels of risk according to both presence of diabetes and socioeconomic status. The compound risk of being a person with diabetes and working in an occupation identified as having low socioeconomic status is clearly revealed in the 11.62 rate ratio for people with diabetes in elemen- tary occupations. Conclusion: The results show liability for early retirement among people with diabetes is especially acute for employees in lower socio-economic groups.The relative risks revealed in our study are especially concerning for the fact that there is evidence to suggest levels of social inequality in prevalent T2DM are on the increase. If current trends continue, therefore, there will, both proportionally and numerically, be more individuals with diabetes with a relatively high risk of disability retirement. Targeting primary, secondary and tertiary prevention to the groups that need it most thus seems an essen- tial strategy in attempts to prolong the working lives of individuals. Diabetologia (2014) 57:[Suppl1]S1–S564 S 432 1 C PS 088 Lifestyle and delivery of care 1050 Early evaluation of the ANAIS (Alimentación Normal Con Ajuste De Insulina) programme, a Spanish version of DAFNE R. Sánchez Hernández1,2, Y. López Plasencia1,2, D. Alvarado Martel1,2, J. Rodríguez Cordero1, A. Jiménez Ortega1, A. Carrillo Domínguez1, F. Nóvoa Mogollón1,2, A. Wägner1,2; 1Endocrinology and Nutrition, Hospital Insular de Gran Canaria, 2Instituto Universitario de Investigaciones Biomédicas y Sanitarias. Universidad de Las Palmas de Gran Canaria, Spain. Background and aims: The DAFNE (Dose adjustment for normal eating) programme promotes a flexible diet and has led to a reduction in clinical ad- missions and improved quality of life (QoL) in type 1 diabetes (T1D). ANAIS is a translation and adaptation of DAFNE to Spanish cultural and dietary patterns. Our aim was to evaluate the programme, using both clinical and patient-centred variables. Materials and methods: This randomized, controlled, parallel group trial in- cluded patients with >2 years of T1D, age >18 years and HbA1c 7-12%. Par- ticipants were randomized (2:1 ratio) to participate immediately in a 5-day outpatient course (immediate ANAIS) or to do so after one year (delayed ANAIS). The primary endpoint was glycaemic control (HbA1c), measured 3, 6 and 12 months after the programme. We also evaluated the number of hy- poglycemic episodes, weight, insulin dose, diabetes-related QoL (EsDQOL), treatment satisfaction (DTSQ), fear of hypoglycemia (MH-15), anxiety (STAI), depression (BDI), diet flexibility (open question) and achievement of self-defined goals (scale from -2 (much less than expected) to +2 (much more than expected)). Comparisons were made (Student‘s t, Wilcoxon‘s test and chi-square). Results: So far, 80 patients (48 immediate ANAIS) have been included, 45% men, aged 34.6 (SD 11.35) years, diabetes duration 12.5 [range 3-47] years, with a BMI: 24.9 (3.9) kg/m2 and a daily insulin dose of 0.75 (0.24) UI/Kg. In the intervention group, there was a 0.2-0.3% non-significant reduction in HbA1c at 3,6 and 12, less hypoglycaemia at 6 months (2 [0-16] vs 1 [0-8]/ week, p=0.045), and a trend towards a reduction in long acting insulin dose at 12 months (31.08 UI (12.5) vs 28.92 UI (12.6), p=0.08). No changes in weight were seen. Quality of life improved at 3 (96.7+/-19.9 vs 82.4+/-22.4 points, p S 433 1 C evaluate the sources of variability such as food-insulin balance and the tim- ing of meals across days. Decreasing variability or IQR may reduce the risk of hypo or hyperglycemia. Conclusion: Analysis of AGP prandial periods can reveal the glycemic effect of food and insulin. AGPs can refine nutrition and insulin therapy relative to type and amount of carb, amount of dietary fat and protein, timing and amount of prandial insulin, and meal times. Clinical Trial Registration Number: NCT01713348 Supported by: Abbott Diabetes Care 1053 Perioperative diabetes regulation and Protocol Compliance: the PC study J. Hermanides1, J.A.W. Polderman1, F.A. de Groot1, F. Holleman2, B. Preckel1; 1Anesthesiology, 2Internal Medicine, Academic Medical Centre, Amsterdam, Netherlands. Background and aims: Peri-operative diabetes regulation differs between hospitals. Compliance with regard to peri-operative glucose regulation pro- tocols is notoriously low and evidence with regard to efficacy is lacking. There are, however, convincing studies demonstrating that improving post- operative diabetes control decreases complications. We therefore investigated whether implementing a modified and stricter peri-operative diabetes proto- col improved compliance and peri-operative glucose regulation. Materials and methods: We performed a prospective observational cohort study from March 2013 through July 2013, comparing compliance and glu- cose regulation following protocol A (conventional protocol) with protocol B (strict protocol). Protocol A prescribed a Glucose-Insulin-Potassium infu- sion, with 8 IU of insulin. Protocol B prescribed a Glucose-Insulin infusion, with 1/8 IU of the normal daily insulin dose. A pre-defined treatment algo- rithm to treat hyperglycaemia, glucose > 10 mmol/l, was added. In May 2013, we switched from protocol A to protocol B. Analyses were performed with the Mann Whitney U test and multivariate regression analyses. Results: In total, 192 patients were included in protocol A and 183 in proto- col B, mean age was 64 years and 54% was female, 10,4% of patients (n=20) had diabetes type 1, 86,5% (n=166) type 2, 3.1% (n=6) other. There were no differences between the patient characteristics in protocol A and B. The me- dian compliance (IQR) was 67% (57-83) for Protocol A and 71% (57-83) for Protocol B, p=0.04. Differences between peri-operative glucose values are shown in graph 1. Adjusted for age, gender, ASA status, DM type, emergency procedures, fasting glucose and compliance, protocol B (red line) improved the proportion of patients that reached the target of < 10 mmol/l (OR 0.36, 95% CI 0.19-0.70, p< 0.01) before discharge to the ward. Conclusion: Implementing a stricter peri-operative diabetes protocol only slightly improved compliance, but significantly and relevantly improved post- operative glucose control before discharge from the recovery. 1054 The burden of “serial non-adherence” in type 2 diabetic patients C. Frois1, K. Dea1, D. Ling1, J. Dunn2, M. Baron3; 1Analysis Group, Inc., Boston, 2VRx Pharmacy Services, Salt Lake City, 3Intarcia Therapeutics, Inc., Boston, USA. Background and aims: Although medication adherence is recognized as a key concern for the treatment of Type 2 diabetes patients, there has been limited assessment of the scope of “serial non-adherence,” i.e. non-adherence to current diabetes regimen following non-adherence to prior therapies (diabetes or non-diabetes related), its predictability, and impact on patient/ healthcare burden. The aim of this research was to characterize the health- care burden associated with non-adherence with second-line anti-diabetic medication and to identify its key predictors, including the role of serial non- adherence. Materials and methods: A retrospective cohort study was conducted using the Truven Health MarketScan® US Commercial health insurance database for the period 2008-2012. Patients were identified as first-line metformin users by requiring no prior anti-diabetic drug history going back at least 12-months before the date of the first metformin prescription. Patients in this group who subsequently initiated second-line (post-metformin) thera- py were identified and selected for study. Diabetic regimen adherence was measured using the proportion of days covered (PDC), calculated as the total number of non-overlapping days of supply of any antidiabetic medication one year following the initiation of second-line therapy divided by 365 days. Adherence for selected non-diabetic medications, including lipid-lowering drugs and anti-hypertensives, was measured in a similar way during the base- line period. Non-adherence was defined as PDC S 434 1 C 1055 Do type 2 diabetes patients achieve therapeutic goals with medication adherence? A.B. Garcia, E. Lopez Gonzalez, S. Milrad, A. Luongo, S. Houssay, M.L. Ruiz, FRADYC II Group; Hypertension Committee, Argentinean Diabetes Society, Buenos Aires, Argentina. Background and aims: In developed countries, adherence to drug treatment in chronic diseases hovers around 50 %. Improving adherence is key to effec- tively address the treatment of people with diabetes. Objectives: To determine the percentage of medication adherence and its relationship to the achieve- ment of A1c, blood pressure and lipids goals in people with T2DM. To assess which factors are associated with lower medication adherence. Materials and methods: A study was conducted in 31 medical centers spe- cialized in diabetes in Argentina (12/ 2010- 05/2011), with a clinical and laboratory evaluation. The Morisky-Green test was used to assess adherence and WHO -5 for well-being and depression. Were taken as therapeutic targets A1c S 435 1 C PS 089 Met and unmet needs in diabetes care 1058 Does clinical inertia vary by personalised HbA1c goal? A study of predictors and prevalence of clinical inertia in a US managed care setting J. Lin1, S. Zhou2, W. Wei2, C. Pan3, M. Lingohr-Smith1, P. Levin4; 1Novosys Health, Flemington, 2Sanofi US, Inc., Bridgewater, 3Pro-Unlimited, Inc., Boca Raton, 4Model Clinical Research, Baltimore, USA. Background and aims: Clinical inertia (CI) in type 2 diabetes mellitus (T2DM) care is defined as a failure to intensify treatment in patients with inadequate glycaemic control. There is a gap in understanding patterns of CI, particularly when considering the recent move towards personalized gly- caemic targets rather than the standardized HbA1c goal of < 7.0%. This study applied 3 different definitions of CI to a T2DM cohort from a large US man- aged care database to investigate the prevalence and predictors of CI. The 3 definitions of CI were based on different HbA1c goals: (1) HbA1c < 7.0%; (2) a modified target defined by Ismail-Beigi; and (3) Healthcare Effectiveness Data and Information Set (HEDIS) HbA1c targets. Materials and methods: Eligible adult T2DM patients were identified from the Optum Impact National Managed Care Database™ (IHCIS) between January 1, 2008-December 31, 2012. The index date was defined as the date of the first above target HbA1c (the trigger HbA1c) during the study period. Patients were required to have continuous health plan coverage for 6 months prior to index (baseline period) and for 6 months after index (response pe- riod). CI was defined as patients not receiving treatment intensification, de- fined as adding another oral antidiabetes drug (OAD), or initiating insulin or a glucagon-like peptide-1 receptor agonist within the response period. Predictors of CI were identified from the baseline period and evaluated by multiple logistic regression. Results: The prevalence of CI was the highest for Definition 1, although the differences between the definitions were not substantial; 72.8%, 70.4%, and 70.6% for Definitions 1, 2, and 3, respectively. Common predictors of CI in- cluded higher age, higher baseline OAD usage, and a more recent index year, indicating a higher risk of CI over time (Table). Patients at a lower risk of CI included those with certain comorbidities, increased baseline non-diabetes medication use, an endocrinologist visit, and a higher index HbA1c value. Differences between the predictors were also observed between the 3 defini- tions. Patients with baseline neuropathy, retinopathy, and a Charlson Comor- bidity Index of 2-3 were less likely to experience CI according to Definition 1 only. Baseline myocardial infarction was identified as a predictor of CI for Definition 1 only. Conclusion: These data suggest that the prevalence of CI in a US managed care setting was high and increased over the study period. The prevalence and predictors of CI were generally similar across all 3 study populations with different HbA1c targets; however, there were some significant differences, particularly in the comorbidity-related predictors identified for Definition 1. This study provides some insights into clinical patterns of CI that could aid targeting of relevant T2DM subpopulations for improvement. Supported by: Sanofi U.S. 1059 Association of metabolic control and chronic complications with self-care behavior in Chinese type 2 diabetic patients: a nationwide, multi-centre, cross-sectional survey Y. Yang1, Z. Sun1, X. Guo2, L. Yuan3; 1Department of Endocrinology, Medical School, Zhongda Hospital, Institute of Diabetes, Southeast University, Nanjing, 2Department of Endocrinology, Peking University First Hospital, Beijing, 3Department of Endocrinology, Medical School, West China Hospital, Sichuan University, Chengdu, China. Background and aims: The purpose of this survey was to study the associa- tion of metabolic control and complications, with attitudes towards diabetes and self-care behavior in type 2 diabetic patients in China. Materials and methods: Adults with type 2 diabetes from 50 medical cent- ers across China covered 29 administrative divisions were eligible. The third version of the Diabetes Attitude Scale (DAS-3) and summary of Diabetes Self-Care Activities (SDSCA) were utilized to assess attitude towards diabe- tes and self-care behavior, respectively. Adequate metabolic control (hemo- globin A1c S 436 1 C Conclusion: These data demonstrated that metabolic control and diabetic complications was directly associated with self-care behavior but not attitude. It suggests that the diabetes educators in China should pay more efforts on diabetes skill training towards self-care behavior for patients. 1060 Identifying consistent inconsistency in network meta-analyses: an illustration in type 2 diabetes D.A. Scott, N. Hawkins; Health Economics & Epidemiology, ICON Clinical Research UK Ltd, Oxford, UK. Background and aims: Network meta-analyses (NMA) combine direct and indirect evidence across a connected network of trial comparisons to provide estimates of comparative efficacy for multiple treatments. They have been em- braced by reimbursement agencies and recommended by health technology assessment bodies. A key concern is the comparability of treatment effect es- timates from different trials. Where there is both indirect and direct evidence for one or more comparisons (‘loops’ in the network) it is possible to evaluate empirically the ‘consistency’ of the network. Inconsistencies between direct and indirect estimates may be alleviated by adjusting for trial differences us- ing regression analysis or by excluding studies. However, as the networks of trial evidence become more complex, and potentially include multiple ‘loops’ it becomes more difficult to interpret evidence regarding ‘inconsistency’. Materials and methods: A variety of alternative methods have been pro- posed to examine inconsistency including (i) node-splitting whereby the difference between the direct and indirect evidence is calculated (where pos- sible) for each comparison in the network, (ii) comparisons of the NMA to an ‘inconsistency’ model where the effect estimates for each treatment com- parison in the network is allowed to be independent, (iii) a method which looks at inconsistencies between alternative study designs, (iv) plotting re- sidual deviance for individual trial arms within the NMA, and (v) plotting mixed predictive p-values against a uniform distribution. We compare the implementation and interpretation of these methods using a previously pub- lished NMA in type 2 diabetes. In this analysis HbA1c was compared across six treatments (two GLP-1s at different administrations, placebo, and insulin glargine) in a network of 22 studies with multiple ‘loops’. As in the original study, a random effect model controlling for baseline HbA1c was performed. Results: The methods agreed in showing the presence of inconsistency with the network. For example, the inconsistency model showed an improved fit (DIC -62.35) compared to the consistency model (DIC -60.25). The node splitting method identified two treatment arcs as being inconsistent, liraglu- tide 1.8mg vs placebo and liraglutide 1.8mg vs exenatide QW. However, the methods varied in their ability to provide an overall ‘test’ of inconsistency across the network and their ability to identify which parts of the network contain inconsistencies. We highlight that none of the methods alone can identify individual studies as being the cause of inconsistencies and argue that we need to consider the whole structure of the network and the charac- teristics of the studies (in terms of treatments, subjects, and design) within the network. Conclusion: Meta-analysis and decision makers often seek similarity across the network. However, we argue that we may be more confident generalising from treatment effects estimated from a well-connected and heterogeneous network if the estimates within the network prove to be consistent. 1061 A patient centered approach on newly-arrived persons to a diabetes clinic A.C. Paiva, M.J. Afonso, R.T. Ribeiro, L. Serrabulho, J. Susano, J.F. Raposo; APDP - Diabetes Portugal (Education and Research Centre - APDP/ERC), Lisbon, Portugal. Background and aims: It‘s recognized that person-centred therapeutic edu- cation and group education with active methodologies promote experiences sharing, conviviality and stimulates learning among participants. The clinic introduced two structured programs for self-management education (DSME) addressed to newly-arrived patients covering a wide range of Diabetes Care in an integrated way. It‘s a comprehensive program of patient-centred care designed to increase autonomy, promote better adherence to treatment, and thus better metabolic control. We aim to perceive the programme’s practical feasibility and people‘s adherence. Materials and methods: Programme 1: lasts for 04:30h;-During this period the person performs several tests: blood samples, EKG and retinography. A nurse performs a foot screening with risk assessment and foot care education. Then the person participates in two group sessions: a session with nurse guid- ance addressing pathophysiology of diabetes, relating the various important aspects in treatment and self-control, as well as doubts clarification and a final session about healthy eating, its importance for the metabolic control and lipid profile, and role of exercise in controlling diabetes (guided by a dieti- cian/nutritionist). The average time between the program and the 1st medical consultation at the clinic is 4 weeks. Programme 2: lasts for 3 months, divided in three group education sessions before the diabetes individual medical con- sultation (med). Sessions are guided by a facilitator using an IDF approved education tool, which provides an interactive verbal and visual learning expe- rience, allowing groups engagement in an open and meaningful debate about diabetes. Sessions are divided by themes: the 1st (S1) leads to a reflection on their role in disease‘s self-management, the 2nd (S2)covers general concepts for healthy eating, and the 3rd (S3) is a physical activity session with a gym teacher. Patients are selected based on their age (50-80 y) and HbA1c(< 10%), they are invited to attend programme 1 or 2 according to their convenience. Results: Programme 1: A sample of 300 people (February to September 2013) with 60.4±10.3 years of age, an initial mean HbA1c of 8.7±1.5% and BMI mean of 28,2±4,6kg/m2 were analysed. No consistent changes were observed in terms of BMI or HbA1c in this group between the session and the medi- cal consultation. Programme 2: We analysed a sample of 231 people (same period of time), with 68.3±8.8 years of age and an initial mean HbA1c of 9.1±0.7% and BMI mean of 35,3 ± 3,2 kg/m². The drop-out rate was 10,8% at session 2 and 82% at session 3. No consistent changes were observed in BMI between the various groups. In terms of HbA1c it was observed a tendency of decrease between S1 and medical consultation directly related to the number of sessions attended (∆A1c - Session 1: -0,27%; Session 2: -0,59 %; Session3: -0,85 %) Diabetologia (2014) 57:[Suppl1]S1–S564 S 437 1 C Conclusion: It‘s well known that active methods are a fundamental tool in group training. Here, the consistent decrease in HbA1c (programme 2) achieved independently of weight loss, hints to the impact of sharing solu- tions among peers by boosting diabetes acceptance, well-being and develop- ment of autonomy with DSME. The longer duration of this program also ena- bles a slower integration of knowledge and skills in the daily life. However, the high drop-out before the exercise session advises us to consider alterations on program implementation, further encouraging patients participation. 1062 Relation between quality of care indicators of diabetes and prediction of hospitalisation and mortality for heart failure L. Policardo, G. Seghieri, P. Francesconi; Agenzia Regionale Sanità, Florence, Italy. Background and aims: Heart failure (HF) is a major complication of diabetes especially as a consequence of ischemic heart disease. With the aim of evalu- ating whether a panel of quality of care indicators, as suggested by current guidelines, are related to the risk of hospitalization or subsequent mortality for HF, this study has retrospectively followed up, along a period of eight years (from 2005 to 2012) a cohort of diabetic patients living in Tuscany, a region of centre of Italy. Materials and methods: The database used for this investigation was ob- tained from linking three datasets: the first concerning all hospital discharges with main diagnosis of heart failure (ICD-9-CM 402 to 428) from all Tuscan hospitals over the period 2005 to 2012, the second was the general population registry of all inhabitants of Tuscany and the third a dataset containing the registry of all known diabetic patients from Tuscany. This latter gave infor- mation about whether patients did perform annual assessment of HbA1c, eye examination, serum lipids, creatinine and microalbuminuria. In addition, from this dataset it was possible to evaluate whether patients were given at least 2 annual drug prescription of ACE-inhibitors (ACEI), aspirin (ASA) or lipid-lowering drugs (LIPID). Results: On a total of 95,205 diabetic patients (47,762 males and 47,443 fe- males), followed up over eight years, we counted 4,494 hospitalizations for HF (2,131 in males and 2,363 in females). After using a Cox proportional hazard model, the hazard ratio (HR) of hospitalization for HF, adjusted for age and Charlson co-morbidity index was inversely related with the annul execution of HbA1c, microalbuminuria or of lipid profile both in men (HR:0.923;0.896- 0.952(95%CI), 0.972;0.949-0.995 and 0.895;0.870-0.920), and in women (HR.0.908;0.883-0.934, 0.968;0.945-0.991 and 0.900;0.878-0.923). The an- nual evaluation of serum creatinine, was on the contrary a positive predic- tor of risk (HR:1.306;1.260-1.353 in males and 1.359;1.315-1.405 in females). Annual eye examination was associated with a decrease in the risk of hospi- talization for HF only among females (HR:0.957;0.937-0.977 among women and HR:0.986;0.964-1.008 among men). Finally, the composite indicator including the prescription for at least five years of ASA, ACEI and LIPID was significantly related with an augmented risk of HF-related hospitaliza- tion in both males (HR:1.285;1.130-1.456) and females (HR: 1.192;1.028- 1.375). After HF hospitalizations 997 all-cause deaths were recorded and, by using a Cox model adjusting for age, sex and Charlson index, the survival after HF-hospitalization was significantly predicted by a low Charlson index (HR:0.928;0.864-0.996) and, again, by a lower number of annual evaluations of serum creatinine (0.935;0.886-0.985). No other indicator was significantly correlated with the prediction of post-hosptalization fatal events Conclusion: In this cohort of Tuscan diabetic patients, over a eight-year pe- riod the annual execution of HbA1c, microalbuminuria, lipid profile and eye examination were inversely related with the risk of hospitalization for HF. Prescription of ACEI, ASA and lipid-lowering drugs was, on the contrary, associated with a significant increase in predicted hospitalization risk. Finally, post-hospitalization mortality risk was predicted by a higher Charlson index and by a higher rate of annual serum creatinine evaluations. 1063 Quality of care in type 1 diabetes in Italy: focus on gender differences V. Manicardi1, G.T. Russo2, E. Torlone3, M. Calabrese4, M.R. Cristofaro5, M.R. Improta6, P. Li Volsi7, A. Maffettone8, M.F. Mulas9, A. Napoli10, C. Suraci11, A. Ceriello12, G. Lucisano13, A. Nicolucci13, M.C. Rossi13; 1Dept. of Int. Medicine, E.Franchini Hospital, Montecchio, Reggio Emilia, 2University of Messina, 3University of Perugia, 4Hospital of Prato, 5Cardarelli Hospital, Campobasso, 6University of Naples, 7Hospital of Sacile (PN), Sacile (PN), 8Hospital Monaldi, Naples, 9Hospital Brotzu, 10Sapienza University, Roma, 11S.Pertini Hospital, Roma, Italy, 12Idibaps, Barcellona, Spain, 13Mario Negri Sud, Santa Maria Imbaro (CH), Italy. Background and aims: We evaluated the quality of care according to gen- der in type 1 diabetes (T1D) in 300 diabetes outpatient clinics participating to “The Annali AMD” initiative. The Italian Association of Diabetologists (AMD) identified a set of quality indicators used for benchmarking activities since 2006, including process and intermediate outcome measures, as well as indicators of treatment intensity/appropriateness. A quality of care summary score (Q score), based on a combination of process and outcome indicators (range 0-40), was also calculated. Materials and methods: We report here clinical data of 28,802 T1D patients (54.5% men; 45.5% women), collected during 2011 and extracted from elec- tronic medical records. Results: Both men and women received similar evaluation for process in- dicators (HbA1c, Lipids, BP, microalbuminuria, foot and eye examination), whereas the outcome indicators and treatment intensity/appropriateness dif- fered according to gender (Table). No gender differences as to age (F: 45±16 years; M: 45±17) and diabetes duration (F: 19±13 years; M: 18±13) were noted. Men were more often smokers than women (31.8% vs. 22.7%) and showed a higher BMI (25±4 vs. 24±4 kg/m2), but F showed higher prevalence of BMI lt 18 or gt 35. T1D women showed poorer metabolic control and more often did not receive lipid-lowering agents in spite of high LDL-cholesterol levels than men, while men had poorer blood pressure control, and higher percentage of micro/macroalbuminuria. However, the overall quality of care, as estimated by Q score, was similar in both genders. Regarding the therapeu- tic approach (MDI vs CSII), the percentage of women treated with CSII was higher than men (19,6% vs. 13.8%); women were slightly younger (42.2±16.9 vs 45.1±16.9) for the same duration of disease (18.3±12.5 vs 18.1±13.2). A greater proportion of subjects treated with CSII reached the HbA1c level be- low 7.0% but women experienced more difficulties in achieving the HbA1c goal even if treated with CSII (HbA1c < 7.0%: 25.1% F vs 31.2 % M ). Conclusion: Despite a similar quality of care, T1D women still show a poor- er metabolic control, with any type of treatment, and were less intensively treated for LDL-C, while men show a worse cardiovascular profile in terms of blood pressure, microalbuminuria, BMI and smoking habits. Diabetologia (2014) 57:[Suppl1]S1–S564 S 438 1 C 1064 A global study of unmet need for glycaemic control and predictor factors among patients with type 2 diabetes mellitus D. Raccah1, E. Chou2, S. Colagiuri3, Z. Gaàl4, F. Lavalle5, A. Mkrtumyan6, E. Nikonova7, N. Tentolouris8, J. Vidal9, M. Davies10; 1Department of Diabetology, University Hospital Sainte-Marguerite, Marseille, France, 2Sanofi R&D, Bridgewater, USA, 3Boden Institute, University of Sydney, 4Josa Andras Teaching Hospital Department of Medicine, Nyíregyháza, Hungary, 5Hospital Universitario, Facultad de Medicina UANL, Monterrey, Mexico, 6Moscow State Medical and Stomatological University named after Evdokimov, Moscow, Russian Federation, 7Global Medical Affairs, Diabetes Division, Sanofi, Bridgewater, USA, 8University of Athens Medical School, Greece, 9Endocrinology and Nutrition Department, Hospital Clinic, Barcelona, Spain, 10Diabetes Research Centre, University of Leicester, UK. Background and aims: Information on the efficacy of various treatments for type 2 diabetes mellitus (T2DM) in achieving glycaemic control is available globally from a number of different sources. This study utilized data from randomized controlled trials (RCTs), clinical trial registries (CTRs) and elec- tronic medical records (EMRs) to identify unmet need for improved prandial glycaemic control in patients with T2DM following initiation of basal insulin therapy in Europe (EU), Asia Pacific (APAC), the US, and Latin America (LATAM). Materials and methods: Different levels of evidence were used as available for each country/region, including RCTs (EU, LATAM, and APAC), CTRs (CREDIT 4-y, 9 countries mainly EU; ALOHA 0.5-y, Japan; Asia-FINE, 11 Asian countries), and EMRs (Germany IMS-DA, UK THIN, and US GE). We evaluated hyperglycaemia status by categorizing as ‘well controlled’ (defined as endpoint fasting plasma glucose [FPG] at target [defined as FPG S 439 1 C PS 090 Individualised care 1066 Trends over 8 years in quality of care provided by Italian diabetes clinics to elderly patients with type 2 diabetes R. Candido1, M.C. Rossi2, M.A. Pellegrini3, G. Felace4, M. Boemi5, M. Scardapane2, A. Nicolucci2, P. Di Bartolo6, C. Giorda7; 1Diabetic Centre, A.S.S. 1 Triestina, 2Department of Clinical Pharmacology and Epidemiology, Consorzio Mario Negri Sud, S. Maria Imbaro (CH), 3Diabetes and Metabolism Unit, AOU Santa Maria della Misericordia, Udine, 4Diabetes and Metabolism Unit, O.C. di Spilimbergo, Spilimbergo (UD), 5Diabetes Unit, Department of Medicine, INRCA-IRCCS, Ancona, 6Diabetes Unit, A.Usl della Romagna, Ravenna, 7Diabetes and Metabolism Unit, ASL TO5, Chieri (TO), Italy. Background and aims: The prevalence of diabetes in people over 75 years approaches 30%, while at least 25% of all patients attending Italian outpa- tients diabetes clinics are older than 75 years. Clinicians who manage older people with diabetes require special skills to provide high-quality care. In the context of a continuous quality improvement initiative promoted in Italy by Associazione Medici Diabetologi (AMD Annals initiative), we evaluated the trends over 8 years in the quality of care provided by diabetes clinics to elderly patients (i.e. > 75 years). Materials and methods: Overall, 300 diabetes clinics (about half of all clinics in Italy), all using electronic medical record systems, extracted data relative to the years 2004-2011 . The proportion of patients with at least one value registered during each year (process measures), the percentage of patients reaching specific favorable or unfavorable targets (intermediate outcome measures), and rates of use of drugs were evaluated. In addition, a quality of care summary score (Q score) was calculated. The Q score, ranging between 0 and 40, is based on process and outcome indicators (HbA1c, blood pres- sure, LDL-cholesterol, microalbuminuria) and is closely related to long-term outcomes in diabetic patients. Results: Over the years, there was an increase in the percentage of patients aged > 75 years (19.9% in 2004 vs. 27.2% in 2011) and in the prevalence of male patients (42% in 2004 vs. 46% in 2011). As compared to 2004 we ob- served after 8 years a slight increase in the mean age (79.9±3.8 vs. 80.5±4.1) and in the duration of diabetes (13.8±10.5 vs. 14.2±10.9). Table 1 shows qual- ity of care indicators. Conclusion: Care provided by diabetes clinics to elderly patients shows a significant improvement over the years. Elderly patients are more frequently monitored for blood pressure, lipid profile and microalbuminuria, show bet- ter intermediate outcomes and are less often treated with sulphonylureas. In addition we observe a better global quality of care with a significant increase in the Q score through the years. 1067 Comparing the use of patient level data to an average patient profile within a type 2 diabetes simulation model P. McEwan1,2, H. Bennett2, T. Ward2, K. Bergenheim3; 1Swansea Centre for Health Economics, Cardiff, 2Health Economics and Outcomes Research Ltd, Monmouth, UK, 3Global Health Economics and Outcomes Research, AstraZeneca, Molndal, Sweden. Background and aims: To accommodate patient heterogeneity and complex treatment pathways, type 2 diabetes models employ simulation techniques beyond the basic Markov process. Yet averages are commonly relied upon when defining patient populations and treatment effects; in turn clinicians may struggle to relate the results of such modelling to the full spectrum of patients they see in practice. This study investigates the value of fully model- ling between patient variation by comparing patient and cohort level model inputs within a published simulation model, based on the UKPDS68 out- comes equations. Materials and methods: Anonymised UK patient data was obtained from The Health Improvement Network (THIN) describing 2,251 patients initi- ating dual therapy. Simulations were initialised by applying either the aver- age patient profile to the cohort, or by individually replicating each patient’s profile, followed by the collation of model outputs over all replications. The impact of utilising patient level baseline and treatment effect data was com- pared to the average patient profile through evaluation of total costs, benefits and complication rates, predicted over a medium-term horizon of 20 years. Results: On average patients were aged 63.36 (±11.14) at baseline, with the following risk factor profile; HbA1c 8.39 (±1.23) %, total cholesterol 4.18 (±0.92) mmol/L, systolic blood pressure 135.07 (±14.76) mmHg and weight 89.85 (±19.01) kg. The mean treatment effect on HbA1c was a reduction of 1.01 (±1.23) %. Over 20 years, fewer macrovascular and microvascular events (-82/1,000 patients) and higher all-cause mortality (+17/1,000 patients) were predicted when using patient level data inputs compared to the average pa- tient profile. Differences in the simulated frequency and timing of deaths were driven primarily by variation in baseline age and led to fewer estimat- ed life-years (-0.66), quality adjusted life-years (QALYs) (-0.59) and costs (-£551) per patient. Patients estimated to have both lower costs and higher QALYs than those associated with the average patient profile were younger, with higher HbA1c and cholesterol and lower blood pressure at baseline. Conclusion: Modelling results differ depending on the use of patient level or average cohort model inputs. Patient level data may provide insight into the type of patients in whom therapy is likely to be most beneficial. Furthermore, it enables the accurate simulation of correlations between patient characteris- tics and treatment effect, which are rarely accounted for as part of a standard probabilistic sensitivity analysis. Supported by: AstraZeneca, BMS 1068 Is there an evidence base for the clinical features used to differentiate type 1 from type 2 diabetes? A systematic review of the literature B.M. Shields1, J.L. Peters2, C. Cooper2, J. Lowe2, B. Knight1, R. Powell3, C.J. Hyde2, A.T. Hattersley1; 1NIHR Exeter Clinical Research Facility, University of Exeter Medical School, 2Peninsula Technology Assessment Group, University of Exeter Medical School, 3Research Design Service, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK. Background and aims: Clinicians predominantly use clinical features to dif- ferentiate Type 1 from Type 2 diabetes but no guidelines quote an evidence base for which features are the most discriminatory. Whilst Type 1 patients are thinner, diagnosed younger, and rapidly go onto insulin, the best cut offs of these criteria are not provided. Misclassification of diabetes is widespread (7-15% of cases) and will result in patients receiving the wrong treatment. The key difference between Type 1 and Type 2, that determines treatment, is that patients with Type 1 diabetes develop absolute insulin deficiency. We aimed to systematically review the literature to identify which clinical criteria could be used to discriminate Type 1 and Type 2 diabetes, using insulin defi- ciency as a gold standard. Materials and methods: The search strategy took the form of: (terms for diabetes) AND (terms for C-Peptide). 14 databases including MEDLINE and EMBASE were searched. All diagnostic accuracy studies, published since 1979, using clinical criteria to predict insulin deficiency (defined by C- peptide concentrations) were included. There was no restriction on race, age, language, or country of origin. Data synthesis was largely descriptive. Results: 10,917 abstracts were screened, and 231 full texts reviewed. 10 ref- erences were identified for final inclusion. Studies varied by age, race, year, and proportion of participants who were C-peptide negative, prohibiting for- mal meta-analysis. Consistent predictors across studies were age at diagnosis (the most discriminatory feature; mean ranking of predictors by discrimi- natory ability within studies=1.4), insulin treatment/time to insulin (mean rank=1.7), and BMI (mean rank=2.6). Discriminatory cutoffs were age at diagnosis 30-40 years, time to insulin 1-2 years, and BMI 27-28kg/m2 (>66% sensitivity and specificity for all). BMI added little over age at diagnosis and/ or time to insulin ( S 440 1 C Conclusion: Studies of the accuracy of clinical criteria to define Type 1 dia- betes are surprisingly few given the importance for treatment of patients. De- spite finding only 10 studies, and considerable heterogeneity between studies, age at diagnosis (35 years) and time to insulin treatment(1.5 years) were con- sistently the most discriminatory. BMI, despite being widely used in clinical practice, adds little to these 2 criteria. Supported by: NIHR 1069 Admission albumin levels are associated with increased risk for hypoglycaemia during hospitalisation as well as poor 1-year survival, among patients with diabetes mellitus E. Leibovitz1,2, M. Boaz3,2, J. Wainstein4; 1Interal Medicine E, Wolfson Medical Center, Holon, 2Sackler school of Medicine, Tel Aviv University, 3Research and Epidemiology, Wolfson Medical Center, Holon, 4Diabetes clinic, Wolfson Medical Center, Holon, Israel. Background and aims: Among non-critically ill patients, there are detrimen- tal effects of hypoglycemia during hospital stay. Hospitalization-associated hypoglycemic events may occur in patients with diabetes mellitus that are prone to develop it. We studied the association between admission hypoal- buminemia and the risk of hospitalization-associated hypoglycemia, among patients admitted to internal medicine departments. Materials and methods: In this retrospective analysis of electronic medical records, we included all 2599 patients with diabetes mellitus and documented admission albumin levels (Mean age 71.7±13.0 years, 48.4% males, 65.9% of all patients with diabetes), admitted to internal medicine departments dur- ing 2009. All glucose measurements were computerized using an institutional glucometer. Patients were categorized into 4 groups according to hospitali- zation-associated hypoglycemia (yes/no) and admission alumin levels below 3.5 g/dL (yes/no). Results: Patients with hypo-albuminemia had higher rates of hypoglycemia and severe hypoglycemia during the admission despite similar HbA1c and average glucose control during the admission. Only creatinine (OR 1.237, 95% CI 1.098-1.392, p S 441 1 C 1071 Overtreatment in elderly patients ≥75 years with type 2 diabetes and renal disease A. Penfornis1, J.-F. Blickle2, B. Fiquet3, S. Dejager3; 1Jean Minjoz Hospital, Besançon, 2Strasbourg University Hospital, 3Novartis Pharma SAS, Rueil Malmaison, France. Background and aims: Few data exist regarding management of patients (pts) with type 2 diabetes mellitus (T2DM) over 75 years (y), who frequently have renal impairment (RI). This real-life study aimed to assess the treatment of T2DM in elderly patients with chronic kidney disease (CKD), a uniquely fragile population that poses specific challenges. Materials and methods: Observational cross-sectional study: 3704 pts with T2DM were recruited by 968 physicians in France. Data from 980 diabetic pts ≥75 y with CKD are presented in this sub-analysis. Results: Mean age was 81 y (range 75-101), 56% were male, with BMI 28.5 kg/m², while mean estimated glomerular filtration rate (eGFR) was 43 mL/ min/1.73m². 20% of patients had severe RI (eGFR S 442 1 C 1073 Randomised trial of a type 2 diabetes community case worker intervention reduces hospitalisations and emergency visits over 3 years in a collaborative care model R.O. Yeung1, Y. Zhang1, A.O. Luk1, Y. Cheung2, R. Ozaki1, H. Chung1, J.C.N. Chan1, W. So1; 1Medicine, Chinese University of Hong Kong, 2Ma On Shan Family Medicine Clinic, Hong Kong. Background and aims: Collaborative care between family medicine doc- tors and specialists with periodic review at a diabetes centre reduced major clinical outcomes and mortality compared to management by family doctors alone. Further, increased contact time with support staff who promoted self- management and reduced negative emotions has improved metabolic control and reduced acute health care utilization. We hypothesized that provision of a community care worker (CCW) in the setting of collaborative care further improves metabolic control and reduces acute health utilization in patients with type 2 diabetes. Materials and methods: In a collaborative care model, patients with type 2 diabetes attending a community-based family medicine clinic underwent annual comprehensive assessment (CA) at a specialist diabetes centre for risk stratification with personalized report and decision support using the web- based Joint Asia Diabetes Evaluation portal and were randomized to either additional support by a trained CCW (JADE+CCW) or JADE alone (JADE). Guided by the personalized report, the CCW counselled patients after the medical consultation to reinforce treatment adherence, self-care, and pro- vide emotional support. Using intention-to-treat analysis, primary outcome was change in glycemic control after 3 years. Secondary outcomes included cardiometabolic control (changes in blood pressure and LDL-C), as well as acute health care utilization as measured by number of hospitalizations, total length of stay (TLOS), and number of emergency visits over 3 years. Meta- bolic control was measured at CAs. Hong Kong’s public health care system services the majority of the population, so utilization measures were retrieved from the Health Authority database that captured all visits to public medical institutions within 3 years of enrolment. Negative binomial regression was used to ascertain the incident rate ratio (IRR) of hospitalization and emer- gency visits between groups, as well as rate ratio (RR) of TLOS. Results: Of 661 patients recruited, 332 were randomized to JADE and 329 to JADE+CCW. At entry, 46% were male, with mean±SD age of 60.3±10.9 years, diabetes duration of 6.8±6.0 years, and HbA1c of 6.8±1.1%. The me- dian number of CAs was 4 in both groups over 3 years, and those in the in- tervention group saw the CCW 12.2±4.3 times. After 3 years, HbA1c deterio- rated in both groups (JADE 0.44% (95%CI 0.32- 0.55%) and 0.48% (95%CI 0.37- 0.70) in JADE+CCW (P=0.635 between groups)), whereas LDL-C improved in both groups. Systolic blood pressure improved in JADE+CCW from baseline (133.1 to 129.5mmHg, P S 443 1 C tured approach in diabetes day clinics in the treatment of adult patients with diabetes in comparison to periodical visits in out-patient clinics. Materials and methods: A total of 280 adult diabetic patients participated in this study, 140 patients were treated in the diabetes day clinic, and 140 con- trols were treated as out-patients. There was no significant difference between groups considering age, sex, diabetes type, and diabetes duration at study start. The study group underwent structured theoretical and practical edu- cation about diabetic nutrition, weight control, physical activity, acute and chronic diabetic complications and pharmacological management. They re- ceived psychological assistance, and training in specific skills in the diabetes day clinic. The control group received standard care in the out-patient clinic. Routine biochemistry relevant to glycemic control (e.g. fasting and postpran- dial glycemia, HbA1c, lipidogram) was determined, and - if necessary - ther- apy was adjusted in all patients. BMI and routine biochemistry were repeated after 3 months. Statistical analysis was performed by using Student’s t-test for numerical data, and chi-square test for qualitative data. Results: In the study group, both fasting (10,68±3,89 vs. 8,34±2,49 mmol/L, p S 444 1 C on care experience and patient-oriented outcomes on populations with dif- ferent background sets of values and preferences. Clinical Trial Registration Number: NCT01861756 Supported by: EFSD grant supported by AstraZeneca 1078 Group-based self-management support leads to more adequate exercise behaviour in recently diagnosed type 2 diabetes patients A.L. van Puffelen1, M. Rijken1, G. Nijpels2, G.E.H. Rutten3, F.G. Schellevis1,2; 1NIVEL, the Netherlands Institute for Health Services Research, Utrecht, 2Department of General Practice & Elderly Care Medicine, EMGO Institute for Health and Care Research, VU University Medical Center, Amsterdam, 3Julius Center for Health Sciences and Primary Care, UMC Utrecht, Netherlands. Background and aims: Adequate self-care behaviours and lifestyle changes are considered essential, though challenging elements in the management of type 2 diabetes mellitus (T2DM). Previous research indicated that illness perceptions of patients and partners are important precursors for health be- haviours. In addition, interventions based on the Common-Sense Model of Self-Regulation (CSM) seem effective in enhancing health outcomes in vari- ous (chronic) conditions. We therefore developed a group-based self-man- agement support programme for recently diagnosed T2DM patients (1-3 years post-diagnosis) and their partners, based on the CSM. In this study, we want to investigate the short and long-term effectiveness of this new course on self-care behaviours in T2DM. Materials and methods: Randomised controlled trial with a pre-test (T0), post-test (T1) design and follow-up after six months (T2). T2DM patients were selected on the basis of medical records of participating general practices in different regions in the Netherlands and, after informed consent, randomly allocated to the intervention (four course sessions) or attention control (sin- gle information meeting) condition. Self-care was assessed with three items of the revised Summary of Diabetes Self-Care Activities (SDSCA) measure, assessing exercise, diet (both 0 - 7 days) and smoking (yes/no) during the past week. Exercise and diet were dichotomised into adequate (≥ 5 days) and non-adequate behaviours. Effectiveness on self-care was established by logis- tic regression analyses adjusting for demographic characteristics (age, gender and educational level) and pre-test scores on self-care. Results: A total of 167 patients participated in the study: 81 in the interven- tion and 86 in the control condition. 146 patients returned the T1 question- naire (drop-out rate 13%). Participating patients had an average age of 64 years (SD = 10.10, 27 - 83 years) and 56% were male. At baseline, 43% of the patients reported adequate exercise, 77% adequate dietary and 82% non- smoking behaviours. Preliminary results showed a significantly higher pro- portion of self-reported exercise behaviours in the intervention group (47%) at T1 (OR = 2.42, CI = 1.07 - 5.50), compared to the control group (32%). No significant differences between the intervention and control group were found on (changes in) reported dietary (76% vs 72%) and non-smoking be- haviours (87% vs 74%) at T1. Conclusion: In the short term, a self-management support programme based on the CSM appears to be effective in enhancing adequate (self-reported) ex- ercise behaviours in recently diagnosed T2DM patients. Results on long-term effectiveness of the course are currently investigated and will be presented during the conference. Clinical Trial Registration Number: NTR3302 Supported by: Dutch Diabetes Research Foundation 2009.70 1079 Diabetes care process performance using the alphabet strategy compared to the national audit data: practice of evidence based medicine (POEM) 2013 J.D. Lee1, P. Saravanan2, V. Patel3; 1Diabetes Centre, George Eliot Hospital NHS Trust, Nuneaton, 2Metabolic & Vascular Health, 3Education and Development, Warwick Medical School, Coventry, UK. Background and aims: The Alphabet Strategy is a diabetes care checklist based on a mnemonic: A for advice with regard smoking cessation, ideal weight attainment, regular exercise; B for blood pressure; C for cholesterol profile and creatinine care; D for diabetes glycaemic control; E for yearly eye exam; F for regular year foot exam; and G for guardian drugs with regard to aspirin, ACE inhibitors, and statins. Such a management strategy has been demonstrated to be effective in ensuring performance of care processes and attainment of diabetes target parameters. We performed an audit of the notes of all patients with Type 2 diabetes attending the diabetes outpatient depart- ment using the Alphabet Strategy as an audit template to determine adher- ence to the checklist. We compared these results to the UK’s recently pub- lished National Diabetes Audit 2011/12. Materials and methods: The Diabetes Outpatient Clinic register was con- sulted for the names of patients currently attending the department. Relevant demographic and diabetes data were extract from information attained at most recent clinic visit. Case notes, electronic letters and the computerised pathology reporting system were reviewed for outstanding care processes. Care performance within 15 months of the most recent clinic visit was reg- istered as positive documentation. Data collection was performed over three months between October to end of December 2013. Statistical analysis was performed using chi-squared test. Results: Data was available from 551 patients. Mean age of the whole cohort was 63 years. 18.7% of people were more than 74 years of age. Men com- prised 59.2% of the study group, and as a group were of similar average age to females at 63 years. Compared to the NDA, POEM achieved better per- formance of care processes except for annual foot exam. For target process achievement, POEM achieved better control in total cholesterol and target blood pressure, but fared worse in glycaemic control. Conclusion: The use of the Alphabet Strategy continues to help achieve per- formance of essential diabetes care processes. 1080 Cost-effectiveness of centralised and partly centralised care compared to usual care for patients with type 2 diabetes A.A.W. van der Heijden1, T.L. Feenstra2, M.C. de Bruijne1, C.A. Baan3, G.A. Donker4, J.M. Dekker1, G. Nijpels1; 1The EMGO Institute for Health and Care Research, VU University Medical Center, Amsterdam, 2UMC Groningen, 3RIVM, Bilthoven, 4Nivel, Utrecht, Netherlands. Background and aims: Due to an ever increasing number of type 2 diabetes patients, innovations to control the increasing health care use and costs are needed. Results of diabetes care programs on the costs or (cost-) effectiveness are heterogeneous. The aim of this study is to compare the cost-effectiveness of two diabetes care models with usual care for type 2 diabetes patients from the societal perspective. Diabetologia (2014) 57:[Suppl1]S1–S564 S 445 1 C Materials and methods: An economic evaluation was performed alongside a clinical trial. In two distinct regions of the Netherlands, two diabetes care models were implemented with different levels of centralized organizational structures. One of them was centralized care (CC) with a central organization and coordination of the care between all care providers and the use of a cen- tral database. Patients receive an annual extended diabetes assessment at the Diabetes Care Centre, in addition to the care by patients’ general practitioner (GP). GP’s receive feedback about their performance. Partly centralized care (PC) focuses on adherence to type 2 diabetes guidelines. An online clini- cal database is used to monitor mean values of risk factors. All assessments were performed in patient’s GP practice. Usual care (UC) has a decentralized organisation structure and patients’ GP is responsible for the diabetes care. Clinical outcome measure was risk of a coronary heart disease (CHD) calcu- lated with the UKDPS risk engine. Cost-effectiveness analysis was performed from the societal perspective comparing patients receiving CC (n=313), PC (n=293) and UC (n=485) during one year of follow-up. Missing costs and effects data were imputed using multiple imputation. Differences in costs, ef- fects and cost-effectiveness between the diabetes care groups were analysed using bootstrapping techniques. Results: Differences in changes in CHD risk over 12 months of follow-up between the three groups were statistically insignificant and clinically irrel- evant. Compared to UC, health care costs during the follow-up period were lower in CC (-1300 (95% CI: -2300 to -570)) and PC (-960 (95% CI: -1890 to -100)). Costs from the societal perspective showed the same trend, although not statistically significant. Conclusion: Clinical outcomes did not differ between the different care models. Lower health care costs were observed in (partly) centralized care compared to usual care, mainly due to substitution of secondary health care use by primary health care use. This suggests that centralizing the diabetes care results in equal outcomes at lower health care costs. Clinical Trial Registration Number: ISRCTN66124817. Supported by: ZonMw and Unive 1081 The effectiveness of the Diabetes Shared Care Program for diabetes-related avoidable hospitalisations in Taiwan: a nationwide population-based study E. Kornelius1,2, C.-N. Huang1,2, J.-Y. Chiou3, Y.-S. Yang1,2, Y.-L. Lu1,2, C.-H. Peng4; 1Department of Endocrinology and Metabolism, Chung Shan Medical University Hospital, Taichung, 2Institute of Medicine of Chung-Shan Medical University, 3School of Health Policy and Management, Institute of Medicine of Chung-Shan Medical University, 4Institute of Biotechnology, HungKuang University, Taichung, Taiwan. Background and aims: The Diabetes Shared Care Program was implemented in 2001 by Taiwan’s Ministry of Health and Welfare to improve the quality of diabetes care. The efficacy of this program is unknown and we aimed to evaluate whether Diabetes Shared Care Program participants had lower fre- quency of avoidable hospitalizations in Taiwan. Materials and methods: We collected nationally representative data from Taiwan’s National Health Insurance Research Database. The dataset com- prised 120,000 patients who were newly diagnosed with type 2 diabetes in 1999. The analysis included the patients‘ follow-up data until December 31, 2011. We designed a case-control study consisting of patients with avoid- able hospitalizations as cases and selected two age-, gender-, and avoidable hospitalization duration-matched controls by risk-set sampling. We further compared the hypoglycemia risk between Diabetes Shared Care Program participants and non-participants. We used the conditional logistic regres- sion to estimate the crude and adjusted odds ratio (OR). Results: A total of 2,377 avoidable hospitalizations cases and 4,754 matched- controls were identified. After adjusting for potential confounders in the stratified analyses, Diabetes Shared Care Program participants had a signifi- cantly lower frequency of all diabetes-related avoidable hospitalizations (OR, 0.15; 95% CI 0.13-0.17). A similar trend was found for short-term complica- tions (OR 0.13; 95% CI 0.08-0.20), long-term complications (OR 0.13; 95% CI 0.11-0.16), uncontrolled diabetes (OR 0.21; 95% CI 0.15-0.30), and lower- extremity amputations (OR 0.06; 95% CI 0.03-0.13). Conclusion: The Diabetes Shared Care Program decreased the frequency of all diabetes-related avoidable hospitalizations in Taiwan. PS 092 Screening and risk factors for gestational diabetes mellitus 1082 Screening for gestational diabetes mellitus in resource constrained settings V. Mohan1, R. Unnikrishnan1, B. Bhavadharini1, M. Mahalakshmi1, K. Maheswari1, G. Kalaiyarasi1, R. Uma2, R.M. Anjana1, M. Deepa1, B. Malanda3, A. Belton3, A. Kayal3; 1Diabetology, Madras Diabetes Research Foundation & Dr. Mohan‘s Diabetes Specialities Centre, 2Gynaecology, Seethapathy Clinic, Chennai, India, 3International Diabetes Federation, Brussels, Belgium. Background and aims: The International Association of Diabetes in Preg- nancy Study Groups (IADPSG) criteria for screening for gestational diabetes mellitus (GDM) requires 3 venous samples drawn in the fasting state. This is difficult in resource-poor settings for two reasons: women find it difficult to come to the clinic in the fasting state and also obtaining three blood samples is difficult due to limited resources. We therefore assessed the usefulness of a non-fasting capillary blood glucose (CBG) as a preliminary screening test to reduce the number of women who would need to be referred for a definite diagnostic test done in the fasting state. Materials and methods: Pregnant women (n=1031) attending antenatal clin- ics in Tamil Nadu in Southern India underwent a CBG test 2 hours after a 75 g glucose load administered irrespective of time of last meal (non-fasting). Participants were requested to come back within the next 2 to 3 days for a fasting OGTT, and both IADPSG and WHO (1999) criteria were used for diagnosis of GDM. The optimal sensitivity and specificity of the non-fasting 2 hour CBG value to identify GDM diagnosed by IADPSG and the WHO (1999) criteria were determined. Results: A non-fasting 2 hour CBG value of 126 mg/dl (6.9 mmol/l) had sen- sitivity and specificity of 64.6% and 63% respectively to identify GDM diag- nosed by IADPSG criteria, but 40% of all pregnant women would need to be referred for the fasting OGTT. For the WHO 1999 criteria, the CBG cutpoint of 144 mg% (7.9 mmol/l) had a sensitivity and specificity of 87.9% and 86.9% respectively and only 19.1% of the women would have to be referred for the fasting OGTT. Conclusion: A 2 step screening procedure for GDM using a non-fasting 2 hour CBG as the initial screening test and use of WHO (1999) criteria would help to minimize the number of women who would need to be referred for a diagnostic OGTT in the fasting state and thus make screening for GDM feasible in resource constrained countries. Supported by: IDF, Abbott Fund 1083 Gestational diabetes: what is the impact of the adoption of the criteria of IADPSG? M. Bachaoui1, K. Benharrat1, A. Rahou1, L. Benali1, A. Ouslim2, F. Ayad1, S. Aribi1, M. Belhadj1; 1University Hospital Establishment (EHU), Internal Medecine-Diabetology, 2University Hospital Center (CHU), Obstetrics and Gynecology, Oran, Algeria. Background and aims: We have always performed a systematic screening for gestational diabetes (GDM). The recommendations made in 2010 by the In- ternational Association of Diabetes and Pregnancy Study Groups (IADPSG) give the possibility for each country to choose whether the screening should be universal or selective on the basis of risk factors assessment (adjusted ac- cording to the prevalence of diabetes in the population). We have chosen to continue systematic screening. The aim of our study was to assess the level of risk in our screened population, and the maternal-fetal morbidity associated with GDM depending on the presence or absence of risk factors Materials and methods: We retrospectively analyzed the data of all pregnant women admitted in day-hospital in the service of our hospital between June 1, 2010 and December 31, 2012 for screening for GDM. We conducted a sys- tematic screening between 24 and 28 weeks of gestation by OGTT with 75 g glucose (IADPSG criteria). We evaluated the frequency of risk-factors: Age ≥ 35 years, BMI ≥ 25 kg/m2, history of diabetes in first degree relatives, previ- ous GDM or macrosomic child. We compared the incidence of maternal and Diabetologia (2014) 57:[Suppl1]S1–S564 S 446 1 C fetal complications between the group with risk-factors (RF +) and the group without risk-factors (RF -). Results: Of 1680 women screened, 330 had a GDM, an estimated prevalence of 19.6%. Among them 52 (15.8%) showed no RF : average age 28.3 years, mean BMI 23.6 kg/m2 ; and 278 (84.2%) had at least one RF : 128 (46%) had age ≥ 35 years, 208 (74%) BMI ≥ 25 kg/m2, 139 (50%) a history of GDM, 70 (25%) a history of macrosomia and 35 (12%), family history of diabetes. The maternal-fetal prognosis of the group without RF compared to the one with RF only differs in macrosomia: 7.3% vs 15.6% (p S 447 1 C p S 448 1 C associated with reduced insulin (PRG=18.69±2.22, GDM=4.05±2.34, PP=8.11±3.56, and NP=3.78±2.04 μU/mL; ANOVA p=0.0005, PRG vs. GDM p=0.01, PRG vs. PP p=0.05, PRG vs. NP p=0.01) and c-peptide (PRG=3.36±0.64, GDM=0.24±0.10, PP=2.01±0.74, and NP=1.66±0.44 ng/ mL; ANOVA p=0.0042, PRG vs. GDM p=0.01) levels. Conclusion: This study shows, for the first time, lower β-cell death in women with GDM, suggesting that β-cell death is not a main contributor to the de- velopment of GDM. 1090 Genetic variants associated with type 2 diabetes and obesity better predict gestational diabetes than traditional risk factors J. Hryniewicka1, M. Zbucka-Kretowska2, N. Wawrusiewicz-Kurylonek3, W. Bauer3,4, M. Szamatowicz1, B. Telejko3, M. Kuzmicki1, J. Szamatowicz1, M. Gorska3, A. Kretowski3,4; 1Department of Gynecology, 2Department of Reproduction and Gynecological Endocrinology, 3Department of Endocrinology, Diabetology and Internal Medicine, 4Centre for Clinical Research, Medical University of Bialystok, Poland. Background and aims: There is increasing evidence that gestational diabetes (GDM) shares genetic risk with type 2 diabetes (T2DM) and that 50 % of women with prior GDM develop T2DM in the next few years. Recent meta- analyses have shown that T2DM risk variants may be associated with GDM through different mechanisms, including impaired β-cell function (CD- KAL1, IGF2BP2, KCNQ1, KCNJ11, MTNR1B), insulin resistance (TCF7L2), and abnormal glucose utilization (GCK). So, in the present study we evalu- ated the risk of GDM development associated with genetic variants increas- ing the susceptibility for obesity and T2DM. Materials and methods: Four hundred and seventy four pregnant women, including 103 diagnosed as having GDM according to the JADPSG criteria and 371 controls with normal glucose tolerance, were genotyped for 65 SNPs, identified in genome wide-associated studies as risk variants for T2DM and obesity. Results: On the basis of genotype case-control differences we found an as- sociation between GDM and four obesity gene variants (BAT2 [rs2260000] - OR=2.07 [1.05-4.1], p=0.03; MTCH2 [rs10838738] - OR=1.97 [1.002- 3.88], p=0.045; FAIM2 [rs7138803] - OR=1.95 [1.15-3.34], p=0.012; CRY2 [rs11605924] - OR=0.56 [0.34-0.92], p=0.021), and three T2DM risk variants (TCF7L2 [rs7901695] - OR=1.67 [1.16-2.4], p=0.03; SLC30A8 [rs11558471] - OR=1.67 [1.16-2.4], p=0.007; CDKAL1 [rs10946398] - OR=1.62 [1.02-2.56], p=0.036). In logistic regression model comprising all studied SNPs, adjusted for relevant clinical data, GDM was significantly predicted by BMI before index pregnancy (OR=1.15 [1.07-1.23], p=5.2x10-5) and 4 studied SNPs: CDKAL1 (rs10946398) - OR=1.67 (1.02-2.1), p=0.04; GCK (rs4607517) - OR=2.64 (1.31-5.3), p=0.007; ADRA2A (rs10885122) - OR=2.88 (1.36-6.1), p=0.003 and KCTD15 (rs29941) - OR=2.17 (1.2-3.9), p=0.007. Conclusion: Our results suggest that genetic variants associated with T2DM and obesity better predict GDM than the patient’s age, the history of previous births and BMI before pregnancy. The mechanisms related to GDM risk seem complex and comprise the effect of numerous genes involved in insulin secre- tion/function and glucose homeostasis. Supported by: Grant No. N N407 141937 from the State Committee for Scien- tific Research 1091 MTNR1B genetic variability is associated with gestational diabetes in Czech women D. Vejrazkova1, P. Lukasova1, M. Vankova1, O. Bradnova1, J. Vcelak1, V. Cirmanova1, K. Andelova2, H. Krejci3, B. Bendlova1; 1Institute of Endocrinology, 2Department of Mother and Child Care, 3Department of Obstetrics and Gynecology, 1st Faculty of Medicine, Charles University and General University Hospital in Prague, Prague, Czech Republic. Background and aims: The gene MTNR1B encodes a receptor for melatonin, the main regulator of sleep cycle and circadian rhythm. Melatonin receptors are expressed in the brain and also in human pancreatic β-cells. This find- ing implies that genetic variants in the gene might affect glucose tolerance. Meta-analysis confirmed that rs10830963 shows the most robust association. The aim of our study was to assess and compare the SNP rs10830963 in the Czech GDM patients and in control normoglycemic women without history of GDM. We also aimed to study relations between the risk confering allele G of the SNP and biochemical as wel as anthropometric characteristics in both groups. Materials and methods: Our cohort of subjects consisted of 880 women, 458 of them were diagnosed with GDM (age 34,0±6,12 years; BMI 24,3±4,93 kg/ m2) and they met the 0.5-1.5 year interval after the childbirth. 422 were nor- moglycemic control women without history of GDM (age 34,8±15,09 years; BMI 23,7±4,18 kg/m2). We evaluated broad biochemical (3-hour oGTT, lipid profile, thyroid and steroid hormones, liver enzymes) and anthropometric (BMI, WHR, waist circumference, fat distribution, total body fat content) parameters in all participants. We used the ABI TaqMan SNP Genotyping Assays to genotype for rs10830963. Statistical analysis was conducted using the NCSS 2004 software. Results: In accordance with available literature data, the risk allele G was in our study significantly more frequent in the GDM group (38.3 % vs 29.4 % in controls, test power 0.96; OR 1.49 CI 95% [1.22; 1.82] pOR=0.0001). De- spite of higher frequency, the G allele in the GDM group was not associated with any markers of glucose metabolism nor with anthropometric data. In contrast, controls showed significant association of the allele G with fasting plasma glucose (median 4.95 mmol/l in GG genotype vs. 4.60 mmol/l in CG as well as in CC genotypes; p=0.002 and p S 449 1 C PS 093 Gestational diabetes mellitus management 1092 Gestational diabetes mellitus and health care cost: short- and long-term association A. Danyliv1,2, P. Gillespie1, C. O‘Neill1, F. Dunne2,3, ATLANTIC DIP (Diabetes in Pregnancy) Research Team; 1School of Business and Economics, 2School of Medicine, Clinical Sciences Institute, 3Galway Diabetes Research Centre, National University of Ireland Galway, Ireland. Background and aims: There is an extensive body of literature showing that gestational diabetes mellitus (GDM) diagnosed on WHO criteria is associ- ated with maternal, ante- and neonatal complications. Yet, little is known about the health care cost associated with GDM beyond the period immedi- ately after the delivery and maternity care cost associated with GDM on the International Association of Diabetes and Pregnancy Study Group (IADPSG) criteria. In this paper, we assess the association of the GDM diagnosis on the IADPSG criteria with the health resource utilization during maternity care and in longer term, two to five years after the index pregnancy. Materials and methods: Maternity care utilization was measured for a sam- ple of 658 women drawn from the Atlantic Diabetes in Pregnancy (ATLAN- TIC DIP) collaborative network who have had pregnancy two to five years previously. Annual health care utilization two to five years after the index pregnancy was assessed for a sub-sample of 348 women who returned a follow-up questionnaire. Irish cost weights were applied to the health care resource utilization to assess the cost of care. The difference in resource utili- zation and cost between women with GDM and normal glucose tolerance in the index pregnancy was assessed in a series of uni-variate and multivariate analyses. Results: The independent effect of GDM is approximately additional € 817.6 to the maternity cost and additional € 680.5 to the annual health care cost two to five years after the delivery. However, at the group/population level the effect might be higher due to inherent background differences of the women with GDM and NGT, especially worse body mass indicators. The excess in the maternity care cost is caused by higher likelihood of caesarean section and neonatal unit admissions in women with GDM. For the follow-up health care cost, most of the excess is related to higher chances of hospitalizations and outpatient unit visits not related directly to diabetes care. Diabetes care (diabetes nurse, dietitian, diabetes day centre, insulin and oral glucose agents, and blood sugar tests) constitutes smaller but still substantial part of the cost increase. Conclusion: Thus, our results suggest that GDM on IADPSG criteria is as- sociated with increased cost of maternity care and health care two to five years after the index pregnancy. This excess is not so much caused by the glucose tolerance treatment and monitoring routines as by more intensive use of hospital services, indicating worse health status of GDM women. Hence, this diagnosis objectively represents a serious condition altering health care cost around and beyond pregnancy. These findings provide information for further cost-effectiveness studies of the programs of GDM screening, treat- ment and monitoring in Irish setting. Supported by: HRB of Ireland, GA No. ICE/2011/3 1093 The effect of professional continuous glucose monitoring on glycaemic control and hypoglycaemia in insulin-requiring gestational diabetes mellitus S.S. Paramasivam1, A.T.B. Tan1, S.P. Chan1, P.C. Tan2, S.Z. Omar2, J. Ratnasingam1, L. Ibrahim1, L.L. Lim1, R.A. Razak1, F.Z.M. Shah1, S.R. Vethakkan1; 1Endocrinology, 2Obstetrics and Gynaecology, University of Malaya, Kuala Lumpur, Malaysia. Background and aims: Reducing hyperglycemia in gestational diabetes mellitus(GDM)improves pregnancy outcomes and reduces perinatal morbid- ity. Continuous Glucose Monitoring (CGM) is ideal for monitoring glucose levels where tight glycemic control without hypoglycaemia is required for a short period of time. While CGM is known to improve glycemic control/ pregnancy outcomes in women with pre-gestational diabetes the same has not been established in GDM. Aims: To determine if professional CGM im- proves glycemic control with less hypoglycaemia in insulin-requiring GDM. Materials and methods: In this prospective, open-label trial 24 women with insulin requiring GDM, gestation S 450 1 C (85.3±17.5/85.3±22.9 kg), BMI (32.0±5.4/32.2±7.2 kg/m2) and GC record- ed during the previous 2 weeks: fasting glucose value was 5.3±0.7/5.3±0.6; postprandial glucose values were: 6.3±0.8/6.3±0.7, 6.6±0.8/6.4±0.6 and 6.8±0.8/6.8±0.9 mmol/L, respectively post breakfast, lunch and supper. At delivery, in the Met-Gly group, 8 women were taking Met only (844±265 mg daily), 14 were taking Met (1179±153 mg) and Gly (3.9±1.9 mg), 10 were on Met-Gly+insulin (1333±250 mg, 8.6±2.2 mg and 12.7±9.9 units) and 3 were on insulin alone. In the 13 women taking insulin (37%), injections were started 4.2±2.1 weeks after initiation of Met-Gly treatment. There was no dif- ference in insulin doses between women taking insulin alone in the Met-Gly group and women in the insulin group: breakfast: 7.0±4.2/11.3±9.0, lunch: 8.5±4.9/9.6±8.1, supper: 11.0±4.2/10.3±6.8, bedtime: 11.0±7.1/18.7±15.1 units. During the 2 weeks prior to delivery, no difference was observed in GC between groups (fasting: 4.7±0.3/4.8±0.3; postprandial 5.8±0.4/5.9±0.5; 5.8±0.5/5.9±0.5; 6.0±0.5/6.1±0.5 mmol/L). At delivery, we found no dif- ference in the number of caesarean sections (9/8), neonatal’ birth weight (3360±389/3227±570 g), gestational age (38.7±1.1/38.4±1.5 weeks) or neo- natal hypoglycemias (21/15). Conclusion: Our data suggest that a majority of women with GDM could benefit from hypoglycemic agents: a combination of mild doses of Met-Gly successfully controlled GC in >60% women with GDM, with neonatal out- comes that are comparable to those of neonates born to women with GDM on insulin therapy. Clinical Trial Registration Number: NCT01215331 Supported by: FRQS 1095 Metformin in gestational diabetes and fasting glucose as a predictor of treatment failure M.O. Khin1, S. Gates2, P. Saravanan1,3; 1Division of Metabolic and Vascular Health, 2Clinical Trial Unit, Warwick Medical School, University of Warwick, Coventry, 3Department of Diabetes & Endocrinology, George Eliot Hospital, Nuneaton, UK. Background and aims: Gestational Diabetes (GDM) is common and affects up to 18% of pregnancy. Proper glycaemic control is the key to the reduction of GDM-related maternal and neonatal morbidity and mortality. The current guidelines suggest insulin or metformin only after dietary failure. Almost half of metformin-treated GDM need supplementary insulin. Identifying the characteristics of women who fail metformin may help to define optimal therapeutic strategy in GDM. Materials and methods: This was a historical cohort of GDM in a District General Hospital, UK between 2009 and 2012. GDM was diagnosed by 75 g OGTT test between 24-28 weeks of gestation with fasting level ≥6.1 mmol/l and/or 2-hour post-prandial(PP) level ≥ 7.8 mmol/l. Treatment targets were: pre-meals 4-5.5 and 1-hour post meals < 7.8 mmol/l. The treatment was es- calated if 2-3 readings per week were above target. Daily metformin dose was 500-2000 mg. Non-parametric tests were applied appropriately to compare the treatment groups. Logistic regression and receiver operator curve (ROC) were performed to identify the predictors of metformin failure. Results: Of the 299 being diagnosed and treated, complete treatment data were available in 228. Out of 228, 31(14%) were diet(D) alone, 141(62%) were on metformin(M) and 46(15%) on insulin(I) until delivery. 10 stopped metformin for gastrointestinal intolerance and were excluded from analy- sis. Compared to D group, despite older age(29 vs 32 years, p S 451 1 C pertension. 43 of the women with GDM (10,2%) presented hypertension during the pregnancy. The time of diagnosed of hypertension was 27 ± 4,2 weeks. The prevalence of overweight /Obesity (estimated by BMI before preg- nancy) was significant higher in the group of GDM women with hyperten- sion (mean BMI=28,86±4,5 kg/m2) during the pregnancy as compared to the GDM women without hypertension(mean BMI= 28,86±4,5 kg/m2), 72% in the group with hypertension vs. 23% in the GDM women without hyper- tension (p S 452 1 C PS 094 Pregnancy outcomes in gestational diabetes mellitus 1100 Improved outcome of women with gestational diabetes mellitus C. Nilsson1, D. Ursing2, H. Strevens3, M. Landin-Olsson2; 1Department of Paediatrics, Institution of Clinical Science, Helsingborg, 2Department of Endocrinology, 3Department of Obstetrics and Gynaecology, Institution of Clinical Science, Lund, Sweden. Background and aims: The number of women with gestational diabetes mel- litus (GDM) during pregnancy is increasing around the world, with a 10- 100% increase in some race/ethnicity groups. GDM affects 1-14% of all preg- nant women and is more common in populations with high frequency of type 2 diabetes. Among known risk factors for GDM is higher age, overweight, prior GDM and/or family history of diabetes. There are complications for the mother and child associated with GDM, both during and after pregnancy. With better maternity care and glucose control during pregnancy we expect improvement in outcome. The aim of this study was to compare pregnancy outcome of women with GDM 2013 against women with GDM 1995-1998. Materials and methods: Since 1995 in our region in Sweden, all pregnant women are tested with a 2-hour oral glucose tolerance test (OGTT) consist- ing of 75 g glucose in solution as a general screening for GDM in the 28th gestational week. The cut off value for GDM in Sweden is ≥10.0 mmol/l in capillary plasma glucose and is based on the European Association for the Study of Diabetes (EASD) recommendations. Women who receive a GDM diagnosis are transferred to our specialised maternity unit with a team of ob- stetrician, diabetologist, midwife and dietician. In 2013, 115 women received GDM diagnosis and as comparisons we choose all women who had GDM during 1995-1998 (n=115) in our region. The women’s medical journals from their GDM pregnancy were studied retrospectively. Duplex pregnancies (n=2 2013, n=3 1995-1998) as well as infants born before 37 weeks of gestation (n=10 2013, n=11 1995-1998) and after 41 weeks of gestation (n=0 2013, n=3 1995-1998) were excluded from the analysis regarding the newborn. Results: When comparing women who had GDM 2013 against women with GDM 1995-1998, there was no significant difference between the groups re- garding age of the mother, ethnicity, family history of diabetes, gestational length, HbA1c, insulin treatment during pregnancy, height of the mother, length of the newborn and apgar score at 5 and 10 minutes of the newborn. However, first weight of the women during GDM pregnancy 2013 was sig- nificantly higher than the weight of women with GDM 1995-1998, 71 kg (43- 119) (n=114) and 64 kg (43-133) (n=111; p=0.007) respectively. However, there was no significant difference in weight of the mother at delivery. For women with GDM 2013 weight at delivery was 80 kg (54-130) (n=49) and for women with GDM 1995-1998 weight at delivery was 78 kg (55-138) (n=77; p=NS). Birth weight of the child in GDM pregnancies 2013 was 3508 g ±487 (n=81) and in GDM pregnancies 1995-1998 3741 g ±589 (n=94; p=0.005). The percent of caesarean delivery was 24% in the GDM pregnancies 2013 and 17% in the GDM pregnancies 1995-1998 (p=0.04). Conclusion: Even though women with GDM 2013 weigh more when they start the pregnancy there is no difference in weight at delivery compared to women with GDM 1995-1998. This is also reflected on the newborn, that 2013 had significantly lower birth weight but with the same gestational length as 1995-1998. We believe that this is due to a more active and intense treat- ment of women with GDM during pregnancy together with higher frequency of caesarean delivery. Prevention of large infants is crucial to avoid complica- tions during delivery. 1101 Gestational diabetes: perinatal outcomes compared to background population and pregestational diabetes J.C. Wiebe1,2, B. Vega Guedes3,2, A. López Alonso3, O. Ramírez García3, F. Cabrera Morales3, L. Valle Morales3, J.A. García Hernández3,2, A.M. Wägner1,2; 1Endocrinology Dept, Complejo Hospitalario Universitario Insular-Materno Infantil de Gran Canaria, 2Instituto Universitario de Investigaciones Biomédicas y Sanitarias, Universidad de Las Palmas de Gran Canaria, 3Gynaecology and Obstetrics Dept., Complejo Hospitalario Universitario Insular-Materno Infantil de Gran Canaria, Las Palmas de Gran Canaria, Spain. Background and aims: Prevalence of gestational diabetes (GD) ranges be- tween 5-14% and is associated with an increased risk of maternal and fetal complications. The aim of this study was to assess the prevalence of GD and its outcomes in our population and to compare perinatal results with those of the general population (GP) and pregestational diabetes (PD). Materials and methods: All patients with GD (NDDG criteria) who deliv- ered at our centre (reference for our region) between January 2011 and De- cember 2012 were identified, and their records were reviewed and compared to those with pregestational diabetes and the background population. Statisti- cal analysis was performed using Microsoft Excel and SPSS for Windows. For frequency comparisons between groups, chi-squared was used. A bilateral p 25 kg/m2) than the GP (62.6% vs 42.9% p S 453 1 C between GDM and NGT. Logistic regression was used to model relationships between maternal and infant outcomes and BMI as a categorical variable ( 30) in the two groups. Results: 5806 women were identified, 887 with GDM and 4919 with NGT. Baseline characteristics were similar between the groups. Women with GDM had a higher risk of polyhydramnios (OR= 3.06; 95% CI 1.72-5.44) and C- section (OR=1.32; 95% CI 1.06-1.66) compared to women with NGT. A BMI >30 further increased the risk of C-section (OR =2.76 95% CI 2.20-3.46). C-section rate in women with BMI >30 was twice as common in GDM com- pared to NGT women (60.3% vs 31.6%, P30 (OR=2.21 95% CI 1.55-3.16). Infants of mothers with GDM had a higher risk of hypogly- cemia (OR 6.39 95%CI 3.34-12.3) and congenital malformations (OR= 1.77 95%CI= 1.37-2.29) compared to infants of NGT mothers. BMI did not im- pact on this any further. Rates of LGA were greatest with BMI > 30 (OR 1.76 95% CI 1.44-2.16). Neonates of GDM mothers were twice likely to be admit- ted to the neonatal unit compared to neonates of NGT mothers (18.5% vs 9.1%, P30. Thus these women need to be monitored and treated in the high intensity multi disciplinary hospital environment and are not suit- able for ongoing less intensive care in the primary care or community setting. Supported by: HRB 1103 Analysis of pregnancy outcomes in gestational diabetes mellitus based on the new IADPSG criteria compared to the former criteria M. Egloff1, K. Gariani1, S. Prati1, J. Philippe1, M. Boulvain2, F.R. Jornayvaz1; 1Division of Endocrinology, Diabetology, Hypertension and Nutrition, 2Division of Obstetrics, Geneva University Hospitals, Switzerland. Background and aims: Results from the Hyperglycemia and Adverse Preg- nancy Outcomes (HAPO) study led the International Association of Diabetes and Pregnancy Study Groups (IADPSG) to propose new diagnostic criteria for Gestational Diabetes Mellitus (GDM). The new IADPSG criteria have lower glucose concentration thresholds, with an increase in GDM prevalence as a consequence. In our institution, these criteria were implemented in Sep- tember 2011. The aim of this study was to determine if the use of the new diagnostic criteria led to changes in the management and in the occurrence of pregnancy complications among women diagnosed with GDM compared to the old diagnostic criteria. Materials and methods: We conducted a retrospective study of women with GDM followed at a single multidisciplinary university center. We included all consecutive women between July 2009 and December 2010 (group 1) di- agnosed using the old criteria (O’Sullivan followed by a 100 g oral glucose tolerance test) and between January 2012 and April 2013 (group 2) using the revised criteria (IADPSG consensus, 75 g oral glucose tolerance test in the fasting state). All women were instructed to adapt their diet and record self-monitoring of blood glucose according to international guidelines. No oral antidiabetic drugs were used. Parameters assessed were: percentage of women requiring insulin therapy; and pregnancy complications: caesarean section, macrosomia (birth weight >90th percentile) and preeclampsia. Man- agement guidelines have not been modified between the two periods. Results: We included 286 women with GDM, 129 in group 1 and 157 in group 2. Mean age was similar between groups (34.0±5.4 (group 1) versus 33.0±5.2 (group 2) [years], P=0.12), as were BMI (25.2±5.3 versus 26.1±5.1 [kg/m2], P=0.17) and weight gain during pregnancy (13.3±5.7 versus 13.2±6.9 [kg], P=0.94). There was a trend - although not significant - towards a lower per- centage of women requiring insulin therapy in group 2 (55.0% (group 1) ver- sus 43.3% (group 2), P=0.06) However, we did not find a significant difference in the number of cesarean section (38.9% versus 43.1%, P=0.55), the number of babies born with macrosomia (37.6% versus 33.0%, P=0.74) and the occur- rence of preeclampsia (1.6% versus 1.3%, P=1.0). Conclusion: Our study indicates that, when using the new IADPSG criteria, we include women with a less severe GDM, as less women required insulin in addition to dietary advice. There was no significant difference in caesarean section, macrosomia and preeclampsia. 1104 Excessive weight gain and pregnancy outcomes in pregestational and gestational diabetes J. Sastre1, F. Del Val1, C.M. Cortes1, A.M. Luque1, M. Olivar1, A. Marco1, V. Peña1, A. Pantoja2, M. Veganzones3, J. López1; 1Endocrinology, 2Pediatrics, 3Obstetrics, Complejo Hospitalario de Toledo, Spain. Background and aims: Women with diabetes mellitus (DM) during preg- nancy are at higher risk of adverse maternal and neonatal outcomes. Exces- sive gestational weight gain (GWG) is a potential risk factor for these adverse results. In 2009 the Institute of Medicine (IOM) established recommenda- tions for appropriate GWG in non-diabetic women. We try to examine if excessive GWG, using IOM recommendations, in pregnancies with pregesta- tional diabetes mellitus (PGDM) or gestational diabetes(GDM) is associated with higher adverse pregnancy outcomes. Materials and methods: We performed a retrospective study of 2773 single- ton pregnancies in women with DM [[259 PGDM (176 type 1 DM, 83 type 2 DM), 2514 GDM]. Maternal weight and body mass index (BMI) had been recorded pre-pregnancy and at the time of delivery. GWG was calculated and compared with IOM guidelines to assess if the upper limit per BMI category was breached. Examined maternal outcomes included pre-eclampsia, ges- tational hypertension and cesarean delivery. Fetal outcomes included large for gestational age (LGA), macrosomia and neonatal morbidity. Multivariate analyses were performed and odds ratio calculated using a logistic regression analysis adjusted with age, parity, ethnicity, pre-pregnancy BMI, and neonatal sex. Results: 1. Women with PGDM: 46,2% demonstrated excessive GWG. In the excessive GWG group a higher percentage were overweight (60,2%) or obese (56,2%, p< 0,01). There were no significant differences in GWG between type 1 and type 2 DM. Excessive GWG was associated with higher odds for LGA (OR 2,5 , CI 1,4-4,7 p S 454 1 C comparable lower levels of locus of control attributed to chance, doctors and other people. The level of anxiety and depression was generally low without any differences between the groups. As compared to women with appropri- ate for gestational age infants (n=86), women with large for gestational age infants (n=59) had similar scores for locus of control, anxiety and depression, and health-related quality of life. High levels of internal locus of control were seen both in women with large for gestational age and appropriate for gesta- tional age infants (30.0 (4.5) vs. 29.8 (5.3), p=0.85). Conclusion: Women with pregestational diabetes who delivered preterm were characterised by worse mental quality of life in early pregnancy. Supported by: EFSD / LifeScan 1106 Pregnancy outcome in depressive and GDM subjects in Bangladesh: a hospital based comparative study K. Natasha; Community Medicine, Institute of Health and Society, Faculty of Medicine, Oslo, Norway. Background and aims: In Bangladesh the prevalence rate of depression in adult population is only 4.6%. Gestational Diabetes Mellitus (GDM) is also alarming here but the actual prevalence rate of both diseases in pregnancy is still lacking, though they are common and result in serious consequences for mother and foetus. To our knowledge there has been little research address- ing the association between them in pregnancy and the outcome. These data are deficient even in South Asia let alone Bangladesh. This study was the cor- ollary part of one study (which tried to find out prevalence of depression and associated factors, among GDM subjects and to compare with Non GDM) to reveal the outcome of pregnancy with depression and GDM and compared. Materials and methods: A total of 748 pregnant women participated in the study. They were followed up from their 1st visit (not included after 13th week) to 1st week after delivery for at least 3 checkups. Depressive symptoms was scored following MADRS scale (0-12=not, 13-19=mildly, 20-34 moder- ately, 35-60=severely - depressed). Blood glucose was measured on every visit following WHO criteria; GDM was diagnosed within 24th to 28th weeks. Delivery procedure, Birth weight and APGAR score at 1st and 5th minute were assessed for the neonate. Results: Prevalence of depression among pregnant women was 12.69%. The rate was higher in GDM ( n 366) subjects (21.73%) with mean age 28.34 years than NGDM (n 382) subjects (7.73%) with mean age 27.17 years. Over all mean depressive score was higher at 3 stages in GDM group. Rate of caesar- ean section, number of live birth, and birth weight was higher but APGAR score at 2 stages were lower in GDM group. Mean age, Parity and birth weight of baby was higher but mean education years, mean APGAR scores of babies at both time period was lower in depressed groups (all 3 stages). Pregnant who were depressed specially in last trimester seem to have more rate of CS. Though gestational age at delivery did not vary much among the depressed and non depressed group. Chi square test and ’t tests’ proved there was signif- icant association with depressive scores with gestational age and birth weight. Physical Exercise and Weight gain rate is under analysis Conclusion: The important finding of this study was the elevated prevalence rate of depression in pregnancy which was greater than assumed. And preg- nancy outcomes are strongly associated with GDM and Depression. Develop- ing countries do not focus much on mental health but it is becoming ultimate necessity for future. Supported by: Iver Helles Foundation, Norway 1107 Atlantic DIP: diabetes in pregnancy is the most significant risk factor for caesarean delivery L. Carmody, A.M. Egan, O. Kgosidialwa, B. Kirwan, F. Dunne; Galway Diabetes Research Centre, National University of Ireland Galway, Ireland. Background and aims: Diabetes mellitus (DM) is a common medical disor- der of pregnancy. Caesarean delivery (CD) is associated with increased cost and maternal morbidity. Our aim was to determine the influence of DM on risk of CD in our cohort of patients enrolled in the Atlantic Diabetes in Preg- nancy (DIP) study. Materials and methods: Women with singleton pregnancies of 23 weeks or more estimated gestational age were identified and 6526 unique deliveries included. This included 4688 (71.8%) pregnancies with normal maternal glu- cose tolerance (NGT), 282 (4.3%) with pregestational diabetes mellitus type 1 and type 2 (PGDM) and 1556 (23.9%) with gestational diabetes mellitus (GDM). GDM was defined according to IADPSG criteria. Backwards step- wise logistic regression was utilised to evaluate the independent influence of selected variables on CD risk including diabetes mellitus (DM), maternal smoking and obesity, nulliparity, hypertension, ethnicity and macrosomia. Results: In total, 1968 (30.16%) pregnancies resulted in CD. In the setting of NGT pregnancies there were 1196 (25.51%) with CD, 181 (64.18%) with PGDM and 591 (37.98%) with GDM (p S 455 1 C PS 095 Postpartum outcomes 1108 Evaluating haemoglobin A1c and metabolic syndrome versus oral glucose tolerance test in postpartum diabetes screening M. Tome, M.J. Picon, J. Alcaide, I. Cornejo, M. Molina, F. Tinahones; University Hospital Virgen de la Victoria, Malaga, Spain. Background and aims: Gestational diabetes mellitus(GDM) increases the risk of type 2 diabetes (DM-2). Guidelines of clinical practice recommend postpartum screening with oral glucose tolerance test (TTOG). Compliance rates of women performing this test are low because this is an uncomfortable test. We investigated the reliability of other postpartum screening test in or- der to increase the compliance of women in postpartum. Materials and methods: We evaluated 419 women with a history of GDM in postpartum screening using hemoglobin A1c (A1C), HDL cholesterol, triglycerides, blood pressure and waist circumference and compared them with TTOG one year after delivery. We classified these patients into normal or pathological carbohydrate metabolism according to TTOG, A1C (follow- ing ADA criteria) and diagnostic criteria for the metabolic syndrome (MS) attending to ATPIII Expert Panel. Results: Based on the TTOG 257 women (61,3%) had normal carbohydrate metabolism and 162 women (38,7%)had a pathological status: 69 (16,5%) im- paired fasting glucose (IFG), 43 (10,3%) impaired glucose tolerance (IGT), 30 (7,2%) IFG+IGT and 14 (4,7%) had diabetes. According to the A1C cri- teria, 331 (77,9%) had normal carbohydrate metabolism and 84 (19,8%) had a pathological status: 82 (19,3%) prediabetes and 2 (0,5%) had diabetes. At- tending to MS, 341 (80,2%) were normal and 64 (15,1%) had at least 3 diag- nostic criteria for MS. Conclusion: A1c and MS diagnostic criteria underestimate carbohydrate metabolism reassessment in women with a history of GDM compared to the gold standard TTOG. 1109 The role of HbA1c for detection of diabetes and abnormal glucose tolerance after pregnancy with gestational diabetes mellitus R. Claesson, M. Ekelund, C. Ignell, K. Berntorp; Department of Clinical Sciences, Malmö, Lund University, Malmö, Sweden. Background and aims: Recently, HbA1c has been proposed as a diagnostic tool to identify individuals with undiagnosed diabetes, or who are at risk of diabetes. Women with gestational diabetes mellitus (GDM) represent a high risk group for type 2 diabetes development. According to Swedish national guidelines life style intervention and follow-up of these women after preg- nancy have high priority, but it is not made clear by which measures these women should be followed. The aim of the present study was to compare the performance of HbA1c with established glucose criteria during an oral glucose tolerance test (OGTT) and to assess HbA1c as a screening test for undiagnosed diabetes and pre-diabetes after pregnancy with GDM. Materials and methods: Glucose homeostasis was reevaluated 1-5 years after delivery in 140 women with prior GDM by means of OGTTs and si- multaneous HbA1c measurements. Glucose tolerance was defined according to World Health Organization criteria. HbA1c ≥48 mmol/mol was used for diabetes diagnosis and HbA1c 39-47 mmol/mol to define high risk women (pre-diabetes). Results: Mean ± SD for age and body mass index of included women were 35.4 ± 5.6 years and 26.6 ± 2.3 kg/m2, respectively. A median (interquartile range) of 26 (21-60) months had elapsed since their GDM pregnancy. Based on the OGTT 62 women (44%) had normal glucose tolerance, 50 (36%) pre- diabetes, and 28 (20%) diabetes, as compared to 114 (81%), 21 (15%) and 5 (4%), respectively, using HbA1c criteria. The agreement between diagnoses resulting from HbA1c and OGTT criteria was estimated by constructing cross tables and calculation of the κ coefficient (κ). The consistency in classifying diabetes versus non-diabetes was 82% (115/140) and κ 0.194, indicating poor agreement. The corresponding figures for the classification of any degree of abnormal glucose tolerance were 59% (82/140), κ 0.227. Combining HbA1c with fasting glucose criteria improved the agreement to fair (κ 0.596), but was no better than using the fasting glucose test alone (κ 0.599). The area under the receiver operating characteristic curve for HbA1c and fasting glucose to detect any degree of abnormal glucose tolerance was 0.708 and 0.834, respec- tively. Diagnostic accuracy was assessed using sensitivity, specificity, positive predicted value (PPV) and negative predictive value (NPV). Proposed cut points of HbA1c had low sensitivity (30%) and modest NPV (52%) to de- tect any degree of abnormal glucose tolerance, whereas specificity and PPV were high. The combined use of HbA1c and fasting glucose criteria improved the performance (sensitivity 67%, specificity 95%, PPV 95%, NPV 69%), but was similar to the use of the fasting glucose test alone. The latter identified 63% (49/78) of the women with pre-diabetes or diabetes in the study cohort. However, combined with a lower cut point of HbA1c (≥31 mmol/mol), an additional 59 women were identified, among whom 36% had pre-diabetes or diabetes based on post glucose load hyperglycemia. Conclusion: Proposed thresholds of HbA1c had low diagnostic sensitivity. Combined with a fasting glucose test the performance was no better than using a fasting glucose test alone. Considering that early detection of pre-di- abetes is of outmost importance in these women to prevent the development of diabetes, combining a fasting glucose test with a lower cut point of HbA1c may be an alternate approach to select women for an OGTT to identify those with isolated post glucose load hyperglycemia. 1110 Glucose intolerance after a recent history of gestational diabetes K. Benhalima1, L. Leuridan1, P. Calewaert1, R. Devlieger2, J. Verhaeghe2, C. Mathieu1; 1Endocrinology, 2Obstetrics & Gynecology, University Hospital of Gasthuisberg, Leuven, Belgium. Background and aims: The prevalence and the risk factors for glucose in- tolerance in early postpartum remain unclear in different populations. Our aim was to evaluate the uptake of our current screening strategy postpartum, the prevalence and the risk factors for glucose intolerance in women with a recent history of gestational diabetes (GDM). Materials and methods: Retrospective analysis of the medical files of our university hospital from 01-01-2010 till 31-12-2013 of women with a recent history of GDM diagnosed with the Carpenter & Coustan criteria. Since 2010 all women with a history of GDM are advised to undergo a 75g oral glucose tolerance test (OGTT) between 6-24 weeks postpartum. Indices of insulin sensitivity (the Matsuda index and the reciprocal of the homeostasis mod- el assessment of insulin resistance, 1/HOMA-IR) and an index of beta-cell function, the Insulin Secretion-Sensitivity Index-2 (ISSI-2) were calculated based on the OGTT during pregnancy and postpartum. Multivariable logis- tic regression was used to adjust for confounders. Results: Over a 4 year period, 231 women were identified with a recent his- tory of GDM. Of all women, 21.4% (46) did not attend the scheduled post- partum OGTT. Compared to women who received an OGTT postpartum, women who did not attend the postpartum OGTT smoked more often be- fore pregnancy (10.9% vs 2.4%, p=0.021), had more often a previous history of GDM (21.7% vs 10.7%, p=0.042), less often breastfed (56.5% vs 75.7%, p=0.011) and had a lower insulin sensitivity based on the OGTT during pregnancy [the Matsuda index 2.2 (1.5-3.1) vs 2.8 (1.9-3.8), p=0.033 and 1/HOMA-IR 0.015 (0.010-0.023) vs 0.020 (0.012-0.032), p=0.030]. Of all women (169) who received an OGTT postpartum, 40.8% (69) had predia- betes (14.9% impaired fasting glucose, 32.5% impaired glucose tolerance and 7.1% both impaired fasting and impaired Glucose tolerance) and 5.3% (9) had overt diabetes. Compared to women with a normal OGTT postpartum, women with prediabetes or diabetes were more often overweight (39.7% vs 25.3%, p= 0.009) or obese at first prenatal visit (27.4% vs 19.5%, p=0.037), were more often multiparous (44.0% vs 26.6%, p=0.028), had a higher glu- cose challenge test [168.0mg/dl (153.0-190.0) vs 159.0mg/dl (149.5-175.5), p=0.007], had an earlier diagnosis of GDM [gestational weeks 26.0 (25.0- 28.0) vs 27.0 (25.0-29.0), p=0.030], had an higher median fasting [94.5mg/ dl (84.2-101.7) vs 88.0mg/dl (81.0-99.0), p=0.006] and 2 hour glucose value during pregnancy [175.0mg/dl (162.0-198.0) vs 168.0mg/dl (158.0-181.0), p=0.003] and were more often treated with basal-bolus insulin injections (52.0% vs 17.4%, p=0.032). Women with glucose intolerance postpartum also had a lower beta-cell function and lower insulin sensitivity, remaining signifi- cant after adjustment for age, BMI, ethnicity, breastfeeding, contraception, multiparity and corticoid treatment [ISSI-2 in pregnancy 1.5±0.5 vs 1.7 ±0.4, p=0.029; ISSI-2 postpartum 1.5 (1.2-1.9) vs 2.2 (1.8-2.6),p=0.020; Matsuda index postpartum 3.8 (2.6-6.2) vs 6.0 (4.3-8.8), p=0.021). Conclusion: Glucose intolerance is frequent in early postpartum and these women have a lower beta-cell function and lower insulin sensitivity. One fifth Diabetologia (2014) 57:[Suppl1]S1–S564 S 456 1 C of women did not attend the scheduled OGTT postpartum and these women have an adverse risk profile compared to women who received the OGTT postpartum. Supported by: Scholarship of FWO Vlaanderen for KB, CM and RD 1111 The risk of future glucose intolerance in women with history of gestational diabetes O. Yilmaz1, N. Dinccag2, S. Celik2, S. Ozel3, C. Idiz2, F. Turker2, A.K. Uzum2, N. Gul2, G. Yenidunya2, I. Satman2; 1Internal Medicine Department, 2Endocrinology and Metabolism Division, 3Bioistatistics and Medical Informatics Department, Istanbul University, Turkey. Background and aims: In women with gestational diabetes mellitus (GDM) abnormal glucose metabolism normalizes soon after delivery. However, the history of GDM predisposes to carbohydrate intolerance in the future. The aim of this study was to evaluate risk factors predicting to future development of diabetes mellitus (DM) in women with a history of GDM. Materials and methods: 221 women with history of GDM were enrolled into the study. A 75 g OGTT was performed within 3 months following delivery (n=181) or 11±2 years after delivery(n=40) . Based on ADA criteria partici- pants were classified into 3 groups as normal glucose tolerance; NGT (group 1), IGT and/ or IFG ( group 2) and DM (group 3) We evaluated the relation- ship between status of carbohydrate metabolism after delivery and possible risk factors such as: age-at- gestation; BMI; family history of DM; poor ob- stetric history; gestational week, A1C and OGTT glucose levels at diagnosis of GDM; weight gain and insulin requirement during index pregnancy. Results: As shown in the table below, 16.2 % of participants had any degree of glucose intolerance after delivery. Gestational age of women who progressed to prediabetes or diabetes was lower than the others. At diagnosis OGTT glucose levels were higher and family history and poor obstetric history was much more higher in group 2 and 3. Women who developed later carbohy- drate intolerance had insulin requirement during pregnancy. Conclusion: Age-at-gestation, having poor obstetric history, higher glucose levels (fasting, 1st, 2nd hr) at diagnosis of GDM and insulin requirement dur- ing pregnancy are crucial predictors of development of glucose intolerance in the future. Those women required careful and more frequent follow-up after delivery. 1112 Biomarkers of endothelial dysfunction in relation to impaired carbohydrate metabolism early after pregnancy with gestational diabetes mellitus C.S. Göbl1,2, L. Bozkurt3, T. Prikoszovich3, A. Tura4, G. Pacini4, A. Kautzky-Willer3; 1Department of Obstetrics and Gynecology, Division of Obstetrics and Feto- Maternal Medicine, Medical University of Vienna, 2Department of Internal Medicine III, Division of Endocrinology and Metabolism, Gender Medicine Unit, Medical University of Vienna, 3Department of Internal Medicine III, Division of Endocrinology and Metabolism, Gender Medicine Unit, Medical University of Vienna, Austria, 4Institute of Biomedical Engineering, Metabolic Unit, National Research Council, Padova, Italy. Background and aims: The diagnosis of gestational diabetes mellitus (GDM) implicates an increased risk for the later development of type 2 diabetes. It was previously shown that metabolic reassessment early after delivery identi- fies females with particularly high risk. However, the relation to endothelial dysfunction, which might represent an early precursor of atherosclerosis and cardiovascular disease in females after pregnancy with GDM, is less well un- derstood. Materials and methods: 108 women with previous gestational diabetes (pGDM) and 40 controls were included 3-6 months after delivery and un- derwent specific metabolic assessments including a frequently sampled in- travenous glucose tolerance test (FSIGT) and an oral glucose tolerance test (OGTT) to derive parameters of insulin sensitivity and β-cell dysfunction in addition to dynamics of glucose, insulin, C-peptide, proinsulin and amylin. The area under the curve was calculated by the trapezoidal rule including measurements of fasting, as well as 30’, 60’, 90’ and 120’ post load concentra- tions. OGTTs were repeated in females with pGDM over 10 years of follow- up to identify subjects with diabetes manifestation. Circulating ICAM-1 (intracellular-adhesion-molecule-1), VCAM-1 (vascular-cell-adhesion-mol- ecule-1) and E-selectin, representing biomarkers of endothelial dysfunction were assessed at baseline and annually over five years after index pregnancy. Results: Adhesion molecules and E-selectin were significantly related to in- sulin sensitivity estimated from the FSIGT (SI-FSIGT): ICAM-1: r=-0.23, p=0.009; VCAM-1: r=-0.22, p=0.011; E-selectin: r=-0.21, p=0.018. Fractional polynomial regression models revealed further associations between ICAM- 1 and AUC-glucose (p=0.003), AUC-proinsulin (p=0.005) and AUC-amylin (p=0.017) as well as between E-selectin and AUC-glucose (p=0.013) and BMI (p=0.004) after adjustment for the degree of insulin resistance. Moreover, ad- hesion molecules remained significantly elevated in pGDM subjects as com- pared to the control group in a multivariable model including SI-FSIGT, age and BMI. 21% of females after pregnancy with GDM developed diabetes dur- ing the follow-up period. Longitudinal analysis revealed significantly higher ICAM-1 and E-selectin levels in subjects with progression to overt diabetes. Conclusion: Biomarkers of endothelial dysfunction are related to impaired carbohydrate metabolism in subjects with recent history of GDM. This ob- servation might indicate the later development of overt cardiovascular dis- orders in this specific risk collective. Follow-up examinations with clearly defined cardiovascular endpoints are recommended to perform an accurate risk stratification. Supported by: Austrian Science Fund (P14515 MED) to AKW 1113 Non-alcoholic fatty liver disease is prevalent in non-diabetic women with previous gestational diabetes mellitus regardless of their glucose tolerance S. Foghsgaard1,2, L. Vedtofte1, C. Andersen1, E.S. Andersen1, L.L. Gluud1, C. Strandberg3, T. Buhl4, E.R. Mathiesen5, P. Damm6, J.A. Svare7, F.K. Knop1, T. Vilsbøll1; 1Diabetes Research Center, 2Department of Biomedical Sciences, Panum Institute, 3Department of Radiology, 4Department of Nuclear Medicine, Gentofte Hospital, University of Copenhagen, 5Center for Pregnant Women with Diabetes, Department of Endocrinology, 6Center for Pregnant Women with Diabetes, Department of Obstetrics, Rigshospitalet, University of Copenhagen, 7Department of Obstetrics and Gynaecology, Herlev Hospital; University of Copenhagen, Denmark. Background and aims: Gestational diabetes mellitus (GDM) increases the risk of developing type 2 diabetes later in life. Furthermore, type 2 diabetes and obesity are associated with non-alcoholic fatty liver disease (NAFLD). Diabetologia (2014) 57:[Suppl1]S1–S564 S 457 1 C NAFLD is defined as a condition with excessive fat accumulation in the liver and absence of significant alcohol consumption. We examined glucose tol- erance and fat distribution in obese, non-diabetic women with prior GDM. Materials and methods: Based on a 75g-OGTT, non-diabetic women with previous GDM were classified as having normal glucose tolerance (NGT) or prediabetes (impaired fasting glucose and/or impaired glucose tolerance). Insulin resistance (IR) was assessed by HOMA-IR. All women underwent an ultrasound scan of the liver to detect fat accumulation, blood sampling, and a dual energy X-ray absorptiometry (DXA) scan. The android-to-gynoid fat-ratio and total visceral fat mass were calculated based on the DXA scan. A questionnaire was used to evaluate alcohol consumption habits. Results: Women with previous GDM were included (n=70; age: 38±5 years (mean±SD); BMI: 31±5 kg/m2; HbA1c: 34±4 mmol/mol (5.4±0.4%); HO- MA-IR: 1.2±0.6; duration since GDM: 5±2 years). Twenty four women had NGT (fasting plasma glucose (FPG) 5.4±0.3 mM; 2h-OGTT plasma glucose (PG): 6.7±1.4 mM) and 46 women had prediabetes (FPG 5.5±0.5 mM; 2h- OGTT PG: 9.2±1.1 mM). NAFLD was diagnosed in 5 (21%) women with NGT and 15 (33%) with prediabetes). Glucose tolerance status (NGT or prediabetes) did not coincide with NAFLD (p=0.39). Multivariable logis- tic regression analyses showed that NAFLD was positively associated with HOMA-IR (p=0.035), BMI (p=0.004), C-peptide (p=0.038), mass of vis- ceral fat (p=0.004), the android-to-gynoid fat-ratio (p=0.027), and plasma levels of alanine aminotransferase (p=0.018) and aspartate aminotransferase (p=0.028). Conclusion: NAFLD is prevalent in non-diabetic women with previous GDM regardless of their glucose tolerance status, but is associated with HOMA-IR, high BMI, high C-peptide levels, android fat distribution, and elevated plasma liver enzymes. Clinical Trial Registration Number: EudraCT 2012-001371-37 Supported by: Unrestricted grant from Novo Nordisk as an IIS-protocol 1114 Incretin function in adult offspring of women with diabetes in pregnancy L. Kelstrup1, T.D. Clausen2, E.R. Mathiesen3,4, T. Hansen5,6, J.J. Holst7, P. Damm1,4; 1Department of Obstetrics, Center for Pregnant Women with Diabetes, Copenhagen, Denmark, 2Department of Gynecology and Obstetrics, Hilleroed, 3Department of Endocrinology, Center for Pregnant Women with Diabetes, Copenhagen, 4Faculty of Health Sciences, University of Copenhagen, 5Marie Krogh Center for Metabolic Research, Copenhagen, 6Faculty of Health Sciences, University of Southern Denmark, Odense, 7University of Copenhagen, The Novo Nordisk Foundation Center for Basic Metabolic Research, Copenhagen, Denmark. Background and aims: Fetal exposure to maternal diabetes is associated with increased risk of pre-diabetes and type 2 diabetes (T2DM) in the offspring. The pathogenesis of T2DM seems to involve dysfunction of the incretin hor- mones glucose-dependent insulinotropic polypeptide (GIP) and glucagon- like peptide-1 (GLP-1), as well as hyperglucagonemia. Our aim was to inves- tigate the levels of the incretin hormones GIP, GLP-1 and glucagon in adult offspring exposed to intrauterine hyperglycemia. Materials and methods: A cohort of 587 Caucasian offspring, without known diabetes was followed up at the age of 18-27 years. We included two groups exposed to maternal diabetes in utero: offspring of women with gesta- tional diabetes mellitus (O-GDM) or type 1 diabetes (O-T1DM). Two refer- ence groups were included: offspring of women with risk factors for GDM, but normo-glycemia during pregnancy (O-NoGDM) and offspring from the background population (O-BP). The subjects underwent a 75-g oral glucose tolerance test (OGTT) with venous blood samples at 0, 30, 120 min. Results: Significantly lower levels of GLP-1 in the fasting state was found in the 2 diabetes-exposed groups (O-GDM and O-T1DM) compared to O-BP (p=0.032 and 0.004 respectively). The levels of glucagon during OGTT (time=30 min.) showed a tendency towards higher level in O-GDM com- pared to the unexposed groups (O-NoGDM and O-BP). No association be- tween levels of GIP and exposure status was found. Conclusion: Reduced levels of GLP-1 in the fasting state and increased levels of glucagon during OGTT may contribute to the increased risk of glucose intolerance among adult offspring born to women with diabetes during preg- nancy. PS 096 Pregnancy in type 1 diabetes 1115 Hyperglycaemia remains the major risk factor of major foetal complications in pregnant women with type 1 diabetes and good glycaemic control J. Skupien, K. Cyganek, B. Katra, I. Janas, P. Witek, M.T. Malecki; Department of Metabolic Diseases, Jagiellonian University Medical College, Krakow, Poland. Background and aims: In women with type 1 diabetes pregnancy is associ- ated with increased risk of adverse outcomes. The aim of this study was to ascertain the association of glycemic control, estimated with HbA1c, with se- lected serious adverse pregnancy outcomes, such as preterm labor, congenital malformations low birth weight and stillbirths in a group of pregnant women with very good metabolic control. Materials and methods: The study cohort comprised 510 women with type 1 diabetes. They received intensive diabetes management during the period of pregnancy in an academic outpatient reference center between 1998 and 2012. We excluded multiple pregnancies (n=9), miscarriages (before 20th week of gestation, n=40), cases with missing outcomes and lost to follow- up (n=6). Of the remaining 455, 183 women (40.2%) entered the intensive management program before conception (pregnancy planning). HbA1c was measured in 3 pregnancy trimesters, in a laboratory implementing internal and external quality assurance protocols, with an assay (high performance liquid chromatography) that was adjusted to DCCT standard. Results: The median HbA1c in the 1 st, 2nd and 3rd trimester was 6.5%, 5.7% and 5.6%, respectively. Only 155 (34.1%) women in the 1st trimester, 44 (9.7%) in 2nd and 39 (8.6%) in 3rd had HbA1c 7.0% or higher. In spite of that, we observed 70 (15.4%) preterm labors (two between 20th and 26th week of gestation). There were 34 (7.5%) cases of low birth weight (below 2500 g), 10 of them occurred in term newborns. There were also 10 (2.2%) stillbirths and 28 (6.1%) congenital malformations in live births. Additional 4 cases of congenital malformations occurred in stillborn fetuses. The strongest predic- tor of preterm delivery was the 3rd trimester HbA1c: relative odds of preterm delivery per 1% increase in HbA1c was 1.46, p=0.004. First and 2 nd trimester HbA1c had weaker impact: odds ratios were 1.20 (p=0.06) and 1.35 (p=0.01), respectively (both per 1% increase in HbA1c). Of note, 97% preterm labors took place after 26th week of gestation. Interestingly we did not observe sig- nificant relationship between low birth weight and HbA1c, while the odds of this complication increased by 1.48 (p=0.02) per each additional 5 years of maternal age. The 1st trimester HbA1c (but not 2 nd or 3rd) was associated with the risk of stillbirths: relative odds per 1% increase of HbA1c was 1.70, p=0.005. Similarly, 1st trimester HbA1c, but not 2 nd or 3rd, was associated with the risk of congenital malformations (relative odds per 1% increase of HbA1c- was 1.41, p=0.005). Although pregnancy planning resulted in 0.8% (6.8 vs. 6.0%) reduction of median HbA1c in the 1 st trimester and 0.3% (5.8 vs. 5.5%) in the 3rd, it was not sufficient to significantly lower the risk of adverse out- comes (odds ratio 0.80, p=0.35 for the composite outcome of preterm deliver- ies, stillbirths and malformations). Conclusion: Despite very good glycemic control the risk of the examined adverse outcomes of pregnancy in women with type 1 diabetes remains high. In this study preterm labors were associated predominantly with the 3rd tri- mester HbA1c but less with earlier glycemic exposure. Hyperglycemia during the first trimester of pregnancy, the period of embryonic organogenesis, was associated with increased risk of congenital malformations and stillbirths. These results emphasize the need for maintaining normoglycemia through- out the whole course of pregnancy. 1116 Fixed rate intravenous insulin infusion for euglycaemic diabetic ketoacidosis to a pregnant type 1 diabetic with good foetal outcome K. Tofeec, L. Pichaipillai, C. Bond, S. Ramtoola, M. Littley; Diabetes and Endocrinology, Royal Blackburn Hospital, Lancashire, UK. Background and aims: Diabetic Ketoacidosis (DKA) is a life threating condi- tion during pregnancy. It carries high mortality (up to 30%) for both infant and mother. Ketosis has been implicated in foetal distress and causes adverse neurological outcome. During pregnancy, euglycemic diabetic ketoacidosis requires prompt recognition and treatment. The current national guideline in UK recommends using a fixed dose of insulin infusion (0.1units/kg/hr), Diabetologia (2014) 57:[Suppl1]S1–S564 S 458 1 C titration of fluids to keep this infusion continuously running and continu- ation of long acting insulin during management of DKA. We report a case of euglycemic DKA in a pregnant Type 1 diabetic who received fixed rate insulin infusion with complete recovery. Materials and methods: A 24 year old type 1 diabetic for 10 years attended the antenatal clinic in a District General Hospital for antenatal management of her 2nd pregnancy. Her antenatal period had been uneventful till 24 weeks, with a satisfactory diabetes control (HbA1c of 50 mmol/mol) on a basal bolus regimen. She presented to the antenatal clinic with history of nausea, recur- rent vomiting, abdominal pain and headache for preceding 6 days. She had stopped taking her short acting insulin but continued with her basal insu- lin. Clinically she was dehydrated. Investigations showed plasma glucose 6.5 mmol/l and blood ketones 3.9. Venous blood gas showed acidity (PH) 7.21, bicarbonate 9 and lactate 0.9. Rest of the biochemical investigations were un- remarkable. Cardiotocography for fetal monitoring was normal Results: A diagnosis of euglycemic diabetic ketoacidosis was confirmed, fixed dose of intravenous insulin with dextrose infusion was initiated as per Joint British Diabetes Societies (JBDS) guidelines. She was transferred to HDU for close monitoring. No precipitating factors for DKA was found. She made an excellent recovery and discharged home in 2 days. She delivered by caesarean section at term with good foetal outcome. Conclusion: Euglycemic DKA is poorly understood in non-specialist prac- tice, but it is important to recognize the condition early. Poor oral intake due to vomiting and nausea as well as pregnancy were the only contributing fac- tors in our case. Fixed rate intravenous insulin has been part of the DKA pathway (JBDS Guidelines) only recently along with bedside measurement of metabolic changes. The rationales for this recommendation are:faster resolu- tion of ketosis with higher insulin and glucose concentrations thereby rapid recovery with reduction in the length of stay, and significant delay in the nor- malisation of pH with a sliding scale protocol. Euglycemic DKA, although uncommon, still continues to be perceived in clinical practice.Type 1 diabetic patient presenting with nausea and vomiting and found to have a normal glucose may still have life-threatening ketoacidosis, and an evaluation of their acid/base status is still revealed.Identification of such presentation timely is crucial and also the need for early fixed rate intravenous insulin infusion along with dextrose which will prevent significant morbidity for both the mother and the foetus like in our case. 1117 Higher gestational weight gain is associated with increasing offspring birth weight independent of maternal glycaemic control in women with type 1 diabetes E.R. Mathiesen1,2, A.L. Secher1,2, C.B. Parellada1,2, L. Ringholm1,2, B. Ásbjörnsdóttir1,2, P. Damm1,3; 1Center for Pregnant Women with Diabetes, 2Department of Endocrinology, 3Department of Obstetrics, Rigshospitalet, Copenhagen, Denmark. Background and aims: To evaluate the association between gestational weight gain and offspring birth weight in singleton term pregnancies of women with type 1 diabetes. Materials and methods: One-hundred-and-fifteen consecutive women re- ferred S 459 1 C 1119 Impact of parity on glycaemic control, cardiovascular risk factors and diabetes-related chronic complications: a multicentre study in Brazil C.A. Negrato1, M.B. Gomes2; 1Internal Medicine, Bauru‘s Diabetics Association, Bauru, 2Diabetes Unit, State University of Rio de Janeiro, Brazil. Background and aims: Pregnancy in women with preexisting diabetes can predispose these women to some diabetes-related chronic complications or accelerate the course of these complications if they are already present. The aim of this study was to determine the impact of parity on glycemic con- trol, cardiovascular risk factors and diabetes-related chronic complications in women with type 1 diabetes in Brazil. Materials and methods: This was a multicenter cross-sectional study con- ducted between December 2008 and December 2010 in 28 public clinics in 20 cities from the four Brazilian geographic regions. Data were obtained from 1,532 female patients, 59.2% Caucasian, aged 25.2 ± 10.6 years. The diabetes mean duration was 11.5 ± 8.2 years. Patient information (clinical factors and number of pregnancies) were obtained through a questionnaire and a chart review. Number of pregnancies were stratified in five groups: group 0 (no pregnancy), group 1 (one pregnancy), group 2 (two pregnancies), group 3 (three pregnancies) and group 4 (≥four pregnancies). Results: The comparison between the patients stratified according to parity showed that patients from groups 0, 1 and 2 were younger than patients from group 4 (p S 460 1 C PS 097 Neuropathy: biomarkers and mechanisms 1122 Cardiovascular autonomic tone relation to obesity stage, metabolic syndrome, advanced glycation end products and other metabolic parameters in prediabetes R. Dimova, T. Tankova, N. Chakarova, G. Grozeva, L. Dakovska; Medical University Sofia, Clinical Center of Endocrinology, Bulgaria. Background and aims: The aim of the study was to assess cardiovascular autonomic function (CAF) at different stage of obesity and in the presence of metabolic syndrome (MetS), and its relation to metabolic parameters and advanced glycation end products (AGEs) in subjects with prediabetes - im- paired fasting glucose (IFG) and impaired glucose tolerance (IGT). Materials and methods: A total of 148 subjects - 91 females and 57 males (mean age 50.4±13.6years, mean BMI 32.4±8.8kg/m2) with prediabetes - 83 with IFG, 29 with IGT and 36 with IFG+IGT, divided into 5 groups according to BMI: 15 normal weight, 43 overweight, 48 obesity class I, 25 class II, and 17 class III, and into 2 groups according to the presence of MetS - 118 with MetS and 30 controls were enrolled. Glucose tolerance was studied during OGTT. Anthropometric indices, blood pressure and serum lipids were measured. AGEs were evaluated by skin autofluоrescence (AGE-Reader). Body compo- sition was estimated by impedance analysis (InBody 720). CAF was assessed by ANX-3.0 using frequency-domain analysis during standard clinical tests - deep breathing, Valsalva and standing from a seated position. Results: No statistically significant difference in CAF was established be- tween the groups according to BMI and the presence of MetS, as well as be- tween subgroups according to glucose intolerance. In subjects with prediabe- tes sympathetic and parasympathetic tone at rest and during applied clinical tests showed negative correlation with visceral fat area, AGEs, and HbA1c and did not demonstrate significant correlation with BMI, waist circumfer- ence, lipid profile and blood pressure. Conclusion: Obesity stage and MetS are not associated with CAF deteriora- tion in prediabetes on their own. Central obesity, HbA1c, and AGEs accu- mulation appear to be the main determinants of autonomic imbalance and increase cardiovascular risk at the early stages of glucose intolerance. Supported by: Medical University Sofia 1123 Decreased beta cell function is associated with cardiovascular autonomic neuropathy in newly diagnosed type 2 diabetic patients X.b. Yang1, H.r. DENG1, W. Xu1, Y.h. Zhu1, S.d. Lin2, H.z. Yang3, D.h. Cai4, J.p. Weng1; 1Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, 2First Affiliated Hospital of Shantou University Medical College, 3Guangdong Provincial People’s Hospital, 4Zhujiang Hospital, Southern Medical University, Guangzhou, China. Background and aims: Cardiovascular autonomic neuropathy (CAN) is closely associated with increased mortality in diabetic patients. Several risk factors of CAN have been clearly clarified. However, the impact of beta-cell function on CAN, especially in newly diagnosed type 2 diabetic patients, is still controversial. Therefore, this study aimed to investigate the association between beta-cell function and CAN in newly diagnosed type 2 diabetic pa- tients. Materials and methods: 90 newly diagnosed type 2 diabetic patients (68.9% male, mean age: 45 years old) were enrolled. Standard battery of Ewing tests, including Valsalva manoeuvre (Valsalva ratio), the deep breathing test of expiration-to-inspiration ratio and the lying to standing test among the heart rate tests, and the orthostatic hypotension test, were used to identify CAN. Based on the results from Ewing tests, subjects were then further divided into two groups: diabetic patients with CAN (CAN+ group) or without CAN (CAN- group). Fasting glucose, insulin and c-peptide were measured. HO- MA-B and HOMA-IR were calculated. Results: The positive rate of CAN in newly diagnosed type 2 diabetic pa- tients was 22.2%. Compared with CAN- group (n=70), CAN+ group (n=20) had significantly lower fasting insulin, c-peptide and HOMA-B [median (interquartile range): 5.34(3.50~8.57)mu/L VS 8.75(4.92~13.87) mu/L, P=0.045; 0.51(0.39~0.65)nmol/L VS 0.75(0.51~0.98)nmol/L, P=0.14; 15.75(9.52~36.36) VS 32.28(18.67~61.04), P=0.012, respectively]. No sig- nificant difference was observed in HOMA-IR between groups. Fasting c- peptide was correlated with valsalva ratio(r=0.24, P= 0.043) and the lying to standing test (r=0.26, P=0.023). Logistic regression analysis revealed that HOMA-B was an independent protective factor of CAN in this population (OR: 0.97, 95%CI: 0.94~0.99, P=0.027). Conclusion: Prevalence of cardiovascular autonomic nerve is high in newly diagnosed type 2 diabetic patients. Decreased beta-cell function is associated with CAN in this population. Supported by: STPPGDP, China (2006A36001001) 1124 Autonomic dysfunction is associated with loss of postprandial glycaemic control in newly diagnosed type 2 diabetes patients T.F. Dejgaard1, J. Fleischer2, S.L. Cichosz2, P. Hoeyem2, E. Laugesen2, P.L. Poulsen2, J.S. Christiansen2, L. Tarnow2,3, T.K. Hansen2; 1Steno Diabetes Center, Gentofte, 2Institute of Clin. Medicine, Aarhus University and Dept. of Endocrinology and Internal Medicine, Aarhus University Hospital, 3Nordsjaellands Hospitaler, Hilleroed, Denmark. Background and aims: Maintaining tight blood glucose control by minimiz- ing the magnitude and frequency of glucose excursions and especially post- prandial glucose control is essential in achieving an ideal glycaemic control. Glycaemic variability may play a significant role in the development of dia- betic neuropathy e.g. through activated oxidative stress. The aim of the pre- sent study was to investigate the association between autonomic dysfunction and postprandial glucose control in newly diagnosed type 2 diabetes patients. Materials and methods: A total of 81 patients diagnosed within the last 5 years with type 2 diabetes according to WHO-criteria were consecutively re- cruited for this observational cohort study. All patients were equipped with a continuous glucose monitor (CGM) for three days and performed self-moni- tored blood glucose measurements (SMBG) at least four times per day during this period. Afterwards they were all tested for cardiac autonomic neuropathy (CAN) using conventional cardiovascular reflex tests (response to standing, deep breathing and Valsalva manoeuvre) and divided according to the pres- ence of CAN. Postprandial glucose measures were calculated using CGM for each group based on the time of SMBG before main meals. Results: The cohort consisted of 38 women and 43 men with non-insulin- treated type 2 diabetes, a mean age of 58 (±11) years, a BMI of 30 (±4), and an HbA1c of 6.6 (± 0.7) %. CAN was present in 22 subjects (9 women and 13 men). The groups with and without CAN were similar with respect to blood pressure, HbA1c, cholesterol levels and had comparable smoking habits. The nocturnal glucose drop and dawn glucose levels were significantly higher in the group with CAN as compared with patients without CAN (7.4 versus 6.5 mmol/L, P=0.017) and (7.9 versus 7.2 mmol/L, P=0.045). The breakfast premeal glucose was not different between patients with or without CAN, but peak glucose after breakfast was significantly higher in patients with CAN (11.4 versus 9.7 mmol/L, P=0.009), and the autonomic dysfunction group had significantly larger excursions 0.5-1.5 hour post meal (P S 461 1 C standing the complex interaction between postprandial glucose excursions and the development of autonomic neuropathy. Clinical Trial Registration Number: NCT00674271 1125 Artery stiffness parameters in patients with cardiovascular autonomic neuropathy and type 2 diabetes mellitus V.A. Serhiyenko, V.B. Segin, S. Ajmi, A.A. Serhiyenko; Endocrinology, National, Lviv, Ukraine. Background and aims: Artery stiffness parameters are known as a poten- tially applicable atherosclerotic risk marker and strongest predictors of car- diovascular morbidity and mortality. This study was aimed to investigate the parameters of arterial stiffness in patients with type 2 diabetes mellitus (DM) and cardiovascular autonomic neuropathy (CAN). Materials and methods: The study involved 65 patients with type 2 DM of age 54.7±3.8 years, median BMI 28.5±0.9 kg/m2 and HbA1c level 7.4±0.6%, among them 12 patients without cardiovascular diseases (CVD) and CAN, 14 patients with mild, 18 - moderate and 21 severe CAN. Control - 12 healthy volunteers. The diagnosis of CAN was based on the results of 5 standard cardiovascular tests, 24-hour HRV. The 24-hour blood pressure profiles, aorta (AIxao) and brachial augmentation index (AIxbr), pulse wave veloc- ity (PWV) and ambulatory arterial stiffness index (AASI) were assessed by TensioMedTM Arteriograph 24 (Hungary). Statistics: ANOVA. Results: We found that development of mild CAN in patients with type 2 DM is associated with statistically significant increase of artery stiffness indices (AIxao +22.1%, AIxbr +41.9%, AASI +16.7%, p S 462 1 C ticipants in the lowest quartile of serum ficolin-3 had a significantly in- crease risk of DPN compared with those belongs to the highest quartiles. Compared with Quartile1(referent), patients in Quartile2(OR, 2.76; 95%CI, 1.56-4.88; P S 463 1 C amined whether excessive GlcN-Hex flux influences on Schwann cell survival in vitro and peripheral nerve function in vivo. Materials and methods: Schwann cell line (IMS32) maintained in DMEM supplemented with 10% FBS in 5% CO2 was stimulated with 10 mM GlcN. Western blot was conducted for the evaluation of apoptosis to identify cleaved caspase3 (CC3) and for the expression of modified glycosylation (O- GlcNAC). Hex activation was sepraretely examined in the fractionated mito- chondoria and cytosole of IMS32 exposed to GlcN. Specific siRNA for Hex I was applied to see the effects of the Hex I on GlcN-Hex pathway on Schwann cell survival. To see the effects of GlcN on Hex expression in the peripheral nerve in vivo, male C57Bl/6J mice, 12 weeks of age were administered (i.p.) with 2 g/kg GlcN and nerve conduction velocity was measured. A half of the animals were pretreated with inosine (100 mg/kg, i.p.), as an ATP donor, one hour prior to GlcN administration. Normal control mice were given saline alone to serve for comparison. Results: In vitro experiment revealed that GlcN exposure elicited increased CC3 expression in Schwann cells. Concurrently, O-GlcNAc expression was increased with GlcN exposure, but it was not proportional to the extent of apoptosis. Hex I expression in mitochondria fraction was markedly increased after GlcN exposure irrespective of glucose concentration. Knock down of Hex I by siRNA suppressed the expression of GlcN-innduced CC3 activation and caused prolonged cell survival. Mice injected with GlcN exhibited en- hanced Hex expression of mitochondrial fraction and delayed nerve conduc- tion, which was prevented by pretreatment with inosine. Conclusion: GlcN exposure induced Schwann cell death with Hex activa- tion in vitro and nerve conduction delay in vivo. Inhibition of Hex activation by siRNA in vitro and inosine pretreatment significantly inhibited GlcN- induced Schwann cell death or neuropathic changes. These findings suggest that GlcN-Hex activation may have a pathogenetic role in diabetic neuropa- thy. PS 098 Mechanisms and outcomes of neuropathy 1130 Glycaemic variability as measured by continous glucose monitoring and diabetic cardiovascular autonomic neuropathy in patients with type 2 diabetes J. Ji-Eun, O. Sewon, J. Sang-Man, K. Jae Hyeon; Endocrinology and Metabolism, Samsung Medical Center, Seoul, Republic of Korea. Background and aims: It is still not clear whether glycemic variability con- tributes to the development of diabetic autonomic neuropathy. We investi- gated the association between glycemic variability determined by continuous glucose monitoring (CGM) and diabetic cardiovascular autonomic neuropa- thy in the patients with type 2 diabetes. Materials and methods: A total of 184 patients with type 2 diabetes with estimated glomerular filtration rate of >30ml/min/1.73m2 who wore a CGM device for 72 hours were enrolled. As parameters of glycemic variability, the standard deviation (SD) around the mean and the coefficient of variation (CV, SD/mean) were obtained from the CGM data. All subjects were inves- tigated for five standard cardiovascular reflex tests according to the Ewing’s protocol. Diabetic cardiovascular autonomic neuropathy was defined as the presence of at least two abnormal tests or autonomic neuropathy points ≥2. Results: Among the enrolled patients, 97 had diabetic cardiovascular au- tonomic neuropathy (52.7%, neuropathy group). SD (49.76±19.28 vs 46.05±21.24, P=0.22) and CV (28.63±8.24 vs 28.33±9.88, P=0.81) did not differ between neuropathy and control group. The hemoglobin A1c (HbA1c, %) was higher in neuropathy group ( 8.43±1.48 vs 7.95±1.32, P=0.02) than in the control group. The duration of diabetes was longer in neuropathy group than in control group (12.49±7.52 vs 6.72±6.72 year, P=0.048). In binary lo- gistic regression analysis, CV and SD failed to contribute to the final model. Conclusion: We failed to show an independent association between glycemic variability as measured by CGM and diabetic cardiovascular autonomic neu- ropathy in the patients with type 2 diabetes. 1131 Effects of diabetic peripheral neuropathy on the use of vision during walking J.C. Handsaker1, S.J. Brown1, C.N. Maganaris2,1, F. Bowling3, D.E. Marple-Horvat1, A.J.M. Boulton3, N.D. Reeves1; 1Manchester Metropolitan University, 2Liverpool John Moores University, 3University of Manchester, UK. Background and aims: Due to a decrease in sensory feedback, patients with diabetic peripheral neuropathy (DPN) are expected to use vision as a way of detecting foot position during walking, restricting their ability to visually identify upcoming obstacles. By examining how people visually acquire (first look at) targets during walking, we can elucidate to what extent diabetes pa- tients use their eyes to identify foot position, and plan subsequent stepping during walking. Materials and methods: Twelve participants (4 with diabetic peripheral neu- ropathy [DPN], 4 with non-neuropathic diabetes [D], and 4 healthy controls [C]) negotiated a stepping walkway, stepping on irregularly placed targets as accurately as possible whilst walking at a natural gait velocity. The timing of horizontal eye movements during stepping were measured using an eye- tracking device, with respect to foot-target contact. Mean group differences were analysed using a one-way ANOVA (p S 464 1 C an attempt to increase stepping accuracy. Patients with neuropathy also take longer to look from one target to another, which may indicate the presence of motor neuropathy in the extra-occular muscles. The increased time to look between targets may therefore be an explanation for the later acquisi- tion of the subsequent target. This visual gaze strategy may restrict the ability of patients with neuropathy to identify upcoming obstacles, increasing the chances of tripping and falling. Supported by: EFSD 1132 Influence of cardiovascular autonomic neuropathy on the next fate of type 1 diabetic patients; 10 years follow up S. Lacigova, J. Brozova, D. Cechurova, J. Tomesova, Z. Rusavy; Medical Dept.I., University Hospital, Pilsen, Czech Republic. Background and aims: Cardiovascular autonomic neuropathy (CAN) is considered to impair morbidity and mortality of diabetic patients (DM). We did not believe in pessimistic results of some studies. Aim of our observation- al retrospective study was to answer to these questions: 1. In what parameters differed DM with CAN (CAN+) and without CAN (CAN-) at baseline of observation? 2. Did the presence of CAN impair morbidity in next 10 years? 3. Was there the higher mortality among CAN+? What were the reasons of mortality? Materials and methods: During the years 2002-2004 we evaluated CAN in 278 type 1DM (Ewing battery). CAN+ involved 111, CAN- involved 167 pa- tients. In spite of positive test, majority of patients were asymptomatic. We have compared the control of DM, weight, presence of microangiopatic com- plications (retinopathy, nephropathy, peripheral neuropathy, diabetic foot) and cardiovascular disease and other risk factors (hypertension, hyperlipo- proteinemia and smoking) between groups (CAN+ vs CAN-) at baseline and 10 years after. Median (interquartil range), Wilcoxon non paired test, contin- gency table, mono- and multifactorial regression and multivariate Regression step-wise were used for statistical evaluation. Results: 1. Group CAN+ was older (47 vs 33 years; p S 465 1 C 1135 Insulin resistance associated interhemispheric coordination deficits in type 2 diabetes mellitus patients: a resting-state study W. Xia, S. Wang, P. Wang, Y. Yang, R. Huang, R. Cai, H. Sun; Affiliated Zhongda Hospital of Southeast University, Nanjing, China. Background and aims: Type 2 diabetes mellitus (T2DM) has an inextri- cable link with mild cognitive impairment (MCI) and Alzheimer’s disease (AD). T2DM shares several same pathogenesis with AD as well as MCI, which might be mediated by insulin resistance. This study aims to investigate whether decreased interhemispheric coordination exists in T2DM patients using resting-state functional magnetic resonance imaging (rs-fMRI). If so, we further examine whether interhemispheric coordination deficits are as- sociated with the insulin resistance among these patients. Materials and methods: We compared the interhemispheric resting state functional connectivity of 32 T2DM patients and 30 age-, sex-, and educa- tion-matched healthy controls using rs-fMRI, computed using a recently proposed measurement named ‘‘voxel-mirrored homotopic connectivity (VMHC)’’. Pearson’s correlation coefficients were measured to detect the rela- tionship between rs-fMRI information and clinical data. Results: Compared with the healthy controls, the patients showed significant decreases in VMHC in several brain regions within default mode network (DMN). Additionally, the VMHC values in the middle temporal gyrus (MTG) and superior frontal gyrus were found to be correlated inversely with the score of the Trail Making Test-B (r = -0.404, p = 0.027; r = -0.544, p = 0.002, respec- tively) among the T2DM patients. Most importantly, log homeostasis model assessment-insulin resistance (HOMA-IR) had negative correlations with the VMHC values in the MTG among patients (r = -0. 528, p = 0.003). Conclusion: Patients with T2DM suffer disturbed interhemispheric correla- tion in several DMN regions, especially in the MTG, which is associated with insulin resistance. We suppose insulin resistance might play a not negligible role in the cognitive dysfunction in T2DM through damaging the MTG func- tion. Clinical Trial Registration Number: ChiCTR-ONRC-13003095 Supported by: NSFC 1136 Increased grey matter volume loss in diabetic neuropathic subjects with depression S. Tesfaye1, R. Gandhi1, I.D. Wilkinson2, D. Selvarajah3; 1Academic Department of Diabetes and Endocrinology, Sheffield Teaching Hospitals NHS Foundation Trust, 2Academic Department of Radiology, 3Department of Human Metabolism, University of Sheffield, UK. Background and aims: Using functional magnetic resonance imaging (fMRI), we have demonstrated increased neuronal activation within regions of the brain involved in emotional pain processing. We have also recently reported a significant volume loss in somatosensory cortex peripheral grey matter in patients with both painful and painless diabetic peripheral neu- ropathy (DPN). As several studies have strongly linked chronic mood disor- ders (depression and/or anxiety) that are very common in patients with DPN, with brain atrophy, we investigated the relationship between brain volumes and mood disorders in subjects with DPN. Materials and methods: 24 subjects with type 1 diabetes [No-DPN (n=8), Painful DPN (8) and Painless DPN (8)] underwent detailed neurophysiologi- cal assessments to quantify severity of DPN [NIS(LL)+7tests]. All subjects underwent volumetric (0.8x0.8x0.8mm3 resolution) brain MRI at 3T. Images were analysed using FSL (fMRIB, Oxford). Symptoms of depression were as- sessed using the Hospital Anxiety and Depression Scale (HADS-D). Results: There were no significant differences in age (p=0.69) and dura- tion of diabetes (p=0.57) between groups. As expected subjects with pain- ful [9.6(6.6)] and painless [10.2(7.3)] DPN had significantly greater Neu- ropathy Composite Scores compared the No-DPN group [1.85(1.3); p=0.02]. Moreover, subjects with painful DPN had significantly greater depression (HADS-D; p S 466 1 C out neuropathy. Most important, significantly more male patients (25.1%) than female (20.8%) reported DN complications - foot ulcers, gangrene and amputations (p S 467 1 C 13% (to 84.4% of patients on the Diabetes register). In 2012/13, 17% of pa- tients were classified as moderate or high risk, which requires regular com- munity podiatry input. However, the actual patient activity indicated that only 8% of such patients with diabetes were seen - a gap of over half. Based on population studies of prevalence of foot risk, it would have been expected that as much as 20-30% diabetes individuals would have a moderate or high classification score. A greater disparity between hospital activity and Primary Care was seen in active diabetes foot disease. Only 0.6% of patients were coded as active, while 4% of patients were actually being seen in the two local acute Trusts’ hospital foot clinics - a 6-fold difference. Conclusion: The DMI found that only one in four patients identified to have moderate and high foot risk currently receives the desired foot follow-up. Furthermore, there was a substantial difference in active diabetes foot burden as perceived by Primary Care to that in fact seen at hospital clinics. Incorrect coding and risk stratification may have considerable impact on future com- missioning of diabetes foot services as well as personnel. Improvements in coding can potentially lead to more efficient and responsive foot health ser- vices. The challenge of non-integrated care for diabetes foot patients has been met by DMI by bringing together all healthcare providers and commissioners to devise a joint strategy across sectors. This includes a locally adapted traffic light risk screening tool, structured referral forms and standardised clinical letters, now in ubiquitous use across the host CCGs. Criteria for bi-direction- al patient referral and discharge across community and hospital settings have now been implemented. Supported by: DMI 1140 Barriers to inpatient diabetic foot examination A.J. Welch, A.I. Eyre-Brook, E.J. Rogers, P. Verlekar, R.J.C. Guy; Diabetes, Hampshire Hospitals NHS Trust, Basingstoke, UK. Background and aims: Despite NICE guidelines and Putting Feet First, only 38% of patients with diabetes receive a foot assessment in England and Wales within 24 hours of admission (National Diabetes Inpatient Audit 2013). A lo- cal audit has demonstrated an increase in foot examinations from 13 to 53% following education and introduction of a diabetic foot sticker in the acute medical admissions proforma. Our aim was to identify the major barriers to diabetic foot examination and to explain why compliance was so low. Materials and methods: A self-administered questionnaire was designed to reflect knowledge, training, practice, confidence and barriers to diabetic foot examination. The survey was given to 110 doctors, 4 diabetic specialist nurses and 3 medical students involved in acute medical admissions. Results: The response rate was 71%: 51% were doctors in training, 30% con- sultants and 9% medical students/diabetic nurses. Although 96% thought foot examination important, only 56% would do this routinely. The other 44% required a further clinical indication before they would perform an ex- amination, such as a diabetic foot ulcer or sepsis. The greatest barriers to examination were: time constraints (86%), occlusive dressings (78%), for- getting (69%), lack of equipment (64%), lack of training (56%) and patients being too ill (41%). Foot odour was a problem for 16% and foot phobia in 11%. 69% admitted to examining less than half of their admissions and 94% examined less than half of patients‘ footwear. This was despite 60% highlight- ing poor footwear as one of the greatest risk factors for ulceration. 31% of respondents have had no formal foot examination training, including 44% of all consultants. The preponderance (45%) of training was undertaken at medical school, the quality of which was only rated as good or excellent by 39%. Little teaching was organised on the wards (18%), in hospital (14%) or online (6%). Confidence was expressed in diagnosing vascular disease (86%) and neuropathy (84%), but staff lacked confidence with acute Charcot joints (73%), neuro-ischaemic ulcers (52%), acute osteomyelitis (49%) and neuro- pathic ulcers (47%). Staff thought the greatest risks for foot ulceration were: poor blood sugar control (80%), ill-fitting footwear (60%), peripheral neu- ropathy (54%) and peripheral vascular disease (49%). Respondents grossly overestimated the number of patients presenting with a foot ulcer leading to a new diagnosis of type 2 diabetes (median 15%). However, the prediction of amputation rates in diabetic patients with an ulcer over a 20 year period was more accurate: median 20%. Conclusion: Despite awareness that diabetic foot examination is important, there continues to be unwillingness to perform this crucial examination un- less clinical indications are present. Lack of time, barrier dressings, forgetting, equipment shortages and poor training account for most incomplete exami- nations. Staff lack confidence in their ability to diagnose acute Charcot joints and osteomyelitis and a significant number, especially consultants have had no formal training. Simple local interventions have the potential to signifi- cantly increase compliance with foot examinations. Regular hospital-based teaching should be made more widely available with refresher courses for consultants. Increasing resources, improving access to equipment, encour- agement to remove occlusive dressings and reminders in admission booklets are all fundamental to improving the foot care of diabetic patients admitted to hospital. 1141 Nonvisual foot examination for people with visual impairment A.S. Williams1, B.D. Caldwell2, J.S. Kawalec2; 1Nursing, Case Western Reserve University, Cleveland, 2Kent State University College of Podiatric Medicine, Independence, USA. Background and aims: Because people with both diabetes and visual im- pairment have high risk for ulcers and amputation, prevention is a top pri- ority. Usual care in diabetes self-management education is to teach visually impaired people to seek sighted assistance for regular foot examination, yet clinical experience suggests that this advice is seldom heeded. One possible solution is to teach use of the nonvisual senses for foot self-examination. The purpose of this pilot study was to compare the efficacy, acceptability, and fea- sibility of nonvisual foot examination, using the senses of touch and smell, with usual care (examination of feet by a sighted person). Materials and methods: Fifty-two visually impaired adults with type 2 dia- betes received comprehensive diabetes self-management education, with the experimental and comparison groups receiving different foot examination instructions. The experimental group was taught nonvisual foot examina- tion, i.e., use of their own hands for a systematic tactile inspection of feet for cuts, swelling, irregularities, or warmth, and use of the nose to detect unusual odors. The comparison group received usual care instructions for visually im- paired persons, i.e., to have someone with good eyesight to check their feet. All kept large print or tactile diaries of how frequently foot examination was done at home and by whom; this was reported monthly to the research team in a phone call. All had baseline podiatric evaluations and follow-up evalua- tions at 3 and 6 months, with extra visits if medically necessary. Participant- reported foot symptoms and podiatrist-confirmed new foot problems were recorded at each visit following the baseline evaluation. Focus groups were held after the 6 month visit. Results: Both the experimental and comparison Groups had their feet checked a mean of about 22 times per month, with no significant difference between groups(p = .69). However, people in the experimental group fol- lowed the instructions they were given (i.e., checked their own feet) much more frequently (M = 1.03) than people in the comparison group did so (i.e., had someone else check their feet) (M = .56), F(1, 45) = 41.42, p < .001. This result is consistent with qualitative information gathered through the focus groups. Many in the experimental group reported checking their own feet frequently, as they were taught, with occasional backup from sighted people. In contrast, many in the comparison group reported being reluctant to fol- low the instructions they had been given and ask for help checking their feet. They said they checked their own feet „as best they could“, using whatever vision they had or using self-invented tactile methods, and got help when it was not too embarrassing or difficult. At podiatrist visits, people in the experimental group reported a greater number of symptoms per visit (M = Diabetologia (2014) 57:[Suppl1]S1–S564 S 468 1 C 1.61) than people in the comparison group (M = 1.13), F(2, 50) = 64.07, p < .001, indicating a greater effectiveness of their home foot check methods. Conclusion: Visually impaired people in this study who were advised to ask someone with good eyesight to check their feet did not do so regularly. Many invented their own method of self-checking in whatever way they are able. These invented methods were not as effective for detecting symptoms of foot problems at home as systematic nonvisual foot examination. Systematic methods of nonvisual foot examination should be routinely taught to visually impaired people with diabetes. Supported by: NIH R21 NR012513 1142 Relationship between gait alterations and microvascular chronic complications in type 2 diabetic patients E. Iacopi1, R. Giannarelli1, A. Coppelli1, D. Martelli2, G. Lamola2, M. Venturi2, C. Chisari2, S. Del Prato1, A. Piaggesi3; 1Department of Endocrinology and Metabolism, 2Department of Neuro-Rehabilitation, 3Diabetic Foot Section - Department of Endocrinology and Metabolism, University of Pisa, Italy. Background and aims: We evaluated gait alterations and their correla- tions with microvascular chronic complications in type 2 diabetic patients (T2DM). Materials and methods: Thirty-six T2DM (M/F: 27/9; age: 63±10 yrs; dia- betes duration: 12±11 yrs; BMI: 29.2±5.6 Kgm2; HbA1c 8.1±0.9%) attending our outpatient clinic, were divided into 3 Groups: Group 1 (n=12) with no diabetic neuropathy (DN) and foot ulcerations (FU); Group 2 (n=10) with DN and no FU; Group 3 (n=15) with DN and non-infected, non-ischemic FU. We analyzed biomechanical alterations of lower limbs by motion analysis system (BTS Elite Clinic, BTS Bioengineering, Milan, Italy). Spatial-temporal and kinematics data were collected through photogrammetric infrared cam- eras while kinetics data with two forces plates. Data were correlated with pa- tients‘ microvascular chronic complications; in particular, all patients were examined with indirect and direct retinoscopy and two non-stereoscopic 45° retinal photographs for each eye. The presence and severity of retinopathy (DR) was determined according to the Eurodiab Study classification. Results: Step Width (SW) was greater in Group 2 (240.9±47.5 mm) and Group 3 (271.6±41.7 mm, p S 469 1 C [163 (76-246) vs 254 (165-366)PU, p=0.001]. Additionally, we found nega- tive correlation between THmax and CRP concentration (Rs=-0.33, p=0.03), TG concentration (Rs=-0.46, p=0.002) and skin AF (Rs=-0.32, p=0.04) and positive correlation between THmax and HDL cholesterol level (Rs=0.58, p S 470 1 C PS 100 Management of foot ulcers 1147 A longitudinal cohort of people presenting with diabetic foot ulcers in northern England N. Holman1, P.J. Chadwick2, J. McAdam2, S. Haycocks2, B. Young3; 1Institute of Cardiovascular and Medical Sciences, University of Glasgow, 2Podiatry Department, Salford Royal Foundation NHS Trust, 3Diabetes and Endocrinology, Salford Royal Foundation NHS Trust, UK. Background and aims: It is known that people with diabetic foot ulcers have high morbidity and mortality but there is little contemporary data ‘real world’ data on this group of people outside specific research trials. This population based study examines the characteristics and outcomes of people with inci- dent diabetic foot ulcers in a longitudinal cohort of people presenting with diabetic foot ulcers in Salford in the north of England. Materials and methods: Since 2001 data on all people with an incident dia- betic foot ulcer in Salford has been recorded in an electronic clinical system as part of routine clinical practice. This database has been linked with death registration records to identify all deaths up to June 2013. Regression models were created to examine factors associated with ulcer healing, ulcer recur- rence and mortality. Results: Between 2001 and 2012 there were 8028 incident cases of diabetic foot ulcers among 2937 people. There has been no change in age at pres- entation (mean 68.6 years, median 70 years) or the proportion of patients that were male (59.4%). 2.4% of incident ulcers were deep enough to involve the bone, 24% were accompanied by cellulitis and 32.7% were in people with peripheral vascular disease. The proportion of people with a previously healed foot ulcer reduced from 67.2% in 2001 to 44.6% in 2012 (p S 471 1 C Conclusion: These findings indicate that baPWV is closely associated with the healing time of diabetic ulcers. We suggest that the measurement of sys- temic arterial stiffness may help to identify ulcers at risk of poor healing in chronic diabetic foot ulcers in patients with type 2 diabetes. 1150 The association between Neuropad testing with foot ulceration in diabetes N. Tentolouris1, N. Papanas2, G. Panagoulias1, D.S. Tesic3, Z. Kamenov4, C. Manes5, S. Bousboulas6, E. Jude7; 11st Department of Propaedeutic and Internal Medicine, Athens University Medical School, Laiko General Hospital, 22nd Department of Internal Medicine, Democritus University of Thrace, Alexandroupolis, Greece, 3Medical Faculty, University of Kragujevac, Serbia, 4Medical University, Sofia, Bulgaria, 5Diabetes Centre, General Hospital Papageorgiou, Thessaloniki, 63rd Department of Internal Medicine and Diabetes Centre, General Hospital of Nikaia Agios Panteleimon, Piraeus, Greece, 7Diabetes Centre, Tameside General Hospital, UK. Background and aims: Foot ulceration in patients with diabetes is a serious complication associated with increased morbidity, mortality and healthcare cost and is the main cause of amputation. The prevalence of foot ulcers is 4% to 10%, and the annual population-based incidence is 1.0% to 4.1%. Pre- vention of foot ulceration and consecutively amputation begins with identi- fication of those at risk. Well-established risk factors for foot ulceration are previous foot ulceration and lower extremity amputation, long duration of diabetes, poor glycemic control, severity of diabetic neuropathy, foot deform- ities and visual impairment. Cross-sectional data have shown that dryness of the skin of the feet assessed by either sympathetic skin response or Neuropad testing has been associated with foot ulceration in patients with diabetes. In addition, Neuropad testing has a high performance for the diagnosis of dia- betic peripheral neuropathy and is proper for self- testing. The aim of the pre- sent prospective multicenter study was to examine the association between Neuropad testing with foot ulceration in patients with diabetes. Materials and methods: A total of 308 patients with diabetes (155 females and 153 males; 280 with type 2 diabetes; mean age 62.8 ± 11.3 years; mean diabetes duration 12.4 ± 9.7 years) with no history of foot ulceration were recruited in the study from the year 2005 until the year 2012. At baseline par- ticipants were evaluated for neuropathy status using the neuropathy disability score (NDS). Patients with NDS 0-2 were considered as having no neuropa- thy, those with NDS 3-5 as having mild neuropathy and those with NDS ≥6 as having severe neuropathy. In addition Neuropad testing was performed and the results were evaluated as normal or abnormal based on complete colour change of the test after 10 min of application. Results: At baseline, 148 patients (48.1%) did not have neuropathy, 82 (26.6%) had mild neuropathy and 78 (25.3%) had severe neuropathy. Neuro- pad testing was normal in 128 (41.6%) and abnormal in 180 (58.4%) patients. The mean follow-up was 5.5 ± 2.5 years. During this time, 55 (17.9%) pa- tients developed foot ulcers. At baseline, patients who developed foot ulcers were older (p=0.03) and had longer diabetes duration (p=0.01). After adjust- ment for age, gender and duration of diabetes, abnormal Neuropad testing at baseline was associated with increased odds (OR, 95% confidence intervals) for foot ulceration [4.2 (1.8-9.8)]. Similarly, the adjusted OR of NDS≥6 vs. NDS S 472 1 C Conclusion: Overall, significantly more isolates are reported from tissue samples than swab samples in patients with infected diabetic foot ulcers. This has potential implications for choice of sampling technique in practice. Clinical Trial Registration Number: ISTRCN 52608451 Supported by: NIHR HTA Programme 1153 Conservative management of neuropathic heel ulceration with calcaneal osteomyelitis and avulsion fracture in a cohort with diabetic foot disease R. Fikri1, V. Bravis1, W.M. Gedroyc2, P. Rosenfeld3, R.A. Gibbs4, D.G. Samarasinghe5, N. Oliver1, J. Valabhji1; 1Department of Endocrinology, Diabetes and Metabolism, 2Department of Radiology, 3Department of Orthopaedics, 4Department of Vascular Surgery, 5Department of Microbiology, Imperial College Healthcare NHS Trust, London, UK. Background and aims: The complex of neuropathic heel ulceration, calca- neal osteomyelitis (OM) and calcaneal fracture in people with diabetes is rare and challenging to manage, especially when complicated by tendon avulsion. We describe 4 patients in whom a conservative non-operative approach re- sulted in limb salvage. Materials and methods: All patients were managed in the multidisciplinary diabetic foot clinic. Specialist input was provided by diabetologists, podia- trists, radiologists, microbiologist, vascular surgeons, orthopaedic surgeons, neurologist and orthotists. Multidisciplinary management included, where appropriate, antibiotics, wound debridement, vascular intervention, vacuum pump therapy and heel offloading. Patients were closely monitored clinically and radiologically for wound healing, resolution of OM and preservation of ankle function. Results: Four patients were included. Their demographics and management are summarised in table 1. Two subjects had multidrug resistant organisms identified. Resolution of complex heel ulceration with calcaneal fracture and tendon avulsion is ongoing with 9 (6-11) months of follow up to date and a similar duration of antibiotics (9 (8-9) months). Vascular surgical interven- tion was required in 2 out of 4 subjects and adjunctive vacuum pump dress- ings were used in all subjects. Ankle function was at least partially preserved in all subjects. Conclusion: This is the first description of a cohort of patients with diabetes, heel ulceration and calcaneal OM complicated by an avulsion fracture man- aged conservatively. Vacuum pump therapy was applied, not only to soft tis- sue, but also to bone, with progressive wound healing. Despite tendon avul- sion, 3 of 4 patients have preserved ankle function and there were no minor or major amputations. Conservative management so far has been successful despite the presence of multiple resistant organisms in 2 of the 4 subjects. This complex can occur following an ulcer or trauma, requires prolonged therapy and has serious physical, psychosocial and biomechanical implica- tions. There is no consensus on conservative or surgical management and functional limb salvage is often unachievable resulting in a major amputa- tion. Whilst this is a small case series, this challenging diabetes foot problem is too rare to examine by randomised controlled trial. These 4 cases, managed successfully by a highly skilled multidisciplinary team provide a template for management. 1154 The efficacy of removable offloading devices to heal plantar foot ulcers in diabetic patients J.J. Van Netten1, J.G. Van Baal1, A. Kottink1, M. Hutten1, E. Manning1, M. Spraul2, S.A. Bus1,3; 1Department of Surgery, Hospital Group Twente, Almelo, Netherlands, 2Matias Spital, Rheine, Germany, 3Academic Medical Center, Amsterdam, Netherlands. Background and aims: Adequate offloading is required for healing neuro- pathic plantar foot ulcers in patients with diabetes. Guidelines recommend non-removable offloading as primary treatment for these ulcers. However, in clinical practice removable offloading devices are more commonly used, mainly for practical reasons. The aim of this study was to assess the efficacy of three commonly used removable offloading devices on success in healing of plantar neuropathic foot ulcers in diabetes. Materials and methods: A total 60 diabetic patients (48 male, mean age 62.5 years, 87% type 2) with a neuropathic non-infected, non-ischemic plantar foot ulcer were randomized to one of three treatment modalities: a bivalved total contact cast (BTCC), a Mabal Cast shoe (MABAL), or a forefoot offloading shoe (FOS). Patients were followed until healing or until 20 weeks, whichever came first. Primary outcomes were percentage healing in 12 and 20 weeks time. Results: Foot ulcers were located at the hallux (n=24), first metatarsal head (n=21), other metatarsal heads (n=13) and toes (n=2). Forty-nine of the 60 foot ulcers were classified as small (< 2 cm2), ), 11 as large (> 2 cm2). Healing rates within 12 and 20 weeks according to intention-to-treat were 60% and 63% for BTCC, 60% and 83% for MABAL, and 70% and 80% for FOS (non- significant between conditions, p=0.703 and p=0.305 for 12 and 20 weeks respectively). Per protocol healing rates were 57% and 69% for BTCC, 67% and 87% for MABAL, and 77% and 88% for FOS (non-significant between conditions, p=0.519 and p=0.374 for 12 and 20 weeks respectively). Conclusion: Healing rates were not significantly different between the three removable devices, but the off-the-shelf FOS condition showed higher heal- ing percentages than the two casting conditions at 12 weeks. Healing rates for the BTCC were substantially lower than previously found for non-removable total contact casts (~90% healing rates), while healing rates for the other two devices are comparable for those previously found for similar removable of- floading devices. A lack of forced adherence is suggested to be an explanatory factor in lower healing rates found compared to non-removable offloading, which stresses the importance of continuous pressure relief in ulcer healing. Supported by: Interreg III 2_EUR_II_2=60 Diabetologia (2014) 57:[Suppl1]S1–S564 S 473 1 C PS 101 Diabetic eye disease 1155 Ophthalmic biomarkers of diabetic neuropathy: does neural dysfunction precede vascular changes? M. Tavakoli, M. Ferdousi, I.N. Petropoulos, H. Fadavi, A. Marshall, A.J.M. Boulton, R.A. Malik; University of Manchester, UK. Background and aims: To evaluate structural- functional changes in the neural level of retina and cornea in type 1 diabetic (T1DM) patients without clinical evidence of retinopathy. Materials and methods: 30 T1DM patients (Age: 48±2 yrs; Duration diabe- tes: 28±3 yrs, HbA1c: 7.6±0.5 %) with no evidence of retinopathy and 15 aged matched healthy controls underwent detailed neurological (including Neu- ropathy Deficit Score (NDS)) and ophthalmic assessment including measure- ments of global and sectorial retinal nerve fibre layer (RNFL) thickness using Spectral Domain OCT (SD-OCT) (SPECTRALIS, Heidelberg Engineering) and retinal ganglion layer function (Flicker-Defined-Form (FDF)) using Hei- delberg High Edge Perimetry (HEP). Corneal nerve fibre length (CNFL) and density (CNFD) were assessed using a corneal confocal microscope (Heidel- berg HRT III) and corneal sensitivity was quantified using non- contact cor- neal aesthesiometer (NCCA). Results: There was a significant reduction in the global and sectorial RNFL (P=0.01) in diabetic patients compared to controls which correlated with the severity of peripheral neuropathy measured by NDS) (r= -0.403, P=0.01). FDF (P=0.01), CNFD (P=0.01), CNFL (P=0.01) and corneal sensitivity (P=0.01) were significantly reduced in T1DM compared to control subjects and they were significantly correlated with NDS.A greater proportion of T1DM patients showed abnormal retinal structure (RNFL-61%) and function (FDF-52%) compared to corneal structure (CNFD-32%) and function (NCCA-20%). Conclusion: Changes at the level of nerve fibre layers were observed in both the retina and cornea. However, the prevalence of abnormality was higher at the retina but the severity was more pronounced in the cornea especially in those patients without neuropathy. The results of this preliminary study suggest that structural changes in the cornea (Peripheral nerves) and in the retina (central nervous system) may occur in parallel and are correlated with functional changes. Supported by: NIH R105991 1156 Early detection of neurodegenerative changes in diabetes: systematic review and meta-analysis E.E.B. De Clerck1, J.S.A. Schouten1, T.T.J. Berendschot1, R.M.M. Nuijts1, H.J.M. Beckers1, M.T. Schram2, C.D.A. Stehouwer2, C.A.B. Webers1; 1Ophthalmology, 2Internal Medicine and Cardiovascular Research Institute Maastricht (CARIM), Maastricht University Medical Centre, Netherlands. Background and aims: The onset of diabetic neuropathy is difficult to detect by electrophysiology. The earliest damage occurs in the small nerve fibers. These fibers can be evaluated by thermal and pain perception, which have limited reproducibility, and by skin biopsy, which is invasive. A more reliable and feasible approach could be the imaging and quantification of neurode- generative changes in the diabetic eye. Aims: To systematically assess ocu- lar neurodegenerative changes in adults with type 1 or type 2 diabetes, their relationship with diabetic retinopathy and neuropathy, and whether these changes are already present in an early stage of diabetes. Materials and methods: This meta-analysis was conducted according to the PRISMA statement and the Cochrane collaboration. The databases searched were MEDLINE, the Cochrane Controlled Trials Register and EMBASE (1968-2013). Language restrictions were English, French, German or Dutch. Only studies of neurodegenerative changes in the retina, the optic nerve head or the cornea in adults (>18 years) with type 1 or type 2 diabetes mellitus were included. One author reviewed studies for inclusion, quality and risk of bias. The second author checked the extracted data. Results: 41 out of 4051 articles were included (3043 diabetic and 2021 non- diabetic individuals). 35 neurodegenerative parameters were analyzed. Com- pared with non-diabetic individuals, diabetic patients with any level of sever- ity of diabetic retinopathy, had a mean reduction in thickness of at least 2 µm in 5 retinal layers: the mean ganglion cell/inner plexiform layer in the pe- ripheral area and the inner plexiform layer, the ganglion cell layer, the mean outer nuclear layer/inner segments and the mean ganglion cell/inner plexi- form layer thickness in the pericentral area of the macula. In the optic nerve head, the highest reduction was observed for the superior retinal nerve fiber layer thickness (superior RNFL: -7.98 µm, 95%CI -12.46 to -3.50). In the cor- nea, the nerve branch density (NBD) and the nerve fiber density (NFD) were lower by 12.80/mm2 and 7.45/mm2, respectively (95% CI -14.56 to -11.04 and -9.88 to -5.01). In patients without diabetic retinopathy the pericentral outer nuclear layer and inner segment were reduced by 2.45 µm (95% CI -4.76 to -0.14). In the optic nerve head the superior and mean RNFL thickness were decreased by 3.24 and 2.37 µm respectively (95% CI -7.31 to 0.83 and -7.32 to 2.59). In the cornea the NBD (-7.77/mm2 , 95% CI -12.20 to -3.34) and the NFD (-3.57/mm2, 95% CI -7.12 to -0.02) were lower. In patients with any severity of diabetic polyneuropathy, the NBD and the NFD of the cornea were decreased by 20.84/mm2 (95% CI -23.92 to -17.77) and 16.32/mm2 (95% CI -19.26 to -13.37). Ocular changes were already present in patients with- out diabetic polyneuropathy (NBD: -9.99/mm2,95% CI -15.26 to -4.72; NFD: -7.80/mm2,95% CI -12.32 to -3.28). Conclusion: This systematic review establishes a correlation between ocu- lar neurodegenerative changes and diabetes. Structural changes are already present in diabetic individuals without diabetic retinopathy or neuropathy. Imaging and quantification of ocular neuronal tissues gives the opportunity to study diabetic neuronal changes and to early detect diabetic neuropathy. 1157 An evaluation of different methods of normalising fovea thickness measurements from optical coherence tomography instruments K.M. Gooding1,2, A.C. Shore1,2, R. Ling3, E. Agardh4; 1Diabetes and Vascular Medicine, University of Exeter Medical School, 2NIHR Exeter Clinical Research Facility, Royal Devon and Exeter NHS Foundation Trust, 3West of England Eye Unit, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK, 4Department of Clinical Sciences, Ophthalmology, Lund University, Malmo, Sweden. Background and aims: The assessment of macular thickness by optical co- herence tomography (OCT) is routinely used in clinical practice. However, interpretation is hampered by the lack of a universal clinical scale, partly due to variations in the segmental algorithms between devices. For example the Stratus, Cirrus and Topcon OCTs all measure thickness to different depths: inner/outer photoreceptors interface, mid and inner limit of the retinal pig- ment epithelium respectively; resulting in a systematic difference when com- paring measurements between devices. Previous work have proposed the log- arithmic transformation of normalised data, and then utilising step changes to define clinically significant changes. Using this approach fovea data has been normalised to fovea thickness as measured by Stratus (200µm) and mid- range normal fovea thickness as measured by spectral domain (SD) OCTs (250µm) in healthy eyes. Utilising these normalisation values may still result in a systematic difference when comparing measurements from different de- vices. Alternative approaches are to either customise the normalisation factor to each device or to normalise data to a ‘gold standard’ before log transforma- tion. As part of the SUMMIT consortium this study aims to explore different normalisation procedures for combining data from different OCT devices. Materials and methods: 43 individuals were recruited (11 controls, 32 with diabetes: 14, 15, 5 with no retinopathy, non- and proliferative retinopathy respectively). Macular thickness was assessed in all participants on both Cir- rus and Topcon-1000 OCT devices (512x128 scanning protocol). Raw data comparisons demonstrated that, as expected, Topcon measurements were consistently lower than with the Cirrus in all quadrants (mean (SD) differ- ence: 7.9(0.5)%, p S 474 1 C 1158 Longitudinal association of retinopathy and glucose metabolism in a Japanese population T. Nakagami1, K. Takahashi1, C. Suto2, A. Hirose1, Y. Uchigata1; 1Diabetes Center, Tokyo Women‘s Medical University, Tokyo, 2Department of Opthalomology, Saitama-ken Saiseikai Kurihashi Hospital, Kuki City, Japan. Background and aims: The epidemiological data for diabetic retinopathy have been extensively analyzed, however, limited studies have assessed the development of retinopathy across different glucose levels ranging from normal glucose tolerance to diabetes. Some studies have reported that retin- opathy increased the risk of vascular disease and incident diabetes. Here, we assess the 6-year incidence for retinopathy across different glucose categories and the association of retinal signs detected at baseline with development of diabetes in a Japanese population. Materials and methods: Subjects were 3,580 among 5,200 health check-up examinees (2,353 men, mean age: 51±8 years old) who underwent the test of non-mydriatic 45° digital fundus photography. The retinal and glucose status were followed up for 6 years in 181 and 3,399 individuals, with and without retinopathy, respectively, at baseline (mean follow up: 5.7 years). Individuals were classified into five categories according to fasting plasma glucose (FPG) ( S 475 1 C Conclusion: Within each of these two programmes the stratification enables categorisation of patients into those with very low of progression to STDR and those at high risk. This information could be used to inform decisions about screening intervals. Supported by: NIHR HTA 1161 Male sex, blood pressure and heart rate are associated with the risk of diabetic retinopathy in normoalbuminuric type 1 patients M. Tomić1, T. Bulum1, K. Blaslov1, S. Ljubić1, S. Kaštelan2, L. Smirčić-Duvnjak1; 1Vuk Vrhovac University Clinic Merkur Clinical Hospital, 2Dubrava Clinical Hospital, Zagreb, Croatia. Background and aims: The pathogenesis and the onset of diabetic retinopa- thy (DR) are still insufficiently understood. It is presumed that retinopathy and nephropathy occur simultaneously and that the severity of retinopathy equals that of renal disease in diabetes. The aim of this study was to assess the prevalence of retinopathy in normoalbuminuric type 1 diabetic patients and evaluate the risk factors for its development and progression in these patients. Materials and methods: A total of 223 normoalbuminuric type 1 diabetic patients (116 male / 107 female) with normal renal function (glomerular fil- tration rate ≥60 ml/min/1.73m2) were included in this study and followed for 48 months. Basic and anthropometric parameters assessed were sex, age, diabetes duration and body mass index (BMI). Glycated haemoglobin (HbA1c), HDL and LDL cholesterol, triglycerides and serum creatinine were determined using routine laboratory methods. Glomerular filtration rate (GFR) was estimated using CKD-EPI formula. Urinary albumin excretion rate (UAE) was measured from at least two 24-hr urine samples and deter- mined as the mean of 24-hr urine collections. Blood pressure was measured with a mercury sphygmomanometer and resting heart rate using a standard 12-lead ECG, after a 10-minute resting period. Ophthalmologic examination included binocular indirect slit lamp fundoscopy and color fundus photogra- phy after mydriasis of two fields (macular field, disc/nasal field) of both eyes according to the EURODIAB retinal photography methodology. Possible risk factors for the development and progression of DR were examined in back- ward stepwise Cox‘s multiple regression analysis. Results: Patients were 38.2 ± 10.3 years old with mean diabetes duration of 17.2 ± 9.1 years. Mean/median values of BMI (24 (18 - 37) kg/m2), HDL cho- lesterol (1.7 ± 0.4 mmol/L), triglycerides (1.02 ± 0.6 mmol/L), systolic (120 (80 - 180) mmHg) and diastolic blood pressure (80 (60 - 110) mmHg), serum creatinine (70.1 ± 12.2 µmol/L), UAE (9.8 (1.3 - 29.0) mg/24h) and eGFR (106.2 ± 15.1 ml/min/1.73m2) were within normal range for diabetic patients, whereas HbA1c (7.0 ± 1.4 %) and LDL cholesterol (2.8 ± 0.8 mmol/L) were slightly elevated. At baseline, 156 (70%) patients had no retinopathy and 67 (30%) had nonproliferative diabetic retinopathy (NPDR). After 48 months, 24 patients (10.7%) developed NPDR or progressed to proliferative diabetic retinopathy (PDR). From the 156 patients with no retinopathy at the begin- ning of the study 15 (9.6%) progressed to NPDR, while from the 67 patients with NPDR at the beginning of the study 9 (13.4%) progressed to PDR. Male sex (HR 2.9, CI 1.04-8.3, p=0.04), systolic blood pressure (HR 1.03, CI 1.01- 1.06, p=0.02), UAE (HR 1.14, CI 1.06-1.23, p=0.001) and resting heart rate (HR 1.03, CI 1.01-1.07, p=0.03) were significantly associated with the devel- opment and progression of retinopathy. Conclusion: The results of this study suggest that diabetic retinopathy may develop and progress in type 1 diabetic patients even without a coexisting renal disease. This points to the need for close monitoring of normoalbumi- nuric type 1 diabetic patients aimed at early detecting, preventing or limiting the progression of retinopathy, especially in men with higher UAE, systolic blood pressure and higher resting heart rate. 1162 Incidence of severe diabetic retinopathy and all-cause mortality among migrants in Denmark J. Oya, M.E. Jørgensen, B. Carstensen, G.S. Andersen; Clinical Epidemiology, Steno Diabetes Center, Gentofte, Denmark. Background and aims: Global migration has been increasing and it has been reported that migrants from non-European countries are usually diagnosed with diabetes more commonly and at younger age than European counter- parts. Consequently, they may have higher rates of diabetic complications and death. Diabetic retinopathy (DR) is a common microvascular complica- tion of diabetes and may progress to proliferative retinopathy and macular oedema if exposure of hyperglycemia and hypertension prolongs. Reports from UK and US showed that African and Asian migrants had higher preva- lence of DR and especially sight-threatening retinopathy, than UK and US born people with diabetes. It has been considered that the difference probably reflects worse glucose and blood pressure control, lower frequency of eye ex- amination and genetic background. However, longitudinal data on incidence of DR using nationwide data among migrants are lacking. Materials and methods: Nationwide clinical data from the Danish Diabe- tes Database for Adults (Dansk Voksen Diabetes Database) on >17 years old diabetes patients followed since 2005 was linked through the personal ID number with data from the National Patient Register for identification of retinopathy diagnoses, the Cause of Death Register for information on deaths and the Central Personal Register for information on country of origin. We classified patients according to country of origin into six groups: Denmark, other Europe, Sub Saharan Africa, Middle East (includes North Africa), Asia and America (includes Oceania). From 73,572 patients with complete infor- mation, follow-up of 56,148 patients (mean age 58±15 years, duration of dia- betes 8.6±9.3 years) and free from severe DR at baseline were analysed. We estimated event rates and hazard ratios (HR) for incidence of severe retinopa- thy and all-cause mortality between geographical regions. Results: At baseline, median age, duration of diabetes and systolic blood pressure were lower and median HbA1c was higher in migrants from Sub Saharan Africa, Middle East and Asia compared to diabetes patients of Dan- ish origin. During 215,565 person-year of follow-up, 6,348 patients suffered an incident of severe DR and 2,400 patients died. Compared to Danish borns, migrants from Middle East and Asia had higher risk of severe retinopathy (HR, 1.15 [95% confidence interval (CI) 1.04-1.27] and 1.24 [1.03-1.49], respectively) after adjustment for age, sex, type of clinic, body mass index, smoking status, and diabetes type, duration and control. The associations re- mained after further adjustment for blood pressure and lipids. Mortality rates were lower among those with other European, Sub Saharan, Middle Eastern, Asian or American origin compared to Danish borns after adjustment for age and sex (HR; 0.71, 0.26, 0.51, 0.55 and 0.52, all significant except America). Conclusion: Migrants with diabetes from the Middle East and Asia showed higher risk of incidence of severe DR compared to Danish borns, but substan- tially lower mortality rates. Diabetologia (2014) 57:[Suppl1]S1–S564 S 476 1 C PS 102 Mechanisms of microangiopathy: experimental 1163 Topical administration of a GLP-1 agonist prevents retinal neurodegeneration in experimental diabetes R. Simó1,2, P. Bogdanov1,2, L. Corraliza1,2, C. Solà1, C. Hernández1,2; 1Diabetes and Metabolism Research Unit, Institut de Recerca Hospital Universitari Vall d‘Hebron, 2CIBERDEM, Barcelona, Spain. Background and aims: Retinal neurodegeneration is an early event in the pathogenesis of diabetic retinopathy (DR). Therefore, treatments aimed at preventing or arresting retinal neurodegeneration has been recently pro- posed as a new strategy to reduce the incidence and progression of DR. There is evidence that glucagon-like peptide-1 (GLP-1) exerts a neuroprotective effect in the central nervous system but there is scarce information on this issue in the retina. The aim of the study was to explore the effectiveness of a GLP-1 analogue (liraglutide) administered topically (eye-drops) in prevent- ing retinal neurodegeneration in db/db mice, an experimental model of type 2 diabetes. Materials and methods: For this purpose we evaluated a total of 24 diabetic mice (db/db) aged 8 weeks that were randomly assigned to daily oral treat- ment with liraglutide (6 mg/ml) (n=12) or vehicle (n=12) for two weeks. Twelve non-diabetic mice (db/+) were used as control group. Retinal neuro- degeneration was evaluated by measuring glial activation (immunofluores- cence and Western blot) and apoptosis (TUNEL, cell count in ganglion cell layer). Functional abnormalities were assessed by electroretinography (ERG). Results: We observed that diabetic mice presented significantly higher glial activation and apoptosis than age-matched non-diabetic mice. The diabetic mice treated with eye drops of GLP-1 analogue presented a significant de- crease of both glial activation and rate of apoptosis than diabetic mice treated with vehicle. Furthermore, a significant improvement of ERG parameters was observed. Conclusion: Topical administration of a GLP-1 analogue prevents retinal neurodegeneration induced by diabetes. Our findings open up a new phar- macological strategy targeted to early stages of DR. Supported by: SAF2012-35562 and PI13/00603 1164 Early-outgrowth bone marrow cells protect the retina of experimental model of type 2 diabetes by improving SIRT-1 signalling D.A. Duarte1, K.C. Silva1, A. Papadimitriou1, M.A.B. Rosales1, E.B.M. Peixoto1, R.E. Gilbert2, J.B. Lopes de Faria1, J.M. Lopes de Faria1; 1Faculty of Medicine, State University of Campinas, Brazil, 2Keenan Research Centre of the Li Ka Shing Knowledge Institute, University of Toronto, Canada. Background and aims: Cellular therapy has been studied in cardiovascu- lar disease but has rarely been investigated in eye diseases. Sirtuins (SIRTs), a family of deacetylases, is thought to be sensitive to oxidative stress. This study hypothesized that cell therapy using early outgrowth cells (EOC) could protect the diabetic retina through antioxidant means, thus improving the SIRT1 pathway. This study investigated the possible therapeutic effects of cells derived from healthy (db/m) and diabetic (db/db) animals on diabetic retinopathy (DR. Materials and methods: Mice aged 8-weeks and db/db were randomized to receive a unique intravenous injection of PBS or 0.5 × 105 db/m EOCs or 0.5 × 105 db/db EOCs. Four weeks later, the animals were euthanized and the eyes enucleated. For in vitro study, an EOCs-conditioned medium (EOC- CM) was generated from db/m and db/db EOCs cultures. The rat Müller cells (rMCs) were exposed for 24 h to normal (NG), high glucose (HG) combined, or with neither db/m nor db/db EOC-CMs. Results: In diabetic rats, there was an increase of DR and oxidative damage markers accompanied by an increase in NOX4 expression leading to detri- ment of SIRT1 protein. This was followed by lysine-310-p65-NFκB acetyla- tion. The treatment with cells from db/m significantly reduced this damage, but the treatment with cells from db/db mice fully restored these alterations to normal levels. The rMCs exposed to HG displayed GFAP and VEGF ex- pression upregulation accompanied by an increase in NOX4 expression, ROS levels and acetyl-lysine-310-p65-NFκB. Protein expression and activity of SIRT1 were markedly reduced in diabetic milieu conditions. The treatment with both EOC-CMs prevented all these abnormalities, but db/db EOC- CM resulted in full restoration to NG conditions. The presence of EX-527, a SIRT1-specific blocker, or siRNA SIRT1 abolished the observed effects in both treatments. Conclusion: This study demonstrates that the endocrine capacity of EOCs is effective in improving the retinal SIRT1 pathway, thus protecting the retina from diabetic milieu insult. This is one evidence that regenerative therapy for DR using EOCs or its released factors could be a promising tool for patients with diabetic eye disease. Supported by: FAPESP and CNPq 1165 Attenuation of reactive gliosis in retinas by PARP-1 inhibitors treatment under diabetes in rats M. Guzyk, A. Tykhomyrov, T. Kuchmerovska; Palladin Institute of Biochemistry, Kiev, Ukraine. Background and aims: Diabetic retinopathy (DR) is a multifactorial disease, and persistent hyperglycemia appears to be a major contributor to its devel- opment. Gliosis is a hallmark of retinas neurodegeneration. The exact mo- lecular mechanisms which contribute to development of diabetes-induced retinas neuropathy are not completely understood. The study was aimed to estimate the Poly(ADP-ribose)ylation and reactive gliosis under DR and ef- fect of Poly(ADP-ribose) polymerase-1 (PARP-1) inhibitors, 3-aminobenza- mide (3-AB) and nicotinamide (NAm) on these processes. Materials and methods: All studies were carried out after 10 weeks of diabe- tes (streptozotocin, 70 mg/kg of body weight, i. p.) in male Wistar rats treated 2 weeks with or without 3-AB (30 mg/kg/day, i. p.) and NAm (100 mg/kg/ day, i. p.). The proteins content was assessed by Western blot and immuno- histochemically. Results: The levels of glial fibrillary acidic protein (GFAP) were significantly increased in retinas at diabetes as result of glial reactivity. In addition, ex- pression of immunoreactive products of GFAP degradation (lower molecular weight polypeptides, in the range from 47 to 37 kDa), another representa- tive indicator of reactive gliosis, was observed. Marked reduction of GFAP levels was seen in retinas of diabetic rats treated with 3-AB or NAm, though content of retinas GFAP did not reached to the control. Noteworthy, 3-AB and NAm administration completely prevented increase in GFAP degrada- tion. Increased GFAP immunoreactivity was observed in retinas sections of diabetic rats compared with control, which is in parallel with data obtained by Western blot. Immunohistochemically, supplementation of diabetic rats with PARP inhibitors counteracted glial activation in retinas. Intensity of glial response appeared to be correlated directly with PARP-1 activation. Western blot analysis demonstrated significant upregulation of PARP-1 expression by about 1.5 in diabetic retinas compared to nondiabetic group. At the same time PARP-1 fragmentation was exacerbated more than 2-fold in diabetic rats as compared to control. PARP-1 inhibitors administration to diabetic rats normalized PARP-1 expression in retinas. However, effect of NAm was more profound than 3-AB not only on protein expression but also on protection of PARP-1 from its fragmentation due to its antioxidant property and improv- ing of proinsulin biosynthesis through increase of NAD level that is respon- sible for the cellular ATP which is necerely for glycolysis and mitochondrial respiration. Poly(ADP-ribosyl)ated proteins (pADPs) were detected primar- ily in range from 72 to 130 kDa of the retinas protein spectrum, with several minor bands at 17 to 55 kDa. Both 72 to 130 kDa and 26 to 37 kDa pADPs expression was increased in diabetic rats as compared to control and was es- sentially prevented by 3-AB and NAm treatment. Conclusion: Thus, PARP-1 inhibitors can be potential drugs for treatment of DR through improvement of PARP-1 fragmentation and modulation of reactive gliosis that maintain homeostasis and normal neuronal function in retinas under diabetes. Diabetologia (2014) 57:[Suppl1]S1–S564 S 477 1 C 1166 p62 is not only a marker of autophagy but also a possible target against retinal Müller cells apoptosis under diabetic milieu conditions C. Montemurro, V.H. de Oliveira Amancio, D. Andreazzi Duarte, A. Papadimitriou, J.B. Lopes de Faria, J. Mendoza Lopes de Faria; Faculty of Medical Sciences, University of Campinas (Unicamp), Renal Pathophysiology Laboratory, Investigation on Complications of Diabetes, Brazil. Background and aims: Inhibition of autophagy, a catabolic process by which cells degrade intracellular components inside lysosomes, induces degenera- tive changes in mammalian tissues that resemble those associated with aging, and was described to be altered in retinal cells under diabetic conditions. Apoptosis, or programmed cell death, appears to play an important role in the pathogenesis of diabetic complications. The aim of this study was to in- vestigate the interaction between autophagy and apoptosis of retinal Muller cells under diabetic conditions in vitro. Materials and methods: Rat Muller cells (rMCs) were cultured in normal glucose (NG; 5.5 mM), high glucose (HG; 30 mM) and under oxidative stress conditions (H2O2; 1 mM) for 24, 48, 72h, with and without N-Acetyl Cysteine (NAC), autophagosome formation inhibitor, 3 Methyladenine (3MA; 1mM) and PERK inhibitor I, GSK 2606414. The markers of autophagosome forma- tion, Beclin1, and autophagosome-lysosome fusion, p62, were measured by Western blot (WB). Intracellular reactive oxygen species (ROS) were evaluat- ed by 2´-7´ dichlorofluorescein (DCFH(2)) assay and apoptosis was assessed by TUNEL and extrinsic apoptotic pathway by caspase 8 activity. Results: In rMCs, HG and H2O2 increased Beclin1 and p62 levels (p S 478 1 C 1169 Thiamine metabolism abnormalities contribute to the progression of diabetic nephropathy K. Kankova1, K. Kuricova1, V. Dvorakova1, L. Pácal1, Z. Marcanova1, J. Svojanovsky2, J. Olsovsky2, J. Belobradkova3, J. Rehorova3; 1Dept. of Pathophysiology, Masaryk University, 2Dept. of Internal Medicine, St. Anne University Hospital, 3Dept. of Gastroenterology, Faculty Hospital Brno-Bohunice, Brno, Czech Republic. Background and aims: Pentose phosphate pathway (PPP) represents poten- tially protective pathway in hyperglycaemia since it can process glycolytic intermediates and thus decrease production of reactive dicarbonyls (meth- ylglyoxal) and reactive oxygen species in mitochondria. Transketolase (TKT) is the rate-limiting enzyme of non-oxidative branch of PPP whose activity depends on thiamine diphosphate (TDP) - an active form of thiamine (vit. B1) - as a cofactor. Thiamine supplementation was shown to prevent devel- opment and progression of diabetic nephropathy (DN) in animal model of diabetes. Thiamine is delivered to the cell via specific thiamine transporters 1 (encoded by SLC19A2) and 2 (SLC19A3) and phosphorylated by thiamine pyrophosphiokinase (TPK). We have previously shown that plasma levels of thiamine are dominantly influenced by GFR, however increasing plasma thi- amine in subjects with decreased renal function are not paralleled by increase of intracellular TDP. The aim of the current study was to analyse relationship between plasma and erythrocyte parameters reflecting the thiamine status and progression of DN, cardiovascular morbidity and mortality. Additionally, since genetic variability in SLC19A2 and SLC19A3 loci may potentially affect activity of thiamine transport, common SNPs in those genes were detected and tested for their contribution to the progression of DN. Materials and methods: Prospective observational cohort study comprised a total of 273 type 2 diabetics with variable stage of DN at baseline followed for a median of 39 [IQR 21 - 59] months. Following end-points were considered: (1) progression of DN by stage, (2) major cardiovascular event (MCVE, non- fatal and fatal myocardial infarction or stroke, limb amputation, revasculari- sation) and (3) all-cause mortality. Plasma and erythrocyte TDP was detected using HPLC. TKT activity was determined by kinetic method. Genotyping of 6 SNPs (3 in the SLC19A2 locus (rs1983546, rs7522245, rs6656822) and 3 in the SLC19A3 locus (rs13025803, rs4973216, rs7567984)) was performed by RT-PCR. Time-to-event analysis was carried out to ascertain contribution of thiamine, TDP and SNPs to studied end-points. Results: Cumulative incidence of DN progression was 22.9%, CVE 8.2%, and ACM 19.8%. Significant differences in DN progression and all-cause mortal- ity were ascertained for plasma TDP tertile groups (both P0.05). Conclusion: Our results indicate that abnormalities of thiamine metabolism induced by diabetes, especially intracellular deficit of active TKT cofactor and impaired activity of TKT contribute to the progression of diabetes-associated morbidity or mortality. Elucidation of molecular mechanisms responsible for decreased intracellular availability of TDP is warranted. Supported by: IGA MZČR NT/11405 1170 Defective mitochondrial turnover in experimental diabetic nephropathy G.C. Higgins1,2, T.-V. Nguyen1, S.A. Penfold1, P.M. Robb1, G. Ramm3, K.E. White4, R.W. Bilous4, M.E. Cooper1,5, J.M. Forbes6, M.T. Coughlan1,5; 1Diabetic Complications, Baker IDI Heart and Diabetes Institute, 2Department of Biochemistry & Molecular Biology, Monash University, 3Biochemistry & Molecular Biology, Monash University, Clayton, Australia, 4EM Research Services, Faculty of Medical Sciences, Newcastle University, Newcastle, UK, 5Department of Medicine, Monash University, Melbourne, 6Glycation & Diabetes, Mater Medical Research Institute, South Brisbane, Australia. Background and aims: Diabetic nephropathy (DN) is the major cause of end stage renal disease in the Western world. Defects in mitochondrial bio- energetics are evident in DN and are thought to initiate renal impairment. Accumulation of fragmented mitochondria are found in the renal cortex in experimental diabetes, suggesting that in tandem with a shift in dynamics, mitochondrial clearance mechanisms may be impaired. The process of mi- tophagy is the selective targeting of damaged or dysfunctional mitochondria to autophagosomes for degradation through the autophagy pathway. Here, we aimed to determine if there was an impairment in mitophagy and changes in mitochondrial dynamics in the kidney in DN. Materials and methods: Markers of mitophagy and mitochondrial dynamics were followed in the renal cortex from mice rendered diabetic with the beta cell toxin streptozotocin (STZ). Tissue from the renal cortex was studied by western immunoblot for changes in mitophagy and mitochondrial dynam- ics changes, while mitochondrial bioenergetics were performed to measure mitochondrial function. Electron microscopy was also undertaken to follow changes in mitochondrial morphology, between control and diabetic popu- lations. Human renal biopsies, taken from both type 1 and control patients (taken from kidney donor patients) were studied by electron microscopy for evidence of changes in mitochondrial morphology and autophagy/mi- tophagy. Results: Bioenergetics were impaired in mitochondria isolated from renal cortex, with a decline in ATP content, increased reactive oxygen species and depolarisation of the inner mitochondrial membrane. Increased recruitment of Drp-1 to mitochondria was observed in diabetes, with a 3-fold increase in Drp-1 translocating to mitochondria (n=6, P S 479 1 C PS 103 Mechanisms of nephropathy: experimental 1171 Nrf2 overexpression inhibits mechanical stretch-induced fibronectin expression L. Gnudi, A.A. Hayward, S. Duggan, K.L. Price, J. Pan, C. Dessapt, R.C.M. Siow, G.E. Mann; Cardiovascular Division, King‘s College London, London, UK. Background and aims: Mesangial cells are a primary target for haemody- namic perturbations in hypertensive glomerulopathies (e.g. diabetes). Oxida- tive stress plays a central role in the pathogenesis of diabetic microvascular complications, and the transcription factor Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) has been proposed as a central player in the cellular anti- oxidant response. The aim of this study was to evaluate, in cultured human mesangial cells (HMCs), the role of Nrf2 in the stretch-mediated antioxidant response and fibronectin secretion. Materials and methods: HMCs were exposed to 10% average elongation (S) or static conditions (NS) up to 48h. Cellular superoxide was assessed with Cytochrome-C reduction assay, while mitochondrial superoxide was investi- gated using MitoSOX Red. Protein and mRNA expression were assessed re- spectively with western imunoblotting and real time PCR after normalisation against housekeeping genes. Fibronectin secretion was assessed by ELISA and normalised against cell number. Results: S increased superoxide production (p=0.01), which was paralleled by an upregulation of the antioxidant enzyme heme oxygenase-1 (HO-1) both at the mRNA and protein level (p=0.01). The increase in superoxide was not accompanied by an increase in p22phox, p67phox or Nox4 mRNA expres- sion, whilst mRNA expression of gamma-glutamate cysteine ligase catalytic subunit was upregulated (p=0.05), but not mRNA levels for NAD(P)H qui- none oxidoreductase-1 or glutathione peroxidase. Stretch-mediated increase in superoxide production was significantly inhibited by the NADPH-oxidase inhibitors apocynin (10 mcM) and diphenyleneiodonium (100 mcM), and by the antioxidant rotenone (25nM), suggesting a mitochondrial origin of oxi- dative stress in stretched cells. Inhibition of stretch-mediated superoxide pro- duction was paralleled by absent stretch-induced HO-1 protein upregulation. Transfection of HMCs with dominant negative Nrf2 resulted in inhibition of stretch-induced HO-1 upregulation (p=0.05), but had no effect on basal and stretch-induced (24 and 48h) fibronectin secretion when compared to control (GFP-transfected) cells. Conversely overexpression of wild type Nrf2 resulted in a significant HO-1 up regulation (p=0.01), which was paralleled by a reduction in both basal and stretch-induced fibronectin (p S 480 1 C 1174 Global toll-like receptor 4 knockout attenuates features of diabetic nephropathy I. Jialal1, S. Devaraj2; 1Internal Medicine, UC Davis Medical Center, Sacramento, 2Pathology, Baylor College of Medicine, Houston, USA. Background and aims: Type 1 Diabetes Mellitus (T1DM) is a pro-inflam- matory state with increased Toll-like receptor (TLR) activity. Inflammation is crucial in diabetic vascular complications including diabetic nephropathy (DN). We tested the effect of a global deficiency of TLR4 on renal inflamma- tion, fibrosis and podocytopathy. Materials and methods: Streptozotocin (STZ) was used to induce diabetes in wildtype (WT) and TLR4-knockout (TLR4KO) mice. Control (C) and dia- betic groups (STZ-WT and STZ-TLR4KO) mice were euthanized at 17 weeks and plasma and kidneys collected. Results: Compared to C, the STZ-WT group had significantly increased macrophage and TLR4 immunostaining in the kidney, significant increases in MyD88, Interferon Regulatory Factor 3, Nuclear Factor Kappa B activity, Tumor Necrosis Factor Alpha, Interleukin 6, and MCP-1 protein levels; all these parameters were significantly decreased in the STZ-TLR4KO compared to STZ-WT mice. Compared to C, there were significant increases in fibro- sis markers including Collagen 4,Laminin and Transforming Growth Factor Beta in STZ-WT which were significantly decreased in the STZ-TLR4KO versus the STZ-WT. Podocyte numbers and podocin protein were decreased in the STZ-WT versus C and these were increased in the STZ-TLR4KO mice. Conclusion: In conclusion, a global genetic deficiency of TLR4 also amelio- rates renal inflammation, fibrosis and podocytopathy and could be important in DN. Supported by: JDRF 1175 Alkalinization attenuates diabetes-induced albuminuria and glomerular injury through reduced oxidative stress T. Senmaru1, M. Fukui1, M. Tanaka1, H. Okada2,1, T. Fukuda1, K. Mitsuhashi1, N. Kitagawa1, T. Kimura1, Y. Hashimoto1, N. Nakanishi1, M. Yamazaki1, G. Hasegawa2,1, N. Nakamura1; 1Department of Endocrinology and Metabolism, Kyoto Prefectural University of Medicine Graduate School of Medical Science, 2Division of Metabolism, Nephrology and Rheumatology, Japanese Red Cross Kyoto Daini Hospital, Japan. Background and aims: Patients with diabetes frequently have chronic meta- bolic acidosis. Metabolic acidosis traditionally defined as a decrease in blood bicarbonate (HCO3-) concentration contributes to the progression of dia- betic nephropathy, leading to end-stage renal failure. However, it is not well known whether alkalinization could prevent the progression of diabetic ne- phropathy, and if so, its mechanism has not been elucidated. In this study, we investigated the effects of alkalinization by potassium-sodium citrate on the pathogenesis and progression of diabetic nephropathy. Materials and methods: Diabetes was induced using streptozotocin in male C57BL/6 (WT) mice at 7 weeks of age. Diabetic mice were divided into two groups: alkalinization (Alkali+) and non-alkalinization (Alkali-). Alkali+ was allowed free access to water containing potassium-sodium citrate (0.2% po- tassium citrate and sodium citrate hydrate) for 16 weeks. Results: Blood HCO3- concentration is higher in Alkali+ than in Alkali- (25.1 ± 1.8 vs. 19.5 ± 1.5 mmol/L, p < 0.05). Urinary albumin excretion was decreased in Alkali+ compared to Alkali- (45.6 ± 7.1 vs. 99.8 ± 14.1 μg/ day, p < 0.05). Diabetic mice exhibited glomerular hypertrophy and mesan- gial expansion as compared with WT mice. Histological and morphomet- ric analysis of periodic acid-Schiff stained sections revealed that an increase of mesangial area was milder in Alkali+ than in Alkali- (26.2 ± 1.1 vs. 36.5 ± 1.0 %, p < 0.0001). In addition, the number of podocytes and the num- ber of apoptosis cells was decreased in Alkali+ compared to Alkali-. On the other hand, alkalinization did not affect tubulointerstitial fibrosis (Sirius red stain). The degree of oxidative stress assessed by 4-hydroxynonena (4HNE) immunostaining in the glomerular area was decreased in Alkali+ compared to Alkali-. Similar to the degree of 4HNE expression, the mRNA expression of p67-phox was decreased in Alkali+ compared to Alkali-. Furthermore, the mRNA expression of TGF-β and vascular endothelial growth factor was down-regulated in Alkali+. It is conceivable that the decreased expression of these fibrogenic cytokine contributes to the prevention of advanced glomeru- lar injury in Alkali+. Conclusion: These results indicate that alkalinization attenuates diabetes- induced albuminuria and glomerular injury. Reduction of oxidative stress may be mainly involved in these changes. Alkalinization could be a useful treatment for the progression of diabetic nephropathy. 1176 Zinc exerts anti-fibrotic action by inhibiting TGF-beta1-induced type I collagen and fibronectin in human renal proximal tubular cells Y. Takiyama, M. Maeda, K.K. Atageldiyeva, T. Yanagimachi, J. Honjo, Y. Fujita, Y. Makino, M. Haneda; Department of Medicine, Asahikawa Medical University, Japan. Background and aims: Hyperglycemia, chronic hypoxia and TGF-β1 con- tribute to the development of diabetic nephropathy.The aim of this study is to investigate whether novel key factors contribute to the pathophysiological mechanism of diabetic nephropathy. Materials and methods: We analyzed microarray data using Affymetrix GeneChip (Rat Gene 1.0 ST Array) with microdissected juxtamedullary proximal tubules in a rat model of type 2 diabetes, Zucker Diabetic Fatty (ZDF) rats. Then, we studied the regulation of the expression of the candidate using human renal proximal tubular epithelial cells (HRPTECs). Results: A total of 27,342 transcripts were analyzed, among them, 47 were upregulated (>2.0-fold) in the ZDF rat compared with the lean con- trol rat. One of them which were upregulated in diabetic renal tubules was metallothionein(MT)-1. MT is a cysteine-rich protein with low molecular weight, and act as an antioxidant against the toxicity of metals, ischemia, and ROS. MT originally has a pivotal role in the regulation of the metabolism of zinc. Previous studies reported that diabetic patients showed decreased plasma level of zinc, and zinc might be involved in diabetic nephropathy. Therefore, to explore the interaction of MT and zinc in diabetic nephropathy, we next studied the regulation of MT expression, and the effects of zinc on MT expression. High glucose (25.5 mM) failed to affect MT-1 mRNA ex- pression in HRPTECs, and insulin (100 nM) significantly enhanced MT-1 mRNA levels. Hypoxia (1% O2) slightly increased MT-1 mRNA levels. 2.5 ng/ ml TGF-β1 significantly decreased MT-1 mRNA levels compared to that of the control (0.62 ±0.16 vs. control, p S 481 1 C could reduce microalbuminuria and provide renal protective effects by im- proving endothelial dysfunction and uncoupling of glomerular VEGF-NO axis in streptozotocin-induced type 2 diabetic rats. Materials and methods: Wistar rats were randomly divided into a normal control (NC) group, a diabetic nephropathy (DN) group induced by high-fat feeding and streptozotocin, diabetic rats injected with adenovirus-expressed adiponectin AD-AdipoQ-IRES-EGFP (AD-AdipoQ, DA group), and diabet- ic rats injected with AD-IRES-EGFP (AD-IRES, DI group). Blood and urine samples were collected. Endothelium-dependent vasodilatation (EDV) of the aorta was measured. Renal tissues were collected for CD34 immunohisto- chemistry. Glomerular NO and VEGF levels were measured by the Griess reaction and western blot testing, respectively. Results: STZ-injected rats in the DN group had severe hyperglycemia and high levels of insulin (P < 0.01). This was reduced by intravenous injections of AD-AdipoQ (glucose, 11.3 ± 5.5 mmol/L vs 21.2 ±3.3 mmol/L; insulin, 16.70±2.40 uU/ml vs 20.91±2.78 uU/ml, P < 0.01). Serum hs-CRP and MDA levels increased and serum adiponectin levels decreased significantly in the DN group compared with the NC group (P < 0.01). Intravenous injections of AD-AdipoQ reduced serum hs-CRP and MDA levels and increased se- rum adiponectin levels in diabetic rats (hs-CRP, 1.85±0.56 mg/L vs 3.13±0.38 mg/L; MDA 2.44 ± 0.28 umol/L vs 4.80 ± 0.59 umol/L; adiponectin, 273.50 ± 21.20 ug/L vs 94.97 ± 25.12 ug/L, P < 0.01). Albumin-to-creatinine (ACR) was significantly higher in rats in the DN group than those of the NC group [180.35 (113.54-233.29) mg/g vs 13.15 (8.93-17.95) mg/g, P < 0.01]. Injec- tions of AD-AdipoQ significantly reduced the ACR in diabetic rats. [72.0 (46.18-99.41) mg/g vs 180.35 (113.54-233.29) mg/g, P < 0.01], indicating an improvement of early diabetic nephropathy. Severe EDV impairment was ob- served in the DN group, which was improved by AD-AdipoQ. CD34 expres- sion in the glomeruli was also enhanced in diabetic rats, indicating increased proliferation of glomerular endothelial cells [mean optical density (IOD/ area): 0.145 ± 0.015 vs 0.073 ± 0.007, P < 0.05]. However, treatment with AD-AdipoQ partly improved the increased proliferation of endothelial cells in glomeruli (IOD/area: 0.113 ± 0.012 vs 0.145 ± 0.015, P < 0.05). Diabetic rats showed increased glomerular VEGF levels and reduced NO levels (P < 0.05). This uncoupling of the VEGF-NO axis was partially improved by AD- AdipoQ (P < 0.05). There were no significant differences in these parameters between the DN and DI groups (P > 0.05). Conclusion: Adiponectin reduces the degree of microalbuminuria and has renal-protective effects by improving endothelial dysfunction and uncou- pling the glomerular VEGF-NO axis in early diabetic nephropathy. Supported by: National Natural Science Foundation of China (Grant NO. 81300688) 1178 Heat shock proteins role in the biochemical and molecular interplay leading to diabetic nephropathy A.I. Serban1, L. Stanca1, O.I. Geicu1,2, F. Barbuceanu1, A. Dinischiotu2; 1Preclinical Sciences, University of Agricultural Science and Veterinary Medicine, 2Biochemistry and Molecular Biology, University of Bucharest, Faculty of Biology, Bucharest, Romania. Background and aims: Heat shock protein (Hsp) expression is correlated with oxidative stress, yet few studies focus on Hsp’s modulation in diabetes. The objective of this study was to highlight the link between Hsp levels and the biochemical changes leading to diabetic nephropathy, using the human embrionary kidney cells HEK293. Materials and methods: Cells were treated for 12, 24 and 48 h with 200 μg/ ml glycated-bovine serum albumin (AGE-BSA) or BSA. Reduced glutathione (GSH), advanced glycation lipid products (ALEs) levels and enzyme activi- ties were evaluated by spectrophotometric methods. Superoxide dismutase (SOD) was also assessed by zymography. The receptor for advanced glycation end products (RAGE), Hsp 27, 60 and 70 expressions were evaluated by west- ern immunoblotting and real-time PCR. Results: AGE-BSA induced a time dependent RAGE upregulation. Maximal increases of gene expression by 2.25 fold and protein expression by 2.01 fold were registered after 48 h. Increased expression of RAGE has been docu- mented in chronic renal disease and diabetic complications. Probably the up- regulated RAGE stimulated pro-inflammatory signaling and reactive oxygen species (ROS) production. Increased ROS contribute to GSH depletion by 70 % after 48 h, which hinders the cellular antioxidative and anti-AGE activities. SOD activity upregulated with AGE-BSA exposure time and increased by 33 % after 48 h. Zymogram analisys revealed a similar pattern. Cu/Zn-SOD had the most important contribution for the SOD activity after 12 h of exposure. Probably the initial superoxide anion source is not mitochondrial but cyto- solic, although later, Mn-SOD increased, indicating mitochondrial damage. After 24 h and 48 h, ALEs arose by 1.75 and 2.29 fold, suggesting oxidative damage is amassed with time. Catalase activity increased over 200 % after 12 h of AGE exposure and remained over 50 % higher than controls later on. This early increase denotes the presence of high hydrogen peroxide amounts, probably generated by NADPH oxidase rather than SOD activity. AGE ex- posure boosted glutathione peroxidase activity by 50 % after 48 h. This en- zyme could be involved in the detoxification of lipid hydroperoxides that are most abundant after 48 h. Glutathione reductase and glucose 6-phosphate dehydrogenase upregulated by 35 % after 48 h. These enzymes are sensitive to ROS, and probably their increase was supported by Hsp70 activity. Hsp 70 gene and protein expressions augmented in a time dependent manner, increasing by 2.07 and 2.78 fold after 48 h. Hsp 70 expression seems a reliable indicator of the biochemical alterations described mainly by the increase of RAGE, ROS and ALE. Hsp 27 gene and protein expressions increased after 12 h of treatment by 1.93 and 1.52 fold, but decrease to 0.45 and respectively 0.16 fold after 48 h. Hsp 60 analysis revealed a similar profile. Hsp 27 and 60 de- crease after 48 h suggests the possibility of pro-apoptotic pathways activation. Conclusion: ROS induced by AGE - RAGE activation employ antioxida- tive mechanisms, which are themselves susceptible to oxidative damage. Al- though Hsps protective roles can improve some of the AGE-induced effects, therapeutic approaches based on Hsp overexpression must take great care not to target anti-apoptotic Hsps, because inhibiting AGE-promoted apoptosis could invoke cancerous transformation. Supported by: CNCS-UEFISCDI, project number PN-II-RU-TE-2012-3-0034 1179 Lentiviral vector-mediated FoxO1 overexpression inhibits extracellular matrix protein secretion under high glucose conditions in mesangial cells F. Guo, G. Qin, Q. Wang, Y. Zhou, L. Wu, X. Ma, F. Huang; Division of Endocrinology, The First Affiliated Hospital of Zhengzhou University, China. Background and aims: The excessive accumulation of ECM proteins in the mesangium of glomerulus results in glomerulosclerosis and contributes to the genesis and development of diabetic nephropathy (DN). The bioactivity of Forkhead transcription factor O1 (FoxO1) is attenuated in renal cortex under high glucose conditions, which may be associated with the progres- sion of DN. High glucose conditions also upregulate the expression level of ECM proteins by activating the transforming growth factor β (TGF-β)/Smad pathway in MCs. The study is designed to investigate whether FoxO1 is in- volved in the redundant secretion of extracellular matrix (ECM) proteins in mesangial cells (MCs) under high glucose conditions. Materials and methods: Constitutively active FoxO1 recombinant lentivi- ral vectors (LV-CA-FoxO1) and GFP lentiviral vectors (LV-NC-GFP) were constructed. MCs were transfected with either LV-CA-FoxO1 or LV-NC-GFP cultured in high glucose medium (25 mmol/L), compared with untransfected cells cultured in high or normal glucose medium (5.6 mmol/L). Functional analysis was performed by real-time PCR, western blotting, immunofluores- cence and flow cytometry to evaluate the expressions of FoxO1, ECM com- ponents and TGF-β signaling pathways. Results: High glucose conditions increased p-FoxO1 protein levels and the ratio of p-FoxO1/FoxO1, thereby attenuating bioactivity of endogenous FoxO1, which was accompanied by marked increases in the expression of ECM components, including FN and Col I. High glucose conditions also ac- tivated TGF-β-mediated pathways by elevating the latent and mature protein levels of TGF-β1, in accordance with the increase of TGF-β1 mRNA levels. The expressions of TGF-βRI and TGF-βRII-as well as the expression and phosphorylation of Smad3-were both increased under high glucose condi- tions, thereby promoting signal transduction of TGF-β-mediated pathways. Immunofluorescence analysis indicated that high glucose conditions in- creased p-Smad3 staining in the nucleus, consistent with an overall increase in p-Smad3 expression. By contrast, FoxO1 expression markedly increased and the ratio of p-FoxO1/FoxO1 decreased after LV-CA-FoxO1 transfection, accompanied by reductions of FN and Col I mRNA levels. The mRNA level of TGF-β1 as well as its latent and mature protein levels both reduced with overexpression of transfected CA-FoxO1, while TGF-βRI and TGF-βRII lev- els decreased simultaneously. Although mRNA and protein levels of Smad3 remained unchanged after transfection, the phosphorylation of p-Smad3 was blocked and the ratio of p-Smad3/Smad3 declined to the level of which under normal glucose conditions. In addition, overexpression of CA-FoxO1 attenu- Diabetologia (2014) 57:[Suppl1]S1–S564 S 482 1 C ated nuclear staining of p-Smad3, in concert with faintly observed immuno- fluorescence in the cytoplasm, indicating lower basal expression of p-Smad levels and diminished transcriptional activity. Conclusion: High glucose conditions increased the expression of fibronectin and collagen I by activating TGF-β/Smad3 pathway in mesangial cells, while lentiviral vector-mediated overexpression of constitutively active FoxO1 blocked this effect by suppressing the activation and signaling of the TGF-β/ Smad3 pathway. FoxO1 may play a vital role in attenuating ECM protein secretion and alleviating the pathological changes associated with diabetic nephropathy. 1180 The experimental study of the effects of FoxO1 on the podocyte injury in diabetic rats G. Qin, Y. Zhang, Q. Wang, Y. Zhou, X. Ma; The First Affiliated Hospital of Zhengzhou, China. Background and aims: Diabetic nephropathy (DN) is the chief reason for end-stage renal disease (ESRD). The podocyte injury of DN is often char- acterized by foot process fusion and effacement, podocytes getting smaller, negative charge becoming less, and podocytes detaching from glomerular basement membrane (GBM) and discharging in the urine. Nowadays, the importance of podocyte injury in pathogenesis of DN and the occurrence of proteinuria and glomerular sclerosis has been gradually acknowledged. We have previously demonstrated that the activity of forkhead transcription factor O1 (FoxO1) was decreased in the renal cortex of DN rats, and discov- ered that compared with DN rats without treatment, the expression of FoxO1 phosphorylation was reduced , the activity of FoxO1 was increased and the podocyte injury was improved in the renal cortex of DN rats with treatment. So it is inferred that the activity of FoxO1 may affect the podocyte injury in DN rats. Our study was performed to investigate the effects of FoxO1 on the podocyte injury in diabetic rats. Materials and methods: Streptozotocin-induced diabetic rats were divided into three groups: diabetic rats (group d), rats transfected with empty lenti- viral vectors (DM plus LV-empty-FoxO1 group, served as group a) and rats which were transfected with constitutively active FoxO1 recombinant len- tiviral vectors (DM plus LV-CA-FoxO1 group, served as group b). The rats which received an injection of diluent buffer served as normal control (group c). 2,4,8 weeks after the infection, the levels of urine albumin, blood glucose, blood urea nitrogen and serum creatinine were measured. Real-time PCR and Western blotting were performed to measure the mRNA and protein ex- pressions of FoxO1, podocalyxin, nephrin, desmin, COL4A3, COL4A5 and FoxO1 phosphorylation (p-FoxO1) level of renal cortex in rats. Moreover, light microscope and electron microscope were used to observe the structure changes of glomerulus and podocytes. Results: Compared with group a and d, in group b, there was a significant increase in the mRNA and protein expressions of FoxO1,and a distinctly decrease in the levels of urine albumin, blood urea nitrogen and serum cre- atinine of rats (except ones in two weeks) (all p S 483 1 C identified from the registry as controls. We performed genome-wide associa- tion study using the Illumina 610Quad genotyping array. Validation was per- formed by de novo genotyping using a MassARRAY platform on additional samples from the Registry. Results: After quality control on the SNP data, we analyzed a total of 446,343 SNPs. We identified 8 genomic regions which showed suggestive association with increased risk or protection against incident diabetic kidney disease (p S 484 1 C to non-diabetic. HbA1c was increased after 6, 12, 18 and 24 weeks of DM (p S 485 1 C 1188 High-fat meals induce systemic cytokine release without evidence of endotoxaemia-mediated cytokine production from circulating monocytes or myeloid dendritic cells C.L. Fogarty1, J.K. Nieminen2, L. Peräneva1, M.I. Lassenius1, C. Forsbloom1, P.-H. Groop1, O. Vaarala2, M. Lehto1, FinnDiane Study Group; 1Folkhälsan Research Center, 2National Institute for Health and Welfare, Helsinki, Finland. Background and aims: Dietary fat is known to increase circulating levels of bacterial lipopolysaccharide (LPS) and proinflammatory cytokines in healthy subjects. Earlier studies have found increased intestinal permeability in patients with type 1 diabetes (T1D), suggesting that patients may display a more pronounced postprandial increase in circulating proinflammatory TLR4 agonists-LPS and triglycerides. While it has been long known that circulating inflammatory markers are predictors of progressive renal failure, we have recently shown that LPS activity levels are strongly associated with the components of the metabolic syndrome and the development of diabetic nephropathy in patients with T1D. Increased TLR4 responsiveness together with postprandially elevated circulating LPS and triglyceride levels may play a key role in the promotion and progression of inflammation, thereby in- creasing the risk of developing diabetic complications. The aim of the present study was to investigate the acute effects of two sequential high-fat meals on circulating cytokines and the LPS responsiveness of innate immune cells in the context of T1D. We hypothesized that high-fat diet-induced endotoxemia would increase cytokine release from peripheral innate immune cells, which could in turn both induce systemic inflammation and modulate TLR4 re- sponsiveness, particularly in patients with T1D. Materials and methods: Eleven patients with T1D and eleven controls re- ceived two fatty meals: breakfast (8:00, 65 g fat) and lunch (12:00, 42 g fat). The patients were studied as part of the Finnish Diabetic Nephropathy (FinnDiane) Study. Fasting (8:00) and postprandial (14:00) blood samples were taken and incubated with or without E. coli LPS. Cytokine levels in se- rum, and cytokine responses in monocytes and mDCs were measured using flow cytometry Results: At fasting, diabetic myeloid dendritic cells (mDCs) exhibited high- er LPS-induced IL-6 and IL-1β production than controls. Postprandially, patients with T1D and controls showed significant increases in the serum concentrations of eight inflammatory cytokines (IL-6, TNF-α, IL-1β, IFN-α, IL-10, IFN-γ, IL-12, and MIP-1β) without concomitant increase in serum LPS activity. Serum cytokine production was similar in both groups. No post- prandial change was seen in the IL-6, TNF-α, or IL-1β production of mDCs or monocytes. Conclusion: Our results show that in whole blood, diabetic dendritic cells are hyperresponsive to in vitro LPS stimulation. Furthermore, in patients and controls alike, acute high-fat meals increase circulating cytokines but have no effect on serum LPS activity levels or cytokine production in circulating mDCs or monocytes. Our results suggest that postprandial increases in se- rum cytokine levels are neither mediated by circulating endotoxins nor the activation of circulating innate immune cells. The production of high-fat meal induced inflammatory markers is conserved between patients and con- trols and most likely regulated at the tissue level. Supported by: Wilhelm and Else Stockmann Foundation (P.-H.G., M.L., C.F.) 1189 Effects of incretin on human mesangial cells loaded with apo E2 remnant lipoproteins and on nephropathy in diabetic patients M. Eto, M. Saito; Clinical Medicine, Ohu University, Koriyama, Japan. Background and aims: In JDCS (Japan) and UKPDS 74, hypertriglyceri- demia was detected as one of the risk factors for diabetic nephropathy. In- creased remnant lipoproteins underlie hypertriglyceridemia. We first found that apo E2 is a genetic factor for diabetic nephropathy, and that increase of plasma TG/remnant lipoproteins caused by apo E2 contributes to nephropa- thy. Remnant lipoproteins are taken up by mesangial cells, and then cause renal damage. The frequency of apo E2 is approx.10% in Japan and western countries. Diabetic nephropathy is characterized by the accumulation of ex- tracellular matrix protein in the glomerular mesangium. Incretin, which is widely used in T2DM patients, reduces plasma TG/remnant lipoproteins as well as glucose. There is little information about the effect of incretin on ne- phropathy. In the present study we examined the effect of GLP-1 on TGF-β and type IV collagen expression in human mesangial cells (HMCs) loaded with apo E2 remnant lipoproteins, and examined whether sitagliptin which increases incretin is effective to improve nephropathy in T2DM patients with apo E2 allele. Materials and methods: Remnant lipoproteins were isolated from plasma of a patient with T2DM and apo E2/2 genotype by ultracentrifugal method. HMCs loaded with remnant lipoproteins were incubated for 24 h with GLP-1 at the concentration of 0, 50, 100 and 500 nmol/L. To evaluate the expres- sion of TGF-β , type IV collagen and GLP-1 receptor mRNA in HMCs, RT- PCR procedure was performed. T2DM with apo E3/3 (n=38 ) and apo E3/2 genotype (n=7) were studied. Patients were treated with sitagliptin over 10 months. HbA1c, plasma TG and remnant cholesterol were compared before and after sitagliptin therapy. Remnant cholesterol was determined as RLP- cholesterol . Urinary albumin-to-creatinine ratio (ACR, normal range S 486 1 C PS 105 Diabetic nephropathy: clinical 1190 Performance of creatinine and cystatin C based CKD-EPI equations to estimate glomerular filtration rate in type 2 diabetes A.A. Soares1, R. Boff1, J.D. Machado1, L.S. Rodrigues1, E.G. Camargo1, J.L. Camargo2, S.P. Silveiro1; 1Universidade Federal do Rio Grande do Sul, 2Hospital de Clínicas de Porto Alegre, Brazil. Background and aims: Diabetes kidney disease (DKD) is the worldwide leading cause of end-stage renal disease. Current guidelines recommend annual screening with urinary albumin measurement and estimation of glomerular filtration rate (GFR) by equations. The aim of this study was to evaluate the accuracy of creatinine and cystatin C based Chronic Kidney Dis- ease Epidemiology Collaboration (CKD-EPI) equations to estimate GFR in patients with type 2 diabetes (DM) as compared to healthy adults. Materials and methods: Cross-sectional study included healthy adults and DM patients with GFR higher than 60 mL/min/1.73 m². GFR was es- timated by CKD-EPI creatinine (CKDEPIcreat), CKD-EPI cystatin C (CK- DEPIcystC), and CKD-EPI creatinine-cystatin C (CKDEPI-CC) equations in comparison with GFR measured by a reference method (51Cr-EDTA). Each group (DM vs healthy) was stratified according to ages above and be- low 45 years and analyzed separately. Serum creatinine was measured by Jaffe traceable method (Advia 1800, Siemens Healthcare) and serum cystatin C was evaluated by immunoturbidimetric method (Dako, Cytomation). Ac- curacy (percentage of estimated GFR within 30% [P30] of measured GFR), bias (mean difference between measured and estimated GFR), and precision (standard deviation of bias) were assessed. Agreement was evaluated accord- ing to Bland & Altman analyses. Results: One hundred healthy adult subjects and 84 DM patients, aged 38±14 (18-86) and 59±19 (31-82) years, respectively, were evaluated. Measured 51Cr-EDTA GFR was 112±19 (64-160) and 104±27 (60-184) mL/min1.73 m², in healthy individuals and in DM patients, respectively. For the healthy group with less than 45 years of age, the measured 51Cr-EDTA GFR was 117±19 and the CKDEPIcreat, CKDEPI-CC and CKDEPIcystC were 117±13, 110±11 and 104±12 mL/min/1.73 m², respectively. Only CKDEPIcreat was in agreement with the reference method (P=0.894). In the DM group, the measured 51Cr-EDTA GFR (126±41 mL/min/1.73 m²) was different from all equations: CKDEPIcreat, CKDEPI-CC and CKDEPIcystC equations were, respectively 98±16, 95±16 and 93±18 mL/min/1.73 m². The Figure depicts the GFR mean values as well as the precision (SD) of the older individuals. In this set, the reference GFR method did not agree with any equation, in both groups. In healthy subjects older than 45 years, the accuracy (P30) was 94%, 91% and 85% for CKDEPIcreat, CKDEPI-CC and CKDEPIcystC, respective- ly. In DM patients older than 45 years, accuracy was 71%, 68% and 55% for CKDEPIcreat, CKDEPI-CC and CKDEPIcystC, respectively. Conclusion: All CKD-EPI equations underestimated GFR in older adults, more markedly in type 2 DM patients. In younger subjects, the creatinine based equation presented the best performance. Supported by: CNPq, FAPERGS and FIPE 1191 Albuminuria predicts progression to end-stage renal disease in diabetic patients T. Babazono, H. Murata, A. Toyonaga, N. Yoshida, S. Takemura, M. Takagi, N. Yoshida, K. Toya, I. Nyumura, K. Hanai, N. Tanaka, Y. Uchigata; Tokyo Women‘s Medical University School of Medicine, Tokyo, Japan. Background and aims: Albuminuria in diabetic patients has been repeatedly shown to be a crucial risk factor for renal outcomes; however, previous stud- ies examining these associations were limited to short observational period or surrogate markers including decline in GFR. There are few large cohort studies that observed patients until reaching end-stage renal disease. We con- ducted this study to clarify the association between albuminuria and hard renal endpoint in diabetic patients. Materials and methods: This was a single-center historical cohort study consisting of adult Japanese diabetic patients who had both serum creatinine concentration and albuminuria data measured within 90 days at our hospital between 1995 and 2012, baseline estimated GFR (eGFR) ≥ 15 mL/min/1.73 m2, and those who were observed at least 90 days. According to the baseline urinary albumin-to-creatinine ratio (ACR), patients were classified into the following 7 groups; S 487 1 C Conclusion: The current data suggest that prevalence of albuminuria may begin to increase in the higher range of IFG, independent of traditional risk factors. 1193 Arterial stiffness is more associated with albuminuria than decreased glomerular filtration rate in the development of type 2 diabetic nephropathy Y. Nam1, S. Kim1, C. Lee2, S. Park1, J. Kim1, W. Kim1, Y. Jeon1, B. Kim1, I. Kim1, S. Son3; 1Department of Internal Medicine, Pusan National University Hospital, 2Department of Internal Medicine, Busan St. Mary’s Medical Center, 3Department of Internal Medicine, Bongseng Memorial Hospital, Busan, Republic of Korea. Background and aims: Arterial stiffness, a clinical atherosclerosis marker, was associated with the progression of chronic kidney disease (CKD) includ- ing diabetic nephropathy. Each albuminuria and decline of glomerular func- tion rate (GFR) are being recognized as complementary rather than obliga- tory markers for diabetic nephropathy. The aim of study was to evaluate the association of arterial stiffness with albuminuria and/or GFR in the develop- ment of type 2 diabetic nephropathy. Materials and methods: This study was an analysis from participants of The Relationship between Cardiovascular disease and Brachial-ankle Pulse Wave Velocity in Patients with Type 2 Diabetes (REBOUND) Study, which was a prospective multi-center cohort study recruited from 8 general hospitals in Busan, Korea. Type 2 diabetic patients aged 30 years and more were recruited and measured brachial-ankle pulse wave velocity (baPWV) as noninvasive marker for arterial stiffness. The patients were categorized into 4 groups ac- cording to albumin-to-creatinine ratio (ACR, normoalbuminuria and albu- minuria) and eGFR ( S 488 1 C py compared to those who had smoked 0-5 py. HR was 1.4 (0.9-2.1) and 1.3 (0.9-1.8) for those who had smoked 5-8.5 and 12.5-30 py. Conclusion: Increase in pack-years of smoking is associated with progression of DN, but the relationship is not linear but rather bimodal. We conclude that the first peak might indicate an increased risk of a subset for the patients who are genetically predisposed and the second due to harmful effects of smoking. Supported by: Academy of Finland, Wilhelm and Else Stockmann Foundation 1196 Reduction of microalbuminuria by calcium channel blockers in patients with type 2 diabetes mellitus and hypertension Y.-C. Hwang1, S. Oh1, K.-H. Yoon2, B. Cha3, K.-W. Lee4, H. Jang5, K. Min6, C. Chung7, M.-K. Lee1; 1Samsung Medical Center, Seoul, 2Seoul St. Mary‘s Hospital, 3Yonsei University College of Medicine, Seoul, 4Ajou University School of Medicine, Gyeonggi-do, 5Seoul National University Bundang Hospital, Gyeonggi-do, 6Eulji General Hospital, Seoul, 7Yonsei University Wonju College of Medicine, Republic of Korea. Background and aims: It has been suggested that reno-protective effects are different among calcium channel blockers. The aim of this study was to com- pare anti-albuminuric effect of different types of calcium channel blockers in patients with type 2 diabetes. Materials and methods: A multicenter, randomized, open label, active-con- trolled study was conducted in seven hospitals in Korea. Inclusion criteria were age ≥18 years, hypertension treated with angiotensin receptor blocker or angiotensin-converting enzyme inhibitor, type 2 diabetes with HbA1c ≤8%, and persistent microalbuminuria. A total of 74 patients were randomized into the cilnidipine 10mg group (n = 37) or the amlodipine 5mg group (n = 37). Patients were assessed at baseline, 12 weeks and 24 weeks after treatment. Results: Compared with cilnidipine group, amlodipine group showed greater diastolic blood pressure reduction at 24 weeks after treatment (P=0.03). In the cilnidipine group, urine albumin to creatinine ratio (ACR) was significantly reduced after 12 weeks (-53.0 ± 123.2 mg/g, P S 489 1 C and functional abnormalities and sympathetic hyperactivity are potential causative factors. Hypoglycaemia has been linked to sudden death in insulin treated diabetes. To explore the relationships between; 1. glycaemic variation and heart rate, rhythm and QTc interval 4hrs before, during (3 to 4hrs) and 4hrs after haemodialysis (HD) 2. changes in serum electrolyte levels during HD and heart rate, rhythm and QTc interval. Materials and methods: In our ongoing study, we undertook week-long con- tinuous glucose monitoring and Holter monitoring, for 1 to 3 weeks in insu- lin deficient patients. Serum electrolytes and 12 lead ECGs were obtained at the beginning, middle and end of 3 HD sessions during the 1st study week. 10 insulin deficient patients (7 female, mean age 50.6±9.8 yrs) were monitored, during 72 HD sessions. Results: Mean glucose levels during HD were significantly lower than pre- HD (8.4±3.6 vs11.2±4.5mmol/L, p S 490 1 C PS 106 Clinical: hypertension in diabetes 1200 Is home blood pressure reporting in patients with type 2 diabetes reliable? S. Majima1, M. Fukui1, S. Matsumoto1, E. Ushigome1, K. Matsushita1, H. Iwase1, T. Senmaru1, M. Hamaguchi1, M. Asano1, M. Yamazaki1, G. Hasegawa2,1, N. Nakamura1; 1Department of Endocrinology and Metabolism, Kyoto Prefectural University of Medicine, Graduate School of Medical Science, 2Division of Metabolism, Nephrology and Rheumatology, Japanese Red Cross Kyoto Daini Hospital, Japan. Background and aims: Hypertension, as well as hyperglycemia, is well known as an important risk of cardiovascular and renal disease in patients with type 2 diabetes. Home blood pressure (HBP) monitoring is a valuable tool in the management of hypertension. To our knowledge no study has in- vestigated the reliability of HBP measurements in patients with type 2 diabe- tes. The aim of this study was to evaluate the reliability of HBP reporting and determine the factors that affected it in patients with type 2 diabetes. Materials and methods: A cross-sectional study was conducted in 280 con- secutive patients with type 2 diabetes recruited from the outpatient clinic at our university. Patients were requested to measure triplicate morning and evening HBP using a digital automatic BP monitor and to record these meas- urements in a logbook over a 2-week period. The patients were not informed of the storage function of their BP monitor. Results: Among a total of 280 patients, 155 (55.4%) were taking antihyper- tensive medication. The concordance rate between the self-reported data in the logbook and the stored data in the monitor was 78.6%. Although 65 of the patients (23.2%) had 100% “concordant data” that were matched the stored data in the monitor and 144 (51.4%) had more than 90% concordant data, 44 (15.7%) had 50% or less concordant data. The self-reported data were signifi- cantly lower than the stored data (mean of morning systolic BP: 129.8 ± 15.8 vs. 130.6 ± 16.2 mmHg, p < 0.0001). In addition, logbook BP was less variable than monitor BP (SD of morning systolic BP: 7.9 ± 3.2 vs. 9.8 ± 3.5 mmHg, p < 0.0001). The most frequent erroneous data (55.8%) were “selected data” that were randomly selected from multiple measurements in the monitor, the second most erroneous (23.3%) were “fictional data” that could not be found in the monitor, and the third most erroneous (11.6%) were “shifted data” that were stored in the monitor on another day but for the same time. The con- cordance rate correlated significantly with HbA1c (r = -0.127, p = 0.0333), HDL cholesterol (r = 0.131, p = 0.0286), triglycerides (r = -0.121, p = 0.0429), mean of morning systolic monitor BP (r = -0.121, p = 0.0443), mean of even- ing diastolic monitor BP (r = -0.127, p = 0.0386) and SD of morning systolic monitor BP (r = -0.129, p = 0.0310). Multiple regression analysis indicated that the independent explanatory variables concerning the concordance rate were HbA1c (β = -0.156, p = 0.0149) and current smoking (β = -0.165, p = 0.0184). Urinary albumin excretion also correlated with the SD of morning systolic BP in the monitor, but not with that in the logbook. Conclusion: Patients with type 2 diabetes sometimes report erroneous HBP, especially those with poor glycemic control or a smoking habit. As a result of this inaccurate reporting, HBP control obtained from logbooks may appear better than that obtained from a monitor. A device capable of automatically saving BP data may be helpful for obtaining accurate measurements of HBP that will assist in providing optimal treatment for hypertensive patients with type 2 diabetes. 1201 Predictive value of markers of vascular damage for renal outcome in type 2 diabetes and essential hypertension R. Bruno1, A. Salvati2, K. Raimo2, M. Barzacchi2, E. Vitolo2, L. Ghiadoni2, A. Solini2; 1Institute of Clinical Physiology, CNR Pisa, 2Dept of Clinical and Experimental Medicine, University of Pisa, Italy. Background and aims: We have shown that renal vasodilating response to nitrates (dynamic renal resistive index, DRIN) is an early vascular alterations of type 2 diabetes (T2DM) and essential hypertension (EH) already present in normoalbuminuric individuals. In this study we prospectively evaluated whether this parameter and other markers of systemic vascular damage are able to predict microalbuminuria (MA) onset and renal function decline in these patients. Materials and methods: We studied 65 individuals (27 T2DM and 38 EH) following them prospectively. Renal resistive index (RI), DRIN (% change in RI after glyceryl trinitrate, GTN 25 μg sl), endothelium-dependent (flow- mediated-dilation - FMD) and independent (response to GTN) vasodilation in the brachial artery, carotid-femoral pulse wave velocity (PWV), and aug- mentation index (AIx) were assessed. At the follow-up visit, MA onset was defined with urinary albumin-creatinine ratio (UACR)>30 mg/g; any reduc- tion in estimated glomerular filtration rate (eGFR, CKD-EPI formula) was also considered as an endpoint. Results: All patients were treatment-naïve at enrollment, whereas at the follow-up visit 62% were taking antihypertensive drugs and 37% were treated with anti-hyperglycemic agents. After a follow-up period of 4.1±0.6 years, mean eGFR (CKD-EPI) decreased from 89.0±14.4 to 86.4±13.0 ml/ min1.73m2, whereas UACR increased from 5 (0-29) to 9 (0-407) mg/g. Ac- cording to our definition, 19 individuals developed MA and 12 a reduction in eGFR. At enrollment, patients who would develop MA tended to be older (60.4±9.3 vs 53.8±10.4 years, p=0.07) and carrying more frequently diabe- tes than their counterparts. Among vascular parameters, RI (0.63±0.04 vs 0.59±0.06, p=0.04), DRIN (-6.4±8.9 vs -10.2±6.3 %, p= 0.09) and PWV (9.5 ±1.3 vs 7.9 ± 1.5 m/s, p= 0.003), were worse at baseline in those who would develop MA during follow-up. Conversely, AIx (24±11 vs 22±13 %, p=0.93), FMD (3.9±1.9 vs 5.6±3.9 %, p=0.27) and GTN (5.1±2.8 vs 5.6±3.4 %, p=0.85) were similar in the two groups. In T2DM patients there was a significant in- crease (p S 491 1 C with the deterioration of GFR. Also in average 16.2% of DM and 13.1% on non-DM moved each year towards the next and more severe stage of CKD (p=0.051). For initial GFR >90 mL/min/1.73m2, 24% of DM and 18% of non-DM showed a reduction per year < 10% of the previous GFR value (qui square, p=0.049). Conclusion: A progressive deterioration of renal function for each next year is frequent in diabetics and non-diabetics with hypertension. Beyond ageing, renal deterioration may be particularly dependent on BP control particularly at nighttime, on certain therapies and on highly abnormal glucose control. 1203 A systematic review and meta-analysis of the effects of diuretics and beta-blockers on glycaemic control in diabetes mellitus J.A. Hirst, A.J. Farmer, B.G. Feakins, J.K. Aronson, R.J. Stevens; Department of Primary Care Health Sciences, University of Oxford, UK. Background and aims: Although there are reports that beta-adrenoceptor antagonists (beta-blockers) and diuretics can affect glycaemic control in peo- ple with diabetes mellitus, there is little information on how blood glucose concentrations are likely to change and by how much. Materials and methods: We systematically reviewed the published literature to identify randomised controlled trials (RCTs) in which glycaemic control was studied in people with diabetes mellitus taking either beta-blockers or diuretics. We pooled end-point data on HbA1c and fasting blood glucose (FBG) concentrations using inverse variance fixed effects meta-analysis. Results: We retrieved 3864 papers and found 10 studies (15 comparisons in- volving 1889 participants) of beta-blockers and 11 studies (12 comparisons involving 312 participants) of diuretics to include in the meta-analysis. One study included comparisons of both beta-blockers and diuretics, giving a to- tal of 20 included trials. Beta-blockers increased fasting blood glucose con- centrations by a mean of 0.64 mmol/l (95% CI 0.24 to 1.03) and HbA1c by 0.75% (95% CI 0.30 to 1.20), corresponding to 8.2 mmol/mol (95% CI 3.3 to 13.1) compared with placebo. Three trials (four comparisons) that studied the beta-blocker propranolol showed a larger increase in FBG concentra- tions than three trials that used other beta-blockers (1.26 mmol/l, 95% CI 0.62 to 1.89 compared with 0.25 mmol/l, 95% CI -0.25 to 0.76). Diuretics increased FBG by 0.96 mmol/l (95% CI 0.29 to 1.63) compared with placebo and HbA1c by 0.24% (95% CI -0.17 to 0.65), corresponding to 2.6 mmol/mol (95% CI -1.9 to 7.1) compared with placebo. The four trials (5 comparisons) that studied thiazide diuretics gave a larger increase in FBG levels than three trials that used non-thiazide diuretics (1.69 mmol/mol, 95% CI 0.69 to 2.69, and 0.36 mmol/mol, 95% CI -0.55 to 1.27 respectively), but when tested us- ing meta-regression this difference was not significant (p=0.101). There were no significant differences in the numbers of hypoglycaemic events or other adverse events between beta-blockers and placebo in three trials. Conclusion: In this synthesis of data we have quantified the effect of two commonly used types of antihypertensive medications on glycaemic control in diabetes. Both medication types increase fasting blood glucose and HbA1c concentrations in patients with diabetes by moderate but clinically significant amounts. Among beta-blockers, propranolol has the biggest effect, but there have been too few studies to allow a full analysis of other beta-blockers. These data will inform the monitoring and use of beta-blockers and diuretics in patients with diabetes. Supported by: NIHR National School for Primary Care Research 1204 Renoprotective effect of ACE-I and ARB in combination with hydrochlorthiazide in hypertensive type 2 diabetic patients N. Akbarova, N. Rakhimova, B.D. Babadjanov, B. Djanabaev, B. Ikramova; Public Center of Purulent Surgery and Diabetes Complications, Tashkent Medical Academy, Uzbekistan. Background and aims: To compare the effects on albumin excretion rate(AER) of two combined preparations: ramipril/hydrochlorothiazide(Ram/htz) and valsartan/hydrochlorothiazide(Val/htz) Materials and methods: This is a 12 months randomized parallel group study. 130patients with hypertension(140mmHg≤ SBP< 180mmHg and DBP< 110), controlled Type 2 diabetes and albuminuria ≥ 20and < 500μg/ min were randomized to Ram/htz(group1,n=61) or to Val/htz(group2, n=69). A dose up-titration from Ram2,5/htz12,5mg up to Ram5/htz25mg or from Val80/htz12,5mg up to Val160/htz12,5mg was allowed up toW12 to achieve blood pressure (BP) targets of SBP S 492 1 C Conclusion: In patients with HT surgery-induced weight loss was associated with a sizeable decline in BP with a high HT remission range. After weight loss patients with DM maintain a high prevalence of abnormal 24-h BP pat- terns and continue to have a high cardiovascular risk. In addition, endothelial function did not improve. Supported by: (PI07/0124) from the Fondo de Investigación San 1206 Associations of visceral fat area and HOMA-R with both diabetes mellitus and hypertension in a health checkup programme: NingenDock H. Hirose1,2, M. Takayama2, Y. Iwao2, H. Kawabe1, Y. Sugino2; 1Health Center, 2Center for Preventive Medicine, Keio University Hospital, Tokyo, Japan. Background and aims: Although the visceral fat area (VFA) value and ho- meostasis model assessment insulin resistance index (HOMA-R) in patients with metabolic syndrome (MetS) or type 2 diabetes mellitus (DM) are known to be higher, the relationship of VFA and HOMA-R with hypertension (HT) remains unclear. In the present study, we investigated the relationship of blood pressure status with VFA, subcutaneous fat area (SFA), HOMA-R, and insulin secretion capacity (HOMA-ß) in male subjects during an annual health check-up program. Materials and methods: A total of 1737 Japanese male subjects, aged 45-74 years, who underwent the whole body health check-up program ‚Ningen- Dock‘ between August 2012 and October 2013 were enrolled in this study. Informed consent was obtained from each subject, and 9 subjects were ex- cluded owing to predetermined exclusion criteria. The subjects were divided into 4 groups according to glucose tolerance status: normal glucose toler- ance (NGT), preDM, DM, and DM with medication (DM+Tx), and blood pressure status: normotensive (NT), high normal (HN), HT, and HT with medication (HT+Tx). Both VFA and SFA were measured at the umbilical level using computed tomography. Results: VFA (mean±SD) in the PreDM (n=425), DM (n=120), and DM+Tx (n=175) groups were significantly higher than that of the NGT (n=1008) group (117±43, 133±54, and 124±54 vs. 101±44cm², F=31.6 and P< 0.0001 by ANOVA). Furthermore, the VFA of the HN (n=231), HT (n=202), and HT+Tx (n=517) groups were significantly higher than that of the NT (n=778) group (109±43, 118±46, and 128±49 vs. 95±42cm², F=58.8 and P< 0.0001). The HOMA-R (median [Q1-Q3] of the HN, HT, and HT+Tx groups were also significantly higher than that of NT group (1.4 [1.0-1.9], 1.5 [1.0-2.1] and 1.6 [1.1-2.7] vs. 1.2 [0.8-1.8], F=38.0 and P< 0.0001). Multiple logistic regression (MLR) analysis of DM as a dependent variable revealed that low HOMA-ß, high HOMA-R, age, VFA, and smoking were relevant. Moreover, MLR analysis of HT as a dependent variable revealed that age, VFA, alcohol consumption (more than 20g ethanol/day), BMI, and HOMA-R were rel- evant (Table 1). Conclusion: These results suggest that VFA and HOMA-R correlate with HT, as well as type 2 DM and MetS, in Japanese male subjects aged 45-74 years. Supported by: MECSST, Japan, No.23500855 1207 Endoscopic, duodenal-jejunal bypass liner in obese patients with or without type 2 diabetes lowers incidence of metabolic syndrome and improves cardiovascular risk R. Chilton1, D. Kim2, E.G.H. de Moura3, A. Liao4, J.-C. Tetreault4, E. Chiquette4, A. Escalona5; 1University of Texas Health Science Center at San Antonio, San Antonio, 2MetaCon, Inc, La Jolla, USA, 3University of Sao Paolo Medical School, Brazil, 4GI Dynamics, Lexington, USA, 5Catolica University, Santiago, Chile. Background and aims: The endoscopically delivered duodenal-jejunal by- pass liner (DJBL) exhibits robust metabolic effects in obese subjects with or without type 2 diabetes (T2D). Several clinical studies have been conducted in development of the DJBL. We report metabolic and cardiovascular (CV) effects of DJBL in obese subjects with or without T2D across 2 uncontrolled clinical studies. Materials and methods: Forty subjects completed the intended 12 month treatment period (baseline mean ± SEM: age 44.5 ± 1.85 yr, BMI 45.2 ± 1.09 kg/m2, A1C 8.6 ± 0.48% and 6.2 ± 0.26% in subjects with [n=13] and without [n=27] T2D, respectively, 77.5 % females). Most common co-morbid condi- tions were hypertension (47.5%), GERD/Gastritis (12.5% each), and hyper- lipidemia (10.0%). Results: At baseline 35/40 ( 87.5%) subject had metabolic syndrome (as defined by ATPIII criteria) and, using the FraminghamBMI, Framingham Lipid, and UKPDS CV risk engine models, the 10-year CV risk was deter- mined to be 11.9 ± 2.09%, 9.0 ± 1.52%, and 5.9 ± 0.93%, respectively. After 12 months of intended implant time, the overall cohort lost 18.6 % in total body weight on average and 17.7 cm from the waist. Additionally, systolic BP dropped by -7.6 mmHg, LDL-cholesterol decreased by 0.6 mmol/L, and A1c reduction of 2.1 % in the diabetic subgroup. The improvement in CV risk fac- tors translated to a 20-32% reduction in 10-year CV risk (see table) and a 37% decreased in subjects meeting the metabolic syndrome criteria, (20/40 sub- jects (p S 493 1 C PS 107 Predicting complications I 1208 Multifactorial risk factor control in clinical practice and risk of cardiovascular disease in type 2 diabetes: report from the Swedish national diabetes register K. Eeg-Olofsson1, S. Gudbjörnsdottir1,2, B. Eliasson1, B. Zethelius3,4, A.-M. Svensson2, J. Cederholm3; 1University of Gothenburg, 2Centre of Registers in Region Västra Götaland, Göteborg, 3Uppsala University, 4Medical Products Agency, Uppsala, Sweden. Background and aims: Cardiovascular complications are still the major cause of morbidity and mortality in patients with type 2 diabetes (T2D). To lower this risk, guidelines advocate multifactorial risk factor control with fo- cus on glucose control, blood pressure (BP) and blood lipids levels. The aim of this observational study was to assess the effect of change in these risk fac- tors on risk for cardiovascular disease (CVD) and mortality in T2D, also in a subgroup with albuminuria. Materials and methods: 47,020 T2D patients from the Swedish National Diabetes Register, aged 30-70 years, were followed from baseline visits 2003- 2006 until 31 Dec 2009, mean follow-up 5.7 years. In all, mean age 60±8 years and diabetes duration 7±6 years. Reference group (n= 13,005) with stable val- ues or increase in HbA1c, systolic BP (SBP) and the ratio non-HDL-to-HDL- cholesterol (non-HDL:HDL) during the study period was compared with seven groups (n=3766 to n=6509) of decrease in either each or in combina- tions of these risk factors. The reference group with stable or increasing risk factor values had mean HbA1c increasing from 52 to 57 mmol/mol, mean SBP increasing from 135 to 141 mmHg, and mean non-HDL:HDL increasing from 2.9 to 3.3. Groups with decrease in HbA1c showed decrease from mean 60 to 52 mmol/mol, with decrease in SBP showed decrease from 146 to 130 mmHg and with decrease in non-HDL:HDL had decrease from 3.3 to 2.5. Results: Compared with reference group hazard ratios (HR) with 95% con- fidence intervals at Cox regression for fatal/nonfatal CVD were 0.45 (0.40- 0.52) with decrease in HbA1c only, 0.31(0.27-0.36) with non-HDL:HDL, 0.27 (0.23-0.31) with SBP, 0.25 (0.22-0.29)-0.36 (0.31-0.41) with decrease in two of three risk factors combined, and 0.24 (0.21-0.28) with all three risk factors decreasing, all p S 494 1 C 1211 Lipoprotein(a), type 2 diabetes and vascular risk in angiographied coronary patients K.-M. Ebner1,2, A. Vonbank3, C.H. Saely3,2, D. Zanolin1,2, P. Rein3, H. Drexel3,4; 1VIVIT Institute, Feldkirch, Austria, 2Private University of the Principality of Liechtenstein, Triesen, Liechtenstein, 3Academic Teaching Hospital Feldkirch, Austria, 4Drexel University College of Medicine, Philadelphia, USA. Background and aims: Lipoprotein (a) [Lp(a)] especially in young individu- als is an important cardiovascular risk factor. However, data on the long-term vascular risk conferred by Lp(a) in patients with type 2 diabetes (T2DM) are scarce. Materials and methods: Lp(a) was measured in a cohort of 909 consecutive patients undergoing coronary angiography for the evaluation of established or suspected stable coronary artery disease; vascular events were recorded over 10 years. Results: Median Lp(a) at baseline was significantly lower in patients with T2DM (n=260) than in subjects without T2DM (10 [interquartile range 1-34] vs. 16 [1-54] mg/dl; p=0.017). Prospectively, 27.8% of our patients suffered vascular events. Lp(a) proved to be a strong and independent predictor of vascular events in total population with a standardized adjusted hazard ra- tio (HR) of 1.15 [1.03-1.27]; p=0.006) as well as in subjects without T2DM (HR 1.22 [1.10-1.36]; p S 495 1 C to the 1955 birth cohort alone. Therefore, despite improved diagnosis of dia- betes and better medical treatment, type 2 diabetes is still associated with an enormous reduction of life-time. 1215 Predictors of ischaemic heart disease and cerebrovascular attack in late elderly diabetic individuals: lessons from 9.1 years study of 4014 diabetic patients K. Ina1, T. Hayashi1, H. Nomura1, M. Takemoto2, K. Yokote2, A. Araki3, H. Sone4, H. Noto5, M. Noda5; 1Department of Geriatrics, Nagoya University Graduate School of Medicine, 2Division of Diabetes, Metabolism and Endocrinology, Department of Internal Medicin, Chiba University Graduate School of Medicine, 3Division of Diabetes, Metabolism and Endocrinology, Tokyo Metropolitan Geriatric Hospital, 4Department of Internal Medicine, Endocrinology and Metabolism, Niigata University Graduate School of Medicine, 5Department of Diabetes and Metabolic Medicine, National Center for Global Health and Medicine, Tokyo, Japan. Background and aims: High LDL-cholesterol (LDL-C) and glucose levels are risk factors for ischemic heart disease (IHD) in middle-aged diabetic indi- viduals; however, the risk among the elderly, especially late elderly older than 75y.o., is not well known. The aim of this study was to identify factors that predict IHD and cerebrovascular attack (CVA) in the elderly and to investi- gate their differences by age, especially legend effect. Materials and methods: We performed a prospective cohort study (Japan Cholesterol and Diabetes Mellitus Study) with 9.1 years of follow-up. A to- tal of 4,014 patients with type 2 diabetes and without previous IHD or CVA (1,936 women; age 67.4±9.5 years, median 70 years; S 496 1 C 1217 Impact of gender on the risk of coronary atherosclerosis and cardiovascular events conferred by HbA1c in subjects without known diabetes A. Muendlein1, C.H. Saely2,3, A. Vonbank2, D. Zanolin1,3, P. Rein2, H. Drexel2,4; 1VIVIT Institute, Feldkirch, Austria, 2Academic Teaching Hospital Feldkirch, Austria, 3Private University of the Principality of Liechtenstein, Triesen, Liechtenstein, 4Drexel University College of Medicine, Philadelphia, USA. Background and aims: Diabetes confers a larger increase in the relative risk of cardiovascular events among women than among men. Whether gender also affects the association of HbA1c with coronary atherosclerosis and car- diovascular events among subjects without known diabetes is unknown. Materials and methods: We enrolled a large consecutive series of 1479 pa- tients undergoing coronary angiography for the evaluation of established or suspected coronary artery disease (CAD), including 495 women and 984 men who did not have previously known diabetes Significant CAD was diag- nosed in the presence of significant coronary stenoses ≥50%. Prospectively, we recorded cardiovascular events over 4.4±1.2 years. Results: Among women, 36.4%, 56.2%, and 7.4% and among men 44.2%, 46.6%, and 9.1% had HbA1c values of S 497 1 C CKD was defined as an estimated glomerular filtration rate (eGFR) S 498 1 C Conclusion: This study is the first southern European validated population- derived model for predicting 5-year fatal/non-fatal CHD and CVD risk in newly-diagnosed type 2 diabetic patients 1223 Single and joint effects of obesity and of the metabolic syndrome on cardiovascular event risk C.H. Saely1,2, D. Zanolin3,2, A. Vonbank1, P. Rein1, H. Drexel1,4; 1Academic Teaching Hospital Feldkirch, Austria, 2Private University of the Principality of Liechtenstein, Triesen, Liechtenstein, 3VIVIT Institute, Feldkirch, Austria, 4Drexel University College of Medicine, Philadelphia, USA. Background and aims: Obesity is a major risk factor for the metabolic syn- drome (MetS), but some obese individuals do not have the MetS while others have the MetS but are non-obese. We prospectively investigated the single and joint effects of obesity and of the MetS on cardiovascular event risk. Materials and methods: Cardiovascular events were prospectively recorded over 10 years in a large cohort of 1705 patients undergoing coronary angiog- raphy for the evaluation of established or suspected stable coronary artery disease. Obesity was defined as a BMI ≥30kg/m²; presence of the MetS was defined according to the current harmonized consensus definition. Results: From our patients, 827 were non-obese and did not have the MetS, 443 were non-obese but had the MetS, 113 were obese but did not have the MetS, and 322 were obese and had MetS. Cardiovascular event risk was 34.1% in non-obese patients with the MetS. It was significantly higher in this pa- tient group when compared to non-obese subjects without the MetS (25.3%; p S 499 1 C 1225 Association between obstructive sleep apneas and complications of type 2 diabetes C. Langrand1, J. Glerant2, F. Gormand3, J. Guerin4, P. Moulin1; 1Endocrinology, Diabetology and Metabolic Diseases, Hôpital Cardiologique Louis Pradel, 2Pneumology, Hôpital Cardiologique Louis Pradel, Bron, 3Pneumology, Centre Hospitalier Lyon Sud, Pierre Benite, 4Pneumology, Hôpital de la Croix Rousse, Lyon, France. Background and aims: Obstructive sleep apneas are highly prevalent in diabetes subjects. Chronic intermittent hypoxia and sleep fragmentation in- duced by sleep apnea increase sympathetic nervous system activity, oxidative stress and inflammation. These alterations are also involved in the develop- ment of diabetes complications, more especially in microvascular complica- tions. The aim of this study was to establish an association between obstruc- tive sleep apnea severity and complications of type 2 diabetes. Materials and methods: A retrospective observational study was set up in 68 type 2 diabetic patients with obstructive sleep apnea syndrome (OSAS). Respiratory parameters of obstructive sleep apnea and occurrence of diabetes complications were reported between diagnosis of sleep apnea and a mean follow-up of 4 years. Results: Participants were diabetic subjects with high cardiovascular risk. Most of them presented a severe OSAS (73%). Mean nocturnal oxygen saturation and time spent with oxygen saturation < 90% were significantly impaired in presence of microvascular complications compared to absence of microvascular complications (90% vs. 93% and 29% vs. 7% respectively, p S 500 1 C thy, respectively. Patients who progressed to retinopathy had higher mean HbA1c (7.4±0.8% vs 6.8±0.7%, p S 501 1 C (12.9%) and retinopathy (10.1%). Female and male prevalence of all diseases except stroke, fatty liver and retinopathy differed significantly (all p S 502 1 C Conclusion: In conclusion, our findings suggest that low levels of ALT are associated with a high risk of all-cause mortality, in particular of non-cardi- ovascular mortality, in patients with type 2 diabetes mellitus with ALT levels largely within the normal range. Supported by: GlaxoSmithKline 1233 Serum PEDF levels in type 2 diabetes: correlations, predictive power for vascular events and increases by fenofibrate in the FIELD Study A. Jenkins, A. Januszewski, D. Calandro, D. Sullivan, J.-X. Ma, J. Best, R. O‘Connell, S. Blankenburg, J. Simes, R. Scott, M. Whiting, P. Barter, M.-R. Taskinen, A. Keech, the FIELD Study; NHMRC Clinical Trials Centre, Sydney, Australia. Background and aims: Pigment Epithelium Derived Factor (PEDF) is an en- dogenous glycoprotein with potent anti-angiogenic, anti-inflammatory and anti-oxidant effects and links with lipoprotein metabolism. In the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) Study of Type 2 dia- betes subjects we aimed to determine if baseline serum PEDF levels (i) are correlated with traditional and novel vascular risk factors; (ii) are associated with on-trial occurrence of microvascular and cardiovascular complications over 5-year follow-up; and (c) are changed by allocation to (up to one year of) 200 mg daily comicronised fenofibrate. Materials and methods: PEDF levels were quantified by ELISA (Merck Mil- lipore, USA) with CVs S 503 1 C 1235 Circulating resistin levels and mortality risk A. Fontana1, S. Spadaro2, M. Copetti1, B. Spoto3, L. Salvemini4, P. Pizzini3, L. Frittitta2, F. Mallamaci3, S. Bacci5, F. Pellegrini1, V. Trischitta4, C. Menzaghi4; 1Unit of Biostatistics, IRCCS, San Giovanni Rotondo, 2Department of Clinical and Molecular Biomedicine, University of Catania, 3Nephrology, Dialysis and Transplantation Unit, CNR-IBIM, Reggio Calabria, 4Research Unit of Diabetes and Endocrine Diseases, IRCCS, 5Unit of Endocrinology, IRCCS, San Giovanni Rotondo, Italy. Background and aims: Studies concerning the association between circulat- ing resistin levels and mortality yielded mixed Results: some studies reported significant associations, while many others showed negative results. To ad- dress this question, we analyzed data from European subjects with type 2 diabetes and evidence of coronary artery disease (CAD), a prospective cohort followed over time for cardiovascular and all-cause mortality. To assess the consistency of the association between resistin and mortality risk, we also performed a meta-analysis of prospective studies. Materials and methods: We investigated: i) the Gargano Heart Study (GHS) prospective design (n=359 diabetic patients; 81 and 58 incident cases of all- cause and cardiovascular (CV) mortality, respectively); ii) meta-analized in a dose-risk fashion our present data from GHS study and all studies from MEDLINE and EMBASE until February 2014 reporting adjusted hazard ra- tios (HR) of circulating resistin for all-cause or CV mortality. Results: In GHS, the adjusted HRs per 10 ng/ml resistin increment were: 1.45 (95% CI: 1.10-1.91) and 1.52 (95% CI: 1.09-2.12) for all-cause and CV mortalities, respectively. The dose-risk meta-analyses included 7 studies (n=3,699, 997 events) for all-cause mortality and 6 studies (n=4,187, 412 events), for CV mortality. Pooled HRs per 10 ng/ml resistin increment were 1.33 (95% CI: 1.03- 1.72, p=0.028, Q-test p for heterogeneity S 504 1 C 1238 ProBNP strongly predicts future macrovascular events in angiographied coronary patients as well as in those without the metabolic syndrome B. Patsch1, P. Rein2, C.H. Saely2,3, A. Vonbank2, D. Zanolin1,3, H. Drexel2,4; 1VIVIT Institute, Feldkirch, Austria, 2Academic Teaching Hospital Feldkirch, Austria, 3Private University of the Principality of Liechtenstein, Triesen, Liechtenstein, 4Drexel University College of Medicine, Philadelphia, USA. Background and aims: Pro-B-type natriuretic peptide (proBNP) is a prog- nostic biomarker for patients with congestive heart failure as well as in other patient populations. The power of proBNP to predict cardiovascular end- points in patients with the metabolic syndrome (MetS) is unclear and is ad- dressed in the present study. Materials and methods: We measured serum proBNP in 722 patients un- dergoing coronary angiography for the evaluation of stable coronary artery disease (CAD). Significant CAD was diagnosed in the presence of coronary stenoses with lumen narrowing of ≥50%. Prospectively, we recorded vascular events over 3.2±1.2 years. Results: ProBNP was significantly higher in patients with (n=386) than in subjects without significant CAD at baseline (711±1287 vs. 663±1565 pg/ ml; p=0.001). Prospectively, we recorded 121 cardiovascular events. The in- cidence of vascular events significantly increased over tertiles of proBNP in patients with the MetS (10.7%, 18.5%, and 28.8% respectively; p=0.004) was well as in those without the MetS (10.4%, 11.5%, and 22.0%, respectively; p=0.011). Similarily, serum proBNP significantly predicted the incidence of major cardiovascular events after adjustment for age, gender, BMI, smoking, systolic and diastolic blood pressure, LDL cholesterol, HDL cholesterol and the eGFR both in subjects with the MetS (standardized adjusted HR 1.48 [1.21-1.80]; p S 505 1 C PS 111 Biomarkers and complications II 1241 Protective effect of telomere length in patients with type 2 diabetes mellitus E.N. Dudinskaya1, N.V. Brailova1, I.D. Strazhesko1, O.N. Tkacheva1, S.A. Boytsov1, M.V. Shestakova2; 1National Research Centre for Preventive Medicine, 2National Research Center for Endocrinology, Moscow, Russian Federation. Background and aims: Type 2 diabetes mellitus (T2DM) is the primary risk marker influencing risk of cardiovascular disease. It is known that telomere length (TL) shortening is a marker of cell aging and associated with increased arterial stiffness (AS) in patients with T2DM, despite the patient‘s age, lead- ing to the vascular aging. The aim of the study was to compare the vascular and cellular aging in patients with and without T2DM. Materials and methods: TL was assessed by quantitative polymerase chain reaction (PCR) in 98 patients with T2DM (mean age 61±2,6 years) and in 101 healthy patients in mean age of 51±1,8 years. Intima media thickness (IMT) and plaque presence (PP) were determined by ultrasonography in both left and right carotid arteries. AS was appreciated by aortic pulse wave velocity (PWV) measuring by SphygmoCor (AtCor Medical). Results: The median of telomere length (TLL) was 9,75. «Short» telomeres were considered if the telomere length was 9.75. All patients were divided into 4 groups by TL - “long” (T2DM+ (n=57) and T2DM- (n=49)) and “short” telomeres ((T2DM+ (n=41) and T2DM- (n=52)). Comparison of vascular aging parameters in patients with long telomeres showed that the state of vessels in T2DM were as similar as in healthy people - they had a lower arterial stiffness and other signs of vascular ageing : PWV 10,58±0,1 (T2DM+) vs 10,5±0.5 m/s (T2DM- ), p=0,913; IMT 0,904±0,09 (T2DM+) vs 0,77±0,03 mm (T2DM-), p=0,1227; PP 0,886±0,4 (T2DM+) vs 0,782±0,2 (T2DM-), p=0,979. In contrast in patients with «short» TL and DM PWV was significantly higher than in non-diabetic people (15,08±1.3 vs 10,7±0.5 m/s, p=0,0151), there were greater number of plaques (1.125±0,29 vs 0,789±0,22, p=,04); IMT was 0,87±0,1 (T2DM+) vs 0,78±0,1 mm (T2DM-), p=0,1814. Conclusion: In patients with short TL and T2DM the severity of vascular disorders is higher than in healthy people. In contrast, in patients with long TL with T2DM there are no significant differences in the vascular struc- ture as compared with healthy individuals. Despite the presence of T2DM signs of vascular aging are minimal in patients with long TL.Short TL may be regarded as non-hemodynamic components of rapid vascular ageing in patients with T2DM. 1242 Sleep apnoea syndrome and type 1 diabetes: impact on macroangiopathies J.-P. Courrèges1, E. Ghanassia2, A. Iosup1, J.-F. Thuan1, P. Fisher2, H. Bonnaure1, N. Vigier-Simmore1; 1Centre Hospitalier Général de Narbonne, 2Clinique Sainte Marguerite, Sète, France. Background and aims: Sleep Apnoea Syndrome (SAS) is an independent cardiovascular morbi-mortality factor that is undoubtedly enhanced when associated with T1D. The aim of this study is to assess prevalence and profile of macroangiopathies in this situation (T1D + SAS), more precisely blood pressure and potential conduction defects (QT interval). Materials and methods: SAS was found in 28 patients (20.7%) from a 135 T1D cohort consecutively enrolled with Apnoea Hypopnoea Index (AHI) )≥15/h confirmed with polysomnography. Among them, respectively 50% and 50% suffer from severe (AHI≥30/h) or moderate (AHI≥ 15 ≤29/h) SAS. Results: Patients’ characteristics at enrolment were: age 64 ± 12.9 years, dia- betes duration 27.1 ± 13.7 years, 75% of males, HbA1c 8.0 ± 1.0% (39.3% with HbA1c≤7.5%), weight: 75.2 ± 18.8kg, BMI : 26.5 ± 5.8 kg/m2 (50% lower than 25kg/m2, 21.4% between 25 and 26.9kg/m2 et 28.6% higher than 27 kg/m²). Among them, 21 patients (75% of SAS T1D and 15.5% of all T1D) show at least one macroangiopathic damage (MA+): 8 coronaritis, 15 carotid damages, 13 lower limb arteriopathy). MA+ patients were compared to MA- patients on HbA1c (8.0 vs 8.0%), BMI (27.2 vs 24.7 kg/m2), creatinemia (10.4 vs 8.2mg/l), total cholesterol (1.8 vs 2.1g/l), triglyceride level (1.4 vs 0.6g/l), neuropathy frequency (81 vs 57.1%), nicotine addiction frequency (14.3 vs 0%), Hypoxia (AHI : 30.4 vs 23.9/h ; 57.1 vs 28.6% of severe SAS and 42.9 vs 71.4% of moder- ate SAS). We have also studied blood pressure (SBP : 126.5 ± 15.3 vs 126.4 ± 15.5mmHg/DBP : 75.6 ± 7.7 vs 73.6 ± 10.3mmHg), Mean Arterial Pressure (MAP: 93.9 ± 9.2 vs 97.0 ± 3.0mmHg) and Pulse Pressure (PP: 50.9 ± 17.2 vs 55.0 ± 13.8mmHg). The electrocardiogram of MA+ patients shows mean QT and QTC intervals of 401.2 ± 42.5 et 414.9 ± 22.4 mm/s while the one of a 22 patients control group (C) with no SAS and no macroangiopathy shows mean QT and QTC intervals of 382.9 ± 22.5 et 398.8 ± 12.7 mm/s. There is a signifi- cant difference in QTC intervals (p = 0.012). No heart rhythm disorder is ob- served. SBP, DBP, MAP and PP have been compared too (SBP: 126.5 ± 15.3 vs 124.7 ± 11.4 mmHg - NS; DBP : 75.6 ± 7.7 vs 72.5 ± 9.4 mmHg - NS; MAP : 93.9 ± 9.2 vs 89.8 ± 9.4 mmHg - NS; PP : 50.9 ± 17.2 vs 52.2 ± 7.7mmHg - NS). The median QTC interval stands respectively (MA+ vs C) at 409 vs 396 mm/s; the proportion of patients with QTC interval longer than 430 mm/s is 14.3 vs 0%. Conclusion: The frequency of SAS in T1D patients is high and this popula- tion often suffers from macroangiopathies. The study of T1D with SAS and macroangiopathy underlines a high frequency of associated neuropathy. The comparison between this group and a control group with no SAS and no macroangiopathy shows no difference in usual blood pressure characteristics. On the contrary, the study of QT (and particularly QTC) interval shows a significant lengthening. Referring to its potential implication in rhythm dam- age, it has to be systematically assessed in this context. 1243 Presence of type 2 diabetes mellitus significantly modulates the power of thyroid stimulating hormone to predict cardiovascular mortality E. Kinz1,2, A. Vonbank3, C.H. Saely3,2, D. Zanolin1,2, P. Rein3, H. Drexel3,4; 1VIVIT Institute, Feldkirch, Austria, 2Private University of the Principality of Liechtenstein, Triesen, Liechtenstein, 3Academic Teaching Hospital Feldkirch, Austria, 4Drexel University College of Medicine, Philadelphia, USA. Background and aims: Elevated thyroid stimulating hormone (TSH) is associ- ated with an adverse cardiovascular risk profile, especially in patients with type 2 diabetes (T2DM). We investigated the association between TSH and cardio- vascular mortality in patients with T2DM as well as in non-diabetic subjects. Materials and methods: We measured TSH in a high-risk cohort of 1741 consecutive patients undergoing coronary angiography for the evaluation of established or suspected coronary artery disease (CAD). The incidence of vascular events was recorded over 10 years; T2DM was defined according to current ADA criteria. Results: From our patients, 34% suffered vascular events. TSH proved to be a strong and independent predictor of cardiovascular mortality in subjects without T2DM (n=1220; standardized adjusted hazard ratio (HR) 1.11 [1.00- 1.24]; p=0.036), but not in patients with T2DM (n=521; HR 0.99 [0.87-1.14]; p=0.934). An interaction term TSH x T2DM was significant (p=0.039), indi- cating that TSH was a significantly stronger predictor of vascular events in subjects without T2DM than in patients without T2DM. Conclusion: From the data of this prospective cohort study we conclude that presence of T2DM significantly modulates the power of TSH to predict car- diovascular mortality. 1244 Is it possible to improve the prediction of silent myocardial ischaemia in type 1 diabetes? G. Llauradó1, A. Cano2, C. Hernández3, A. Rodríguez-Revuelto4, M. González-Sastre5, J. Puntí6, L. Albert2, R. Simó3, J. Vendrell1, J.M. González-Clemente2; 1Endocrinology and Diabetes Unit, CIBERDEM. Hospital Universitari Joan XXIII de Tarragona. IISPV. Universitat Rovira i Virgili, 2Endocrinology and Nutrition Department, Hospital de Sabadell. Corporació Sanitària i Universitària Parc Taulí (UAB), 3Diabetes and Metabolism Research Unit. CIBERDEM. Institut de Recerca Hospital Universitari Vall d‘Hebron. Universitat Autònoma de Barcelona, 4Nuclear Medicine Department, UDIAT Centre Diagnòstic. Corporació Sanitària i Universitària Parc Taulí (UAB), Sabadell, 5Ophthalmology Department, 6Cardiology Department, Hospital de Sabadell. Corporació Sanitària i Universitària Parc Taulí (UAB), Spain. Background and aims: Cardiovascular disease (CVD) is the major cause of mortality in type 1 diabetes (T1DM). Given the higher prevalence of silent Diabetologia (2014) 57:[Suppl1]S1–S564 S 506 1 C myocardial ischemia (SMI) in this population, the need to screen asymp- tomatic T1DM patients assumes increasing importance. The aims of this study were: 1) to assess the prevalence of SMI in asymptomatic patients with T1DM, 2) to evaluate the clinical predictors associated with its presence, in- cluding the evaluation of microvascular complications and the assessment of arterial stiffness (AS) (as a measurement of pre-clinic arteriosclerosis) and 3) to evaluate the diagnostic value of the previous established criteria for its diagnosis (ADA 1998) and how to improve them. Materials and methods: 80 patients with T1DM (without previous history of CVD) were screened for SMI by stress myocardial perfusion gated-SPECT (Single Photon Emission Computed Tomography). Data on traditional car- diovascular risk factors and microvascular complications were also recorded. AS was assessed by aortic pulse wave velocity (aPWV - Sphygmocor®). Results: The baseline characteristics of the patients included in the study are shown in Table 1. The prevalence of SMI was 12.5%. SMI was associated with the presence of hypertension (OR 4.5 (95%CI 1.1-18.9); p=0.041) and HbA1c value (OR 2.4 (95%CI 1.1-5.0); p=0.023.). The ADA criteria for the screening of SMI showed a good predictive value (OR 8.3 (95%IC 1.7-40.6); p=0.009. AUC 0.73 (0.54-0.91)). However, the addition of HbA1c value (AUC 0.87 (0.75-0.99) vs. 0.73 (0.54-0.93); p=0.034) and the presence of diabetic retin- opathy (AUC 0.86 (0.74-0.98) vs. 0.73 (0.54-0.92); p=0.057) improved the predictive value, thus resulting in the best model of SMI prediction (AUC 0.91 (0.83-0.99) vs. 0.73 (0.54-0.92); p=0.026). aPWV did not add any signifi- cant improvement for SMI identification (p=0.583). Conclusion: The prevalence of SMI in patients with T1DM without previous CVD is 12.5%. Although The ADA criteria for the screening of SMI predicted its presence adequately, adding HbA1c and taking into account the presence of diabetic retinopathy significantly improve the prediction rate. Supported by: BGLV 2012 (ACD); PI12/0095 (ISCIII); GLC - Rio Hortega (CM12/00044. ISCIII) 1245 Integrative analysis of biomarkers for cardiovascular complications through networks of metabolites and proteins R.A. Momo1, F. Vitali2, F. Mulas3, V. Schnecke1, R. Bellazzi2; 1Translational Science, AstraZeneca, Mölndal, Sweden, 2Dipartimento di Ingegneria Industriale e dell‘Informazione, 3Centro Interdipartimentale di Ingegneria Tissutale, Università degli Studi di Pavia, Italy. Background and aims: The SUMMIT project (SUrrogate markers for Micro- and Macro-vascular hard endpoints for Innovative diabetes Tools), seeks to identify genetic and non-genetic biomarkers for prediction of disease pro- gression and risk of vascular complications. A collection of samples from dif- ferent cohorts have been provided by combining a variety of high-throughput techniques encompassing proteomics and metabolomics. Integrative bioin- formatics methods that combine different data sources may provide deeper insights into the molecular interactions underlying complex diseases. Within this context, we constructed a network with proteins and metabolites as nodes whose analysis suggests the most relevant entities involved in the dis- ease pathways. Materials and methods: Starting from a list of SUMMIT-discovered bio- markers, i.e., proteins and metabolites that have been shown to be associated with cardiovascular complications we obtained a disease-related network by combining evidences from different repositories. In detail, STRING, STITCH and MetaCore were used to initially build individual networks that were sub- sequently merged into a combined network associating the discovered mark- ers to other molecules known to interact with them. Topological analysis of this network has been used to identify enriched biological pathways and mechanisms, and to suggest a set of relevant nodes as candidates for pharma- cological therapies. Results: The network representation could elucidate molecules and interac- tions involved in the disease and better define its global picture. While the different repositories can be used separately to build disease networks, merg- ing these networks provided a better model of the whole disease context. The list of pathways identified as significantly enriched (p-value < 0.001) con- tained those related to insulin resistance as well as others that have not previ- ously been identified as potential targets for standard diabetes therapies. For example, the signaling by Insulin-like Growth Factor 1 Receptor (IGF1R) was amongst those identified. This receptor is a negative regulator of insulin sign- aling and currently has not been taken into account for diabetes therapies. However, it has been shown that cardiac overexpression of IGF-1R in mice prevents diabetes-induced cardiac fibrosis and diastolic dysfunction. Conclusion: We believe that higher integration across different systems biol- ogy platforms eases interpretation of discoveries, supports validation of hy- pothesis and ultimately speeds up the drug discovery process. Supported by: IMI, European Commission’s FP7 (the SUMMIT consortium, IMI-2008/115006) 1246 Correlation of inflammatory markers and insulin resistance indexes with the short- and long-term outcome after an acute coronary syndrome in different glycaemic categories A. Koutsovasilis1, A. Papazafiropoulou1, A. Sotiropoulos1, A. Melidonis2, S. Foussas3, T. Peppas1, E. Tamvakos1, S. Bousboulas1; 13rd Internal Medicine Department & Diabetes Center, General Hospital of Nikaia-Piraeus, Athens, 21st Internal Medicine Department & Diabetes Center, 3Cardiology Department, Tzanio General Hospital of Piraeus, Greece. Background and aims: The glucose metabolism disorder is frequent in pa- tients who are admitted after an Acute Coronary Syndrome (ACS). The aim of this study is to examine the correlation of ACS outcome with insulin resist- ance indexes and inflammatory markers in known diabetes patients, newly diagnosed diabetes patients, patients with impaired glucose tolerance (IGT) and normoglycaemic patients. Materials and methods: 536 patients who were admitted at the Cardiology Department within a two-year period were included in the study. All patients underwent a clinical laboratory testing including insulin level in order to measure insulin-resistance indexes (HOMA- Homeostasis Model Assess- ment and QUICKI- Quantative Insulin Sensitivity Check Index) and inflam- matory markers (high-sensitivity C-reactive protein (hs-CRP), white blood cells count (WBC), fibrinogen and erythrocyte sedimentation rate (ESR)). All non diabetic patients went under an oral glucose tolerance test (OGTT) one month after discharge. Study’s end-points were death, a new ACS, arrhyth- mias and acute pulmonary oedema of cardiological origin during hospitaliza- tion and 12 months after the ACS. Results: 199 (37,12%) patients were normoglycemic while 168 (31,34%) were known diabetics, 59 (11%) newly diagnosed diabetic patients and 110 (20,52%) were IGT patients. There was a statistically significant corre- lation of the hs-CRP with the outcome during hospitalization (HR=1.712, 95%CI:1.193-2.458, p=0.004) for all patients. hs-CRP (HR=2.612, 95%CI:1.474-4.628, p=0.001) and HOMA index (HR=1.967, 95%CI:1.184- 3.267, p=0.009) were significantly correlated with the appearance of a major cardiovascular event (MACE) during hospitalization for all patients but only hs-CRP was correlated with MACE for each group separately. Complications during the first 12 months after an ACS were significantly correlated only with the hs-CRP for all patient categories. hs-CRP, HOMA and QUICKI indexes were correlated with MACE in a multivariate analysis for known diabetes patients (HR=1.542, p=0.023, HR=1.482, p=0.039 and HR=0.810, Diabetologia (2014) 57:[Suppl1]S1–S564 S 507 1 C p=0.031 respectively), newly diagnosed diabetes patients (HR=2.401, p=0.035, HR=1.364, p=0.046, HR=0.832, p=0.041) and IGTs (HR=1.354, p=0.023, HR=1.269, p=0.047, HR=0.782, p=0.036) 12 months after the ACS. Association of hs-CRP (HR=2.012, p=0.008), HOMA (HR=2.438, p=0.002) and QUICKI (HR=0.503, p=0.010) with MACE was significant in an univari- ate analysis while only hs-CRP (HR=1.477, p=0.032) remained significant in a multivariate analysis for normoglycaemic patients Conclusion: hs-CRP is the inflammation marker which is correlated with the appearance of complications during hospitalization and 12 months after an ACS, while HOMA and QUICKI indexes are correlated with the appearance of major cardiovascular events during hospitalization as well as the 12-month follow-up period regardless of patient glycaemic profile. The combination of these indexes with hs-CR significantly improves the latter’s prognostic model in particular regarding major cardiovascular events. 1247 Both DHA/AA ratio and absolute DHA levels, but not EPA levels, constitute an independent risk factor for the severity of coronary atherosclerosis in type 2 diabetic patients T. Nakamura, K. Murata, R. Kaneko, K. Ikeda; Omihachiman Medical Community Center, Omihachiman, Shiga, Japan. Background and aims: The concentrations of dietary n-3 polyunsaturated fatty acids (PFAs), such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) modify the incidence of coronary heart disease (CHD). Espe- cially, the effects of EPA on cardiovascular events have been clarified through recent large-scale intervention clinical studies. However, it is still unknown whether DHA may make a contribution to clinical efficacy associated with the risk of CHD. Thus, we investigate which n-3 PFA is more effective on the association of the severity of CHD in type 2 diabetic patients. Materials and methods: 184 type 2 diabetic subjects were nominated as poor blood glucose control from April 2008 to September 2010 and aged 40 to 72 years on this trial. We firstly calculated fasting plasma PFAs, LDL-cholesterol, HDL cholesterol and triglyceride (TG) levels. The subjects were asked about daily intake of different type of fish and n-3 PUFAs intake was calculated using computer food composition software. We excluded the subjects who reported the history of CHD or ECG abnormality consistent with ischemia at baseline survey. Then they had been followed for five years, focusing on the occurrence of major cardiovascular events or ECG abnormality for an observation period, which compromised nonfatal and fatal CHD diagnosed using percutaneous coronary intervention. CHD severity was assessed using the Gensini scoring system. Next, we divided them into 2 groups, who were in the presence of CHD (DM-PCI, N=56) and no episodes of CHD (DM, N=70) and 71 healthy volunteers were nominated (NGT). These values were analyzed using an ANOVA model, Pearson’s test and multivariate logistic re- gression models analyses. Results: The ratios of smokers and subjects with antihypertensive and/or li- pid modifying treatment were not significant between DM-PCI and DM. The average of systolic blood pressure, age, BMI, the concentrations of HbA1c, LDL, HDL and TG were no different between two groups. On the contrary, the concentrations of FPG, HbA1c, LDL, TG and remnant like particle (RLP) were significantly lower in NGT than in DM-PCI and DM. Firstly, we inves- tigated the relation between the basal plasma concentrations of both EPA and DHA and daily fish intake was the positive correlation in the diabetic sub- jects (r=0.77 p S 508 1 C 1249 Soluble urokinase plasminogen activator receptor is elevated in diabetes and associated with diabetic complications in patients with type 1 diabetes S. Theilade1, S. Lyngbaek2, T.W. Hansen1, J. Eugen-Olsen3, M. Fenger4, P. Rossing1,5, J. Jeppesen2,5; 1520, Steno Diabetes Center, Gentofte, 2Department of Medicine, Glostrup Hospital, Glostrup, 3Clinical Research Centre, Hvidovre Hospital, 4Department of Biochemistry, Hvidovre Hospital, 5Aarhus University, Denmark. Background and aims: Soluble urokinase plasminogen activator receptor (suPAR) is a marker of inflammation and endothelial dysfunction. We in- vestigate associations between suPAR and diabetic complications in type 1 diabetes. Materials and methods: From 2009-2011, 667 type 1 diabetes patients and 51 non-diabetic control subjects were included in a cross-sectional study at our center. suPAR was measured with ELISA (ViroGates, Denmark). Dia- betic complications were cardiovascular disease (CVD) (previous myocar- dial infarction, revascularisation, peripheral arterial disease and stroke), autonomic dysfunction (heart rate variability during deep breathing < 11 beats per minute), albuminuria (urinary albumin excretion rate (UAER) ≥ 30 mg/24-hours) or high arterial stiffness (pulse wave velocity ≥ 10 m/s). Adjusted analyses included gender, age, systolic blood pressure, estimated glomerular filtration rate, UAER, HbA1c, total-cholesterol, body mass index, C-reactive protein, antihypertensive treatment and smoking. Results: suPAR was lower in controls vs. patients, controls vs. normoalbumi- nuric patients (< 30 mg/24-hours), normoalbuminuric patients with short vs. long diabetes duration (> 10 years), and increased with degree of albuminuria (adjusted p < 0.001 for all). Furthermore, suPAR levels were higher in patients with vs. without CVD (n = 144 (21.3%)), autonomic dysfunction (n = 349 (59.2%)), albuminuria (n = 357(53.1%)) and high arterial stiffness (n = 297 (47.2%)) (adjusted p ≤ 0.024). Per 1 unit increase in ln-suPAR adjusted odds ratios were for CVD: 2.5 (1.1-5.7), autonomic dysfunction: 2.7 (1.2-6.2), al- buminuria: 3.8 (1.3-10.9) and high arterial stiffness: (2.5 (1.1-6.1) (p ≤ 0.039). Conclusion: suPAR is associated with type 1 diabetes, diabetes duration and complications independently of other risk factors. suPAR may be a novel risk marker in the management of diabetes. Clinical Trial Registration Number: 2009-056; NCT01171248 PS 112 Clinical observations in children 1250 The role of glucagon in post-prandial hyperglycaemia for children 5 years after onset of type 1 diabetes S. Fredheim1,2, M.L.M. Andersen1, S. Pørksen1, L.B. Nielsen1, C. Pipper3, L. Hansen1, J. Johannesen1,2, H.B. Mortensen1,2, J. Svensson1,2; 1Pediatric Department, Herlev Hospital, 2Copenhagen University, 3Department of Biostatistics, Copenhagen University, Denmark. Background and aims: The role of glucagon on glycemic control in type 1 diabetes is debated, and high glucose concentrations have previously been shown to increase glucagon release. We investigated stimulated glucagon and glucagon-like peptide-1 (GLP-1) levels in children during 12 to 60 months after onset of type 1 diabetes. Materials and methods: The study cohort comprised 129 children (66 boys) mean (SD) age at onset; 10.04 (3.9) range 0.6-16.6 years with newly diagnosed type 1 diabetes. Liquid mixed-meal (Boost) stimulated levels of C-peptide, glucagon, glucagon-like peptide-1 (GLP-1), gastric inhibitory polypeptide (GIP), blood glucose (measured 90 min post-ingestion of Boost drink), and HbA1c (DCCT), were analysed centrally 1, 3, 6, 12 and (N=40) 60 months (m) after diagnosis. Robust multivariate regression methods were used for statistical analyses. Results are listed as RR [95% CI] corresponding to a dou- bling of the explanatory variable. Analyses were adjusted for age, sex, diabetes duration and stimulated C-peptide. Results: During the 60 months, post-prandial glucagon levels were highly as- sociated to the rise in stimulated GLP-1 (RR 1.33 [1.19; 1.48], p S 509 1 C egorised as S 510 1 C betes on the cerebrovascular system with a simple, non-invasive method the transcranial Doppler sonography. Materials and methods: We performed transcranial Doppler sonography in 62 type 1 diabetic patients (older than 10 years of age, mean 14.59±2.46 years) treated in our hospital and 46 (mean age 13.66±2.06) healthy children served as controls. We examined the middle cerebral arteries in both sides in lying position during rest by registering the velocities (peak systolic and end diastolic). Than we asked the children to stop their breathing as long as they can (but at least for 30 seconds) to increase the PCO2 which led to vasodila- tation. After a short period of resting period came the decreasing of PCO2 (causing vasoconstriction) by asking the children to take deep breaths for 30 seconds. After collecting these data we calculated the pulsatility index ((sys- tolic velocity-diastolic velocity)/mean velocity)and resistive index ((systolic velocity-diastolic veolcity)/systolic velocity) in all the 3 cases. Results: The Kolmogorov-Smirnov test showed that all of the investigated variables showed normal distribution thus two-tailed t tests could be per- formed to find statistical differences between diabetic patients and control group. After processing the data we can conclude that the diabetic group showed lower values in all investigated variables. The strongest difference (p=0.0214) could be found in the peak systolic velocities after hypervenitla- tion (82.8 cm/s vs. 91.8 cm/s). An other significant (p=0.0401) difference was detected in the peak systolic velocities after hypoventilation (136.6 cm/s vs. 146.5 cm/s), while all other findigs proved to be non-significant. Conclusion: The fact that all measured velocities were lower among diabet- ic children and adolescents suggests that some kind of pathophysiological changes must have been started among these patients and this harmful influ- ence would proceed and would be responsible for the further well-known effects among adults. We assume that the first phase of this process affects the response for the ,,CO2 challenge”, that’s why the peak systolic velocities are influenced after both hypo- and hyperventilation. Besides raising the number of performing transcranial Doppler sonography we are working on inden- tification of factors which can play a role forming the changes among our patients. 1255 40 year follow up of childhood onset diabetes reveals little long-term decline in the cumulative incidence of complications by diagnosis cohort except for renal failure T. Orchard, T. Costacou, R.G. Miller; Department of Epidemiology, University of Pittsburgh, USA. Background and aims: Despite dramatic improvements in life expectancy, morbidity rates in type 1 diabetes (T1D) still greatly exceed those of the gen- eral population. To evaluate trends, by year of diagnosis, in the cumulative incidence of micro- and macro-vascular complications over 22 years of fol- low-up, the prospective, observational, Pittsburgh Epidemiology of Diabetes Complications (EDC) study of childhood-onset T1D study was examined. Materials and methods: Participants were categorized into five calendar year onset cohorts: 1950-59, 1960-64, 1965-69, 1970-74, and 1975-80. Renal failure (RF, self-report), hard coronary artery disease (CAD; MI, fatal CAD and revascularization), blindness (self-report) and amputations (self-report) were determined at 20, 25, 30, 35 and 40 years duration on the complete cohort (n=912). Proliferative retinopathy (PR; fundus photography), con- firmed distal symmetric polyneuropathy (CDSP; clinical exam plus vibratron threshold), symptomatic autonomic neuropathy (SAN; 2+ symptoms and expiration/inspiration ratio S 511 1 C Results: Median (interquartile range) duration of T1D was 6.1 (3.3-9.6) years; mean (±SD) A1c was 8.5±1.8% (69±20 mmol/mol). A1c targets were achieved by 28% of patients (32% of 8-12 y/o, 29% of 13-18 y/o, and 19% of 19-25 y/o). Overall, 351 (5.9%) patients reported diabetic ketoacidosis (DKA) in the past 3 months, with the highest occurrence in 19-25 y/o (6.6% vs 5.6% in 8-12 y/o and 5.7% in 13-18 y/o). Severe hypoglycaemia leading to seizure or loss of consciousness was reported by 162 (2.7%) patients with the highest occurrence in 19-25 y/o (4.1% versus 2.5% in 8-12 y/o and 2.2% in 13-18 y/o). Incidence of acute T1D complications by A1c goal attainment and age are shown (Table). In all age groups, occurrence of DKA rates was higher in those not at A1c target. Rates of microalbuminuria, retinopathy treatment, and neuropathy were highest in 19-25 y/o and in patients not at target. In multivariate analyses, presence of DKA (OR 0.53 [95% CI 0.39, 0.73]) and diabetic neuropathy symptoms (OR 0.47 [95% CI 0.34, 0.66]) were signifi- cantly associated with not attaining A1c target (p S 512 1 C CSII. It was the aim of the trial to analyse the quality of diabetes control and patients‘ ability for a sufficient diabetes self-management using CSII versus injection therapy. Materials and methods: All children/adolescents (n=901, age 11.5±4.0, dia- betes duration 4.0±3.6 years) with type 1 diabetes were enrolled in the trial, admitted to an in-patient rehabilitation during the period 04/2004-10/2010. Results: At the time of hospital admission n=707 patients (78%) had an in- jection therapy (IT), n=194 (22%) had a CSII. Quality of diabetes control (IT vs CSII: HbA1c 8.72±2.26 vs 8.35±1.71%, p=0.09), mean amplitude of daily blood glucose excursions (9.4±3.4 vs 9.8±3.2 mmol/l, p=0.22), inci- dences of acute complications (hypo-/hyperglycaemia) were comparable. Children/Adolescents with CSII had a longer diabetes duration (5.3±3.7 vs 3.6±3.5 years, p< 0.01), better postprandial blood glucose levels (10.0±3.0 vs 12.2±3.3 mmol/l, p=0.03) and a higher frequency of blood-glucose self-tests (45.4±13.3 vs 38.2±11.5 self-tests/week, p< 0.01). Neither there were substan- tial differences in the group of children younger than 6 years: CSII (n=34) vs IT (n=58) (HbA1c 7.44±0.82 vs 7.45±1.21%, p=0.99, mean amplitude of daily blood glucose excursions 10.5±3.4 vs 10.4±3.1 mmol/l, p=0.87, num- ber of hypoglycaemia during the preceeding 4 weeks 15.4±11.1 vs 14.1±11.7, p=0.61). However, CSII offers children and adolescents more flexibility, more effective diabetes self-management and in adolescents a higher quality of life (p 1SD from the mean was considered pathological and with a high risk of developing eating disorder. Results: 44 patients completed the EDE-Q test, 6 of them reported a score > 1SD (6/44, 13.2%) and were considered at risk of developing eating disorders. The logistic regression showed that, among all the anthropometric and meta- bolic variables, the only one that could be considered a predictor of a high score at the EDE-Q test (> 1SD from the mean) was the BMI z-score before the onset of diabetes. The BMI z-score (index of overweight/obesity) before the onset of T1DM was then identified as an independent predictor of a posi- tive EDE-Q test (p S 513 1 C 1262 International comparison of quality of life in children and adolescents with type 1 diabetes A. Lukács1, E. Kalyva2, M. Abdul-Rasoul3, L. Barkai1,4; 1Faculty of Health Care, Department of Theoretical Health Sciences, University of Miskolc, Hungary, 2The International Faculty of the University of Sheffield, City College, Thessaloniki, Greece, 3Faculty of Medicine, Kuwait University, Kuwait, 4Postgraduate Institute of Pediatrics, Medical and Health Science Center, University of Debrecen, Miskolc, Hungary. Background and aims: The ultimate goal of chronic patient care is to main- tain or improve the quality of life. The patients’ quality of life may be affected the successful disease management in short- and long term. There is a lack of researches examining the quality of life in pediatric diabetic population across countries, because of the lack of appropriate validated age- and dis- ease-specific quality of life instruments. Our research aimed to assess the diabetes-specific quality of life in children and adolescents with type 1 diabe- tes from different geographically located countries using the child self- and parent proxy-reports. We looked for the relationship between the glycemic control and quality of life. Materials and methods: A total of 416 youths with type 1 diabetes (212 boys and 204 girls) aged 8-18 years participated in the study (84 Greek, 135 Hun- garian and 197 Kuwait participants). There were no significant differences among the study participants in gender, age and glycemic control. Diabetes- specific quality of life was measured using the Pediatric Quality of Life™ 3.0 Diabetes Module culturally adapted in every country. Glycemic control was expressed by HbA1c. Data were analysed with Chi-square test, paired t-test, one-way analysis of variance. The difference across countries was evaluated using Tukey post hoc test. SPSS 19.0 was used for statistical analysis. Results: There were significant differences among countries both in self- (F(2, 413)=13.17, p S 514 1 C 1265 Comparison of the effect of minodronic acid hydrate and bazedoxifene on biochemical markers of bone turnover in postmenopausal type 2 diabetes women with low bone mass A. Miyoshi, H. Miyoshi, S. Nagai, T. Kondo, H. Nomoto, K. Yamamoto, T. Atsumi; Division of Rheumatology, Endocrinology and Nephrology, Hokkaido University, Sapporo, Japan. Background and aims: Although patients with type 2 diabetes (T2D) are at significant risk for well-recognized diabetic complications, including mac- rovascular disease, retinopathy, nephropathy, and neuropathy, it is also clear that T2D patients are at increased risk for fragility fractures. It is suggested abnormalities in not only bone material strength but also bone microarchi- tecture (bone quality) are risk factors of bone fractures in T2D. It is reported that urinary pentosidine levels predict the future fracture independent of bone turnover and bone mineral density (BMD) without treatment for osteo- porosis. Pentosidine is a surrogate marker of whole advanced glycation end products, which is characterized as non-enzymatic collagen cross-linking. Non-enzymatic cross-links reduce the mechanical and biological integrity of bone and so pentosidine is one of bone quality markers. It has been reported that selective estrogen receptor modulator (SERM) decreases urinary pento- sidine and ameliorates bone quality in postmenopausal women with osteo- porosis. However, effects of osteoporosis medication to bone quality have no proven in T2D with osteoporosis. Thus, we administered bisphosphonate or SERM to postmenopausal T2D women with low bone mass and evaluated the effects of bone turnover including bone quality markers. Materials and methods: We measured femoral neck (FN) and lumbar spine (LS) BMD of postmenopausal women with T2D by dual-energy X-ray ab- sorptiometry and 31 patients with a T score≦-2.0 at the FN or LS in this test were randomized 1:1 to receive minodronic acid hydrate (n=15) or bazedox- ifene (n=16). Changes in born turnover and quality markers, BMD at the FN and LS, and common clinical data involved in T2D were assessed at 6 month. This was a prospected observational study. Primary endpoint was set as the reduction of urinary pentosidine level. A total of 31 patients consented to participating in this study. On average, the patient group was aged 69 ± 7 years and they had been living with T2D for the past 15.5 ± 10.7 years. Their HbA1c was 7.0 ± 0.7% and BMI was 24.0 ± 3.8. Results: There was no difference in urinary pentosidine levels between the groups. No significant reduction was seen in both groups, which urinary pentosidine levels were very high and more than the level considered as a risk level of future fractures. In the other born turnover and quality markers, minodronic acid hydrate significantly decreased serum P1NP levels, urinary DPD and serum ucOC compared with bazedoxifene (-58.6% versus -19.1%, -30.1% versus 1.6%, -55.3% versus -29.1%, respectively). Significantly greater increases in BMD were observed with minodronic acid hydrate treatment at the FN or LS (6-mo treatment difference: 2.3%, FN; 2.4%, LS; p < 0.05) but no changes were observed with bazedoxifene therapy. Bazedoxifene significantly decreased serum non HDL-C levels compared with minodronic acid hydrate (-13.6 mg/dl versus 5.4 mg/dl; p < 0.05). Both treatments didn’t aggravate diabetes conditions. Conclusion: Urinary pentosidine levels did not change in both treatments in this study. Minodronic acid hydrate showed significantly improvement in BMD and bone turnover markers compared with bazedoxifene therapy in also T2D patients. Bazedoxifene may ameliorate lipid metabolism in T2D women with osteoporosis. Clinical Trial Registration Number: UMIN000008783 PS 114 Clinical observations in type 2 diabetes 1266 Clinical features and management of new-onset diabetes mellitus presenting with diabetic ketoacidosis in China: a multicentre, clinic-based study X. Wu1, Y. Xu1, J. Bai2, J. Shaw3, China Initiative for Diabetes Excellence (CIDE) Collaborative Group; 1Department of Endocrinology, First Affiliated Hospital of Nanjing Medical University, 2Department of Epidemiology & Biostatistics, Nanjing Medical University, China, 3Baker IDI Heart and Diabetes Institute, Melbourne, Australia. Background and aims: Diabetic ketoacidosis (DKA) is a serious acute com- plication of diabetes, resulting from severe insulin deficiency and an excess of counter-regulatory hormones. Under certain circumstance, new-onset diabe- tes patients may present with DKA, which is typical for type 1 diabetes mel- litus (T1DM) or ketosis-prone diabetes in western countries. However, there is little information on hospitalization for this kind of diabetes patients in the Chinese population. The present study is to evaluate the clinical features and management of new-onset diabetes mellitus presenting with DKA in China. Materials and methods: A retrospective cohort study of adult and adoles- cents new-onset diabetes patients hospitalized with DKA between 2010 and 2012 were carried out in fifteen tertiary hospitals around China. Clinical and laboratory data were collected. Patients were classified based on clinical fea- tures. Groups were compared for differences in vital statistics and biochemi- cal profiles at presentation. Results: The study cohort comprised 255 new-onset diabetes patients pre- senting with DKA: 115 patients (45.1%) with type 2 diabetes mellitus (T2DM), 98 patients (38.4%) with T1DM, 24 patients (9.4%) with atypical DM (ADM) and 18 patients (7.1%) with Fulminant T1DM. T1DM exceeded T2DM in the age group less than 40 years while T2DM predominated in the age group more than 40 years. Male patients were predominant in T1DM, T2DM and ADM group. The most common precipitating factors were un- known reason (36.9%) and infection (29.0%). The typical symptoms included polyuria/polydipsia (72.9%), nausea/vomiting (67.1%), dehydration (27.1%), variable degree of confusion (26.7%) and abdominal pain (22.4%). More gas- trointestinal symptoms and dehydration were found in T1DM due to severe ketoacidosis. The levels of blood glucose and HbA1c were 25.3±11.1mmol/L and 12.1±3.1% at admission. About 89.0%, 74.4% and 55.9% of patients were evaluated for HbA1c, beta-cell function and autoantibodies for classifica- tion. The complications included electrolyte disturbance (71.4%), arrhythmia (19.6%), hyperosmotic state (14.5%), renal failure (6.7%), shock (3.1%), heart failure (0.8%) and cerebral edema (0.8%). During the first day after admis- sion, total fluid supplement were 3550.4±1904.1ml. Low doses insulin infu- sion lasted for 2.8±5.4 days. The disappearance of urine ketone body took 3.3±2.7 days. Therapeutic regimen was markedly different within the four groups when they discharged from hospitals. Conclusion: New-onset diabetes patients presenting with DKA occur in a spectrum of diabetes types in China, including T2DM, T1DM, ADM and Fulminant T1DM. Its clinical heterogeneity has significant implications for classification and management of diabetes. 1267 Effects of cardiovascular risk factors in urban community in patients with type 2 diabetes mellitus: a 48-months prospective trial X.-L. Zhang1, S.-Y. Yuan1, G. Wan2, L.-X. Zhu1, H.-J. Fu1, M.-X. Yuan1, G.-R. Yang1, S.-F. Pan1, R.-R. Xie1, Y.-J. Lv3, Beijing Community Diabetes Study Group; 1Department of Endocrinology, Beijing Tongren Hospital, Capital Medical University, 2Medical Records and Statistics Department, Beijing Ditan Hospital, Capital Medical University, 3Cuigezhuang Community Health Service Cente, Beijing, China. Background and aims: To assess whether prospective interventions have long-term effects on the risk of diabetes-related macrovascular complications in patients with type 2 diabetes mellitus in urban communities of Beijing. Materials and methods: A total of 3,264 type 2 diabetic subjects (aged 20-80 years) was recruited from 15 community health centers in Beijing in 2008. Diabetologia (2014) 57:[Suppl1]S1–S564 S 515 1 C The subjects were divided into three groups: DM group (n=930), HTN group (n=1397), and CVD group (n=937). By using Framingham risk score (FRS), the subjects in afore three groups are subdivided into risk categories of 20% (high Framingham risk strata). After 48 months, study participants were fol- lowed-up to assess the long-term effects of the interventions. Results: Subjects with CVD in diabetes were more prone to be older, have a longer duration of diabetes, higher systolic blood pressure and diastolic blood pressure than that of DM group (P S 516 1 C Conclusion: About 30% of CAD patients had PAD while about 60% of PAD patients had CAD. Factors common to CAD and PAD were age, smoking habit, nephropathy and retinopathy. Especially PAD is considered clinically worthy of note in people over 75 years old. In our increasingly aging country, we consider it essential to educate the public as well as diabetes patients about the importance of early diagnosis of PAD which carries a poor prognosis. 1270 Outcome reduction with an initial glargine intervention and legacy effects (ORIGINALE) Z. Punthakee1, J. Tyrwhitt1, J. Bosch1, G.R. Dagenais2, R. Díaz3, H. Jung1, A.P. Maggioni4, J. Pogue1, J. Probstfield5, A. Ramachandran6, M.C. Riddle7, L.E. Rydén8, S. Yusuf1, H.C. Gerstein1, ORIGIN Trial Investigators; 1Medicine, Population Health Research Institute and McMaster University, Hamilton, 2Institut Universitaire de Cardiologie et de Pneumologie de Québec, Canada, 3Estudios Clínicos Latino América, Rosario, Argentina, 4Associazione Nazionale Medici Cardiologi Ospedalieri Research Center, Florence, Italy, 5University of Washington, Seattle, USA, 6India Diabetes Research Foundation, Chennai, India, 7Oregon Health and Science University, Portland, USA, 8Karolinska Institute, Stockholm, Sweden. Background and aims: ORIGIN was a 2 x 2 factorial randomised trial as- sessing cardiovascular and other effects of insulin glargine vs standard care and omega-3-fatty acids vs placebo among people with IFG, IGT or early type 2 diabetes and other cardiovascular risk factors. After a median 6.2-year follow-up, insulin glargine had a neutral effect on cardiovascular outcomes and cancers, and reduced the incidence of diabetes in people with IFG or IGT. Omega-3-fatty acids also had a neutral effect on cardiovascular out- comes. ORIGIN participants who consented to the ORIGIN and Legacy Effects Study (ORIGINALE) were followed for more than 2 years after the trial finished to determine the longer term effect of therapy on cardiovascular outcomes and new diabetes. Materials and methods: By the end of March 2014, consenting participants completed up to 2 study visits to ascertain anthropometrics, blood pres- sure, creatinine, A1C, medication use and outcomes. Cumulative incidence, hazard ratios and odds ratios from the time of randomisation to the end of follow-up will be calculated for the ORIGIN trial outcomes. Participants will be analysed in the groups to which they were randomised. Results: At the end of ORIGIN, 10,544 of the original 12,537 participants were alive and followed at 553 sites. 283 sites (comprising 6,497 participants) agreed to participate in ORIGINALE and obtained ethics approval. To date, at least 5721 participants from 274 sites have contributed data comprising ap- proximately 96 new MIs, 63 new strokes, 194 new cardiovascular deaths and 231 new cancer outcomes. Verification of events and data accuracy are ongo- ing. Analyses comparing insulin glargine vs standard care, and omega-3-fatty acids vs placebo will be completed by August 2014. Clinical Trial Registration Number: NCT00069784 Supported by: Sanofi-Aventis 1271 Effect of three insulin regimens on carotid intima-media thickness in patients with type 2 diabetes: the randomised Copenhagen Insulin and Metformin Therapy (CIMT) trial L. Lundby-Christensen1,2, the CIMT trial group, Copenhagen university hospitals; 1Steno Diabetes Center, Gentofte, 2Department of Paediatrics, Copenhagen University Hospital, Hvidovre, Denmark. Background and aims: Post prandial hyperglycaemia may be associated with increased risk of cardiovascular disease in patients with type 2 diabetes. Insu- lin analogue regimens targeting post prandial and/or fasting plasma glucose may therefore influence the risk of cardiovascular disease differently. Carotid intima-media thickness (IMT) is a frequently used risk marker for cardiovas- cular disease. The aim of the Copenhagen Insulin and Metformin Therapy (CIMT) trial was to evaluate the effects of three insulin analogue regimens in combination with metformin or placebo. The metformin/placebo compari- son has previously been reported. This abstract reports the insulin regimen comparisons. Materials and methods: The CIMT trial is an investigator initiated, 3 x 2 factorial, treat-to-target (HbA1c ≤ 7.0% (53 mmol/mol)), multicenter ran- domised clinical trial. 412 participants with type 2 diabetes, HbA1c ≥7.5% (58 mmol/mol), receiving oral antidiabetic agents for at least one year and/or insulin for at least three months, were randomised 1:1:1 to 18 months open label treatment with one of three insulin regimens: insulin aspart biphasic one to three times daily (n=137) versus insulin aspart three times daily in combination with insulin detemir once daily (basal-bolus, n=138) versus in- sulin detemir once daily (n=137). Primary outcome measure was change in mean carotid IMT. Other outcomes were change in HbA1c, weight,and insu- lin dose, and risk of hypoglycaemia and serious adverse events. After multi- ple imputations of missing data on the primary outcome, intention-to-treat analyses adjusting for baseline and stratification variables were performed. Results: 90% (biphasic group), 80% (aspart+detemir group), and 72% (de- temir group) of the participants completed the trial. Mean carotid IMT changed by −0.009 mm (95% confidence interval −0.022 to 0.004, P= 0.17) in the biphasic group, 0.000 mm (−0.013 to 0.013, P=0.99) in the aspart+detemir group, and −0.012 mm (−0.025 to 0.000, P=0.06) in the detemir group. These changes were not significantly different between the groups. HbA1c was sig- nificantly more reduced (P S 517 1 C were shown by J-INDEX (lower in TR_ND than in TR_D and in T2_D, P S 518 1 C PS 115 Liver 1274 Non-invasive predictor of non-alcoholic fatty liver disease in Japanese patients with type 2 diabetes mellitus M. Miyasato1, Y. Murase-Mishiba1, M. Bessho1, M. Miyawaki1, H. Imbe1, C. Tsutsumi1, K. Tanimoto1, A. Imagawa2, J. Terasaki1, T. Hanafusa1; 1First depertment of Internal Medicine, Osaka Medical College, 2Metabolic of Medicine, Graduate School of Medicine, Osaka University, Japan. Background and aims: Individuals with type 2 diabetes mellitus (T2DM) have a higher risk of developing nonalcoholic fatty liver disease (NAFLD) than those without. The serum cytokeratin-18 fragment (CK-18) level has been suggested to be a biomarker of NAFLD, although its usefulness in pa- tients with T2DM is unknown. The objective of the present study was to as- sess the usefulness of the serum CK-18 level as a biomarker for NAFLD in T2DM patients. Materials and methods: The study was divided into two parts. In the first cross-sectional study, a total of 200 patients with T2DM and 58 healthy con- trol subjects were recruited. NAFLD was diagnosed using ultrasonography, and the T2DM patients and nondiabetic control subjects were subdivided into groups with or without NAFLD (DM/NAFLD, DM/nonNAFLD, non- DM/NAFLD and nonDM/nonNAFLD, respectively). We used the scoring system, based on the criteria of hepatorenal echo contrast, vascular blur- ring and liver brightness and deep attenuation, to evaluate the presence of NAFLD. The ultrasonography scores ranged from 0 to 6 points, and a diag- nosis of NAFLD was made for scores of ≥2. In the subsequent longitudinal study, we evaluated the three-month change (Δ) in the CK-18 level and other parameters in 40 T2DM patients with NAFLD. Serum CK-18 levels were measured using ELISA. Results: The median [IQR] level of serum CK18 was 158.4U/L [107.1-291.9] in the DM/NAFLD group, 96.1 U/L [74.1-142.6] in the DM/nonNAFLD group, 172.4 U/L [130.4-278.8] in the nonDM/NAFLD group and 120.4 U/L [97.5-158.1] in the nonDM/nonNAFLD group. The serum CK18 val- ues were significantly higher in the NAFLD group than in the nonNAFLD group among both diabetic (p S 519 1 C Conclusion: The combination of EMPA+LINA was superior to the respective monotherapies in improving insulin sensitivity of the liver and in reducing liver lipid content. These results suggest that the combination of both drugs may be a potential therapy for the most common liver diseases associated with T2D. Supported by: Boehringer Ingelheim 1277 Mosapride citrate improves non-alcoholic steatohepatitis with increased faecal lactic acid bacteria and plasma glucagon-like peptide-1 level in a rodent model H. Okubo1, M. Yoneda2, H. Sakoda3, A. Kushiyama4, M. Fujishiro3, Y. Nakatsu1, T. Fukushima1, Y. Matsunaga1, H. Kamata1, M. Iwashita5, F. Nishimura5, A. Tomoichiro1; 1Department of Medical Science, Graduate School of Medicine, 2Department of Molecular and Internal Medicine, Graduate School of Biomedical and Health Sciences, University of Hiroshima, 3Department of Internal Medicine, Graduate School of Medicine, University of Tokyo, 4Institute for Adult Disease, Asahi Life Foundation, Tokyo, 5Department of Dental Science for Health Promotion, University of Hiroshima, Japan. Background and aims: Several lines of evidence have suggested a role of gut microbiota in the etiology of non-alcoholic steatohepatitis (NASH). Gastro- intestinal motility may influence gut microbiota and NASH subjects report- edly had a prolonged orocecal transit time coexistent with small intestinal bacterial overgrowth. Thus, we investigated effects of the gastroprokinetic agent mosapride citrate (MC) on gut microbiota and the development of NASH using methionine-choline deficient diet (MCDD)-induced NASH ro- dent model. Materials and methods: 6-week-old C57BL/6 mice were divided into three groups, given the normal chow diet (NCD), the MCDD or the MCDD con- taining 10 mg/kg/day MC (MCDD plus MC) for 6 weeks. Results: Gut microbiota analyses revealed that total numbers of bacteria were lower in the MCDD than in the NCD group, but similar to those in the MCDD plus MC group. Calculation of the relative abundance of each strain indicated lactic acid bacteria such as Bifidobacterium and Lactobacil- lus in feces to be specifically decreased in the MCDD group. Interestingly, the reduction in lactic acid bacteria in the MCDD group was reversed in the MCDD plus MC group. Subsequently, HE staining of livers from the MCDD group revealed remarkable NASH development as evidenced by deformity of hepatocytes, large fat droplets and inflammatory cell infiltration, all of which were suppressed in the MCDD plus MC group. Serum ALT levels were also elevated in the MCDD group, but were normal in the MCDD plus MC group. Azan staining revealed marked collagen deposition and mRNA levels of fibrosis markers such as α-smooth muscle actin, tissue inhibitor of metal- loproteinase 1 and transforming growth factor β were elevated in the MCDD group, while these abnormalities were significantly reversed in the MCDD plus MC group. Finally, the molecular mechanism underlying the resistance to NASH development conferred by MC treatment was investigated. The mRNA level of tumor necrosis factor α and the serum concentration of li- popolysaccharide, a possible inducer of hepatic inflammation, were increased in the MCDD group, while these increases were suppressed in the MCDD plus MC group. Glucagon-like peptide-1 (GLP-1) reportedly attenuated the development of NASH. Plasma GLP-1 levels were lower in the MCDD than in the NCD group. Interestingly, plasma GLP-1 levels were increased in the MCDD plus MC group. To evaluate intestinal inflammation, immunostain- ing employing the anti-NFκBp65 antibody was performed. Nuclear NFκBp65 positive cells in the colon were increased in the MCDD group, while the numbers of these cells were reduced in the MCDD plus MC group. Conclusion: MC showed a protective effect against the NASH development induced by MCDD, in which increased fecal lactic acid bacteria and plasma GLP-1 may be involved. Thus, MC may be effective for treating NASH. 1278 PDK2 deficiency reduces hepatic steatosis and insulin resistance in mice fed a high fat diet S.-W. Kim1, K.-U. Lee2, M.-G. Choi3, K.-H. Bae1, Y.-H. Go1, I.-K. Lee1; 1Department of Endocrine and Metabolism of Internal medicine, Kyungpook National University School of Medicine, Daegu, 2Department of Endocrine and Metabolism of Internal medicine, Asan Medical Center, Seoul, 3Department of Endocrine and Metabolism of Internal medicine, Hallym University Medical Center, Chuncheon, Republic of Korea. Background and aims: Hepatic steatosis is rapidly evolving as a major medi- cal problem in the world. It involves a variety of lipid abnormalities includ- ing enhanced fatty acid influx from the adipose tissue, increased de novo lipogenesis,reduced fatty acid oxidation and ketogenesis. The pyruvate dehy- drogenase kinases (PDKs) regulate pyruvate oxidation by controlling the ac- tivity of the pyruvate dehydrogenase complex (PDC). We examined whether PDKs were increased in the liver of HFD-fed mice and how they regulate hepatic steatosis and insulin resistance. Materials and methods: The physiological importance of regulation of PDC activity by PDK isoenzyme 2 was assessed by comparing PDK2 knockout (PDK2 KO) mice with wild type mice fed a high fat diet (HFD) and an isoca- loric low fat diet (LFD). Results: Body weight gain and hepatic steatosis were attenuated by PDK2 deficiency in the HFD fed mice. Fasting blood glucose, serum insulin,and liver pyruvate, lactate, oxaloacetate, citrate, diacylglycerols, and triacylglyc- erols were also reduced. Hepatic glucose production was also reduced and insulin sensitivity was increased in the HFD fed PDK2 KO mice. The hepatic enzyme capacity for fatty acid oxidation and ketogenesis was increased while the capacity for lipogenesis was decreased. In spite of this but consistent with greater PDC activity, the respiratory exchange ratio was higher in the PDK2 knockout mice. Energy expenditure was increased without changes in physi- cal activity. Increased hepatic insulin sensitivity and improved glucose toler- ance correlated with reduced PKCε phosphorylation. Conclusion: The findings support the case for PDK2 as a promising target for hepatic steatosis and insulin resistance. Supported by: NRF-2012R1A2A1A03670452 1279 Pro- and antioxidant status in dependence on vitamin D3 availability in the development of diabetes-induced liver injury D. Labudzynskyi, I. Shymanskyy, A. Mazanova, O. Lisakovskaya, M. Veliky; Laboratory of Medicine Biochemistry, Palladin Institute of Biochemistry of the National Academy of Sciences of Ukraine, Kyiv, Ukraine. Background and aims: Diabetes is known to be associated with overpro- duction of ROS and/or impaired antioxidant defence contributing to the onset and progression of diabetic complications on vessels, retina, kidneys, nerves and liver. Nevertheless, precise mechanism by which oxidative stress could facilitate and accelerate the development of hepatic lesions in diabetes is not fully clarified. Recent studies have shown that vitamin D3 is currently recognized as a potent immunomodulator and antioxidant affecting various inflammatory and autoimmune diseases. The present study was performed to determine the relationship between 25-hydroxyvitamin D3 (25(OH)D3) availability and pro-/antioxidant profile in liver of diabetic mice. Materials and methods: Type 1 diabetes was induced in male C57BL/J6 mice (weighing 25.0 ± 1.5g) by i.p. injection of multiple low dose streptozotocin (40 mg/kg b.w.). Control and STZ-diabetic mice were treated with or without vitamin D3 (15 IU/mouse per os, for 8 weeks). Serum 25OHD3 was assessed by ELISA. The levels of phospho-NF-kB/p65, poly(ADP-ribose)polymerase 1 (PARP-1), poly-ADP-ribosylated and nitrosylated proteins were measured by Western-blot analysis. Intracellular reactive oxygen and nitrogen species (ROS and RNS) production were detected by 2‘,7‘-dichlorofluorescin (DCF) and 4,5-diamino-fluorescein diacetate (DAF-DA) fluorescence respectively using flow cytometry. Hepatic pro-/antioxidant factors and enzymes activity in liver were measured spectrophotometrically. Results: Serum level of 25OHD3, the main circulating metabolite of D3, was shown to be reduced to 23.8±1.9 in diabetes vs. 39.7±2.9 nmol/l in control, that reflects reliably vitamin D3 deficiency (p S 520 1 C tively vs. control, p S 521 1 C and respective control mice were analyzed for expression of HSPs by qPCR and western blot. Hepatocyte-specific reconstitution of HSPs in db/db mice was achieved via adenoviral constructs for Hsp70, DNAJB1 and DNAJA2. Results: In our models of diabetes, we observed downregulations of up to 90% in the hepatic expression of central HSPs such as Hsp70. In contrast, non-diabetic mice on a high-fat diet showed no significant regulation of he- patic HSP expression. In db/db mice, two weeks of hepatocyte-specific recon- stitution of Hsp70 reduced both fasting glucose and HbA1c by approx. 25% (p S 522 1 C PS 116 Vascular calcification 1284 Oestrogen-related receptor gamma mediates vascular calcification through up-regulation of BMP2 Y.-K. Choi1, J.-H. Kim1, M.-K. Kim2, G.-S. Jung1, K.-H. Bae1, M.-J. Kim3, E. Kim4, I.-K. Lee1, K.-G. Park1, M.-G. Choi5, K.-U. Lee6, E. Jung7; 1Division of Endocrinology and Metabolism, Department of Internal Medicine, Kyungpook National University School of Medicine, 2Division of Endocrinology and Metabolism, Department of Internal Medicine, Keimyung University School of Medicine, 3Pooren Meerae Internal Medicine Clinic, 4Division of Endocrinology and Metabolism, Department of Internal Medicine, Daegu, 5Division of Endocrinology and Metabolism, Department of Internal Medicine, Hallym University Medical Cente, Chuncheon, 6Division of Endocrinology and Metabolism, Department of Internal Medicine, Asan Medical Center, Seoul, 7Division of Endocrinology and Metabolism, Department of Internal Medicine, Daegu Catholic University Medical Center, Republic of Korea. Background and aims: Vascular calcification, which refers to ectopic miner- alization in vascular smooth muscle cells, occurs frequently in many diseases such as chronic kidney disease, atherosclerosis, and diabetes. Estrogen re- lated receptor (ERR)γ, a member of orphan nuclear receptor superfamily has diverse roles in regulating homeostatic and metabolic processes. However the role of ERRγ in vascular calcification has not yet been investigated. This study was undertaken to examine the role of ERRγ in vascular calcification. Materials and methods: Rat aortic smooth muscle cells (RASMCs) were cul- tured and vascular cell calcification was induced by treatment with inorganic phosphate and calcium. Next we investigated effects of adenovirus-mediated overexpression of ERRγ on phosphate-induced VSMC calcification and ex- amined whether siRNA-mediated inhibition of endogenous expression of ERRγ or pharmacological inhibition of ERRγ with GSK5182 prevent phos- phate-induced vascular calcification. Results: Along with increased expression of bone morphogenic protein- 2(BMP2), Runx2 and Msx2, ERRγ expression was upregulated during phos- phate - induced calcification. Von-kossa staining showed that adenovirus- mediated overexpression of ERRγ (Ad-ERRγ) in RASMCs accelerated Pi- induced calcification. Ad-ERRγ increased the expression of osteogenic gene including Runx2, OPN and Msx2 but decreased alpha-SMA. ERRγ induced BMP2 transcription through directly binding to its promoter region and increased BMP2 signaling including phosphorylation and nuclear localiza- tion of smad1/5/8. Moreover, inhibition of ERRγ by both siRNA-mediated knockdown of endogenous ERRγ and a selective inverse agonist, GSK5182 attenuated vascular calcification and osteogenic gene expression in vitro and in vivo. Conclusion: This study demonstrated that ERRγ mediates vascular calcifi- cation through up-regulation of BMP2 signaling. These results indicate that inhibition of ERRγ may be a potential therapeutic strategy for prevention of vascular calcification. 1285 Effects of high glucose on the OPG/RANK/RANKL/TRAIL system in the progression of vascular calcification in rat aortic vascular smooth muscle cells Y. Kang, H. Choi, E. Sim, H. Lee, D. Yi, S. Son; Internal Medicine, Pusan National University School of Medicine, Yangsan, Republic of Korea. Background and aims: Diabetes mellitus is frequently complicated by car- diovascular disease, such as vascular calcification and accelerated athero- sclerosis. Recently, it has been known that the OPG/RANK/RANKL/TRAIL system may play a major role in vascular calcification and atherosclerosis. However, the possible effects of long term high glucose stimulation at least 4 weeks on the OPG/RANK/RANKL/TRAIL system in the progression of vascular calcification are less clear. We attempted to evaluate the effect of high glucose on the progression of vascular calcification in rat aortic smooth mus- cle cells (RASMCs) and to detect the expression changes of OPG, RANK, RANKL, and TRAIL for 2 and 4 weeks. Furthermore, we used BMP-7, which has been known to attenuate vascular calcification, to detect the possible changes of OPG, RANK, RANKL, and TRAIL expressions on the calcifica- tion of RASMCs. Materials and methods: The primary cultured RASMCs were stimulated with normal glucose (5.5mmol/L glucose, NG) and high glucose (30 mmol/L glucose, HG) with calcification medium. The mRNA levels and the protein expressions of OPG, RANK, RANKL, and TRAIL were measured by reverse transcription polymerase chain reaction (RT-PCR) or Western blot. Results: The intensity of calcium staining was increased in HG after 2 weeks and more increased and prominent after 4 weeks compared to NG. OPG mRNA and protein expressions were not different after 2 weeks, however, after 4 weeks, OPG expressions were significantly decreased in HG. Regard- ing RANK, RANKL, and TRAIL expressions, there were no differences after 2 or 4 weeks of stimulation. After 4 week of rhBMP-7 co-treatment, the den- sities of calcium stains were attenuated and the total amount of calcium was also decreased. The mRNA and protein OPG expressions were maintained with BMP-7 after 4 weeks of stimulation. There were no differences in the expressions of RANK, RANKL, and TRAIL between with and without BMP- 7 co-treatment. There was no difference in Bax mRNA expression, an apop- totic marker, between with and without BMP-7 co-treatment, however, ALP mRNA expression, a marker of mineralization, was decreased in the presence of BMP-7. Conclusion: Chronic hyperglycemia may enhance vascular calcium depo- sition and high glucose may increase OPG mRNA and protein expression with short-term stimulation, but decrease it with long-term stimulation. The expression of RANKL, RANK, and TRAIL were maintained with long term stimulation. As the OPG expression was decreased, the increased mineraliza- tion activity may be more associated rather than apoptotic activity with the progression of long-term high glucose induced vascular calcification. 1286 The DPP-4 inhibitor linagliptin increases plasma fetuin-A concentrations in a rat model of uraemic calcification C. Reichetzeder1,2, A. Pasch3, K. von Websky1,2, O. Tsuprykov1, T. Klein4, B. Hocher1; 1University of Potsdam, 2Centre for Cardiovascular Research / Institute of Pharmacology, Charité, Berlin, Germany, 3Department of Nephrology, University of Bern, Switzerland, 4Boehringer Ingelheim, Germany. Background and aims: Fetuin-A is a potent inhibitor of ectopic mineralisa- tion. An increase in fetuin-A has beneficial effects in uraemic vascular calcifi- cation, whereas fetuin-A deficiency is associated with soft tissue calcification in mice and humans. Biomarker data in diabetic and non-diabetic rodent models of chronic renal failure suggest that the dipeptidyl peptidase (DPP)-4 inhibitor linagliptin may be able to influence the pathogenesis of vascular calcification. The aim of the present study was to investigate the effect of lina- gliptin on uraemic vascular calcification in nephrectomised rats. Materials and methods: 1,25-dihydroxyvitamin D3 (0.25 μg/kg/day) was used to induce uraemic vascular calcification in 5/6 nephrectomised rats (5/6NxVitD). Rats were allocated to 3 treatment groups: a) sham-operated rats treated with placebo (n=10); b) 5/6NxVitD rats treated with placebo (n=14); c) 5/6NxVitD rats treated with linagliptin (n=14). Rats were treated for 6 weeks. We analysed plasma factors known to be involved in uraemic vascular calcification such as cystatin C, fibroblast growth factor 23 (FGF23), magnesium, calcium, phosphate, and fetuin-A. Blood pressure was moni- tored and at study end, animals were sacrificed; blood was taken and stored at -80°C for analysis. Results: Blood pressure was similar in all groups. 5 of 14 animals died in the 5/6NxVitD placebo group; 3 of 14 animals died in the 5/6NxVitD linaglip- tin group (p>0.05). At study end, magnesium, FGF23, and cystatin C levels were unaffected by linagliptin in 5/6NxVitD rats. The change from baseline in phosphate levels was similar with linagliptin and placebo in 5/6NxVitD rats, whereas the change from baseline in calcium levels was less with linagliptin (-0.062 ± 0.255 mmol) than placebo (0.961 ± 0.439 mmol) in 5/6NxVitD rats (p S 523 1 C 1287 Below-knee arterial calcification in type 2 diabetes: association with receptor activator of nuclear factor kappa B ligand, osteoprotegerin and neuropathy O. Bourron1,2, C.E. Aubert1, S. Liabeuf3, P. Cluzel1, M. Komajda1, S. Jacqueminet1, Z. Massy3, A. Hartemann1,2; 1Pierre et Marie Curie University Paris 06 - Assistance Publique Hôpitaux De Paris, 2CHU Pitié-Salpêtrière, 3Picardie University - CHU Amiens, Paris, France. Background and aims: Calcification of the arterial wall in diabetes con- tributes to the arterial occlusive process at the below knee level. The osteoprotegerin(OPG)/RANKL system is suspected to be involved in the calcification process. The aim of the study was to investigate if there is a link between arterial calcification in type 2 diabetes and 1) the conventional cardio-vascular risk factors, 2) the serum RANKL and OPG levels and 3) neuropathy. Materials and methods: We objectively scored, in a cross-sectional study, infra-popliteal vascular calcification using CT-scan in 198 patients with type 2 diabetes, a high cardio-vascular risk and with glomerular filtration rate > 30mL/mn. Colour duplex ultrasonography was performed to assess periph- eral arterial occlusive disease, and mediacalcosis. Peripheral neuropathy was defined by a neuropathy disability score (NDS) > 6. RANKL and OPG were measured in serum by routine chemistry. Results: Below knee arterial calcification was associated with arterial occlu- sive disease. In multivariate logistic regression analysis, variables significantly and independently associated with the calcification score were age (OR=1.08; 95% CI=1.04-1.13 ; p400, all patients were submitted to history tak- ing, clinical and anthropometric evaluation, laboratory investigations (FBG,2 h PP,lipid pfofile,HBA1C%,serum insulin, albumin creatinine ratio, TSH,free T3, free T4), HOMA-IR, coronary multislice CT. Results: There was a high statistical significant difference between the 4 groups as regards EAT volume ,being the highest( >200 cm3) among Gr 4 ,there was a high statistical significant difference between the 4 groups as regards HBA1C%, (10.23+1.49), Urinary albumin creatinine ratio(UACR) (109.87+52.31 )mg%, (P S 524 1 C 1290 Metabolically unhealthy non-obese subjects are at higher risk of subclinical coronary artherosclerosis than metabolically healthy obese subjects in Korea K.-U. Lee1,2, J. Jang1,2, C. Woo1, M. Lee1, C. Jung1, W. Lee1, E. Koh1,2, J.-Y. Park1, I.-K. Lee3, M. Choi4; 1Department of Internal Medicine, Asan Medical Center, Seoul, 2Metabolism Research Unit, Asan Institute for Life Sciences, Seoul, 3Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, 4Department of Internal Medicine, College of Medicine, Chuncheon, Republic of Korea. Background and aims: Metabolically healthy obesity (MHO) is an emerging phenotype with a cardiovascular disease (CVD) risk between healthy, normal weight and unhealthy, obese individuals. It is not yet established whether this phenotype has higher risk of CVD than nonobese subjects with metabolic abnormalities. In this study, we compared the degree of subclinical coronary atherosclerosis detected by coronary multidetector computed tomography (MDCT) in four groups defined by the state of metabolic health and obesity in an asymptomatic Korean population. Materials and methods: We collected the data of 4,009 asymptomatic sub- jects (mean age, 53.2 yr) who participated in a routine health screening ex- amination at a medical center in Korea. Significant coronary artery stenosis (CAS) defined as >50% stenosis, and coronary artery calcium scores (CACS) were assessed by MDCT. Participants were stratified by BMI (cut-off value, 25 mg/m2) and metabolically healthy state. Results: MHO subjects (n = 589) had a significantly higher prevalence of subclinical coronary atherosclerotic burden (CAS and CACS > 0) compared with metabolically healthy non-obese (MHNO) subjects (n = 1367). Howev- er, MHO subjects had a significantly lower prevalence of subclinical coronary atherosclerotic burden (CACS > 0) compared with metabolically unhealthy nonobese (MUNO) subjects (n = 853) as well as metabolically unhealthy obese (MUO) subjects (n = 1200). The prevalence of subclinical coronary stenosis in MUNO subjects was comparable to MUO subjects. Conclusion: Our data show that MHO might not be a benign disease in terms of coronary atherosclerotic burden. However, MHO subjects had a significantly lower prevalence of subclinical coronary atherosclerotic burden compared with MUO subjects Thus, the presence of metabolic abnormalities may be more important than the presence of obesity in evaluating cardio- vascular risk. 1291 Contributors to mortality in high risk diabetes patients D.W. Bowden1, F.-C. Hsu2, B.I. Freedman3, J. Carr4, A.J. Cox5; 1Center for Human Genomics, 2Biostatistical Sciences, 3Internal Medicine, Wake Forest University School of Medicine, Winston-Salem, 4Radiology, Vanderbilt University School of Medicine, Nashville, USA, 5Griffith Health Institute, Southport, Australia. Background and aims: Individuals with type 2 diabetes (T2D) exhibit sub- stantial differences in mortality risk. Coronary artery calcium (CAC) is a powerful, well documented, predictor of mortality. In this study we examined a set of individuals at high risk for death based on CAC>1000 (40% mortality in 7.5 years) and evaluated a range of clinical measures, including modifiable cardiovascular disease (CVD) risk factors, for insights into risk for mortality. Materials and methods: All-cause mortality was determined in 371 Euro- pean American individuals with T2D and CAC>1000 from a larger ongoing population based cohort study of CVD in people with T2D. After 8.2 ± 3.0 (mean ± SD) years of follow-up differences in CVD risk factors, medication use, and other clinical measures were compared between living (n=218) and deceased (n=153) participants. Cox Proportional Hazards Regression models were used to quantify risk for mortality to appropriately account for time-to- event effects. Results: Mean age at basline was 65.5 years with mean duration of diabetes 12.6 years. The subjects had a mean BMI of 32.0 and were 70.1% male. Death was confirmed through the US Social Security death index. Deceased par- ticipants had a longer duration of T2D (p=0.02). Differences in cholesterol were nominal (HR 2.24 CI 1.09-4.60; p=0.03), but kidney function (HR 4.78 CI 2.38-9.62; p S 525 1 C PS 117 Vascular dysfunction 1293 Caveolin-1 alpha expression is decreased in the aorta of rats with glucose intolerance induced by a low dose of streptozotocin J. Srankova, L. Pivackova, G. Doka, B. Gajdacova, L. Mensikova, E. Malikova, K. Galkova, J. Klimas, P. Krenek; Department of Pharmacology and Toxicology, Faculty of Pharmacy Comenius University, Bratislava, Slovakia. Background and aims: Caveolae are plasma membrane invaginations of en- dothelial cells, which represent a predominant location of endothelial nitric oxide synthase (eNOS), important in regulation of endothelial function. Ca- veolin-1 (cav-1) is the main component of caveolae, but the specific role of its isoforms α and β in regulation of eNOS activity in diabetes and glucose impaired tolerance is still not known. Materials and methods: Impaired glucose tolerance was involved by strep- tozotocin administration (STZ, 25 mg/kg/day, i.p. injections after every 24h, n=10). Control (CON, n=10) received the vehicle. After 8 weeks, oral glucose tolerance test (OGTT) was performed. At the end of experiment (10 weeks after STZ) we measured of vascular reactivity to acetylcholine. The expres- sions of cav-1α, cav-1β, eNOS, hsp90, gp91phox and MnSOD were analyzed by Western blot in aorta samples. The expressions of both isoforms of cav-1 in aorta were also determined at mRNA level by qRT-PCR. Results: We observed normal fasting glucose in both groups (CON 6.15 ± 0.27 mmol/l vs. STZ 6.65 ± 0.97 mmol/l), but STZ group had a significantly increased glycaemia 1h after oral administration of glucose in OGTT (CON 8.06 ± 0.44 mmol/l vs. STZ 16.75 ± 2.58 mmol/l, P S 526 1 C (p=0.004), c-peptide (p S 527 1 C effect of vitamin D supplement on arterial stiffness in patients with type 2 diabetes and vitamin D deficiency. Materials and methods: This study was designed as a prospective, rand- omized, open-label controlled trial. A total of 40 (20 men and 20 women) patients with type 2 diabetes and vitamin D deficiency, defined as serum 25(OH)D levels < 20 ng/mL, were enrolled. A randomized-study group (10 men and 10 women) was administered daily 1000IU vitamin D3 for 6 months, and control group had no interventional drug. After 6 month of intervention, authors estimated the pulse wave velocity and the aortic augmentation index as primary endpoints in both group, and statistically analyzed the variation. Results: A mean age of patients was 56.2 years old, and mean level of 25(OH) D was 11.9 ng/mL. A 25(OH)D level of study group was significantly in- creased for 6 months, which were 25.3 ng/mL in study group and 9.4 ng/mL in control group (P S 528 1 C PS 118 Advanced glycation 1301 Evaluation of a glyoxalase 1 mutant mouse A. Shafie, M. Xue, P.J. Thornalley, N. Rabbani; WMS, CSRL, University of Warwick, Coventry, UK. Background and aims: Glyoxalase 1 plays an important role in the metabo- lism of reactive dicarbonyl metabolites, glyoxal and methylglyoxal, to less reactive products and prevention of dicarbonyl-derived advanced glycation endproduct formation. Glo1 deficient mice and transgenic mice overex- pressing Glo1 provide valuable models to study control of change in extent of dicarbonyl glycation in mammalian systems. The aim of this study is to characterise the genotype and phenotype of the Lexicon Glo1 mutant mouse. Materials and methods: The Glo1 mutant mice (+/- breeding pair), Glo1Lex were obtained from the European Mutant Mouse Archive, Heidelberg, Ger- many. We maintain a colony of Glo1Lex heterozygotes and sibling wild type controls. The mice were produced by Lexicon Pharmaceuticals, Inc, USA. In a C57BL/6 genetic background, mutation was produced by retroviral inser- tion of a DNA cassette between coding exons 1 and 2 (LEXKO-1493). Gen- otyping Forty-four offspring were genotyped from ear punch samples. For PCR, the three pairs of primers were used to discriminate between wild-type: heterozygote and homozygote mutant mice. Phenotyping Nineteen mice, 12 Glo1Lex (+/-) (6 male, 6 female) and 7 wild type control siblings (5 male, 2 female) were sacrificed at 7 months old, tissues (brain, heart, liver, spleen, kidney, pancreas and skeletal muscle) collected and stored at -80°C until analysis. Aliquots of tissue were homogenised and Glo1 enzymatic activity was determined in cytosolic extracts by spectrophotometric assay, protein by Western blot and mRNA by RT-PCR. Protein advanced glycation endproduct (AGE) residues of liver were analysed in 10 Glo1Lex (+/-) (5 male, 5 female) and 10 wild type control siblings (5 male, 5 female) by LC-MS/MS and nor- malized to the corresponding unmodified amino acid residue. Results: In the genotyping of Glo1Lex mutant mice we found only Glo1Lex (+/-) heterozygote and wild type siblings. No homozygous Glo1Lex (-/-) mice have been born to date. No significant impairment in fertility was found for Glo1Lex (+/-) mice. The activity of Glo1 was not significantly different between wild-type controls and Glo1Lex (+/-) mice (P S 529 1 C 1304 Nuclear RAGE, a regulator of DNA integrity in diabetic lung fibrosis V. Kumar1, T. Fleming1, E. Schleicher2, H.-U. Häring2, P.P. Nawroth1; 1Internal Medicine I and Clinical Chemistry, University of Heidelberg, 2Department of Internal Medicine, University of Tübingen, Germany. Background and aims: The receptor for advanced glycation endproducts (RAGE) is a member of the immunoglobulin super-family of cell receptors, whose activation has been suggested to contribute to various pathologies. Re- cent studies have suggested that the localization of RAGE is not restricted to the cell-surface but also inside the nucleus interacting either directly or indirectly with DNA. However, the role that nuclear RAGE services in the nucleus remains unclear. In this study, we have investigated the ligand-inde- pendent localization of RAGE to nucleus and its role as a DNA repair protein. Materials and methods: Pulmonary fibroblasts were isolated from the lungs of RAGE-deficient (RAGE-/-) and aged-matched wild-type (C57BI/6; male) mice. RAGE and DNA repair proteins were detected by immunofluorescence. DNA damage was assessed by single-cell gel electrophoresis (COMET) as- say. Micro-irradiation of cells was performed using a FluoView1000confo- cal microscope, using 355nm laser. Laser settings were chosen to generate detectable double-strand breaks (DSBs) within the path of the laser, without induction of major cytotoxic effects Results: Pulmonary phenotyping in RAGE-/- mice showed that these mice had a reduced lung volume as compared to aged-matched control, consist- ent with a restrictive lung disease, such as fibrosis. This was confirmed by trichrome staining which showed that RAGE-/- mice had a greater accumula- tion of collagen and other extracellular matrix. Analysis of the subcellular localization of RAGE within the lung showed that the majority of the pro- tein was in the nucleus. The localization of RAGE to the nucleus was con- firmed in vitro, using wild-type lung fibroblasts transient transfected with GFP-tagged RAGE. It was observed that RAGE-/- lung fibroblasts has a sig- nificantly altered growth rate and morphology. Further analysis showed that RAGE-/- lung fibroblasts had significantly higher levels of DNA damage as compared to wild-type fibroblasts treated with and without bleomycin. The level of DNA damage could be increased in wild-type fibroblasts by transient transfection with a shRNA vector specific for RAGE. Conversely, transfection of the GFP-tagged RAGE into RAGE-/- lung fibroblasts could significantly reduce the level of DNA damage by ca.80%. This would suggest, that with- in the context of the lung, nuclear RAGE is responsible for maintain DNA integrity. Following laser-induced DSBs, it was found that nuclear RAGE is phosphorylated by ATM, the main protein kinase sensor of DSBs, leading to its association with a variety of repair proteins, including MRN11, Rad50 and Nba1, which are involved in the formation of the MRN complex. It was sub- sequently shown that post-translational modification of RAGE by methylgly- oxal, a reactive metabolite elevated in diabetes, leads to loss of function with respect to DNA resectioning and binding, an effect which was not observed following oxidative modification. Conclusion: Loss of nuclear RAGE in the lung may provide a new mecha- nism for the development diabetic lung fibrosis; inactivation of nuclear RAGE by reactive metabolite driven post-translational modifications would prevent the correct formation and function of the MRN complex, leading to increased DNA damage via a reduced repair capacity, analogous to the mo- lecular phenotype observed in the RAGE-/- mice. Supported by: DFG (BI-1281/3-1 & NA 138 /7-1), DZD/BMBF, Dietmar Hopp Stiftung 1305 Soluble receptor for advance glycation end-products (sRAGE) relates to oxidative stress index in patients with diabetes J. Škrha jr.1, J. Soupal1, M. Kalousova2, R. Mikova1, M. Prazny1, J. Škrha1; 1Third Department of Medicine, First Faculty of Medicine, Charles University in Prague, General University Hospital, 2Institute of Medical Biochemistry and Laboratory Diagnostics, First Faculty of Medicine, Charles University in Prague, General University Hospital, Czech Republic. Background and aims: Oxidative stress is one of the most important factors in the development of vascular damage in patients with diabetes. Novel find- ings support the idea of advanced glycation end-products involvement in the pathophysiology of diabetic angiopathy. The aim of our study was to evaluate relationship between oxidative stress and soluble receptor for advanced gly- cation end-products (sRAGE) in patients with diabetes. Materials and methods: Total 121 persons were included within the study - 45 patients with Type 1 diabetes (T1DM; aged 52 ± 15 yrs), 59 patients with Type 2 diabetes (T2DM; aged 65 ± 11 yrs) and 17 healthy controls (47 ± 15 yrs). Oxidative stress was measured by Free Radical Analytical System (FRAS4; H&D, Italy) and evaluated by two blood tests: reactive oxygen me- tabolites test (d-ROMs) and by biological antioxidative potential test (BAP). Oxidative stress index (OSi) expressing the total oxidative status was calcu- lated (OSi = d-ROMs/BAP). In all patients routine biochemical parameters, glycated hemoglobin (expressed in mmol/mol acc. EFCC), sRAGE, cell adhe- sion molecules (VCAM, ICAM), von Willebrand factor, (micro)albuminuria (expressed as albumin/creatinine ratio UACR), and anthropometrical data were measured. Results: Oxidative stress index was similar between groups (T1DM: 0.25 [0.13 - 0.47], T2DM: 0.25 [0.13 - 0.42], controls 0.24 [0.18 - 0.34]; NS) and also sRAGE did not differ significantly (T1DM: 1352 ± 514, T2DM: 1179 ± 814, controls: 1177 ± 564 ng/l, NS). Significantly lower d-ROMs were ob- served in patients with T1DM and T2DM without albuminuria (UACR S 530 1 C HDL in healthy people. From the correlation coefficient, r = - 0.42, P S 531 1 C 1309 Serum paroxonase-1 activity is more closely related to HDL particle concentration and large HDL particles than to HDL cholesterol in type 2 diabetic and non-diabetic subjects R.P.F. Dullaart1, J.D. Otvos2, R.W. James3; 1Endocrinology, University Medical Center, Groningen, Netherlands, 2LipoScience Inc., Raleigh, USA, 3Internal Medicine, Division of Endocrinology, Diabetology, Hypertension and Nutrition, University Hospital Geneva, Switzerland. Background and aims: We determined relationships of the anti-oxidative enzyme, paraoxonase-1 (PON-1), with high density lipoprotein (HDL) sub- fractions, and tested whether these relationships are stronger than those with HDL cholesterol and apolipoprotein A-I (apoA-I) in subjects with and with- out type 2 diabetes mellitus (T2DM). Materials and methods: Serum PON-1 (arylesterase activity) and HDL sub- fractions (nuclear magnetic resonance spectroscopy) were determined in 67 T2DM patients and in 56 non-diabetic subjects. Results: PON-1 activity, HDL cholesterol and apoA-I were decreased in T2DM (all p S 532 1 C 1312 The synergistic effect of two mtDNA point mutations in complexes of the respiratory chain promote age-dependent hyperglycaemia and mitochondrial ROS production S. Schröder, J. Niemann, F. Koch, C. Johne, M. Tiedge, S. Baltrusch; Institute of Medical Biochemistry and Molecularbiology, Rostock, Germany. Background and aims: Mutations in complexes of the respiratory chain, which are encoded in the mitochondrial genome (mtDNA), can mediate an increase in production of reactive oxygen species (ROS). Thus, mtDNA mu- tations are proposed to evoke mitochondrial dysfunction and to contribute to the pathogenesis of type 2 diabetes mellitus. In this study we investigated ROS production and the expression of antioxidative enzymes in liver and muscle from conplastic mice carrying a mtDNA point mutation in the cy- tochrome c oxidase (complex IV) or a combined mutation in the cytochrome c oxidase (complex IV) and the NADH dehydrogenase (complex I). Materials and methods: Blood glucose was measured in the conplastic mouse strains C57BL/6NTac-mtBPL/1J (NADH dehydrogenase mutation und cytochrome c oxidase mutation, mtBPL), C57BL/6NTac-mtNOD/LtJ (cytochrome c oxidase mutation, mtNOD) and C57BL/6NTac-mtAKR/J (control; mtAKR). For analysis of mitochondrial ROS production MitoSox was injected in living animals. Thereafter mice were killed and tissues were analyzed for accumulated ROS by fluorescence microscopy. Gene expression of the antioxidative enzymes catalase, SOD1, SOD2 was investigated by quan- titative PCR analyses. Results: At the age of 6 months mtBPL mice (7.5 mmol/l) showed signifi- cantly higher blood glucose levels compared to mtNOD mice (6.1 mmol/l) and mtAKR control mice (6.2 mmol/l). In liver and muscle of mtNOD and mtAKR mice an age-dependent increase in ROS was observed between month 3 and 12. In contrast, mtBPL mice showed a significant 20-fold peak increase in ROS at the age of 6 months, which normalized at the age of 12 month. At this time point the ROS level in muscle and live was comparable to mtAKR control mice. In dependence of age mtNOD mice showed a com- parable or higher expression of antioxidative enzymes than mtAKR control mice, whereas in mtBPL mice the expression of all antioxidative enzymes and in particular the expression of catalase was significantly reduced. Conclusion: The coincidence of mtDNA point mutations in the cytochrome c oxidase and NADH dehydrogenase resulted in higher blood glucose levels in 6 month old mtBPL mice. At this age mtBPL mice showed a significant increase of ROS production in liver and muscle. Thus, mtDNA mutations confer high ROS production in liver during adolescence with inadequate adaptation of the antioxidative system. This scenario could be a pathogenic trigger for insulin resistance with progressive ageing and contribute to the development of type 2 diabetes mellitus. 1313 Streptozotocin-induced diabetes does not affect mitochondrial structure and function N. Volk1, T.H. Fleming1, A.S. Hidmark1, E.-M. de Villiers2, P.P. Nawroth1; 1Department of Medicine I and Clinical Chemistry, University of Heidelberg, 2Division for the Characterization of Tumorviruses, Deutsches Krebsforschungszentrum, Heidelberg, Germany. Background and aims: A central mechanism underlying the development and progression of late diabetic complications is increased production of reactive oxygen species (ROS). Mitochondria, and specifically the electron transport chain (ETC), are considered to be the main source for intracellular ROS. During hyperglycemia, ROS production is increased leading to chang- es in the function of mitochondria including decreased activity of the ETC complexes, reduced ATP production, as well as changes in morphology. Un- published data from our research group have shown discrepancies between hyperglycemic conditions and elevated ROS levels. To estimate whether mi- tochondria dysfunction is associated with diabetes, the structure and func- tional properties of mitochondria were studied in the organs of streptozo- tocin-induced diabetic mice, using high-through put screening assays. Materials and methods: Diabetes was induced in healthy, wild-type mice (C57BI/6; male), by low-dose streptozotocin treatment. After four months, the hearts, kidneys, and livers from control and diabetic mice were collected, and mitochondria isolated by differential centrifugation. Purity of the iso- lated mitochondria was determined by electron microscopy (EM) and FACS. Morphology of the mitochondria within the tissue was also assessed by EM. The abundance of ETC complexes was analyzed by western blot (WB) and the activity of the complexes was tested using specific activity assays. Total oxygen consumption, ATP, and ROS production were analyzed during state 3 and state 4 respiration. Results: After four months of diabetes, both blood glucose (158.8 ± 4.1 mg/ dl vs 467.8 ± 75.7 mg/dl, P-value = 0.0002) and HbA1c (2.9 ± 0.1 % vs 9.5 ± 0.4 %, P-value < 0.0001) were significantly increased in the diabetic group, and body weight was significantly decreased (36.1 ± 1.2 mg vs 23.8 ± 2.0 mg, P-value < 0.0001). The mitochondria fraction isolated from the organs of each group had a purity of more than 95 %. No gross morphological changes between control and diabetic groups were observed in either the isolated mi- tochondria or within the tissue. No differences were observed with respect to the abundance of the ETC complexes, ATP, and ROS production. However, during state 3 respiration, both heart- and liver-derived mitochondria from diabetic mice showed increased oxygen consumption. During state 4 respira- tion, oxygen consumption was increased in liver-derived mitochondria from diabetic mice. No changes could be observed at either state 3 or 4 for kidney- derived mitochondria. Conclusion: This study describes a methodology not only for the isolation of high purity mitochondria from tissue, but also high-through put screening of mitochondrial function. Utilizing this methodology, it was found that at four months of streptozotocin-induced diabetes, there were little or no changes in the structure and function of mitochondria. Additional studies are required to determine whether mitochondria dysfunction is observed at later stages of diabetes and is associated with increased ROS production. 1314 Pyruvate kinase isoenzyme M2 shows increased expression in diabetes S. Vittas, S. Kopf, N. Volk, A. Hidmark, D. Oikonomou, T.H. Fleming, P.P. Nawroth; Department of Medicine I and Clinical Chemistry, University of Heidelberg, Germany. Background and aims: Pyruvate kinase (PK) catalyzes the conversion of phosphoenolpyruvate (PEP) to pyruvate, generating ATP as the last step of glycolysis. PK is expressed in 3 different isoforms named M1, M2 and LR. PKM1 and PKM2 show different activity levels, with PKM2 being highly expressed in cancer where enables cells to use aerobic glycolysis instead of oxidative phosphorylation. mRNA expression of the different isoforms (M1- M2) was investigated using specific primers and siRNAs in control and dia- betic material from mouse and human in order to investigate whether PKM isoforms show differential expression in diabetes and how this balance affects late diabetic complications. Materials and methods: Hearts, kidneys, sciatic nerves (SN) and dorsal root ganglia (DRG) derived from 3-6 months control and diabetic mice and hu- man peripheral blood mononuclear cells (pBMCs) derived from healthy aged matched control and type 2 diabetic individuals were used for RNA extrac- tion, cDNA preparation and subsequent real-time PCR using specific prim- ers for the different PKM isoforms (M1-M2). Additionally, in vitro experi- ments were performed by the use of siRNA specific for either both or M2 PK isoform in order to measure metabolic dysfunction with respect to glycolysis and the electron transport chain. Results: PKM isoforms show differential expression in diabetic tissues and cells in compare to the respective controls. PK M1/M2 ratio is significantly decreased in hearts, kidneys, SN and DRG derived from diabetic mice in compare to age matched control mice. The same ratio is also decreased in human pBMCs derived from diabetic individuals in compare to control ones with a positive correlation being found between the levels of PKM2 isoform and the neuropathy symptom score (NSS) in the diabetic group. Conclusion: PK as the last enzyme of glycolysis plays an important role in cell metabolism. This study reveals a differential expression of the two major PK isoforms (M1-M2) in diabetes. The increased PKM2 expression in dia- betic tissue can be correlated with the ability of cells to use aerobic glycolysis instead of oxidative phosphorylation. These findings reveal new aspects in development of late diabetic complications, thereby providing new opportu- nities for therapeutic interventions. Supported by: DFG (BI-1281/3-1 & NA 138 /7-1), DZD/BMBF, Dietmar Hopp Stiftung Diabetologia (2014) 57:[Suppl1]S1–S564 S 533 1 C PS 120 Animal models of complications I 1315 Treatment with divalent copper chelation reverses defective myocellular copper transport in the heart of diabetic rat S. Zhang1, H. Liu1, G. Amarsingh1, C.C.H. Cheung1, S. Hogl1, U. Narayanan1, L. Zhang1, S. McHarg2, J. Xu1, D. Gong1, J. Kennedy3, B. Barry3, Y. Choong1, A.R.J. Phillips1, G.J.S. Cooper1,2; 1School of Biological Sciences, University of Auckland, New Zealand, 2Institute of Human Development, University of Manchester, UK, 3National Isotope Centre, GNS Science, Wellington, New Zealand. Background and aims: Heart disease is the leading cause of death in diabetic patients, and defective copper metabolism could play important roles in the pathogenesis of diabetic cardiomyopathy. This study was to determine how myocardial copper status and myocellular copper chaperon proteins become impaired by diabetes, and how they respond to treatment with the Cu(II)- selective chelator triethylenetetramine (TETA). Materials and methods: Experiments were performed using streptozo- tocin (STZ)-induced diabetic rats with and without TETA treatment. Car- diac function were analysed in isolated, ex-vivo perfused working hearts and myocardial copper content measured using particle-induced x-ray emission spectroscopy coupled with Rutherford backscattering spectrometry. The ex- pression (mRNA and proteins) and/or activity of key protein components involved in copper binding and transport were analysed using combination of RT-qPCR, western blotting, immunofluorescence staining and enzyme ac- tivity assays. Results: Left-ventricular (LV) copper levels and LV function were markedly deficient (~50%, p S 534 1 C 2 vs 5.6±0.4.10-2, P S 535 1 C the simultaneous regeneration of reduced glutathione that occurs following activation of the PPP. Conclusion: A pro-inflammatory stimulus such as IL1β transforms excess glucose into a deleterious agent in VSMC by causing an increase in glucose uptake and its subsequent diversion into the PPP. This action ultimately promotes the pro-oxidant conditions required for the exaggeration of pro- inflammatory pathways. Interestingly, all these pathways were blunted by the blockade of IL1 receptors with anakinra. These findings can improve our understanding of the mechanisms for the development of diabetic vascular complications and suggest the combination of anti-inflammatory strategies together with glycaemic control for a more efficient prevention of such dis- eases. Supported by: SAF2011-24648; SAF2011.28011 1321 Nitric oxide hyperproduction and markers of DNA damage in the early phase of diabetic nephropathy in rat streptozotocin diabetes mellitus model J. Sokolovska1, E. Rostoka1, O. Sugoka2, S. Isajevs3, L. Baumane1, J. Sharipova1, I. Kalvinsh1, N. Sjakste1; 1Latvian Institute of Organic synthesis, Riga, 2University of Latvia, Institute of Biology, Salaspils, 3University of Latvia, Riga, Latvia. Background and aims: Reactive oxygen species (ROS) production due to intracellular hyperglycaemia is an important factor that initiates inflamma- tory pathways leading to diabetic nephropathy. NO hyperproduction in the kidney might lead to nitrosative stress with resulting DNA damage. However, data on NO, nitric oxide synthases (NOS) in the diabetic nephropathy are contradictory. Other enzymes like xantine oxidoreductase (XOR) might be involved. The aim of this work was to study changes and mechanisms of NO production in the kidneys of rats with streptozotocin (STZ) diabetes mellitus model, along with expression of poly ADP ribose polymerase (PARP) and histone gamma H2AX as markers of DNA damage. In order to modify NO production, we have used 1,4-dihydropyridine (1,4-DHP) class compounds, synthesized in the Latvian Institute of Organic Synthesis. Materials and methods: Diabetes mellitus in rats was induced by streptozo- tocin (STZ) 50 mg/kg, i.v. Production of NO in kidneys was monitored by means of ESP spectroscopy of Fe-DETC-NO complex. Different inhibitors of NO synthesis (GdCl3, aminoguanidine, 1400W, allopurinole) have been ap- plied. 1,4-DHP class drugs have been administered 0,5 mg/kg for 3 days per os. Kidney iNOS, eNOS, XOR, PARP and H2AX mRNA and protein expres- sion were detected by qRT-PCR and immunohistochemistry correspond- ingly, 12 days after diabetes mellitus induction and after 3 days of treatment with DHP class drugs. Results: NO production increase in kidneys was stable and potent (from 2.64 ± 0.97 to 15.04 ± 2.04 ng/g tissue). Application of non-selective NOS inhibi- tor aminoguanidine and XOR inhibitor allopurinol resulted in a significant decrease of NO hyperproduction. Further, 1,4-DHP class compounds etafto- ron, cerebrocrast and fenoftoron significantly decreased NO production. iNOS and XOR protein expression was upregulated in STZ rat kidney and decreased under etaftaron treatment (iNOS: control 11 ± 4 cells/mm2, STZ: 29 ± 15 cells/mm2 , STZ+etaftoron; 13 ± 6 cells/mm2 p S 536 1 C Results: Plasma glucose was significantly higher in diabetic groups than in controls (P < 0.05). The RT-PCR detection of the aquaporins (1, 7, 8, 9 &11) considered in this study is illustrated in Fig. 1. The level of expression of AQP 1 in diabetic testis was similar to that of control groups. The levels of expres- sion of AQPs 7, 8, 9 and 11 in diabetic testis were significantly higher com- pared to controls (P < 0.05). Conclusion: In summary, this work describes the expression of AQPs 7, 8, 9, & 11 in testis of diabetic rats were altered compared to controls. Altera- tion of AQPs expression leads to disruption of water homeostasis and affects spermatogenesis. Recent, studies shows that AQPs 7, 8 & 11 plays an im- portant role in spermatogenesis and physiological sperm volume regulation. AQP9 plays a specific role in the transport of water and non-charged solutes in Leydig cells. Furthermore, altered expression and regulation of AQPs have already been demonstrated to be the cause of several disorders of the male reproductive system. The full enlightenment of these molecular interactions and mechanisms may point towards possible therapeutic targets to counter- act diabetes induced male subfertility/infertility. PS 121 Animal models of complications II 1324 Decreased use of experimental animals through optimisation of experimental design for efficacy studies R. Stoop, A.M. van den Hoek, R. Hanemaaijer, S. van Buuren, A.-M. Zuurmond; Metabolic Health Research, Netherlands Organization for Applied Scientific Research (TNO), Leiden, Netherlands. Background and aims: Well-designed studies provide the maximum of in- formation from the minimal number of experimental animals. Aim of this study was to explore whether asymmetric experimental designs can reduce number of animals needed without affecting statistical power of the study. Materials and methods: Data from several efficacy studies in bleomycin- induced lung fibrosis in mice were combined to make tentative estimates of the expected outcomes for newly planned studies. Classically, these types of studies comprise a PBS control group, a bleomycin-induced (bleo) group, and several treatment groups with the same number of animals in all groups. Instead of comparing all groups to each other using ANOVA and post-hoc testing, k + 1 planned comparisons were defined in advance: bleo versus PBS control, and bleo versus k treatments (in total k + 1 comparisons). Since the bleo group appears in both the comparison with the PBS control and the treatment group, changing the size of this group will have most effect on the statistical power. Since the direction of the effects was anticipated, only one- sided t-tests were used. For fixed type I error, statistical power of the Student t-test was calculated for all combinations of group sizes of 2-30 animals. The optimal asymmetric design achieves the desired statistical power for each planned comparison with the minimum of animals needed. Results: We calculated that the best asymmetric designs required about 20- 25% fewer animals than the best symmetric design. Conclusion: This demonstrates that substantial reduction of animal use is possible by smarter choices on group size without loss of power. To aid re- searchers to optimize their experimental design, we developed a freely availa- ble sample size calculator (available at www.tno.nl/3R) to compare symmetric and asymmetric designs. 1325 Twice-daily hyperglycaemic spikes accelerate atherosclerotic lesion formation in C57BL/6 mice Y. Shuto, A. Asai, M. Nagao, M. Kawahara, H. Sugihara, S. Oikawa; Endocrinology, Diabetes and Metabolism, Nippon Medical School, Tokyo, Japan. Background and aims: Increasing epidemiological evidence indicates that postprandial hyperglycemia in individuals with impaired glucose tolerance (IGT) is a risk factor for atherosclerotic cardiovascular disease. A number of studies have demonstrated the possible roles of acute glucose fluctuations, a hallmark of IGT, in atherogenesis in cell culture models; e.g., intermittent high glucose exposure-induced oxidative stress, inflammatory response, and subsequent proatherogenic changes in vascular endothelial cells, smooth muscle cells, and macrophages. However, there is little convincing experi- mental evidence that repetitive glucose fluctuations can actually result in accelerated atherosclerosis in vivo. Hence, the in vivo mechanism underly- ing IGT-induced atherogenesis is not yet fully understood. In this study, we demonstrated the effect of long-term repetitive hyperglycemic spikes on ath- erosclerotic lesion formation in mice. Materials and methods: Female C57BL/6 mice were fed an atherogenic diet (AD; 1.25% w/w cholesterol, 0.5% w/w sodium cholate, 36% energy fat) from 8 to 28 weeks of age. During the 20-week AD feeding period, the mice re- ceived 20% glucose solution (50 mg glucose/mouse) (G) or distilled water (W) twice daily by oral gavage (6 days a week). The post-challenge blood glucose changes were monitored on a day in the week 2, 11, and 20. OGTT and insulin tolerance test were performed in the week 19 and 20, respec- tively. After the AD feeding, serial frozen sections (10 µm thickness, spanned 450 µm) of aortic sinus were prepared and the atherosclerotic lesion size was quantitated from the oil red O-stained area (mean of 9 sections, each sepa- rated by 50 µm). Gene expression levels in the thoracic aorta and the plasma Diabetologia (2014) 57:[Suppl1]S1–S564 S 537 1 C lipid profiles were also analyzed. Data are expressed as mean±SEM (n=8). Differences between G and W were assessed by Student’s t test. Results: Blood glucose levels in G increased by ~5 mmol/l above W at 20 min after the glucose ingestions and returned to similar levels to W within 60 min. There are no significant differences in body weight, plasma lipid profiles (to- tal cholesterol, HDL-cholesterol, non-HDL-cholesterol, triacylglycerols, and NEFA), and blood glucose levels in OGTT and insulin tolerance test between G and W. Aortic sinus atherosclerotic lesion size in G was 3.8-fold greater than that of W (22.2±5.7 vs 5.8±1.8 ×103 µm2, p=0.025). Gene expression levels of an adhesion molecule Icam-1 (1.26-fold, p=0.036) and a macrophage marker Cd68 (1.27-fold, p=0.036) in thoracic aorta were significantly higher in G relative to those in W. Conclusion: These results demonstrate that repetitive hyperglycemic spikes can accelerate atherosclerotic lesion formation in vivo independent of basal blood glucose levels, insulin resistance, or plasma lipid profiles, possibly via proatherogenic changes in arterial wall. The repetitive glucose ingestion in mice on AD would be a simple and useful method for studying the patho- genesis, prevention, and treatment of IGT-induced atherosclerotic vascular disorders. 1326 Cardiac magnetic resonance imaging evaluation after experimental myocardial infarction in streptozotocine induced diabetic rats with high glycaemic variability M. Joubert1, J. Hardouin1, K. Blanchart2, S. Allouche3, A. Manrique3; 1Endocrinology Unit, 2Cardiology Unit, University Hospital of Caen, 3EA4650, Caen, France. Background and aims: During acute myocardial infarction (MI) in type 2 diabetic (T2D) patients, both hyperglycemia and hypoglycemia (at admis- sion and/or during the hospital stay) are associated with poor outcome. In addition, the impact of glycemic variability in these patients is also discussed: several in vitro and in vivo studies have suggested a deleterious effect of gly- cemic variability on endothelial function through an increase in oxidative stress. However, there is actually no data concerning the direct impact of this glucose metric on MI morbidity. We address here this question in a rodent diabetic model. Materials and methods: Twenty-six adult male Wistar rats (315 ± 41 g), synchronized to a reverse 12 hours dark-light cycle were divided in three groups : control (C) (n=7), sham (S) (n=5) and diabetes (D) (n=14). Rats from group D received 70 mg/kg of intraperitoneal (IP) streptozotocin (STZ). Rats from group C and S received vehicle only 70 mg/Kg IP. Group D was divided in two subgroups : diabetes hyperglycemia (Dh) (n=6) and diabetes variability (Dv) (n=8). Glargine was introduced 48 hours after STZ injection : 5 UI/kg/j subcutaneously (SC) at 9:00 AM for Dh rats and 15 UI/ kg/j SC at 9:00 AM for Dv rats. Blood glucose monitoring was performed to confirm that glycemia was maintained above 250 mg/dl in Dh sub-group and presented swings below 150 and above 250 mg/dl in Dv sub-group. After 4 weeks, all rats (except in group S) were subjected to ischemia (30 min of transient coronary artery ligation) followed by reperfusion. Cardiac magnetic resonance imaging (CMR) was performed 1 and 3 weeks (w1 and w3) after surgery to assess left ventricular function [end-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (LVEF), wall motion score (WMS)] and myocardial edema. Results: Mean blood glucose was 140±24, 130±28 and 265±88 mg/dl in group C, S and D (global p S 538 1 C Conclusion: Our results provide evidence that hyperglycemia is an inde- pendent rist factor for atherosclerosis, but it does not promote restenosis, and different mechanisms are involved in the development of atherosclerosis and restenosis. Supported by: NSFC, CMAF 1329 Effect of glucose level on the size of myocardial infarction in pigs S.S. Diemar1,2, A.-S. Sejling1,3, K.K. Iversen1, T. Engstrøm4, J.L. Honge5, N. Tønder1, N. Vejlstrup4, K. Ekström1,2, U. Pedersen-Bjergaard1,2, M. Dalsgaard1, B. Thorsteinsson1,2; 1Department of Cardiology, Nephrology and Endocrinology, Nordsjællands Hospital, Hillerød, 2Faculty of Health Sciences, University of Copenhagen, 3Faculty of Health Sciences, University of Southern Denmark, Odense, 4Department of Cardiology, Rigshospitalet, Copenhagen, 5Department of Cardiothoracic and Vascular Surgery, Aarhus University Hospital, Skejby, Denmark. Background and aims: Patients with diabetes and acute myocardial infarc- tion (MI) have a two- to five-fold increased in-hospital mortality compared with non-diabetic patients. Concomitant hypoglycaemia or hyperglycaemia is suggested to worsen the acute prognosis of MI. Mortality after MI is associ- ated with necrotic damage to the myocardium. Consequently, minimising the necrotic area is warranted. Our aim was to investigate the effect of the acute plasma glucose level on the size of MI in a closed-chest pig model. Materials and methods: 38 non-diabetic Danish/Landrace female pigs were randomised to three different plasma glucose levels: hypoglycaemia (2.1 ± 1.0 mmol/l; 15 pigs (mean±SD)): normoglycaemia (5.8 ± 1.6 mmol/l; 12 pigs) and hyperglycaemia (21.5 ± 2.9 mmol/l; 11 pigs). After 30 minutes of steady state with plasma glucose levels within target (1.8-2.2 mmol/l, 5-7 mmol/l, 22-23 mmol/l, respectively) MI was induced. Using a closed-chest model a balloon catheter was inserted via the internal carotid artery to the left ante- rior descending coronary artery and inflated for 30 minutes. Afterwards the heart was reperfused while the plasma glucose level remained in target for 2 hours. Following sternum split an in-vivo staining with Evans Blue was per- formed. The heart was subsequently taken out and in-vitro stained with TTC to delineate the necrotic tissue from the viable tissue (area at risk (AAR)). Results: Twenty-five (66%) pigs developed ventricular fibrillation (VF) dur- ing the reperfusion period. Of these 7 pigs died. There was no difference in incidence of VF (p=1.0) or death (p=0.22) between the glycaemic groups. No significant differences in infarction size, AAR or infarction/AAR ratio were detectable between the groups. Conclusion: In a closed-chest pig model the acute glucose level did neither affect cardiac morbidity or mortality nor the size of the myocardial infarction. Supported by: Research Foundation of Nordsjællands Hospital, Hillerød, Den- mark 1330 Dipeptidyl peptidase-4 inhibition in a rat model of ischaemia- reperfusion injury may accelerate tubular regeneration but does not improve glomerular filtration rate B. Hocher1, C. Reichetzeder1, K. von Websky1, O. Tsuprykov1, T. Klein2; 1University of Potsdam, 2Boehringer Ingelheim, Biberach, Germany. Background and aims: Dipeptidyl peptidase (DPP)-4 inhibitors have been shown to have protective effects on ischaemia-reperfusion injury (IRI) of the heart and lung. As the DPP-4 enzyme and its substrates are also expressed in the kidney, we studied the effects of the DPP-4 inhibitors linagliptin, vilda- gliptin, and sitagliptin on the outcome of IR-induced acute kidney injury in uninephrectomised rats. Materials and methods: Male Wistar rats obtained from Charles River (Ger- many) weighing 150-170 g were subjected to uninephrectomy. Two weeks later, the remaining kidney was exposed to IRI by clamping the renal artery for 30 min; sham surgery was performed without clamping. The rats (n=10- 14 per group) received DPP-4 inhibitor treatment once-daily with either linagliptin (1.5 mg/kg/day), vildagliptin (8 mg/kg/day), sitagliptin (30 mg/ kg/day), or vehicle via gavage on the 3 consecutive days prior to IRI. An ad- ditional group was treated with sitagliptin until study end in order to inhibit DPP-4 activity during the entire experiment. This group was treated initially with sitagliptin 30 mg/kg/day; after induction of IRI, the dose was adjusted to 15 mg/kg/day because of renal failure. Levels of plasma cystatin C (a bio- marker of glomerular filtration rate) and clusterin (a biomarker of tubular regeneration) were measured at 0, 24, 48, 72, and 168 hours after IRI or sham operation. Results: Levels of active GLP-1 increased 3-4 fold in all DPP-4 inhibitor treatment groups versus the placebo group, up to 24 hours after clamping (p S 539 1 C 1331 Morphometric characteristics of the air-blood barrier in experimental diabetes mellitus O. Pivovarova1, E. Rozova2, B. Mankovsky3; 1State Establishment „Lugansk State Medical University, Lugansk, 2Bogomoletz Institute of Physiology, Kyiv, 3P.L. Shupyk National Medical Academy of Post-Graduate Education, Kyiv, Ukraine. Background and aims: Respiratory diseases contribute significatntly to mor- tality and morbidity in patients with diabetes mellitus. However, the mecha- nisms underlying the lung damage in diabetic patients are not fully under- stood. The air-blood barrier (ABB) performs an important role in function- ing of the respiratory system but his structure in diabetes was not studied in detail. Therefore, the aim of this study was to investigate specific histological features of the ABB in the streptozotocin (STZ)-induced diabetes. Materials and methods: The study was performed on 47 STZ-induced diabet- ic and 43 control male Wistar rats. Diabetes was induced by single intraperito- neal injection of STZ (SIGMA, USA) 60 mg/kg in 0.1 M citrate buffer, pH 4.5. The tissue for the further evaluation was collected after 18 weeks of diabetes. The diabetic rats were not treated with insulin. The morphometric assessment was performed using the electron microscope PEM-125K and the data were analyzed using the software Image Tool Version 3 (USА). The thickness of arithmetic average (τ) and harmonic average (τh) were calculated. Results: We found the significant increase of thicknesses of τ and τh of ABB in STZ-induced diabetic rats compared to control group - τ were 302,7 ± 11,4 nm and 163,4 ± 6,8 nm, τh were - 289,5 ± 12,5 nm and 155,6 ± 5,4 nm, re- spectively, in diabetic and control animals, p 1 C S 540 Author Index A Aasberg, A. 296 Abassi, Z. 801 Abbas, Z. G. 1138 Abdel-Wahab, Y. H. A. 495, 497 Abderrahmani, A. 609 Abdul-Ghani, M. 552 Abdul-Rasoul, M. 1262 Abels, M. 133, 490, 492 Aberer, F. 1005, 979 Abgdauletova, S. 819 Abiko, A. 526, 536 Aboromia, M. M. M. 1289 Abouglila, K. 993 Abreu, C. 910 Abu Seman, N. 355 Achenbach, P. 334 Adachi, H. 1133 Adachi, Y. 766 Adam, K.-P. 142, 548 Adamidou, A. 653 Adamska, A. 222 Adamska, E. 356 Adas, M. 449 Adeli, K. 1308, 921 Adina, M. 278 Admiraal, W. M. 21 Adriaenssens, A. A. 202 Adrian, C. L. 936 Afonso, M. J. 1061 Agakov, F. 11 Agardh, E. 1157 Agbalika, F. 512 Agesen, R. M. 485 Aggelidi, A. 1240 Agha-Jaffar, R. 1044 Agoston, V. A. 746 Agrawal, P. 970 Aharon-Hananel, G. 658 Şahin, I. 941 Ahlkvist, L. 537 Ahlquist, E. 107 Ahlqvist, E. 90 Ahlqvist, M. 673 Ahluwalia, R. 541 Ahmaan, A. J. 949 Ahmad, S. 65 Ahmann, A. 1003, 37, 649 Ahn, C. 873 Ahn, J. 295 Ahn, J. 569 Ahn, S.-Y. 214 Ahn, Y.-B. 10, 547, 652 Ahrén, B. 403, 491, 518, 537, 55, 56, 829 Ahren, B. 831 Ahrén, B. 886 Ai, H. 287, 710 Aikawa, M. 573 Aikawa, N. 929 Ait Djoudi, M. 1227 Aitken, R. J. 450 Ajdukovic, D. 1028 Ajjan, R. A. 993 Ajmi, S. 1125 Akacha, M. 886 Akanuma, Y. 272 Akasheva, D. 95 Akbarova, N. 1204 Akhtar, S. 566 Akimoto, Y. 88 Akirav, E. M. 1089 Akolkar, B. 334 Al-Aissa, Z. 444 Ala-Korpela, M. 1283 Alam, D. S. 236 Alam, S. M. K. 596 Al-Arouj, M. 883 Alba, M. 853 Albano, L. 758 Albareda, M. 1224 Albert, L. 1244 Alberti, C. 1033 Albiero, M. 729 Albornos, C. 910 Albrecht, T. 715 Albrechtsen, A. 103 Alcaide, J. 1108 Alcarraz-Vizán, G. 706 Alcarria, E. 910 Aldington, S. J. 1160 Aldington, S. J. 49 Alenkvist, I. 395 Aletras, V. 988 Alevizaki, M. 1086 Alexander, C. M.. 322 Alexandre, D. 554 Alexopoulos, A.-S. 1038 Al-Hasani, H. 419, 557 Ali, F. M. 303 Ali, L. 654 Ali, W. 1047 Ali, Y. 379 Alkayyali, S. 345 Alkhalaf, A. 1219, 1232 Alla, F. 151 Allard, C. 670 Allen, J. M. 1007, 194, 960, 989 Allingbjerg, M.-L. 485 Allouche, S. 1326 Alluis, B. 932, 934 Almeida, A. 261 Almeida-Pititto, B. 1037 Almgren, M. 743 Almgren, P. 143, 352 Almodóvar, F. 910 Alquier, T. 802 Alsalim, W. 518 Alsifri, S. 481 Alt, M. 748 Altabas, V. 1075 Altshuler, D. 363 Altzibar, J. M. 1222, 1268 Alvarado Martel, D. 1050 Álvarez, C. 138 Alvarez, C. 240 Álvarez, C. 736 Alvarez, M. A. 187 Alvarsson, M. 887 Alves, T. C. 394 Amarsingh, G. 1315 Amati, F. 615, 617 Ambery, P. D. 41, 837, 908 Ambos, A. 841 Ametov, A. S. 927 Amiel, S. A. 1007, 194, 316, 476, 645, 989 Amisten, S. 99 Amols, M. 1263 Amorese, G. 478 Amouzou, C. 586 Amri, E.-Z. 693 An, H. 517 Anarte, M. T. 438 Anastasiadis, G. 631 Anastasiou, E. 1086 Andelova, K. 1091 Andersen, C. 1113 Andersen, D. K. 213 Andersen, E. S. 1113 Andersen, G. S. 1162 Andersen, G. 934 Andersen, H. Ullits. 1025 Andersen, K. R. 2, 818 Andersen, L. 1049 Andersen, M. A. 103 Andersen, M. L. M. 1250 Andersen, S. 198 Anderson, B. 1251, 1257 Anderson, G. 548 Anderson, J. E. 1070 Anderson, S. G. 1029 Andersson, L. 393, 399 Andersson, L. E. 398 Andersson, T. 173 Anděl, M. 810 Ando, T. 422 Andou, K. 513 Andreassen, K. V. 84 Andreazzi Duarte, D. 1166 Andren Aronsson, C. 170 Andres, E. 864 Andrews, Z. B. 672 Andrulli Buccheri, V. 1316 Angeletti, S. 448 Angelidi, A. 1027, 1295 Angelopoulou, A. 1086 Anjana, R. Mohan. 1082 Anouar, Y. 554 Anselmino, M. 125, 688 Antoine, B. 767 Antoine, N. 58 Antonarakis, S. E.. 30 Anwar, T. 596 Aoki, S. 874 Apanovitch, A. 800, 804 Aparicio, S. 520 Arabo, A. 554 Arafat, A. 525 Arafat, A. M.. 653 Arai, H. 663 Arakawa, T. 722 Araki, A. 1209, 1215 Araki, E. 663 Arase, Y. 298 Araszkiewicz, A. 1144 Araya-Ramirez, F. 212 Archibald, L. K. 1138 Ardestani, A. 429 Ardilouze, J.-L. 1094 Arias, P. 657 Aribi, S. 1083 Arnal, J.-F. 225 Arnardottir, N. Y. 63 Arnolds, S. 1034, 989 Aroda, V. R. 145 Aroney, S. A. 594 Aronne, L. 184 Aronson, J. K. 1203 Aronson, R. 1002, 1009, 1012, 481, 955, 999 Arosio, M. 210 Arous, C. 431, 434 Arpacı, D. 941 Arpi, M. 451 Arranz, A. 506 Arrese, M. 982, 983 Arslan, M. I. 596 Arteaga, J. 841 Arteagoitia, J. Mª. 1222 Arteagoitia, J. M. 1268 Arts, I. C. W. 1234 Arzumanov, S. 480 Asahara, S.-I. 418, 454 Asai, A. 1325, 708 Asai, S. 876 Asano, M. 1200 Asante, A. 1263 Ásbjörnsdóttir, B. 1117 Ascher-Svanum, H. 1026 Ashcroft, F. M. 202 Ashfaq, A. 259 Ashley, D. 489 Ashmore, T. 24 ASPIRE In-Home Study Group, 1003, 649 Assefa, B. 653 Astiarraga, B. 688 Aston, C. E. 1118 Astrup, A. 73, 833, 881, 903, 904 Aswendt, M. 335 Atageldieyeva, K. K. 536 Atageldiyeva, K. K. 1176 Atakan, Z. 309 Athanasiadou, E. 1077 Atisso, C. 40 Atkin, S. 182 Atkinson, R. L. 743 ATLANTIC DIP (Diabetes in Preg- nancy) Research Team, 1092 Atmaca, A. 941 Atsumi, T. 1265, 470, 874, 929 A.T. Teixeira, D. 1173 Aubert, C. Elodie. 1287 Augstein, P. 1051, 388 Augustin, T. 979 Aukrust, I. 363 Aust, G. 157 Auster, M. E. 251 Åvall, K. 379 Avdic, T. 196 Aveyard, P. 302 Avignon, A. 586, 640 Avila-Rubio, V. 587 Avó, L. 471 Avogaro, A. 729, 930 Avron, A. 453 Axelsson, A. S. 859 Ayad, F. 1083 Aydın, H. 941 Ayubova, N. Leonidovna. 160 Azar, S. T. 846 Azcutia, V. 1320 Azevedo, H. 261 Azevedo Da Silva, M. 260 Azizi, Z. 389 Aznar, M. 248 Azoulay, L. 892 B Baan, C. A. 1080, 14 Babadjanov, B. D. 1204 Babazono, T. 1191 Bacchiddu, S. 187 Bacci, S. 1235, 343 Bachaoui, M. 1083 Backhose, M. 1152 Bacos, K. 234, 468 Badeau, R. M. 1283 Badenhoop, K. 646 Bader, G. 900 Badin, P.-M. 32, 730 Badrick, E. 246 Badrick, E. L. 294 Bae, E. 35, 416 Bae, J. 873 Bae, J.-H. 593 Bae, J. 221 Bae, J. 858 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 541 Bae, K.-H. 1278, 1284, 238, 591, 867 Bae, M. 569 Bae, S.-J. 64 Baechle, C. 1258 Bærug, A. 293 Bagge, A. 393, 398, 400 Bagger, J. I. 540 Bagger, M. 1042 Baggesen, L. M. 789 Bähler, L. 21, 218, 22, 698 Bähr, A. 918 Bähr, V. 653 Bai, J. 1266 Baik, S. 295 Bailes, Z. 894 Bailetti, D. 357 Bailey, C. 701 Bailey, C. J. 846 Bailey, T. 1003, 145, 649, 994 Bailey, T. S. 949 Baillargeon, J.-P. 1094 Bain, S. 182 Baj, A. 285 Bajaj, H. S. 955 Baker, C. 1308, 921 Baker, D. 100 Bakker, S. J. L. 1219 Bakker, S. J.. L.. 1232, 1237 Bakker, S. J. L. 192 Bakker, S. J.. L.. 8 Bakk-Nurdisány, B. 746 Balaz, M. 167, 180 Baldassarre, D. 223 Baldi, S. 125, 527, 688 Balhuizen, A. 375 Balkau, B. 110, 260, 501 Balo, A. 994 Balti, E. V. 288, 452 Baltrusch, S. 1312, 726, 727, 728 Banakh, I. 388 Banasik, K. 139, 339, 341 Bandala-Sanchez, E. 388 Bandier, M. 145 Banks, P. 803 Banton, T. 13 Banu, I. 1236, 660 Barabash, A. 1168 Barale, C. 913 Baranov, O. 928 Barbaro, M. 299 Barbas, C. 741 Barbetti, F. 357 Barbuceanu, F. 1178 Bar-Dayan, Y. 776 Bardini, G. 1228 Bargellini, I. 1145 Bargiota, A. 988 Barkai, L. 1254, 1262 Barkhof, F. 81 Barna, G. 444 Barnard, K. 1008 Barnett, A. H. 900 Barnett, A. H. 850 Baron, A. 793 Baron, M. 1054 Baron, M. A. 242 Baroni, M. G. 305 Baronti, W. 478 Barosa, C. 592 Barquissau, V. 693 Barrat, C. 1275 Barrès, R. 235 Barres, R. 248 Barrett, A. 359 Barrio, F. 266 Barros, C. Risso. 1037 Barry, B. 1315 Barsotti, E. 688 Barsotti, M. 478 Bartaskova, D. 998 Bartaun, C. 821 Bartek, J. 995 Barter, P. 1233 Barton, J. 172 Barutta, F. 1186 Barzacchi, M. 1201 Basat, O. 1046 Basque Country Sentinel Practice Network, 1222, 1268 Bassols, J. 684 Battaglia, M. 231 Batut, A. 703 Bauer, W. 1090, 356 Bauman, J. 323 Baumane, L. 1321 Baumann, J. 747, 748 Baumann, K. 423 Baumstark, A. 1017 Bayer, M. 230, 411 Baz, B. 1227, 512 Băban, A. S. 1036 Beagley, J. 308 Beatson, C. R. 987 Beaucamps, C. 609 Beck, P. 979 Beck, R. 1251, 1257 Becker, R. H. 953 Beckers, H. J. M. 1156 Beck-Jensen, R. 1121 Beck-Nielsen, H. 1121, 282, 321 Bee, Y. M. 338 Bee, Y.-M. 26 Beer, R. 1231 Begovatz, P. 589 Beguinot, F. 353, 354, 758 Béhe, M. 447 Beijing Community Diabetes Study Group, 1267 Bekaert, M. 269, 763 Belan, V. 167, 180 Belgian Diabetes Registry, 288 Belhadj, M. 1083 Belinfante, E. R. M. 1098 Belinova, L. 735, 771 Bellante, R. 85 Bellary, S. 993 Bellazzi, R. 1245 Belles, C. 695 Belligoli, A. 125 Bellili-Muñoz, N. 1183 Bellini, R. 688 Belo, J. A. 28 Belobradkova, J. 1169 Belton, A. 1030, 1082 Beltramo, E. 51, 52 Bem, R. 1146, 1148 Benaiges, D. 187 Benali, L. 1083 Benard, N. 901 Benbenek-Klupa, T. 951 Bender, C. 472 Bendlová, B. 602 Bendlova, B. 1091 Benedini, S. 924 Benesch, C. 989 Benhalima, K. 1110 Benhamou, P.-Y. 477 Benharrat, K. 1083 Benko, R. 746 Benko, Z. 746 Bennet, H. 375, 466 Bennet, L. 551 Bennett, H. 1067 Benroubi, M. 38 Bensellam, M. 101 Béranger, G. 693 Berencsi, K. 321 Berendschot, T. T. J. 1156 Berensci, K. 282 Beres, N. 746 Bergenheim, K. 1067 Bergenstal, R. 1003, 181, 39, 76, 829 Bergenstal, R. M. 148, 878, 946, 947, 949, 961 Berggreen, C. 377 Berggren, J. 740 Berggren, P.-O. 136, 379 Bergis-Jurgan, N. 15 Bergmann, A. 525 Bergmann, K. 953 Berkeš, D. 34 Berković Cigrovski, M. 1075 Berney, T. 477, 991 Bernjak, A. 633 Berntorp, K. 1109 Berrahmoune, H. 554 Berta, R. 688 Bertolotto, A. 1084 Bertoquini, S. 1202 Bertuzzi, F. 462, 612 Besseiche, A. 396, 498 Bessho, M. 1274, 611 Best, J. 1233 Betsholtz, C. 88 Beulens, J. W. J. 770, 772 Beuzelin, D. 681, 693 Bezy, O. 1332 Bhanot, S. 861 Bhargava, A. 948 Bhatt, D. L.. 183 Bhatt, D. L. 186 Bhatt, D. L.. 483, 641, 877, 887, 930 Bhavadharini, B. 1082 Bhogal, M. 1152 Bian, H. 277, 718, 863 Bian, Z. 274 Bianchi, C. 656 Biden, T. J. 164, 463 Bierhaus, A. 87 Bierwagen, A. 581, 589 Biessen, E. A. L. 753 Bilan, P. J. 759 Bilban, M. 226 Bilen, H. 941 Bilet, L. 623 Billestrup, N. 408 Billionnet, C. 151 Bilo, H. J. G. 1219 Bilo, H. J.. G.. 1232, 1237 Bilo, H. J. G. 192, 193, 314 Bilous, M. 1198 Bilous, R. W. 1198 Bilous, R. W. 1170 Binda, M. 520 Bingley, P. 290 Bingley, P. J. 450 Biondi, G. 464 Biragova, M. 480 Birkeland, K. I. 293, 618 Birkenfeld, A. 861 Bisbal, C. 586 Biswas, S. K. 596 Bitterman, O. 1098 Bizzotto, R. 504 Bjerregaard, P. 103 Bjørkhaug, L. 363 Björne, S. 787 Blackman, A. 184 Blahova, K. 372 Blanchart, K. 1326 Blankenburg, S. 1233 Blasco, V. 571 Blasetti Fantauzzi, C. 1307 Blaslov, K. 1161, 440, 528 Blauw, H. 1011 Blevins, T. 146 Blickle, J.-F. 1071 Blizanowska, P. 1120 Blonde, L. 846 Blonde, L. 112, 42, 836, 847, 898, 962, 973 Blondeau, B. 396, 498 B. Lopes de Faria, J. 1173 Blüher, M. 1276, 144, 662, 709, 765 Blutke, A. 918 Boaz, M. 1069, 637 Bobadilla, L. L. 1185 Bobruff, Y. 545 Bode, B. 1003, 149, 181, 39, 5, 649, 76 Bode, B. W. 41, 78, 959 Boeing, H. 770 Boemi, M. 1066 Boenner, A. 168 Boerman, O. C. 447 Boesch, C. 615 Boff, R. 1190 Bogdanov, P. 1163 Boggi, U. 373, 478 Bogl, L. H. 36 Bohl, M. 733 Bojsen-Møller, K. N. 122, 534 Bolan, E. 132 Bolanos, J. P. 1320 Bolek, T. 43 Bollerslev, J. 296 Bolli, G. 608 Bolli, G. B. 148, 946, 947, 961, 963, 980 Bomans, M. 253 Bonadonna, R. 511 Bonadonna, R. C.. 745 Bonardo, B. 575 Bond, C. 1116 Bondarenko, O. N. 160 Bondor, C. I. 1137 Bonelli, A. L.. 657 Bonfanti, R. 1253 Bonfiglio, A. 905 Böni-Schnetzler, M. 614, 751 Bonnaure, H. 1242 Bonner, C. 609 Bonnet, F. 381 Bonnici, F. 453 Bonnici, F. 877 Bonora, E. 305, 511 Boomsma, D. 139, 339 Boon, P. 629 Boopalan, A. 37 Bordin, S. 550 Borel, C. 30 Bories, L. 477 Borio, L. 51 Borkenstein, M. 232 Bornat, Y. 991 Borucka, K. 1144 Boscari, F. 1019 Bosch, F. 419, 432, 712 Bosch, J. 1270 Boschero, A. C. 424 Bosch-Traberg, H. 182 Bosco, D. 991 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 542 Boselli, L. 511 Bosi, E. 451 Bosma, H. 63 Boss, A. 615 Bossowski, A. 195 Botezelli, J. Diego. 616 Böttcher, Y. 347, 662 Botteman, M. 12 Böttger, B. 12 Botusan, I. R.. 250 Bouam, M. 711 Boucek, P. 479 Bouchard, J. 323 Boudou, P. 1227, 512 Boudoux, J. 642 Boulet, N. 695, 696 Bouloumié, A. 225, 695, 696 Boulton, A. J. M. 1131, 1155, 129, 130 Boulvain, M. 1103 Bound, M. J. 217, 774 Boura, P. 1077 Bourguignon, J.-P. 1043 Bourlier, V. 730 Bourron, O. 1287 Boursiquot, L. 648 Bousboulas, S. 1150, 1246 Boushey, C. J. 270 Boutati, E. 1240, 1295 Boutens, L. 755 Bouzakri, K. 431 Bowden, D. W. 1291 Bowe, J. E. 100 Bowen, A. 384 Bowering, K. 852 Bowers, R. 212 Bowling, F. 1131 Bowling, F. L. 129, 130 Boye, K. 221 Boye, K. S. 839, 907 Boytsov, S. 95 Boytsov, S. A.. 1241 Bozkurt, L. 1112 Bradfield, J. P. 106 Bradnova, O. 1091, 602 Braffett, B. 127 Braffett, B. H.. 128 Braga, L. 432 Bragg, F. 274 Brailov, Y. 95 Brailova, N. V. 1241, 95 Bramlage, P. 206, 304 Branda, M. E. 1077 Brändle, M. 17 Brandslund, I. 282, 321 Brangani, C. 511 Brath, H. 900 Brath, H. 1272 Braun-Reichhart, C. 918 Braunwald, E. 186 Bravis, V. 1153, 290 Brazg, R. 1003, 649 Bréant, B. 396 Breitel, S. 1010 Breitfeld, J. 144, 709 Brekke, L. 111 Brekke, M. 1216 Brereton, M. F. 202 Bretzel, R. G. 887 Breuker, C. 586 Briganti, S. 448 Brink, R. 98 Brinks, R. 1214 Brismar, K. 250, 336, 355, 600, 743 Broccatelli, M. 620 Brodovicz, K. G.. 322 Brodovicz, K. G. 898 Broedl, U. C. 1, 2, 806, 818 Broedl, U. C.. 849 Broedl, U. C. 851, 956 Brom, M. 445, 447 Brønden, A. 532 Brøndsted, L. 936 Broskey, N. T. 615, 617 Brouwers, B. 623 Brouwers, O. 713 Brown, K. K. 982, 983 Brown, R. 864 Brown, S. 1152 Brown, S. J. 1131 Brown, S. J. 129, 130 Brown-Augsburger, P. L. 938 Brownlee, M. 46 Brozinick, J. T. 759 Brozova, J. 1132 Brozova, L. 995 Brozzi, F. 428 Brudi, P. 898 Bruerberg, S. 397 Brüggemann, J. 249 Brugnara, L. 624 Brun, J.-F.R. 477 Brunak, S. 139, 339 Brune, M. 1282 Brunel, N. 642 Brunkwall, L. 62 Brunner, E. J. 61 Bruno, R. 1201 Brunton, S. 1070, 112, 962 Brusko, T. 231 Bruttomesso, D. 1000, 1019, 197 Brychta, R. J. 63 Bucci, M. 121, 177 Buchan, I. 246, 294 Buchanan, T. A. 156 Buchsbaum, M. S. 661 Buchwald, P. 785 Buckingham, B. A. 1000 Buda, P. J. 467 Buday, B. 604 Budiman, E. S. 1018, 990 Buehler, L. 1038 Bueno, P. G. 471 Buettner, C. 669 Bugianesi, E. 237 Bugliani, M. 373 Buhl, E. S. 789 Buhl, T. 1113 Buijs, J. 93 Bujacz-Jedrzejczak, U. 950 Bukau, B. 1282 Buldenko, T. 345 Bulsara, M. K. 1256 Bulum, T. 1161, 1281, 440, 528 Bumke-Vogt, C. 653 Buncová, M. 1146 Buono, P. 1253 Burcelin, R. 225 Burda, V. 912 Burgard, D. 1259 Burger, D. 91 Burkart, V. 415 Burkhardt, B. R. 595 Burlina, A. B. 555 Burns, C. 793 Burton, C. M. 1042 Burtt, N. P. 363 Bus, S. A.. 1154 Buse, J. 793 Buse, J. B. 149, 78, 835 Buss, U. 109, 928 Busse, N. 744 Butte, A. J. 1181 Büttner, P. 157 Button, E. 142 Buyse, M. 498, 717, 767 Buysman, E. 111 Buzkova, J. 680 Buzzetti, R. 210, 451, 742, 872 Buzzigoli, E. 237 Byrne, M. L. 476 Bytoft, B. 1121 C Caballero, A. Enrique. 894 Cabre, J. J. 266 Cabrera Morales, F. 1101 Caccioppoli, C. 1316 Cadavez, L. 200, 706 Cadoudal, T. 703 Cahn, A. 186, 483, 641, 930 Cai, D. hong. 1123 Cai, R. 1135, 666, 667, 72 Cai, X. 1057 Caiazzo, R. 233 Caixàs, A. 190 Calabrese, M. 1063 Calado, F. 813 Calandro, D. 1233 Caldeira, J. 563 Calders, P. 269, 629 Caldwell, B. D. 1141 Calewaert, P. 1110 Calhoun, P. 1000 Califf, R. M. 273 Calle, A. L. 1168 Calle, R. A. 529 Callesen, N. F. 1105 Calvi-Gries, F. 940 Calvi-Gries, F. 110, 113, 943, 973 Calvo, S. 1097 Camargo, E. G. 1190 Camargo, J. L. 1190 Camargo, R. G. 583 Camastra, S. 688 Cambier, D. 629 Camisasca, R. 897 Campagna, G. 451, 742 Campbell, M. D. 621 Campbell, P. 388 Campbell-Thompson, M. 285 Campitelli, M. 354 Campmans-Kuijpers, M. J. E. 770, 772 Campos, S. 506 Canavan, J. B. 701 Candelas, C. 1251, 1257 Candeloro, P. 608 Candido, R. 1066, 18 Canecki Varžić, S. 931 Canney, L. 937 Cano, A. 1244 Cano, M. 1097 Canouil, M. 233 Canovatchel, W. 854 Cantley, J. 98 Cao, C. 529 Cao, J. X. C. 971 Capehorn, M. 1030 Cardella, F. 1253 Cardona, R. 636 Cardona, S. 114 Cardona-Alvarado, M. I. 664 Cardozo, A. K. 430 Careaga, M. 1205 Careaga, M. C. 636 Carella, M. 357 Cariou, B. 145, 147, 640 Carli, F. 237 Carlsson, P. Ola. 174 Carlsson, P.-O. 173 Carlsson, S. 173, 174 Carlton, M. 520 Carmody, L. 1102, 1107, 153 Carneiro, E. M. 424 Caroleo, M. C. 137 Caroli, E. 18 Carot, L. 187 Carpinelli, A. R. 469 Carr, B. 142, 548 Carr, J. 1291 Carr, M. 246, 294 Carr, M. C. 830, 831, 838 Carrat, G. R. J. 134 Carreira, M. 438 Carrillo Domínguez, A. 1050 Carrobio, S. 671 Carstensen, B. 1162, 245 Carstensen, M. 760 Carvalho, D. 813 Carvalho, E. 251 Carvalho, F. 592 Casana, E. 432 Casellas, A. 432 Caserotti, P. 63 Casey, S. L. 1292 Casiraghi, F. 924 Caspar-Bauguil, S. 681 Cassidy, S. A. 627 Castañeda, J. 1002 Castaneda, J. 1012, 999 Castañeda, T. R. 557 Castaño, C. 200, 674, 706 Castel, B. 695 Castellino, G. 905 Casteràs, A. 190 Castillo, K. 1258 Castillo Figueroa, A. L. 699 Castro, J. 910 Catargi, B. 991 Caterson, I. D. 777 Caton, P. W. 262 Catrina, A. I.. 250 Catrina, S.-B. 250 Catton, E. 965 Cauchi, S. 233 Caumo, A. 625, 924 Cauvin, V. 1253 Cavallo Perin, P. 1186 Cavalot, F. 210, 305, 86, 913 Cavé, H. 703 Cea Soriano, L. 315 Cechurova, D. 1132 Cederholm, J. 1208, 196, 317 Cefalo, C. 515, 524 Cefalo, C. M. A. 620, 783 Celeng, C. 80 Celik, S. 1111, 307 Cepeda, M. 910 Cepero-Donates, Y. 228, 691 Cerdas, S. 847 Ceriello, A. 1063 Cerny, M. 998 Cersosimo, E. 74 Cescutti, J. 849 Ceuppens, P. R. 673 Cezaretto, A. 1037 Cha, B. 1196, 1230, 220, 873 Cha, B.-Y. 1226, 374 Chadjichristos, C. 767 Chadt, A. 419 Chadwick, P. J. 1147 Chae, H.-Y. 58 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 543 Chaikiat, A. 319 Chaim, É. A. 123 Chakarova, N. 1122 Chakera, A. J. 359 Chalkiadaki, G. 155 Chaloupek, K. 998 Chambard, J.-C. 693 Chambon, P. 1280 Chan, J. Y. 459 Chan, J. 101 Chan, J. 1040 Chan, J. C. N. 1045, 1073, 1182, 279 Chan, S. P. 1093 Chang, A. 994 Chang, C.-J. 582 Chang, P.-Y. 1197 Chang, S.-A. 1298 Chang, X.-X. 277 Chang, X. 863 Chanon, S. 227 Chanu, B. 1236 Chanut-Vogel, C. 901 Charpentier, G. 642, 648, 832, 840 Chartrel, N. 554 Chase, P. 1000 Chatelas, A. 359 Chaturvedi, N. 93 Chatzi, L. 155 Chava, P. 112, 962 Chavez-Valazquez, A. 552 Chavez-Velazquez, A. 722 Chawla, M. S. 883 Chechele, A. C. 595 Chee, C. 291 Chehimi, M. 227 Chen, D. 500 Chen, F. 1327 Chen, H. 638 Chen, J. 165 Chen, J. 274 Chen, K. Y. 63 Chen, L. 349 Chen, M. 328 Chen, R. 723 Chen, S. 665 Chen, Y. 117 Chen, Y. 274 Chen, Z. 274 Chen, Z. 546 Chen, Z. 710, 919 Cheng, K. 885 Cheng, X. 975, 976 Cherney, D. 823 Cherney, D. Z. 956 Cherrington, A. D. 860 Cherubini, V. 1253 Cherviakova, L. 345 Chess, D. 690 Cheung, C. C. H. 1315 Cheung, K. 279 Cheung, Y. 1073 Chianello, P. 1098 Chiba, Y. 1209 Chien, J. Y. 857, 880 Chien, L.-N. 1197 Chiheb, S. 1275, 508, 660 Chillarón, J. José. 1224 Chillarón, J. J. 187 Chilton, R. 1207 Chin, H.-J. 514, 549 China Diabetes and Depression Study Group, 1040 China Initiative for Diabetes Excel- lence (CIDE) Co, l. 1266 Chiou, H.-Y. 1197 Chiou, J.-Y. 1081 Chiquette, E. 1207 Chisari, C. 1142 Chiu, W.-T. 1197 Chlup, R. 995 Cho, D.-H. 1149, 516 Cho, K. 929 Cho, N. H. 308 Choe, G. 572 Choi, D. 295 Choi, H. 1285 Choi, H.-J. 1167 Choi, I. 933, 972 Choi, K. 295 Choi, M. 1290 Choi, M.-G. 1278, 1284 Choi, S.-K. 386 Choi, S. L. 957 Choi, S. 517 Choi, S.-E. 214, 569 Choi, Y.-K. 1284, 591, 867 Choi, Y. 1230 Chon, S. 512 Choong, Y. 1315 Chou, E. 1064, 486, 639 Choudhary, P. 1007, 194, 476, 645 Choukem, S. P. 1227, 512 Chow, C. Chung. 1182 Chow, E. 633 Chow, K.-C. 279 Chow, L. 638 Chow, S. Z. W. 60 Christen, S. 472 Christen, U. 230, 411, 472 Christensen, K. B. 909 Christensen, M. 532, 540, 856, 923 Christensen, T. E. 48 Christiansen, A. V. 806 Christiansen, J. S. 1124, 282, 321, 974 Christiansen, J. 931 Christiansen, M. 994 Chrousos, G. P.. 281 Chu, C. 893 Chu, J. W. 111 Chu, K.-Y. 164, 458 Chu, P.-L. 644 Chua, L. 935 Chubb, B. 489 Chung, C. 1196 Chung, D.-J. 1149, 516 Chung, H. 1073 Chung, H.-W. 486, 639 Chung, J.-O. 1149, 516 Chung, M. 179 Chung, M.-Y. 1149, 516 Chuter, V. H. 1292 Ciavarella, A. 299 Ciborowski, M. 356, 741 Ciccarelli, M. 758 Ciccodicola, A. 353 Cichosz, S. L. 1124 Ciciliot, S. 729 Cicorelli, A. 1145 Cignarelli, A. 299, 464 Cigrovski Berković, M. 481 Cinek, O. 169 Cinek, O. 366, 372 Cinque, A. 465 Cinti, S. 688 Cintra, A. R. S. 616 Cintra, D. E. 616 Ciociaro, D. 237, 74 Cioli, P. 608 Cione, E. 137 Cioni, R. 1145 Cirkel, D. T. 41, 837 Cirmanova, V. 1091 Cissokho, S. 980 Citko, A. 356, 741 Ciudin, A. 190 Claesson, R. 1109 Claret, M. 674 Clark, A. 202 Clark, D. 1030 Clarke, A. 409 Claudius, B. 184 Clausen, T. D. 1114, 1121 Clausen, T. R. 534 Cleal, B. 1048, 1049 Cleary, P. 127 Cleary, P. A.. 128 Cleary, P. A. 92 Cleary, S. 548 Cleasby, M. E. 213, 577 Clee, S. M. 704 Clemmesen, J. O. 178 Cline, G. W. 394 Cluzel, P. 1287 Coant, N. 465 Cobb, B. 213 Cobb, J. 548 Cobb, J. E. 142 Cobelli, C. 1000, 123, 500, 567 Cocco, L. 451 Cochran, E. 864 Codella, R. 625 Codner, E. 332 Coelho, J. C.. 778 Coester, H.-V. 75, 926 Cohen, G. 425 Cohen, O. 1002, 1009, 1012, 981, 999 Cohen, S. 985 Colagiuri, S. 1064 Colby, A. 201 Colclough, K. 359, 367 Coleman, K. J. 156 Coleman, T. 302 Colenbie, S. 763 Colhoun, H. 158 Colhoun, H. M. 11 Colilla, S. 324 Coll, A. P. 520 Collier, A. 993 Colligiani, D. 125 Collins, J. 1139 Collins, R. 274 Collins, S. C. 670 Colom, C. 1224 Colombo, M. 11 Colomo, N. 438 Conde, S. V. 562 Conde, S. V.. 778 Conget, I. 900 Conget, I. 1002, 1009, 1012, 636, 999 Connor, C. 49 Connor, T. 707 Conradi, A. 764, 765 Conte, C. 515, 524, 619, 620, 783 Contreas, G. 1261 Control to Range Study Group, 1000 Conza, D. 758 Cook, W. 4 Cooper, C. 1068 Cooper, G. J. S. 1315, 601 Cooper, J. G. 147 Cooper, M. E. 1170 Copetti, M. 1235, 343 Coppelli, A. 1142, 1145 Corcoran, M. 601 Cormier, J. 983 Cornejo, I. 1108 Cornes, B. K. 344 Corraliza, L. 1163 Corrêa-Giannella, M. 1183 Corredoira González, L. 1035 Corsi, A. 197 Cortes, C. M. 1104 Cosson, E. 1236, 508, 660 Cossu, E. 451 Costa, B. 266 Costa, S. 197 Costa, V. 353 Costacou, T. 1255 Coster, A. 739 Costigan, T. M. 969 Coudenys, E. 288 Coue, M. 32 Coughlan, M. T. 1170 Coupaye, M. 695 Courrèges, J.-P. 1242 Courtade, J. A. 254 Courty, E. 717 Coutant, D. 935 Cox, A. J. 1291 Cox, D. J. 13 Cozzolino, A. 353 Cragg, R. 49 Craike, P. E. 1292 Cramon, P. 1105 Crenier, L. 1004 Cristofaro, M. Rosaria. 1063 Cristofilidis, G. 1027 Cristol, J. 586 Crooke, R. M. 379 Crowe, S. 3, 823 Crown, W. H. 111 Crowther, D. 1231 Cruz-Bravo, M. 910 Csitári, G. 647 Cuddihy, R. 111 Cuggia, M. 1010 Cui, X. 857, 880 Cui, Y. 188 Cuizon, C. 921 Cum, S. 18 Cunha, D. A. 204 Curella, V. 745 Curran, J. 1256 Curry, E. S. 156 Cussac-Pillegand, C. 508, 660 Cyganek, K. 1115, 1120 Cymeryng, C. B.. 657 Cyrille, C. 508 Czeczor, J. K. 707 Czernichow, P. 381 Cziraky, M. J. 111 Czudkova, E. 686 Czupryniak, L. 1159, 950 D Dace, W. 202 Dadan, J. 741 Dadson, P. Y. 597 Dafoulas, G. E. 988 Dagan, S. 453 Dagenais, G. R. 1270 Dagnelie, P. C. 96 Dahl, D. 146 Dahlbäck, B. 408 Dahlhoff, M. 918 Dahl-Jørgensen, K. 1216 Dahlström, U. 44, 45 Daiber, A. 819 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 544 Dain, M.-P. 113, 943 Daka, B. 599 Dakovska, L. 1122 Dalal, M. 1070, 112, 962 Dalantaeva, N. 441 Dalenbäck, J. 576 Dalla Man, C. 123, 500, 567 Dallongeville, J. 901 Dalsgaard, M. 1329, 917 Dalva, M. 364 Damilano, F. 1332 Damm, P. 1105, 1113, 1114, 1117, 1121, 408 Danchin, N. 113, 940 Danesh, J. 65 Daniel, B. 425 Daniele, F. 462, 612 Daniele, G. 552, 74 Danis, D. 369 Danne, T. 1013, 1251, 1257, 964 D’Annunzio, G. 1253 Danse, W. A. 595 Danyliv, A. 1092 Daoudi, A. 648 Da Porto, A. 18 Daraki, V. 155 Dardano, A. 59 da Silva, A. S. R. 616 da Silva Xavier, G. 349 Davalli, A. 462 Davanger, S. 618 Davidson, H. W. 452 Davidson, J. 887 Davidson, J. A. 854 Davies, M. 1064, 181, 182, 76 Davies, M. J. 39 Davis, E. A. 1256 Davis, K. 112, 962 Dayan, C. 290 Dayeh, T. 234, 468 DCCT/EDIC, 127 DCCT/EDIC Research Group, 128 Dea, K. 1054 Deacon, C. F. 217 Deacon, C. F.. 928 de Andrade, M. 158 Dear, A. E. 916 De Beaufort, C. 1043 De Berardis, G. 154, 890 De Block, C. 288 De Bruijn, K. M. J. 247 de Bruijne, M. C.. 1080 de Bruin, T. 118 DeBruin, T. W.. 805 de Castro, J. J. 900 DeCastro, A. 661 De Clerck, E. E. B. 1156 Decochez, K. 288 De Cosmo, S. 343 deCourten, B. 180 Dedov, I. 443 Dedov, I. Ivanovich. 160 Dedov, I. I. 607 Deeb, L. C. 964 Deeg, M. A. 969 Deelen, J. W. T. 22 Deepa, M. 1082 Deetman, P. E. 1232 de Faire, U. 223 De-Franco, E. 359 DeFronzo, R. 1, 181, 39, 722, 76, 851 DeFronzo, R. A. 552, 74, 793 de Geus, E. 139 Degrace, P. 711 de Groot, F. A. 1053 Dejager, S. 1071, 325, 886 Dejgaard, T. F. 1124 de Keyser, T. E. 247 Dekker, G. A. 1099 Dekker, J. M. 1080, 14, 208, 258, 275, 342, 942 Dekker Nitert, M. 375, 468 Dela, F. 584 Delalleau, N. 609 de Lange, M. 1008 de la Peña, A. 857, 880, 965 del Campo, E. 624 Deleruyelle, S. 703 Delfolie, A. 75, 926 Del Forno, E. 18 Del Guerra, S. 59, 656 Delizona, M. 988 de Loredo, L. 37 Delorme, B. 474 Del Prato, S. 1084, 1142, 1145, 59, 656, 848, 85, 888, 891, 894 del Re, L. 997 Del Val, F. 1104 De Marinis, Y. 401 Demeester, S. 288, 452 Demil, N. 829 Demizieux, L. 711 de Moura, E. G. H. 1207 de Moura, L. P. 616 DENG, H. rong. 1123 Deng, H. 287 Deng, W. 863 Denig, P. 16 Denis, R. 767 Dennehy, E. 1026 Denroche, H. C. 33 Denyer, G. 777 Denys, C. 970 de Oliveira Amancio, V. H. 1166 De Pascalis, M. 448 DE-PLAN-CAT, 266 de Portu, S. 1009, 999 De Pouvourville, G. 901 Deprez, B. 609 Derave, W. 180 Derevyanko, O. 441 Derlindati, E. 745 Dermitzakis, E. T.. 30 Dermitzakis, E. 434 Derous, D. 558 Deshmukh, H. A. 90 Desiderio, A. 758 De Simone, G. 451 Desmetz, C. 1317 Desmoulins, L. 670 de Souza, C. T. 616 De Souza, E. 150, 937 Dessapt, C. 1171 Destounis, A. 1240 de Toro-Martin, J. 138, 240 de Toro-Martín, J. 736 Detournay, B. 325 D‘Ettorre, A. 154 Devaraj, S. 1174 de Villiers, E.-M. 1313 Devlieger, R. 1110 de Vries, L. 14 DeVries, J. H. 1019 DeVries, J. 1011 Dex, T. 1070 De Zeeuw, D. 207, 799, 8 Dharmadhikari, G. 749 DIA-AID 1 Study Group, 453 Diabetes Distress and Care Regist- ry at Tenri (DDCR, T. 313 Diabetes in primary care Study group MutuaTerrassa, 628 Diacourse Study Group, 1074 Diamant, M. 139, 241, 81, 912 Diamant, M. 1013 Dias, A. C. 424 Díaz, R. 1270 Di Bartolo, P. 1066 Di Bartolo, V. 905 Dibner, C. 570 Di Cairano, E. S. 462, 612 Di Cianni, G. 1084 Dicker, D. 846 Didangelos, T. 7 Diebold, S. 185 Diéguez, C. 166 Diekmann, U. 390 Diemar, S. S. 1329, 632, 917 Dietrich, A. 662 DiGenio, A. 112, 843, 962 Dik, W. A. 1318 Dimitriadis, G. 1240, 1295, 1300, 561, 564, 565, 94 Dimitriadou, A. 1027 Dimova, R. 1122 Dinccag, N. 1111, 307 Ding, L. 971 Ding, Z.-M. 862 Dinischiotu, A. 1178 Dionne, D. A. 499 Di Paola, R. 357 Dirksen, C. 122, 534 Distelmaier, K. 1263 Di Trani, A. 299 Di Vieste, G. 154 Djanabaev, B. 1204 Dmitriev, I. 480 Dmitrieva, R. 764, 765 Do, H. T. T. 498 Dobias, M. 912 Döcke, S. 765 Doehner, W. 525 Dogan, B. 449 Dogan, H. 655 Dogan, M. 406 Doi, K. 634 Dóka, G. 1239 Doka, G. 1293, 1294 Doktorova, T. 763 Doll, W. 1019 Domenger, C. 1251, 1257 Dømgaard, M. 1042 Domsgen, E. 744 Donaldson, J. 908 Donath, M. Y. 614, 751 Dong, F. 1252 Dong, J. J. 1328 Dong, X. 276 Donker, G. A. 1080 Donkin, I. 235 Donn, R. 351 Donnelly, L. A. 257 Donnelly, R. 814 Donsa, K. 979 Dooley, S. 764, 765 Dor, Y. 97 Doran, J. 520 Doree, D. 862 D‘Oria, R. 1316 Doria, A. 357 Dorkhan, M. 173, 174 Dorothée, G. 498 Do Thi Tu, H. 767 Dotta, F. 231, 29 Dotzauer, A. 744 Doubravova, B. 995 Doudi, A. 642 Down, S. 1030 Doyle III, F. J. 1000 Drake III, D. R.. 970 Drapalova, J. 912 Dreßler, M. 662 Drevon, C. A. 618 Drexel, H. 1211, 1212, 1217, 1218, 1220, 1223, 1238, 1243, 762 Drexler, A. 985 Drion, I. 1219 Driss, F. 1183 Drobni, Z. D. 80 Dropout Study Group, 941 Dror, E. 614, 751 Drury, P. L. 1213 Drynda, R. L. 100 Du, H. 463 Du, H. 274 Duarte, D. A. 1164 Duarte, J. M. N. 677 Dubrick, J. L. 27 Dubský, M. 1146 Dubsky, M. 1148 Duda-Sobczak, A. 1022 Dudele, A. 700 Dudinskaya, E. 95 Dudinskaya, E. N. 1241 Dueñas, B. 624 Duff, C. J. 1088 Duffles, E. 1296 Dufner, B. 748 Duggan, S. 1171 Dugravot, A. 260 Duivenvoorde, L. 558 Dullaart, R. P. F. 1309 Dumitraş, D. E. 1036 Dungan, K. M. 40 Dunger, D. B. 1007, 194, 960, 989 Dunn, G. 1029 Dunn, J. 1054 Dunn, R. 127 Dunn, R. L.. 128 Dunne, F. 1092, 1102, 1107, 153 Dupont, P. 648 Dupuis, A. 586 Dupuis, J. 344 Duraj, L. 43 Duran, J. 417 Durand, C. 227 Durkan, E. 548 Dusatkova, P. 366, 372 Dusaulcy, R. 431 Dusek, J. 372 Duteil, R. 508 Dütting, E. D.. 896 Duvigneau, C. 226 Duvnjak, L. 440, 528 Dvořáková, V. 568 Dvorakova, K. 602 Dvorakova, V. 1169 Dworak, M. 886 Dżygało, K. 1021 Dzebisashvili, T. G. 267 Dzien, A. 1039 Dzien, C. 1039 Dzien-Bischinger, C. 1039 Dzygalo, K. 622, 977 E Eaton, J. N. 839 Ebert, T. 761 Ebner, K.-M. 1211 Echevarriarteun, L. 1222, 1268 Eckel, J. 1319, 692 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 545 Eckhart, A. 142 Economopoulos, T. 1295 Eddicks, M. 918 Edelman, S. 1030 Edelman, S. V. 1020 Edmonds, M. E. 1139, 1151 Edner, M. 44, 45 Edouard, T. 703 Edwall Löfvenborg, J. 173 Eeg-Olofsson, K. 1208, 292, 317 Eekhoff, M. 139 Eekhoff, M. M. W. 339 Egan, A. 1102 Egan, A. M. 1107, 153 EGG Consortium, 106 Egido, J. 544 Egloff, M. 1103 Eguchi, H. 502 Ehrmann, D. 1015, 15 Ehses, J. A. 60 Eid, Y. M. 1289 Eide, I. A. 255 Einwallner, E. 226 Eiriksdottir, G. 63 Eizirik, D. L. 204, 428, 430 Ekelund, M. 1109 Ekström, K. 1329, 917 Ekström, N. 320 Elders, P. 275 Elders, P. J.. M.. 208 Elders, P. J. M. 258, 942 Eleftheriadou, I. 1288 Elias, D. 453 Elias, I. 712 Eliaschewitz, F. G. 846 Eliasson, B. 1208, 196, 292, 306, 317, 320, 336, 754 Eliasson, L. 204, 377, 400, 426 Eliasson, M. 336 Eljaafari, A. 227 Ellard, S. 172, 358, 359, 367 Elleri, D. 1007, 194, 960, 989 Elley, C. R. 1213 Ellingsgaard, H. 60 Ellis, B. B. 938 Elomaa, V.-V. 446 Elorabi, H. A. 1289 Elsen, M. 692 El-Sharkawy, M. 1044 Elsner, M. 455 Elsøe, M. 773 Elverson, R. 142 Elvstam, O. 426 Emanuele, F. 905 Endahl, L. 147 Endresen, K. 256 Engel, S. S. 115 Engel, S. S.. 898 Engelen, L. 93 Engstrøm, T. 1329, 917 Enters-Weijnen, C. 481 Eny, K. 283 Şerban, A. G. 1036 Eren, R. 453 Erichsen, L. 120 Ericson, U. 143, 62, 79 Eriksson, J. W. 576, 605 Eriksson, J. 177 Ernst, B. 124 Ernst, P. 892 Escalona, A. 1207, 982, 983, 984 Eschwège, E. 832, 840 Escrivá, F. 138 Escriva, F. 240 Escrivá, F. 736 Esguerra, J. 467 Esguerra, J. L. S. 204, 400, 426 Esmatjes, E. 1205, 636 Espelage, B. 1302 Espinasse, M. 148 Espinosa Carrasco, G. 412 Esposito, R. 353 Essén, S. 375 Esser, M. 168 Esteban, Y. 674 Esterbauer, H. 226 Esteve, D. 695, 696 Esvant, A. 1010 Eszlinger, M. 157 Eto, M. 1189 Eugen-Olsen, J. 1249 Evans, M. 327, 489, 888, 891 Evans, M. L. 1007, 194, 989 Eyre-Brook, A. I. 1140 F Faber, J. 632 Fabre, O. 586 Fadavi, H. 1155 Fadini, G. 729 Fadista, J. 352 Fadista, J. 401 Faerch, K. 61 Færch, L. A. H. 485 Fagan, N. 956 Fagherazzi, G. 772 Fahlbusch, P. 763 Fahrbach, J. L. 40, 907 Faillie, J.-L. 892 Falco, R. 354 Falcone, V. 299 Falcon-Perez, J. 216 Falk, R. S. 293 Fall, T. 576 Fallucca, S. 448 Famulla, S. 932 Fan, C.-P.Steve. 898 Fan, C.-P. 808, 811 Fan, Y. 385, 705 Fanelli, C. G.. 608 Fares, E.-J. 617 Farley, A. 302 Farmer, A. 302 Farmer, A. J. 1203 Farmer, B. 860 Farr, S. 921 Farr, S. L. 1308 Farran, B. 11 Farret, A. 1019 Farrokhi, F. 1038, 114 Fasold, M. 157 Fasshauer, M. 761 Fatima, S. 89 Fauman, E. 107 Faustman, D. L. 91 Fayfman, M. 1038 Feakins, B. G. 1203 Federici, M. 343 Fedor, M. 43 Fedorova, J. 43 Fedou, C. 586 Feelisch, M. 24 Feenstra, H. 193 Feenstra, T. L. 1080 Feigh, M. 875 Feinle-Bisset, C. 530, 538 Fejfarová, V. 1146 Fejfarova, V. 1148 Felace, G. 1066 Feldt-Rasmussen, B. 909 Fellows, G. F. 27, 28 Fenger, M. 1249 Fenici, P. 850 Feodoroff, M. 1195 Ferdinand, M. 1139 Ferdousi, M. 1155 Fernandez, B. O. 24 Fernandez, E. 1320 Fernandez, L. 827 Fernández, R. 1168 Fernández-Agulló, T. 736 Fernandez-Millan, E. 138, 240 Fernández-Millán, E. 736 Fernandez-Miró, M. 1224 Fernández-Pérez, A. 378 Fernández-Real, J. M. 664 Fernandez-Real, J. M. 683, 684 Fernández-Real, J. M. 687 Fernandez-Real, J. M. 716 Fernández-Ruiz, R. 553 Fernandez-Sanchez, M. 571 Ferrannini, E. 125, 142, 3, 504, 527, 688 Ferrannini, E. 501 Ferrannini, G. 3 Ferraro, M. 864 Ferré, P. 34, 396 Ferreira, P. 434 Ferreira, S. M. 424 Ferreira, S. Roberta. Gouvea. 1037 Ferreira, T. 104 Ferrer, J. 102, 349, 97 Fesatidou, M. 767 Feskens, E. J. M. 1234, 141 Fève, B. 498, 717 Feve, B. 767 Fex, M. 375, 466 Fica, S. 790 Fichelle, M. 1043 Figueiredo, H. A. 699 Fikri, R. 1153 Filion, K. 892 Filipponi, F. 373 Findeisen, H. M. 185 Fineman, M. 793 Finer, S. 236 Fink, K. 646 FinnDiane, 1195 FinnDiane Study Group, 1188 Fioramonti, X. 670 Fioretto, P. 197 Fiquet, B. 1071 Firneisz, G. 444 Fischer, A. 934 Fischer, E. 126 Fisher, P. 1242 Fisk, M. 1139 Fjeld, K. 364 Flacke, F. 987 Flanagan, S. E. 172, 358, 359 Flannery, C. A. 572 Flannick, J. 105, 363 Flatt, P. R. 495, 497 Fleiner, H. F. 439 Fleischer, J. 1124 Fleming, T. 1303, 1304, 215 Fleming, T. H.. 1313 Fleming, T. H. 1314 Flint, A. 882 Floderer, J.-B. 991 Florea, B. 1137 Flores, L. 1205 Flores-Le Roux, J. A. 1224, 187 Floyd, A. 584 Focosi, D. 373 Fogarty, C. L. 1188 Fogelstrand, P. 116 Foghsgaard, S. 1113 Folb, L. 590 Foletto, M. 125, 729 Foley, J. 879, 889 Foley, J. E. 886 Folli, F. 462, 612, 924 Fondelli, C. 210, 305, 86 Fonseca, V. 241 Fontana, A. 1235, 343 Fontanillas, P. 105 Fontserè, S. 1224 Forbes, J. M. 1170, 330 Ford, D. 13 Forsblom, C. 107, 1195, 1299, 51, 87 Forsbloom, C. 1188 Forsén, T. 481 Forst, T. 40, 75, 827, 895, 926 for T2D-GENES and GoT2D Consortia, 105 for the Beta Cell Team of FNIH Biomarkers Consorti, u. 500, 529 for the GDS group, 442 for the German BMBF Compe- tence Net Diabetes Mellit, u. 646 for the PCDIAB consortium, 1011 for the VIVID-DME/VISTA-DME Study Investigators, 54 Foster, G. 184 Foster, N. 939 Foufelle, F. 34 Foussas, S. 1240, 1246 Fousteris, E. 1240 Fousteris, V. 1027 Foutris, A. 1027 Fox, P. 74 FRADYC II Group, 1055 Frahnow, T. 682, 734 Framarino, M. 1098 Franc, S. 642, 648 Francesconi, P. 1062 Frandsen, C. S. S. 632 Franek, E. 931 Frank, S. 252 Franke, B. 993 Franks, P. W. 259 Franks, P. W.. 551 Franks, P. W. 65 Fraty, M. 9 Fraunberger, P. 762 Freathy, R. M. 106 Freckmann, G. 1013, 1017, 785, 997 Frederich, R. 872, 877 Fredheim, S. 1250 Fredriksson, M. 787 Freedman, B. I. 1291 Freemantle, N. 110, 113, 940, 943, 973 Freisinger, O. 124 Freisling, H. 770, 772 French, P. 349 Friberg, A. 474 Friborg, S. 282, 321 Friedlander, G. 381 Frielink, C. 445 Frier, B. M. 487 Frigerio, F. 671 Friis, S. 245 Friis, S. U. 519 Frimpter, J. 968 Fritsch, M. 1296, 232 Fritsch, M. J. 581 Fritsche, A. 721 Fritsch Fredin, M. 673 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 546 Frittitta, L. 1235 Froguel, P. 233 Fröhlich-Reiterer, E. 232 Frois, C. 1054 Frongia, A. 1253 Froy, O. 776 Fryburg, D. A. 500, 529 Frystyk, J. 653 Fu, B. 351 Fu, D. 1118 Fu, H.-J. 1267 Fu, Y. 978 Fuchs, A. 12 Fuchsberger, C. 105 Fuentes, M. 1168 Fuentes-Batllevell, N. 664 Fuentes-Batllevell, N. 687 Fujihara, K. 298 Fujikawa, J. 884 Fujimoto, K. 884 Fujimoto, S. 505 Fujimura, T. 392 Fujioka, K. 966 Fujioka, K. 73, 833, 881, 903, 904 Fujioka, Y. 503 Fujisawa, R. 437 Fujishiro, M. 1277 Fujita, Y. 502 Fujita, Y. 1176, 526, 536 Fukaya, M. 430 Fukuda, M. 634 Fukuda, T. 1175 Fukuhara, A. 559 Fukui, K. 559 Fukui, M. 1175, 1200, 189, 781 Fukushima, A. 626 Fukushima, S. 1194, 925 Fukushima, T. 535 Fukushima, T. 1277 Fulcher, G. 974 Fuller, J. H. 93 Fuller, N. R. 777 Fumeron, F. 1183, 89 Fung, A. 5 Funk, B. 574 Furihata, T. 626 G Gaàl, Z. 1064 Gaens, K. H. 753 Gaggini, M. 237 Gagniuc, P. 340, 436 Gahr, A. 1023 Gaitan, J. 991, 992 Gajdacova, B. 1293, 1294 Gal, A. 985 Galajda, P. 43 Galasso, S. 1019 Galderisi, A. 555 Galdon, A. 1097 Galgani, J. E. 507 Galindo Rubio, M. 1035 Galip, T. 941 Galitzky, J. 695, 696 Galkova, K. 1293, 1294 Gallwitz, B. 842, 894 Galstyan, G. 481 Galstyan, G. Radikovich. 160 Galtier, F. 586 Galvanovskis, J. 421 Gamerman, V. 1030 Gancheva, S. 589 Gand, E. 9 Gandhi, R. 1134, 1136, 131 Ganotopoulou, A. 1027 Gans, R. O. B. 193 Gansevoort, R. T. 8 Gantz, I. 115 Gao, J. 328 Gao, L. 382, 383 Gao, X. 165, 277, 328, 651, 718, 863 Garau, V. 59 Garcia, A. 674 Garcia, A. B. 1055 Garcia, M. 717 Garcia, M.-P. 767 García de la Torre, N. 1168 García-Días, D. 332 Garcia-Fontana, B. 587 García Hernández, J. A. 1101 Garcia-Martin, A. 587 Garcia-Ramírez, M. 53 García Rodríguez, L. A. 315 Garcia-Roves, P. 674 García-Rovés, P. M. 553 García-San Frutos, M. 736 Gardner, D. S. L. 338 Gardner-Sood, P. 309 Garg, S. 1003, 649 Garg, S. K. 1118, 149, 987 Garhyan, P. 957 Gariani, K. 1103 Garieri, M. 30 Garofolo, M. 85 Garrido-Mesa, N. 701 Gartlehner, G. 271 Gärtner, D. 662 Gasa, R. 417 Gaspar, J. M. 563 Gaspari, T. 916 Gasparics, R. 647 Gasperikova, D. 180, 369 Gastaldelli, A. 237, 688, 74 Gates, S. 1095 Gaughran, F. 309 Gault, V. A. 119 Gaulton, K. J. 140 Gause-Nilsson, I. 805 Gautier, J.-F. 1227, 832, 840 Gautier, J.-F. 396, 512 Gayle, C. 1139 Gazis, T. 291, 49 Gedroyc, W. M. 1153 Geelhoed-Duijvestijn, P. 481 Geicu, O. I. 1178 Geiger, K. 762 Geiser, J. S. 857, 880 Gejl, M. 224 Geloneze, B. 123 Gembardt, F. 821 Genders, A. J. 707 Genuth, S. 92 George Fantus, I. 919 Georgescu, O. 790 Georgiou, P. 1044 Georgiou, V. 155 Geransar, P. 900 Gericke, G. 1296 Gerlo, S. 428 German Diabetes Study Group, 249 Gerrits, C. 486 Gerst, F. 404, 405, 435, 768 Gerstein, H. C. 1270 Gertow, K. 223 Getahun, D. 156 Ghanassia, E. 1242 Ghela, T. 67 Ghiadoni, L. 1201 Ghotbi, A. A. 48 Giacca, M. 432 Giaccari, A. 515, 524, 619, 620, 783 Giacco, F. 46 Giacomo, Z. 86 Giani, G. 1258, 589 Giannakakos, H. 988 Giannarelli, R. 1142 Giannini, S. 1228 Gibbs, R. A. 1153 Gibson, M. 351 Gidlund, P. 292 Gieger, C. 144 Giegerich, C. 1231 Giglio, R. Vincenza. 905 Gilbert, R. E. 1164 Gilbert, R. G. 594 Gilewska, M. 1159 Gillard, P. 288 Gillespie, K. M. 450 Gillespie, P. 1092 Gillies, H. 489 Gilon, P. 58 Gimenez, M. 636 Ginsberg, B. 1016 Giorda, C. 1066 Giordano, L. 555 Giorgino, F. 1316, 299, 38, 451, 464 Giorgio, M. 729 Girón, R. María. 506 Giroud, M. 693 Gitt, A. K.. 206, 304 Giuliani, C. 1098 Giussani, P. 465 Giusti, L. 59, 85 Gjesing, A. P. 341 Gjoni, E. 465 Gkiata, P. 988 Glazunova, A. 480 Glechner, A. 271 Glerant, J. 1225 Glindorf, M. 1025, 1065 Gloyn, A. L. 359 Gluud, L. L. 1113 Gluud, L. L.. 856 Gluud, L. 539, 606 Gmyr, V. 609 Gnudi, L. 1171, 66 Go, Y.-H. 1278 Gobalakichenane, P. 717 Göbl, C. S. 1112 Godang, K. 296 Godfrey, L. 1306 Godsland, I. 290 Godzien, J. 741 Goeke, B. 918 Goetze, D. 895 Goetze, J. 911 Gogas Yavuz, D. 941 Gögebakan, Ö. 765 Gogg, S. 685 Goh, S. Y. 338 Göke, R. 900 Göke, R. 896 Golay, A. 501 Goldberg Eliaschewitz, F. 641 Goldenberg, R. 843 Goldfracht, M. 481 Goldstein, B. 69 Goldstein, N. 985 GoLEAD, 158 Golici, L. 739 Golicki, D. 977 Gomes, M. B. 1119, 668 Gomes, P. R. L. 550 Gomez, A. 636 Gómez, R. 1168 Gómez Huelgas, R. 841 Gómez-Peralta, F. 910 Gómez-Ruiz, A. 58 Gomis, R. 417, 674, 699 Goncalvesová, E. 1239 Gonder-Frederick, L. 13 Gong, D. 1315 Gonzalez, A. 504 Gonzalez, C. 562 Gonzalez, N. 504 González, N. 544 Gonzalez Albarrán, O. 1097 Gonzalez Albarran, O. 205 Gonzalez-Carril, F. 1222, 1268 González-Clemente, J. M. 1244 Gonzalez-Franquesa, A. 553 González-Gálvez, G. 853 González González, J. G.. 40 González-Sastre, M. 1244 González-Touceda, D. 166 Gooding, K. M. 1157 Goodnow, C. C.. 229 Goodyer, C. G. 27, 28 Gooley, J. L. 1001 Göransson, O. 377 Gorden, P. 864 Gordin, D. 1299 Gordon, J. 327 Goren, I. 252 Goretti, C. 1145 Görgens, S. W.. 1319 Görgler, N. 135 Gormand, F. 1225 Gorostidi, M. 205 Gorska, M. 1090, 286, 337, 356, 741 Gorus, F. K. 288, 452 Goscik, J. 356 Goto, M. 870 Gotthardt, M. 445, 447 Gottlieb, P. A. 987 Gotzamani-Psarakou, A. 7 Goubert, P. 452 Gough, S. 487, 827, 835 Gough, S. C. L. 78 Goulis, D. G. 1077 Goumas, K. 603 Gourdy, P. 225, 832, 840 Głowińska-Olszewska, B. 195 Goyeau, H. 947 Graaff, R. 283 Grabert, M. 1258 Grabner, M. 111 Grabois, F. 864 Gracheva, S. 480 Graciano, M. 469 Grady, M. 1016 Graham, M. J. 379 Grall, S. 670 Grallert, H. 144, 82 Grandjean, P. 212 Grant, S. F. A. 106, 348 Grarup, N. 103, 106 Gravholt, C. H. 1121 Gray, A. M. 273 Green, L. 645 Greenway, F. 73, 833, 881, 903, 904 Greenwood, K. 309 Gregersen, S. 224 Gregersen, S. 733 Greggio, C. 615, 617 Greig, M. 1134 Grekas, D. 7 Grempler, R. 1276 Gresti, J. 711 Grevstad, U. 923 Grey, S. T. 229, 98 Gribble, F. M. 202, 520, 531 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 547 Grieco, F. A. 204, 428 Griffin, J. L. 24 Griffin, P. 77 Grigoropoulou, P. 1288 Grill, V. 397, 439 Grimaldi, S. 1186 Grinstaff, M. 201 Gritzapis, A. 1295 Groenier, K. H. 1219 Groenier, K. H.. 1232, 1237 Groenier, K. H. 192, 193 Grohl, A. 852 Grønbæk, L. 1221 Grøndahl, M. 540 Gröne, H.-J. 1303 Groop, L. 144, 173, 174, 336, 345, 363, 401, 551, 740 Groop, P.-H. 1188, 1195, 1299, 330, 51, 87 Gross, J. 42 Grosse, J. 520 Groth, A. 12 Grozeva, G. 1122 Gruden, G. 1186, 86 Gruenberger, J.-B. 813 Gruetter, R. 579, 677 Grünberg, J. R. 712 Grunberger, G. 827 Grünler, J. 250 Gu, H. F. 355 Gu, H. 919 Guardiola-Lemaître, B. 263 Guariguata, L. 308 Guarino, M. P. 562 Guarino, M. P.. 778 Gudbjörnsdottir, S. 1208, 196, 245, 306, 317, 318, 320 Gudmundsdottir, V. 139, 339 Gudnason, V. 63 Guerin, J. 1225 Guerrasio, A. 913 Guglielmi, C. 448 GUIDANCE, 944 Guilhem, I. 1010 Guillemain, G. 396 Guillou, H. 681 Guilmin-Crépon, S. 1033 Guinness, M. 150 Guinovart, J. J. 417 Guja, C. 340, 436, 790 Gul, N. 1111 Guldbrand, H. 787 Gullberg, B. 62 Gulliford, M. 66 Gundgaard, J. 487 Günther, A. 256 Guo, F. 1179 Guo, H. 949 Guo, X. 1040, 1056, 1059 Guo, Y. 274 Gupta, R. 139, 339 Gurgul-Convey, E. 433 Gusenbekova, D. G. 927 Gusmão, G. 899 Gustafsson, A. 408 Gustafsson, P. 223 Gutsmann, B. 157, 709 Gutwein, P. 411 Guvener, N. 941 Guy, R. J. C. 1140 Guyton, R. A.. 114 Guzik, T. 361 Guzman, C. 1317, 756 Guzman, S. 984 Guzyk, M. 1165 Găvan, N. A. 1137 Gydesen, S. 84 Gyldenløve, M. 788 Gylfe, E. 420, 610 Gysel, T. 629 Gysemans, C. 428 H Ha, E. 214, 569 Haahr, H. 120, 936 Haaijer-Ruskamp, F. M. 16 Haak, T. 1013, 1015, 1023, 1041, 15 Haapaniemi, E. 172 Haaparanta-Solin, M. 19 Haas, J. 732 Haase, C. L. 789 Haase, T. N. 408 Hach, T. 3, 806, 823 Hadarits, O. 444 Hadjadj, S. 1183, 89, 9 Haering, H.-U. 405, 768 Hagiwara, A. 1194 Hagopian, W. 170 Hague, W. M. 1099 Hahm, J. 1273 Hahn, C. 433 Haixia, S. 667 Hajduch, E. 34 Hakkarainen, A. 679 Hakkarainen, A. 36, 680, 718 Hakozaki, A. 967 Halban, P. 432 Halban, P. A.. 30 Halban, P. A. 431 Halban, P. 434 Halbron, M. 1008 Halkos, M. 114 Hall, E. 468 Hall, K. 820 Hallsworth, K. 627 Halpern, A. 73, 833, 881, 903, 904 Hals, I. 397 Haluzik, M. 912 Haluzík, M. 931 Hamada, I. 902 Hamaguchi, M. 1200 Hamam, A. M. 1289 Hamamoto, Y. 884 Hamasaki, A. 522, 535, 766 Hammad, S. M.. 1118 Hammais, A. 329 Hammar, M. 576, 605 Hammarstedt, A. 712, 754 Hamo-Tchatchouang, E. 660 Hamsten, A. 223 Han, F. 1177, 1297 Han, J.-H. 547, 652 Han, J. 842, 844 Han, J. 1056 Han, K. 858 Han, S. 214, 569 Hanafusa, T. 1274, 437, 611 Hanai, K. 1191 Hanas, R. 1251, 1257 Handgraaf, S. 225 Handsaker, J. C. 1131, 129, 130 Haneda, M. 1176, 526, 536 Hanefeld, M. 841 Hanemaaijer, R. 1324 Hannerz, H. 1049 Hannukainen, J. C. 121 Hannukainen, J. C.. 1283 Hannukainen, J. C. 163, 738 Hansen, A. 60 Hansen, C. P. 540 Hansen, G. 875 Hansen, H. S. 780 Hansen, H. B. 875 Hansen, J. 178 Hansen, K. B. 780 Hansen, L. 1250, 4 Hansen, M. 584 Hansen, M. 532, 533 Hansen, P. M. 909 Hansen, T. W. 1249, 198, 48 Hansen, T. 103, 1114, 339, 341 Hansen, T. K. 1124 Hanssen, K. F. 1118, 1216 Hantel, S. 806 Hanyu, N. 509 Hanyu, O. 298, 630 Hara, K. 1032 Hara, K. 719, 794 Hara, S. 298 Harada, N. 522, 535, 766 Harada, T. 522, 535, 766 Harding, J. 245 Hardouin, J. 1326 Hardy, E. 845 Häring, H.-U. 1304, 215, 404, 435, 665, 689, 714, 721 Harjutsalo, V. 1195, 87 Härkönen, T. 330, 333 Harman, E. 655 Harman-Boehm, I. 985 Harney, F. 153 Harreiter, J. 271 Harris, S. 5 Harris, T. B. 63 Harrison, L. C. 388 Harrison, P. L. 566 Hartemann, A. 1287, 151 Hartmann, A. 255, 256, 296 Hartmann, B. 780, 923 Hartwig, S. 557 Haruki, T. 1194 Hasan, M. 654 Hasbak, P. 48 Haseda, F. 437 Hasegawa, G. 1175, 1200, 189 Hasegawa, O. 1269 Hashimoto, H. 870 Hashimoto, N. 1194, 925 Hashimoto, Y. 1175 Hashinaga, T. 719 Hassan, H. 57 Hassan, Z. 100, 99 Hassoun, A. A. K. 883 Hastings, T. 357 Hasygar, K. 494 Hattersley, A. 510 Hattersley, A. T. 1068, 172, 358, 359, 367, 450 Hattori, A. 1187, 914 Hatziagelaki, E. 564, 565 Hatzitolios, A. 7 Haug, C. 1017, 785 Haupt, A. 646 Häuselmann, S. 614 Hausken, T. 538 Havrdova, T. 479 Hawkins, N. 1060 Hawkins, N. S. 839 Hawthorne, W. J.. 388 Hayashi, T. 1215 Hayashino, Y. 1076, 313 Haycocks, S. 1147 Haynes, A. 1256 Hayter, G. A. 990 Hayward, A. A. 1171 Hazell, L. 890 He, J. 743 He, P. 1187, 868, 88 He, S. 328 He, W. 263 He, X. 212 He, Y. 244 Heald, A. H. 1029, 351 Heaton, N. D. 476 Heckermann, S. 1034 Hedenbro, J. 740 Hedjazifar, S. 712 Heffron, H. 520 Hegedűs, Á. 647 Heianza, Y. 298, 630 Heimann, G. 886 Heine, R. 1026 Heine, R. J. 942 Heinemann, L. 997 Heinke, P. 1051, 388 Heinonen, S. 679, 680 Heise, T. 1011, 120, 932, 934, 935, 936 Heiskanen, K. 172 Hekimsoy, Z. 941 Hekkala, A. 171 Helge, J. Wulff. 584 Heller, S. R. 1007, 194, 487, 633, 878, 989 Hellgren, M. I. 599 Hellstrand, S. 62 Helmke, A. 755 Helmy, N. 1275 Hemi, R. 981 Hendel, J. W. 57 Henderson, A. 384 Hendig, D. 1302 Henkel, J. 583 Henninger, A. M. J. 754 Henriksen, K. 84 Henriksen, T. 1118 Henry, R. R. 242, 77 Henry, R. M. A. 96 Herbach, N. 918 Herder, C. 760 Hermanides, J. 1053 Herman Mahečić, D. 1075 Hermann, J. 939 Hermann, J. M. 1231 Hermanns, N. 1015, 1023, 1028, 1041, 15 Hermansen, K. 733 Hernández, C. 1163, 1244, 190, 53 Hernandez, J. 412 Hernandez-Triana, E. 843 Hero, C. 292 Herpel, E. 1303 Herrera, P. L. 60, 614 Herrero, P. 1044 Herrmann, K. 118 Hersløv, M. L. 487 Hertle, E. 1234, 141 Herzig, S. 1282 Hesse, D. 289 Hesselink, M. K. C. 623 Hiatt, W. R. 987 Hidmark, A. 1314 Hidmark, A. S.. 1313 Hiep Nguyen, H. 828 Hietakangas, V. 494 Hietala, K. 51 Higa, M. 496 Higgins, G. C. 1170 Hildebrandt, T. 97 Hilding, A. 743 Hill, M. 735, 771 Himmelreich, U. 335 Hincelin-Méry, A. 75 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 548 Hincelin-Mèry, A. 926 Hindy, G. 143, 79 Hintermann, E. 230, 411, 472 Hirano, H. 402 Hirano, S. 505 Hirano, T. 493, 870, 871, 915 Hirasawa, A. 522 Hiratsuka, A. 203 Hiromura, M. 871, 915 Hirose, A. 1158, 50 Hirose, H. 1206 Hirose, T. 975, 976 Hirota, Y. 952 Hirsch, I. B. 149 Hirshberg, B. 183, 186, 4, 483, 641, 872, 877, 887, 930 Hirst, J. A. 1203 Hirvonen, J. 678, 83 Hishida, A. 161, 598 Hitman, G. A. 236 Hivert, M.-F. 1094 Hjort, R. 173, 174 Hjuler, S. T. 84 Ho, J. S. K. 1182 Hobbs, T. M. 323 Hocher, B. 1286, 1330 Hodson, D. 349 Hodson, D. J. 102 Hoebaus, C. 162 Hoekstra, J. B. L. 21, 218, 22 Hoekstra, J. B.. L.. 483 Hoekstra, J. B. L. 698 Hoekstra, T. 208, 258, 942 Hoell, B. 979 Hoene, M. 689 Hoevelmann, U. 932 Hoeyem, P. 1124 Hofer, M. 1231 Hofer, S. 939, 958 Hoffmann, B. 589 Hoffmann, J. M. 712 Hofman, A. 247 Hogl, S. 1315 Hohenstein, B. 821 Hoitsma, A. J. 314 Højlund, K. 1121 Holen, T. 618 Holková, K. 169 Holl, R. 646 Holl, R. W. 1231, 1258, 939, 958 Hollander, K. S. 801 Hollander, P. 948 Holleman, F. 1053, 21, 218, 22, 698 Höller, E. 232 Höller, V. 124 Hollingsworth, K. G. 627 Holman, N. 1147 Holman, R. R. 273 Holmerova, I. 602 Holst, J. J. 1114, 122, 519, 532, 533, 534, 539, 540, 57, 606, 780, 909, 922, 923 Holst, J. J.. 928 Holstein, A. 484 Holstein, J. 484 Holt, S. 128 Home, P. 110, 113, 838, 940, 943, 973 Home, P. D. 148, 947, 961, 980 Hompesch, M. 120, 661, 866, 933, 937, 972 Hong, E. 517 Hong, E.-G. 374 Hong, H. 295 Hong, S. Min. 280 Honge, J. L. 1329 Honjo, J. 1176, 526, 536 Honjo, S. 884 Honka, H. 219, 678, 738, 83 Honka, M.-J. 1283 Hopkins, D. 309, 476 Horikawa, C. 630 Horikoshi, M. 106 Horman, K. 985 Hornemann, S. 682, 734, 764 Hornum, M. 909 Horowitz, M. 217, 530, 538, 774 Horsburgh, J. C. 1001 Hörtenhuber, T. 1296 Horvath, E. M. 746 Horvath, T. 80 Hosoe, H. 1032 Hossain, I. Ara. 654 Hosseini, M. 51 Hotaling, J. 127 Hou, N. 1177, 1297 Hou, W. Lin. 791 Houben, A. J. H. 96 Houde, G. 1094 Houssay, S. 1055 Houssier, M. 681 Houweling, S. T. 314 Hövelmann, U. 936 Hovorka, R. 1007, 194, 960, 989 Howard, J. K. 701 Hoyer, A. 1214 Hrabĕ de Angelis, M. 404 Hryniewicka, J. 1090 Hsieh, S. 298 Hsu, E. 91 Hsu, F.-C. 1291 Hu, W. 111 Hu, X. 704 Hu, Y. 1126, 1127 Hu, Y. 916 Huang, C.-N. 1081 Huang, F. 1179 Huang, G. C. 99 Huang, N. 1177, 1297 Huang, R. 1135, 666, 667, 72 Huang, S. 1002 Huang, Y. 675 Huckova, M. 369 Hugo, C. 821, 824 Hugtenburg, J. G. 942 Huh, B. 1230 Huh, K. 1230 Huh, K. 873 Hulmán, A. 61 Hummel, S. 170 Humpert, P. 87 Humphreys, T. 906 Humphries, S. 223 Hung, H.-C. 582 Hunt, B. 17 Huovinen, V. 177 Hur, K. 650 Hur, K. 473, 475 Hurwitz, B. E. 504 Hussain, M. 312 Huster, W. J. 146, 948, 969 Hutri-Kähönen, N. 312 Hutten, M. 1154 Hwang, E. 569 Hwang, J.-H. 581 Hwang, S. 933 Hwang, S. 220 Hwang, Y.-.Cheol. 280 Hwang, Y.-C. 1196 Hyde, C. J. 1068 Hyde, C. L. 107 Hyöty, H. 169 Hyoty, H. 285 Hyöty, H. 333 I Iacobini, C. 1307 Iacopi, E. 1142, 1145 Iafusco, D. 1253 Iancu, S. S. 1036 Ibberson, M. 97 Iberl, M. 202 Ibrahim, L. 1093 Icks, A. 939 Ide, R. 1264 Idiz, C. 1111 Idorn, M. 917 Idorn, T. 909 Idris, I. 814 Idzior-Walus, B. 1310 Iglay, K. 322, 808, 811 Ignaut, D. A.. 965 Ignell, C. 1109 Ihara, Y. 418 Ihionvien, K. 746 Iida, K. 952 Iida, K. 522 Iimuro, S. 1209 Ijzerman, R. G. 81 Ikeda, H. 884 Ikeda, K. 1247 ikeda, s.-I. 578 Ikeda, T. 659, 966 Ikehara, S. 61 Ikonomidis, I. 1300, 94 Ikramova, B. 1204 Ilag, L. L. 146, 948, 969 Ilangumaran, S. 228, 691 Iliadis, F. 7 Ilkavets, I. 764, 765 Ilonen, J. 169, 171, 329, 330 Ilyin, A. V. 607 Im, K. 186 Im, S.-S. 593 Imagawa, A. 1274, 437, 502, 559, 611 Imai, S. 781 Imamura, M. 108 Imbe, H. 1274 Imbernón, M. 674 Immonen, H. 1283 Immonen, H. M. 163 Improta, M. Rosaria. 1063 Ina, K. 1215 Inagaki, N. 522, 535, 766, 812 Inaishi, j. 376 Incalza, M. 1316 in cooperation with ESPED, DPV and the GCNDM, 1258 Indusekhar, R. 1088 Ineson, A. J. 409 Ingenwerth, M. 415 Ingerslev, L. R. 235 Ingersoll, K. 13 Ingvast, S. 474 Ingwersen, S. H. 959 Innocenti, M. 1228 Inoue, H. 750 Inoue, H. 418 Inoue, K. 573 Inoue, N. 573 Inverardi, L. 625 Inzucchi, S. E. 898 Ioacara, S. 790 Ionescu-Tirgoviste, C. 340, 436, 790 Iosup, A. 1242 Iotova, V. 964 Iozzo, P. 121, 163, 177, 219, 738, 74 Iqbal, A. 1007, 194 Iqbal, N. 183, 244, 4, 641, 77, 845, 872 Iqbal, N. 825 Iqbal, S. 236 Iraklianou, S. 1240 Irgens, H. U. 370 Irigoien, I. 1222, 1268 Irribarra, V. 982 Irribarra, V. 983 Irribarra, V. 984 Irshad, Z. 68 Irwin, N. 119 Isajevs, S. 1321 Isherwood, G. 1072 Ishibashi, R. 1187, 868, 88 Ishibashi, S. 663 Ishigaki, Y. 967 Ishiguro, H. 630 Ishii, H. 1076, 313 Ishikado, A. 779 Ishikawa, T. 1187, 868, 88 Ishizaki, R. 422 Ishizaki, T. 422 Ishizu, T. 446 Ishizuka, T. 659, 966 Isken, F. 82 Islam, N. 465 Ismail, K. 309, 316 Isomaa, B. 144 Itaya-Hironaka, A. 392 Ito, C. 272 Ito, H. 1209 Ito, Y. 115 Itoh, H. 376 Ivanova, O. 371, 441 Ivanova, O. N. 362 Iversen, K. K. 1329, 917 Iwahashi, H. 502, 559 Iwai, R. 925 Iwamoto, R. 559 Iwamoto, Y. 272 Iwao, Y. 1206 Iwasaki, K. 522, 535, 766 Iwasaki, N. 1264 Iwasaki, Y. 834 Iwase, H. 1200 Iwashita, M. 1277 Iwata, S. 794 Izawa, S. 503 Izumi, T. 422 J Jaber, Y. 508 Jackson, I. 701 Jacobs, J. 1043 Jacobs, N. 1014 Jacobs, S. 270 Jacobs, S. 114 Jacobsen, L. V. 959 Jacobsen, S. H. 122 Jacobson, A. M. 128 Jacobs-Tulleneers-Thevissen, D. 253 Jacqueminet, S. 1287, 151 Jacquin, P. 1033 Jae Hyeon, K. 1130 Jaen, A. 628 Jäger, C. 748 Jager, K. J. 314 Jager-Wittenaar, H. A. 795 Jais, A. 226 Jakovljevic, D. G. 627 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 549 Jakubowicz, D. 776 James, J. E. 488 James, R. W. 1309 Jamieson, P. M. 577 Jamiołkowksa, M. 195 Jamiołkowska, I. 195 Jan, E. 458 Janas, I. 1115 Jang, H. 1196 Jang, J. 1290 Jansa, M. 636 Jansà Morató, M. 1035 Jansen, A. P. D. 258 Jansen, A. P.. D.. 208 Jansen, D. 942 Janssen, M. J. R. 447 Janssens, W. 615 Jansson, P. 352 Jansson, P.-A. 116, 599 Januszewski, A. 1233 Jarlov, H. 836 Jarzebska, N. 821 Jax, T. 1011 Jayawardena, R. 265 Jazbec, A. 1281 Jeandel, L. 554 Jeandidier, N. 640, 961 Jehle, A. W. 751 Jelenik, T. 1172, 557, 581 Jellimann, S. 1043 Jelok, I. 167 Jelsing, J. 519 Jendle, J. 42 Jenkins, A. 1233 Jenkins, A. J. 1118 Jenndahl, L. E. 754 Jennersjö, P. 787 Jennison, C. 257 Jennum, P. 644 Jensen, C. B. 181, 184, 39, 73, 76, 833, 881, 903, 904 Jensen, D. M. 1121 Jensen, J. E. 122 Jensen, J. 618 Jensen, L. 120 Jensen, M. T. 922 Jensen, T. 909 Jenssen, T. 255, 256 Jenssen, T. 296 Jenum, A. Karen. 293 Jeon, H.-J. 1167 Jeon, J. 214, 569 Jeon, W. Seon. 280 Jeon, W. 386 Jeon, W. 297 Jeon, Y. 1193 Jeong, E.-A. 1087 Jeong, H. 569 Jeong, J.-G. 569 Jeong, S. 391 Jeppesen, J. 1249 Jeppesen, P. B. 773 Jeppesen, P. B.. 784 Jepsen, S. L. 57 Jermendy, A. L. 80 Jermendy, G. 483, 80, 930 Jeter-Jones, S. 862 Ji, L. 1040 Jia, W. Ping. 791 Jia, W. 1040, 1126, 1127, 483, 863 Jialal, I. 1174 Jiang, D. 1026 Jiang, G. 1182 Jiang, H. 42 Jiang, X. 723 Jiao, Y. 188 Ji-Eun, J. 1130 Jimenez, V. 432 Jiménez Ortega, A. 1050 Jin, H. 1184 Jin, S.-.Man. 280 Jin, S.-M. 473, 475, 650 Jin, T. 919 Jing, D. Qing. 410 Jirkovská, A. 1146 Jirkovska, A. 1148 Jodar, E. 853 Johal, S. 813 Johannesen, J. 1250 Johannson, B. B. 364 Johannsson, E. 63 Johansen, O. 956 Johansen, P. 817 Johansen, T. 974 Johansson, B. B. 370 Johansson, I. 44, 45 Johansson, P. 822 Johansson, S. 315 Johansson, S. 363, 370 Johne, C. 1312, 727 Johnsen, N. F. 268 Johnsen, S. P. 789 Johnson, J. D.. 385 Johnson, J. D. 458, 499, 704 Johnson, J. D.. 705 Johnson, J. D. 715 Johnson, N. A. 1292 Johnsson, E. 798, 800, 804, 805, 807, 822, 848, 852 Johnsson, K. M. 800, 804 Johnston, B. T. 348 Johnston, D. G. 367 Johnston, D. 1044 Johnston, D. G. 290 Johnston, N. R. 102 Joner, G. 370 Jones, H. T. 566 Jones, J. 592 Jones, J. G.. 166 Jones, K. L. 217, 530, 538, 774 Jones, P. M. 100, 406, 407, 99 Jones, T. W. 1256 Jonsson, A. 339, 341 Jonsson, P. V. 63 Joo, E. 522, 535, 766 Joosten, L. 445 Jørgensen, J. 757 Jørgensen, M. E. 103, 1162 Jørgensen, M. B. 909 Jørgensen, N. 235 Jørgensen, N. R. 644 Jorgensen, N. B. 122 Jørgensen, N. B. 534 Jørgensen, P. G. 922 Jornayvaz, F. R. 1103 Jörns, A. 414 Jorsal, T. 57 Jose, B. 906 Joshi, P. 106 Jothydev, S. 310 Jotic, A. 331 Joubert, M. 1326 Joung, K. 1248 Jover, B. 1317 Ju, C. 1056, 1057 Ju, S. 188 Jude, E. 1150, 993 Jugnee, N. 1044 Juhl, C. B. 632 Jukarainen, J. S. 679 Jukarainen, S. J. S. 697 Jun, H.-S. 737 Jun, J. 475, 650 Jung, C. Hee. 280 Jung, C. 1290 Jung, E. 1284, 591, 867 Jung, E.-D. 238 Jung, G.-S. 1284, 867 Jung, H. 1270 Jung, J. 1273 Jung, J. 1273 Jung, S. 866, 933, 972 Jung, Y.-A. 867 Junker, A. E. 539, 606, 788 Junmei, Y. 1045 Juntti-Berggren, L. 379 Jurczak, M. 861 K Ka, S. 35, 416 Kabeya, Y. 191 Kaden, S. 1303 Kadoguchi, T. 626 Kadono, M. 189 Kadowaki, T. 1032, 272 Kaestner, K. H. 424 Kahl, S. 442, 592 Kahle, M. 109 Kahleova, H. 735, 771 Kahles, F. 185 Kahn, R. C.. 1332 Kähönen, M. 312 Kaiser, G. 404, 405, 435 Kaji, M. 300 Kajikawa, M. 952 Kajita, K. 659, 966 Kajiyama, S. 781 Kakehi, S. 578 Kakino, S. 719 Kaňková, K. 568 Kaksonen, R. 680 Kaku, K. 812 Kalaca, S. 307 Kalaiyarasi, G. 1082 Kalic, R. 1256 Kalousova, M. 1305 Kalsekar, I. 828 Kalvinsh, I. 1321 Kalyva, E. 1262 Kamata, H. 1277 Kameda, H. 929 Kamenov, Z. 1150 Kaminski, M. 433 Kaminski-Hartenthaler, A. 271 Kampe, K. 751 Kampouri, M. 155 Kanazawa, I. 1229 Kaneko, R. 1247 Kaneko, S. 834 Kaneko, S. 752 Kanety, H. 981 Kang, E. 1230, 220, 873 Kang, J. 866, 972 Kang, Y. 1285 Kang, Y. 214, 569 Kangas, A. J.. 1283 Kankova, K. 1169 Kannan, A. 1323 Kanno, A. 454 Kantartzis, K. 721 Kantor, C. 97 Kantor, F. R. 235 Kapantais, E. 782 Kapitza, C. 75, 926, 953, 959 Kaplan, L. M. 982 Kaprio, J. 36, 679, 680, 697 Kapur, R. 481 Kapusta, M. 360 Kapustin, J. 1070 Kara, I. 575 Karabulut, Ü. 1046 Karaca, M. 671 Karachaliou, M. 155 Karagianni, O. 603 Karagiannis, T. 1077 Karajić, A. 992 Karakaya, P. 941 Karalliedde, J. 66 Kararic, M. 333 Karczewska-Kupczewska, M. 176, 211 Karlsson, H. K. 678, 738, 83 Karpe, F. 600 Karsdal, M. A.. 84 Karsidag, K. 307 Kartal, I. 449 Kaspers, S. 823, 956 Kassanos, D. 1096 Kaste, R. 1, 851 Kasteleyn, M. J. 1074 Kaštelan, S. 1161 Kato, H. 427 Kato, H. 876 Kato, K. 191 Kato, M. 503 Katoh, S. 300 Katra, B. 1115, 1120 Katsilambros, N. 1288 Katsogiannos, P. 605 Katsouri, V. 1027 Katsuki, T. 191 Kattan, M. W. 323 Katulanda, P. 265 Katz, L. B. 1016 Kaufman, L. 452 Kauhanen, S. 219 Kaul, K. 1172, 168 Kaur, A. 290 Kautzky, A. 271 Kautzky-Willer, A. 1112, 271 Kawabe, H. 1206 Kawahara, M. 1325 Kawalec, J. S. 1141 Kawamori, R. 578 Kawamoto, T. 454 Kawamura, H. 1187, 868, 914 Kawamura, T. 964 Kawasaki, S. 1269 Kawashima, M. 966 Kawazu, S. 209 Kayal, A. 1082 Kaye, S. 36, 697 Kazazian, C. 767 Kazdova, L. 71, 735, 771 Kazemi, M. 1070 Ke, X. 111 Keech, A. 1233 Keers, J. C. 1031 Keijer, J. 558 Keller, M. 662 Kelly, D. M. 566 Kelstrup, L. 1114 Kema, I. P. 192 Kemper, M. 764, 765 Kempf, C. 901 Kempler, P. 647 Kendall, D. 1026 Kenealy, T. 1213 Křenek, P. 1239 Kenichi, S. 88 Kenna, L. M. 1089 Kennedy, A. 548 Kennedy, J. 1315 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 550 Kenny, H. 548 Kere, J. 172 Kerenyi, Z. 1085 Kern, M. 1276, 709 Kerner, W. 958 Kerr-Conte, J. 429, 609, 744 Kesavadev, J. 310 Keskimäki, I. 245 Kessler, B. 918 Keyhani Nejad, F. 82 Keyhani-Nejad, F. 764 Keymeulen, B. 288, 452 Kgosidialwa, O. 1102, 1107 Khalimon, N. 345 Khatami, H. 968, 970 Khin, M. O. 1095 Khunti, K. 481, 885 Kibbey, R. G. 394 Kido, Y. 418, 454 Kiec-Wilk, B. 360, 368 Kieffer, T. J. 33 Kiens, B. 922 Kieswich, J. 262 Kieć-Wilk, B. 1006 Kikuchi, M. 209 Kilpeläinen, T. O. 344 Kim, B. 1193 Kim, C.-H. 64 Kim, D. 933 Kim, D. 214, 569 Kim, D. 1207 Kim, D.-J. 1192 Kim, E. 1284, 591, 867 Kim, E.-H. 238 Kim, E. 652 Kim, E.-H. 64 Kim, E.-S. 547 Kim, G. 2, 806, 818 Kim, G.-H. 1087, 460 Kim, H. 214 Kim, H. 295 Kim, H.-K. 64 Kim, H. 591 Kim, H. 1248 Kim, I. 1193 Kim, J. 473, 475, 650 Kim, J. 549 Kim, J.-H. 1284, 1298 Kim, J.-Y. 1087, 460 Kim, J. 933 Kim, J. 1193 Kim, J. 280 Kim, J. 866, 972 Kim, J. 1192 Kim, J.-H. 386 Kim, K. Joon. 280 Kim, K. 853 Kim, K.-J. 517 Kim, M.-K. 1226 Kim, M.-J. 1284, 238, 591, 867 Kim, M.-K. 1284, 238, 591, 867 Kim, M. Jae. 179 Kim, M. 873 Kim, N. 295 Kim, N. 295 Kim, S. 1193 Kim, S. 295 Kim, S.-C. 391 Kim, S. 1273 Kim, S.-K. 1230 Kim, S. 547 Kim, S. 295 Kim, S. 47 Kim, S.-W. 1278 Kim, T. 214 Kim, T. 295 Kim, T. 212 Kim, T. 793 Kim, W. 1193 Kim, W.-H. 1087, 460 Kim, Y.-D. 593 Kim, Y. Bum. 179 Kim, Y. 866, 933, 972 Kim, Y.-B. 563 Kim, Y. 873 Kimachi, K. 929 Kimura, A. 402 Kimura, T. 422 Kimura, T. 1175 Kinduryte, O. 964 King, A. 243 Kinoshita, H. 427 Kinugawa, S. 626 Kinz, E. 1243 Kirchsteiger, H. 997 Kirwan, B. 1102, 1107 Kirwin, B. 153 Kiselyov, V. V. 971 Kishimoto, I. 598 Kishimoto, I. 161 Kishimoto, J. 272 Kiss, A. 488 Kiss, L. 746 Kitada, Y. 659 Kitada, Y. 966 Kitagawa, N. 1175 Kitamoto, T. 542 Kitamura, T. 659 Kitano, S. 50 Kitao, N. 470 Kitazato, H. 272 Kitsunai, H. 526 Kivimaki, M. 61 Kiviranta, H. 284 Kjaer, A. 48 Kjær, T. N. 757 Kjær, T. W. 632 Klarlund-Pedersen, B. 352 Kleefstra, N. 1232, 1237, 192, 193, 314 Kleefstra, N. 1219 Kleemann, R. 264 Klein, E. 77 Klein, T. 1276, 1286, 1330, 875 Klimas, J. 1239 Klimas, J. 1293, 1294 Klimes, I. 180, 369 Klingensmith, G. 1252 Klingensmith, G. J. 964 Klingler, C. 665 Klip, A. 759 Klonoff, D. 1003, 649 Kloos, C. 1024, 301, 945 Kloos, C. S. 958 Klös, M. 662 Klöting, N. 1276, 709, 765 Klupa, T. 1006, 360, 368, 951 Klusmeier, N. 1302 Knabbe, C. 1302 Knadler, M. Pat. 957 Knadler, M. 938 Knap, D. 883 Knežević-Ćuća, J. 440, 528 Knight, B. 1068 Knip, M. 169, 171, 284, 329, 330, 333 Knop, F. 856 Knop, F. K. 1113, 519, 532, 533, 539, 540, 541, 57, 606, 788, 909, 922, 923 Knorr, S. 1121 Ko, K. 1192 Ko, S.-H. 10 Kobayashi, K. 1187, 868, 914 Kobayashi, M. 509 Kobe, M. 815 Koblik, T. 360, 368 Koch, A. 185 Koch, C. 206, 304, 895 Koch, F. 1312, 728 Kodama, K. 1181 Kodama, S. 298, 630 Kodama, T. 663 Koezuka, R. 598 Koezuka, R. 161 Koffert, J. 738 Koffert, J. P. 219 Kogevinas, M. 155 Kogot-Levin, A. 658, 725 Koh, E. 1290 Koh, G. 514, 549 Kohashi, K. 915 Koivisto, V. 1299 Kok, F. J. 775 Kokkinos, A. 1288 Kolahdooz, F. 1037 Koliaki, C. 168, 442, 581 Kollár, R. 647 Kolonel, L. N. 270 Kolouskova, S. 366 Kolterman, O. 118 Komada, H. 952 Komajda, M. 1287 Kondo, K. 779 Kondo, M. 724, 779 Kondo, T. 1265, 874, 929 Kondo, Y. 1269 Kone, M. 97 Kong, A. 1040, 1045 Kong, A. P. S. 279 Kong, S. X. 323 Königsrainer, A. 721 Königsrainer, I. 721 Konopka, A. 1263 Konrad, R. J. 969 Konstantinova, E. E. 311 Konvalina, N. 981 Kooi, M. E. 623 Kooijman, M. N. 106 Kooistra, T. 264 Koopman, A. D. 275 Koopman, A. D. M. 326 Kooy, A. 792, 795 Škop, V. 71 Kopecky, P. 912 Kopf, S. 1314 Koponen, J. 284 Koppensteiner, R. 162 Koranyi, L. 604 Kornelius, E. 1081 Korneliussen, T. S. 103 Körner, A. 157 Korsgren, O. 474 Korsholm, L. 78, 835 Koseki, H. 767 Koshiyama, H. 884 Koskinen, S. 177 Kostek, O. 449 Koster, A. 63, 96 Kostev, K. 796 Kotalova, R. 372 Kotani, K. 635 Kothny, W. 816, 879, 888, 889, 891, 893, 902 Kothonas, F. 603 Kottink, A. 1154 Kotwica, A. O. 24 Kotzka, J. 1172 Kou, K. 376 Kou, T. 324 Koukoui, F. 648 Koukoulis, C. 94 Koukoulis, G. 988 Kounaki, S. 299 Kousathana, F. 1300, 94 Koutra, K. 155 Koutsovasilis, A. 1246 Kovacs, P. 144, 157, 484, 662, 709 Kovar, F. 43 Kovatchev, B. 843 Kovatchev, B. 1000, 13 Kowall, B. 1214, 760, 796 Kowalska, A. 622, 977, 986 Koyama, M. 975, 976 Koyanagi, M. 454 Kozawa, J. 502 Kozuka, C. 496 Kracmerova, J. 686 Kragl, M. 419 Kralisch, S. 761 Kramer, A. 314 Kramer, G. 1024, 945, 958 Kramná, L. 169 Krarup, N. T. 103 Krasner, A. 150, 937 Krause, K. 144 Krauß, C. 785 Kreiner-Møller, E. 106 Kreis, R. 615 Krejci, H. 1091 Krejci, J. 995 Krejci, R. 479 Krempf, M. 73, 833, 881, 903, 904 Krenek, P. 1293, 1294 Krentz, A. 661 Krentz, N. A. J. 25 Kretowski, A. 1090, 356, 741 Škrha, J. 1305 Škrha jr., J. 1305 Krieger, K. 126 Krijnen, W. P. 795 Krischer, J. 170 Krishnan, G. 310 Kristensen, H. L.. 784 Kristiansen, V. B. 122, 519, 534 Kröger, J. 270 Kröller-Schön, S. 819 Kröner, C. 785 Kronshage, B. 932 Kroon, E. 253 Kropff, J. 1019 Krüger, J. 709 Kruljac, I. 1075 Krus, U. 375 Kruse, M. 682, 734, 82 Ku, B. 1248 Kubisz, P. 43 Kubo, M. 108 Kuchibhotla, S. 251 Kuchmerovska, T. 1165 Kudo, M. 920 Kudomi, N. 163 Kuehl, M. N. 595 Kuhn, A. 992 Kulkarni, K. D. 1052 Kulkarni, R. N. 515, 524 Kullo, I. J. 158 Kulzer, B. 1015, 1023 Kulzer, B. 1041 Kumar, P. 877 Kumar, S. 145 Kumar, V. 1304 Kume, S. 724 Kun, A. 1085 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 551 Kuniss, N. 1024, 945 Kuno, K. 1187, 868 Kuo, M. S. 11 Kuraeva, T. 371 Kuraeva, T. L. 362 Kurbasic, A. 259 Kurdiova, T. 167, 180 Kuricova, K. 1169 Kuricová, K. 568 Kurihara, Y. 874 Kurioka, S. 869 Kurita, Y. 719 Kuroda, S. 812 Kuroda, Y. 559 Kushiyama, A. 1277 Kushner, R. 181, 39, 76 Kusmartseva, I. 285 Kuzmicki, M. 1090 Kvist, K. 243, 836 Kwon, H. 47 Kwon, H.-S. 1226, 374 Kwon, S. 866, 933, 972 Kyriklaki, A. 155 Kyselovič, J. 1239 Kyselovicova, O. 167 L Laakso, M. 501 Labarbuta, R. 464 Labriola, L. 461 Labudzynskyi, D. 1279 Lacaria, E. 1084 Lachin, J. M. 448, 92 Lacigova, S. 1132 Lacraz, G. 228, 691 Ladefoged, M. 60 Ladent, T. 767 Ladenvall, C. 107, 144, 90 Ladman, G. W.. D.. 1237 Laffel, L. 1251, 1257 Lahav, M. 981 Lahtela, J. 37 Lai, K. B. 58 Laiho, J. E. 285 Laine, A.-P. 329 Laine, J. 19 Laires, P. 808, 811 Lajer, M. 90, 911 Lajoix, A.-D. 1317, 756 Lalic, K. 331 Lalic, N. 331, 481, 501 Lally, J. 309 Lam, E. 935, 957 Lam, V. K. L. 1182 Lamacchia, O. 210, 86 Lambadiari, V. 561, 564, 565 Lambadiari, V. A. 1300 Lambadiari, V. A.. 94 Lambers Heerspink, H. J. 207 Lambers Heerspink, H. J.. 8 Lambert, K. 586 Lammert, E. 419 Lamola, G. 1142 Lampasona, V. 334 Lamprinou, A. 1096 Lams, A. 474 Lancaster, J. 74 Landau, Z. 776 Landini, L. 121 Landin-Olsson, M. 1100, 152, 196 Landis, G. 782 Landman, G. W. D. 1219 Landman, G. W.. D.. 1232 Lang, A. 981 Lang, J. 991, 992 Langbakke, I. 835 Langbakke, I. H. 243 Lange, C. 389 Lange, K. H. W. 632 Lange, K. 530 Langer, T. 1280 Langholz, E. 519 Langin, D. 32, 681, 693, 730 Langlois, M.-F. 691 Langrand, C. 1225 Langslet, G. 854 Länne, T. 787 Lanzoni, G. 625 Lapuerta, P. 803 Larkin, M. L.. 128 LaRosa, S. 612 Laroye, H. 642, 648 Larsen, S. 540 Larsson, C. A. 599 Larsson, M. H. 673 Larsson, N.-G. 133, 1332 Lasch, S. 230, 411 Lassenius, M. I. 1188 Lasseter, K. C. 882 Lau, D. C. W. 73, 833, 881, 903, 904 Laubner, K. 1096, 747, 748 Laugesen, E. 1124 Launer, L. J. 63 Laurens, C. 730 Laurent, D. 661 Laurila, E. 144 Laustsen, C. 773 Lavalle, F. 1064 Lavebratt, C. 743 Lavigne, S. 640 Laviola, L. 1316, 299, 305, 464 Law, G. 799 Lawton, K. 142 Laybutt, D. Ross. 101 Laybutt, D. R. 459 Leal, A. M. O. 471 Leal, E. C. 251 Leal, J. 273 Lebherz, C. 185 Lebl, J. 366, 372 Lebreton, F. 991, 992 Lecce, M. 758 Leclerc, I. 199 Lecube, A. 190 Ledent, T. 717 Ledoux, S. 695 Lee, B.-.Wan. 280 Lee, B.-W. 1230, 220 Lee, C. 1193 Lee, C. 737 Lee, D.-H. 514, 549 Lee, D.-Y. 1087, 460 Lee, D. S. 500, 529 Lee, E. 1230 Lee, E. 1230, 220, 521 Lee, E.-Y. 203 Lee, H. Man. 1182 Lee, H. 1285 Lee, H. 1230, 220, 873 Lee, I.-K. 1278, 1284, 1290, 591, 867 Lee, J. D. 1079 Lee, J. 715 Lee, J.-E. 591, 867 Lee, J.-H. 547 Lee, J. 1030 Lee, J. 866 Lee, J. 1248 Lee, J.-M. 1298 Lee, K.-U. 1278, 1284, 1290 Lee, K. 858 Lee, K.-W. 1196, 214, 569 Lee, K.-M. 386 Lee, L. 388 Lee, M. 1290 Lee, M. 297 Lee, M.-K. 1196, 473, 475, 650, 858 Lee, M.-S. 473, 475, 650 Lee, S. 972 Lee, S.-A. 569 Lee, S.-A. 514, 549 Lee, S. 1002, 999 Lee, S. 47 Lee, S.-H. 1226, 179 Lee, S. 618 Lee, S. 391 Lee, S. 1230 Lee, W. 858 Lee, W.-Y. 297 Lee, W. 1290 Lee, Y. 104 Lee, Y.-H. 1167 Lee, Y.-H. 220 Lee, Y. 873 Lee, Y.-J. 1087, 460 Lee, Y.-M. 866, 933 Lee, Y.-S. 737 Leelarathna, L. 1007, 194, 989 Leenders, F. 421 Lefai, E. 227, 570 Lefebvre, H. 554 Le Gonidec, S. 703 Leguerrier, A.-M. 1010 Lehert, P. 792, 795 Lehmann, R. 665, 689 Lehr, S. 557 Lehrke, M. 185 Lehtimäki, T. 312 Lehtiniemi, P. 446 Lehto, M. 1188 Lei, H. 579 Leibovitz, E. 1069, 637 Leibowitz, G. 183, 887, 930 Leiherer, A. 762 Leiss, V. 714 Leiter, L. A. 145 Leiter, L. A.. 805 Leiter, L. A. 841, 847, 877 Lekakis, J. 1300, 94 Leloire, A. 233 Leloup, C. 670 Le Marchand, L. 270 Lemmers, R. F. H. 1318 Lemmi, P. 1084 Lempainen, J. 329 Lencioni, C. 1084 Lenzen, S. 390, 414, 433, 455, 456 Leonardini, A. 1316, 464 Lepicard, E. 381 Lepore, G. 197 Lepore, V. 154 Lernmark, A. 172 le Roux, C. 73, 833, 881, 903, 904 Le Stunff, H. 465 Leto, G. 451 Letra, L. 261 Leuridan, L. 1110 Levin, P. 1058 Levin, P. A. 111 Levine, A. 982, 983 Levy, B. 1008 Lévy, M. 1008 Levy, S. 370 Levy-Marchal, C. 1033 Lewin, A. 1, 181, 39, 76, 851 Lewis, B. S.. 641 Lewis-D’Agostino, D. 894 Li, A. M. 279 Li, G. 26 Li, H. 276 Li, J. 420, 610, 665, 689 Li, J. 27, 28 Li, K. 543, 588 Li, L. 1127 Li, L. 274 Li, L. 239, 543, 588, 702, 723, 769, 816 Li, M. 150 Li, M. M. 601 Li, M. 769 Li, Q. 791 Li, S. 594 Li, W. 1040 Li, X. 651 Li, X. 319 Li, X. 1056 Li, Y. 4 Li, Y. 273 Li, Y.-L. 1197 Li, Y. 618 Li, Y. E. 759 Li, Z. 142 Li, Z. 808, 811, 898 Li, Z. 710 Liabeuf, S. 1287 Liakos, A. 1077 Liang, H. 710, 919 Liang, S. 335 Liang, W. 264 Liao, A. 1207 Liao, B. Mana. 720 Liao, L. 1328 Liaskos, C. 1288, 631 Liatis, S. 631 Lieberman, N. 981 Liehua, L. 403 Liesa, M. 175, 690 Lievens, S. 428 Lieverse, A. G. 1318 Lièvre, M. 640 Liew, A. 303 Lillleøre, S. K. 73, 903, 904 Lilly, L. 1005 Lim, C. 972 Lim, G. E. 715 Lim, J. 47 Lim, L. Ling. 1093 Lim, W. 1099 Lima, I. S. 179, 563 Lima, M. M. O. 123 Lin, B. 710, 919 Lin, H.-D. 277 Lin, H. 863 Lin, J. 1058, 843 Lin, S. da. 1123 Lin, S. 287 Lin, Y.-P. 1210 Lin, Y.-F. 1197 Lindahl, J. P. H. 255, 256 Lindblad, U. 599 Lindblom, A. 673 Lindegaard, M. 882 Lindi, V. 106 Lindner, J. 912 Lindqvist, A. 740 Lindtner, C. 669 Ling, C. 234, 352 Ling, C. A. 468 Ling, D. 1054 Ling, R. 1157 Ling, Y. 651 Ling, Z. 253 Lingohr-Smith, M. 1058 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 552 Lingvay, I. 739 Linjawi, S. 78 Link, M. 1017 Linker, N. J. 1198 Linn, T. 453 Linneberg, A. 103 Linnebjerg, H. 935, 957 Linnenkamp, U. 308 Liosi, V. 1027 Lipar, K. 479 Lipinska, D. 286, 337, 356 Lippmann-Grob, B. 1008 Lipponen, H. 177 Lips, M. 912 Lipsky, B. 1152 Lisakovskaya, O. 1279 List, J. 822 List, J. F. 6 List, J. F.. 798 List, J. F. 800, 804, 825 Literati-Nagy, B. 604 Little, T. J. 530 Littlefield, L. 212 Littley, M. 1116 Litwak, L. E. 183 Liu, B. 100, 99 Liu, C. 747, 748 Liu, D. 1, 851 Liu, F. 1126, 1127, 791 Liu, H. 1315 Liu, J. 897 Liu, L. 287 Liu, M. 410 Liu, X. 170 Liu, X. 347 Liu, X. 1177, 1297 Liu, Y. 410 Livingstone, R. 581 Livne, A. 425 Li Volsi, P. 1063 Lizárraga-Mollinedo, E. 138 Lizarraga-Mollinedo, E. 240 Lizárraga-Mollinedo, E. 736 Ljubić, S. 1161 Ljubic, S. 1281 Ljung, R. 245 Llabre, M. M. 504 Llauradó, G. 1244 Llorente, M. 910 Loba, J. 1159, 950 Lobbens, S. 233 Lockie, S. H. 672 Lodererova, A. 479 Lodwig, V. 997 LoGerfo, F. W. 251 Loghin, C. 857, 880, 957 Logtenberg, S. J. J. 192, 193 Logtenberg, S. J.. J.. 314 Lohmann, T. 662 Loi, V. 603 Long, A. E. 450 Longo, M. 354 Looker, H. C. 11 Loos, R. J. F. 344 Lopatina, T. 52 Lopes de Faria, J. Mendonça. 1164 Lopes de Faria, J. Butori. 1164, 1166 López, J. 1104 López Alonso, A. 1101 Lopez Gonzalez, E. 1055 López Plasencia, Y. 1050 Lopez-Sendon, J. 186, 483, 930 Lord, G. M. 701 Lortz, S. 455, 456 Lou, Q. 1056 Louchami, K. 335 Louche, K. 32, 730 Loudovaris, T. 388 Loureiro, A. P. M. 1185 Lovász, R. 1254 Lowe, J. 1068 Lowe, J. M. 488 Lowe, R. 236 Lu, F. 1127 Lu, H. 919 Lu, S. 348 Lu, W. 791 Lu, Y.-L. 1081 Lu, Z. 651 Luba, M. 741 Lubina-Solomon, A. 1007, 194, 989 Lucchesi, D. 59, 85 Lucena, C. F. 469 Lucidi, P. 608 Lucisano, G. 1063, 154, 890 Ludovico, O. 357 Ludvik, B. 481 Ludwig, A. 411 Ludwig, T. 82 Ludwig-Galezowska, A. 368 Luijf, Y. M. 1019 Luk, A. 1040, 1045 Luk, A. O. 1073 Luk, A. O. Y. 1182 Lukács, A. 1262 Lukashevich, V. 816, 879, 888, 889, 891, 902 Lukasova, P. 1091, 602 Lukic, L. 331 Lum, J. 1000 Lumeng, C. 226 Luna, E. 970 Lunardi, A. 1145 Lund, A. 540 Lund, M. Taulo. 584 Lund, S. S. 806, 823 Lund, S. 700 Lundbom, J. 589, 680 Lundbom, J. 36 Lundbom, N. 36, 679, 680, 718 Lundborg, E. 673 Lundby-Christensen, L. 1271 Luongo, A. 1055 Lupi, R. 59, 656 Luque, A. M. 1104 Luquet, S. 767 Lutale, J. K. 1138 Lutgers, H. L. 1031, 1098 Lutgers, H. L.. 283, 8 Luzi, L. 625, 731, 865, 924 Luzuriaga, J. 101, 459 Lv, C. 328 Lv, J. 287 Lv, X. 816 Lv, Y.-J. 1267 Lycett, D. 302 Lyden, M. R. 1014 Lyngbaek, S. 1249 Lynn, F. C. 25, 60 Lyons, L. K. 1052 Lyons, T. J. 1118 Lyssenko, V. 336, 345 Lytrivi, M. 1004 M Ma, J.-X. 1233 Ma, J. 816 Ma, R. 279 Ma, R. C. W. 1045, 1182 Ma, R. C.. W.. 183 Ma, x. 546 Ma, X. 1179, 1180 Maahs, D. 939 Mabley, J. G. 67, 70 Macauley, M. E. O. 585 Macedo, M. P.. 166 Macedo, M. Paula. 179, 527 Macedo, M. P.. 563 Macesic, M. 331 Machado, J. D. 1190 Machann, J. 714 Machicao, F. 721 Machin, A. 906 Machlowska, J. 360, 368 MacIsaac, R. J. 1001 Maciulewski, R. 286, 337 Macura, S. 931 Madani, S. 832, 840 Mader, J. 1005 Mader, J. K. 1019, 979, 989 Madsbad, S. 122, 534, 644 Maechler, P. 1280, 671 Małecki, M. 1006, 1310, 951 Małecki, M. T. 1120 Maeda, M. 1176 Maeda, S. 108 Maedler, K. 263, 389, 744, 749 Maegawa, H. 724, 779 Maemura, K. 272 Maessen, D. 713 Maezawa, Y. 1187, 868, 88, 914 Maffeis, C. 1261 Maffettone, A. 1063 Maganaris, C. N. 1131, 129 Maggioni, A. P. 1270 Maggs, D. 982, 983 Magi, R. 104 Magliano, D. J. 245 Magnan, C. 465, 671 Magni, L. 1000 Magyar, W. 1216 Mahajan, A. 104, 105, 341 Mahalakshmi, M. 1082 Mahdaviani, K. 690 Maheswari, K. 1082 Mahfouz, R. 34 Mahurin, A. 212 Maillo, M. Ángeles. 506 Mainou, M. 1077 Mairal, A. 681, 693 Majic Milotic, D. 1281 Majima, S. 1200 Major, L. 864 Makaroff, L. 308 Makbol, K. M. 1289 Makedou, A. 7 Makino, H. 161, 598 Makino, Y. 1176, 526, 536 Makrilakis, K. 1288 Malagón, M. 687 Malaisse, W. J. 609 Malanda, B. 1082 Malecki, M. T. 1115, 360, 361, 368 Malik, R. A. 1155 Malikova, E. 1293, 1294 Malina, M. 372 Malinska, H. 71, 735, 771 Malipatil, N. S. 1029, 351 Mallamaci, F. 1235, 343 Malle, E. K. 229, 98 Mallol, C. 432 Mallory, J. M. 838 Malloy, J. 244, 77, 842, 844, 845 Malm, H. A. 377 Malmegrin, K. F. 471 Malmgren, S. 56 Maltoni, G. 1253 Mamza, J. 814 Manca, E. 18 Mancarella, F. 231 Mancilla, G. 332 Manda, N. 874 Mandøe, M. J. 780 Manes, C. 1150 Manicardi, V. 1063 Mankovsky, B. 1331 Manley, S. E. 303 Mann, G. E. 1171 Manning, E. 1154 Mannucci, E. 1228 Manrique, A. 1326 Mansell, P. 291 Mansfield, T. 798 Mansfield, T. A. 6, 800, 804, 825 Mantzur, C. 725 Manu, C. A. 1151 Mao, M. 1003, 649 Marandola, L. 742 Marangell, L. 1026 Marathe, J. 1099 Maratou, E. 561, 564, 565 Marazuela, M. 506 Marbury, T. C. 882 Marcanova, Z. 1169 Marchal, P.-O. 767 Marchand, M. 1143 Marchesini, G. 841 Marchetti, P. 134, 204, 29, 373, 428, 478 Marco, A. 1104 Marcovecchio, L. 90 Margaritidis, C. 7 Mari, A. 125, 501, 504, 507, 515, 524, 527, 738 Marigliano, M. 1261 Marinelli Andreoli, A. 608 Marinho, R. 616 Marinković, J. 1075 Mark, M. 1276, 802, 824, 875 Markovic, I. 331 Markovic, T. P. 777 Marks, E. 701 Marks, J. B. 504 Marmugi, A. P. 199 Maroccia, M. 946 Marple-Horvat, D. E. 1131 Marques, M.-A. 32, 703 Marrano, N. 464 Marre, M. 1183, 89, 9 Marselli, L. 134, 204, 29, 373, 428 Marshall, A. 1155 Marshall, S. M. 1198 Martelli, D. 1142 Martí, Y. 387 Martin, C. 127 Martin, C. M. 119 Martin, J. 857 Martin, J. A. 938 Martin, P. 118 Martín-Duce, A. 544 Martinell, M. 173, 174 Martinerie, C. 767 Martinez, L. 832, 840 Martinez, M. P. 156 Martínez-Arrieta, F. 544 Martin-Levilain, J. 1280 Martins, D. A. 1128 Martins, F. O. 166 Marucci, A. 343, 357 Marullo, L. 344 Marx, N. 185 Marzi, C. 144 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 553 Masaki, Y. 626 Masania, J. 558 Mascali, L. G. 204 Mashitani, T. 313 Masindova, I. 369 Maskarinec, G. 270 Mason, G. 394 Massy, Z. 1287 Mast, M. R. 208, 258 Mast, R. 942 Mastorakos, G. 782 Masuzaki, H. 496 Matafome, P. 261 Matei, A. 1097 Matejko, B. 1120, 951 Matejko, B. 1006, 360, 361 Mathews, C. 236 Mathews, S. T. 212 Mathiesen, E. R. 1105, 1113, 1114, 1117, 1121, 408, 540 Mathieu, C. 900 Mathieu, C. 1110, 1251, 1257, 428 Matikainen, N. 36 Matsaggos, S. 1240 Matsaggouras, G. 1300, 94 Matsagos, S. 1295 Matsagouras, G. 564, 565 Matsuda, T. 418, 454 Matsuhisa, M. 975 Matsumoto, J. 573 Matsumoto, S. 1200 Matsunaga, K. 365 Matsunaga, K. 422 Matsunaga, S. 630 Matsunaga, Y. 1277 Matsuoka, A. 884 Matsushima, S. 626 Matsushita, K. 1200 Matsushita, M. 496 Matsuyama, Y. 209 Matsuzawa, K. 503 Matsuzawa, Y. 542 Mattaliano, C. 1145 Mattern, Y. 1302, 69 Matthaei, S. 852 Mattheus, M. 885 Matthews, D. 855 Matthews, J. E. 831 Mattiello, L. 913 Mattis, K. K. 202 Mattsson Hultén, L. 116 Matulewicz, N. 176, 211 Matz, M. 423 Maucher, D. 252 Maurovich-Horvat, P. 80 Mavrodi, A. 988 Mavros, P. 898 May, A. M. 770 Mayas, D. 687 Mayerova, V. 686 Mayhue, M. 228, 691 Maymó-Masip, E. 438 Mayo, K. 545 Mayoux, E. 1276, 801, 802, 819, 821, 824 Mayu, T. 598 Maywald, U. 12 Mayzel, Y. 985 Mazanova, A. 1279 Mazza, C. 1251, 1257 Mazza, T. 357 Mazze, R. 1005, 829, 956 Mazzeo, A. 52 McAdam, J. 1147 McAuley, S. A. 1001 McCaffrey, N. 548 McCarthy, M. 105, 139, 339, 341 McCarthy, M. I. 106, 140 McCaughan, G. 165 McCormack, S. 348 McDonald, T. 450 McDonald, T. J. 172, 510 McEwan, P. 1067, 327 McGee, S. L. 707 Mcginty, J. 349 McHarg, S. 1315 McInnes, G. 888, 891 McKeigue, P. 11 McKillop, A. M. 497 McMurray, J. 908 McNeill, B. 46 McNeill, C. 893 McVicker, L. 153 Mechta, M. 235 Meda, P. 991 Medeiros, M. H. G. 1185 Medina, A. S. 466 Medina Benavente, A. 375 Medlej, R. 883 Meek, C. L. 531 Mehana, A. E. 747, 748 Mehio, M. 554 Mehmeti, I. 455, 456 Mehta, R. 814 Mehta, R. 970 Meier, J. J. 541, 75, 926 Meiffren, G. 932, 934 Meigs, J. B. 344 Meikle, P. J. 463 Meinicke, T. 806, 849 Meininger, G. 799, 847, 853, 854, 855 Meisinger, M. 581 Melander, O. 143, 79 Melidonis, A. 1027, 1240, 1246, 1295 Mellerio, H. 1033 Mellett, N. 463 Melo, B. F. 562 Melo, B. F.. 778 Ménard, J. 1094 Menard, M. K. 1118 Menart, B. 442 Mendez, A. J. 504 Mendonca, C. 357 Mendoza Lopes de Faria, J. 1166 Menegazzo, L. 729 Menge, B. A. 75 Menini, S. 1307 Mensberg, P. F. 922 Mensikova, L. 1293, 1294 Mensink, M. 775 Menzaghi, C. 1235, 343 Meraviglia, V. 612 Mercader, J. M. 664, 687 Mercau, M. E.. 657 Mercier, J. 477, 586 Mercuri, L. 357 Merger, S. 958 Merioud, B. 1275 Merkely, B. 80 Merker, L. 806 Merkwirth, C. 1280 Merovci, A. 552 Mestres, O. 190 Metha, A. 312 Metzger, D. 1280 Metzig, C. 54 Meur, G. 134 Meyer, C. 185 Meyer, E. 728 Meyerovich, K. 430 Meyers, J. 112, 962 Mezza, T. 515, 524, 619, 620, 783 Miarka, P. 1310 Miccoli, R. 59, 85 Michael, M. Dodson. 971 Michaelsen, N. B. 918 Mieda, Y. 454 Miedema, I. 795 Miele, C. 353, 354 Mifune, H. 794 Mignone, L. E. 774 Mignot, E. 184 Migrenne, S. 671 Miguel, J. F. 982 Mihic-Necic, B. 747 Mikhed, Y. 819 Miki, T. 203, 521 Mikkola, K. 446 Mikova, R. 1305 Milenkovic, T. 172 Milicevic, Z. 42 Milicic, T. 331 Milinovich, A. 323 Miller, D. M. 838 Miller, M. 638 Miller, M. C. 545 Miller, R. G. 1255 Millwood, I. 274 Milne, N. 1199 Milne, R. 46 Milrad, S. 1055 Min, A.-K. 238 Min, K. 1196 Minami, K. 920 Minder, A.-E. 17 Mineo, I. 559 Miquel, J. 983 Mir, J. 417 Mirošević, G. 1075 Mirra, P. 353 Misawa, H. 1269 Misitzi, A. 1295 Misra, S. 367 Mitanchez, D. 151, 717 Mitchell, L. K. 739 Mitchell, M. 142 Mitchell, R. 349 Mitchell, R. K. 102 Mitrou, P. 561 Mitrou, P. 564 Mitsuhashi, K. 1175 Mitsui, S. 437 Miwa, I. 505 Miyake, H. 1229 Miyasato, M. 1274 Miyata, T. 713 Miyawaki, M. 1274 Miyoshi, A. 1265 Miyoshi, H. 1265, 470, 874, 929 Mizgier, M. L. 507 Mizukami, H. 1129 Mizuno, K. 573 Mizushima, W. 626 Mkrtumyan, A. 1064 Mleczko, J. 216 M. Lopes de Faria, J. 1173 Mlynárová, J. 1239 Mo, E. 652 Mo, E.-Y. 547 Moberg, E. 481 Mobini, R. 116 Mocan, A. S. 1036 Mochizuki, K. 634 Moen, T. 439 Moerman, E. 609 Moffa, S. 524, 619, 620, 783 Mohammed, M. 302 Mohammedi, K. 1183, 9 Mohan, V. 1082 Moharam, S. A. 1289 Möhlig, M. 653 Mohtarin, S. 596 Moisan, C. 901 Mojibian, M. 33 Mokan, M. 43 Mokan, M. 43 Mokhort, T. V. 311 Molina, M. 1108 Molina-Wilkins, M. 74 Moll, A. C. 208, 258 Møller, C. 336 Mollet, I. G. 377 Möllmann, J. 185 Molnes, J. 363, 370 Molsted, S. 268 Moltke, I. 103 Molven, A. 363, 364, 370 Momo, R. A. 1245 Moncada, S. 1320 Monciotti, C. 555 Mondragon, A. 349 Monedero, J. 548 Monk, A. 829 Monnier, V. M. 92 Monsour, H. P. 590 Montalto, G. 905 Montane, J. 200, 706 Montanya, E. 387 Monte, M. 1202 Montemurro, C. 1166, 1173 Montesano, A. 625, 731, 865 Montgomery, M. 101 Montori, V. M. 1077 Moon, S. 652 Moon, S. 547 Moore, M. C. 860 Mootha, V. K. 393 Mor, A. 282, 321 Mora, S. 216 Morales-Santana, S. 587 Moralidis, E. 7 Moran, B. M. 497 Moran, T. 246, 294 Morandi, A. 1261 Morano, S. 305 Moreno, C. 910 Moreno, M. 664 Moreno, M. 683, 684 Moreno, M. 687, 716 Moreno-Castellanos, N. 687 Moreno-Navarrete, J. M. 664 Moreno-Navarrete, J. M. 683, 684 Moreno-Navarrete, J. M. 687 Moreno-Navarrete, J. 716 Moreno-Villegas, Z. 544 Moret, V. 555 Moretti, S. 462, 612 Morgado, C. S. C. 1128 Morgan, N. G. 172, 285, 358, 409 Morgantini, C. 527 Mori, I. 659, 966 Mori, K. 834 Mori, M. A.. 1332 Moriconi, D. 125 Morigny, P. 681 Morimoto, A. 1260 Morimoto, Y. 270 Morino, K. 779 Morita, A. 203 Morita, H. 659, 966 Mørkrid, K. 293 Moro, C. 32, 730 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 554 Morris, A. 105 Morris, A. P. 104, 106, 140, 344 Morris, R. 302 Morris, V. 1151 Morrow, L. 661, 937 Mortensen, B. 57 Mortensen, H. B. 1250 Mosenzon, O. 183, 186, 483, 641, 877, 887, 930 Moses, R. G. 641 Mosig, A. S. 682 Mossé, A. 506 Mota, E. 278 Mota, M. 278 Motté, E. 253 Moulin, P. 1225 Moulton, C. D. 316 Moureira, M. 899 Mourier, A. 1332 Mourouzis, I. 1288 Mousiama, T. 1240 Moutquin, J.-M. 1094 Moyers, J. S. 971 Moysyuk, Y. 480 M.P. Prado, E. 1173 Mraz, M. 912 Mrozińska, S. 951 Muchmore, D. B. 149 Mudi, S. R. 596 Muehlen-Bartmer, I. 946, 949, 961, 980 Mueller, N. 945 Mueller, P. W. 334 Mueller, U. A. 945 Mueller-Wieland, D. 732 Muendlein, A. 1217, 762 Mukai, E. 203 Mulas, F. 1245 Mulas, M. Franca. 1063 Mulder, H. 133, 350, 393, 398, 399, 400 Mulder, M. T. 1318 Müller, N. 1024, 301 Müller, P. 230, 411 Müller, S. 12 Muller, T. 711 Müller, U. A. 1024, 301, 958 Muller, W. 60 Müller-Lühlhoff, S. 557 Mundiany, M. 679 Munoz, R. 982 Muñoz, R. 983, 984 Munoz-Bellili, N. 89 Muñoz-Torres, M. 587 Münzel, T. 819 Mur, T. 628 Murahovschi, V. 764, 765 Murakami-Kawaguchi, S. 392 Murase-Mishiba, Y. 1274, 437, 611 Murata, H. 1191 Murata, K. 1247 Murawaki, A. 503 Mureşanu, C. M.. 1036 Murfitt, S. A. 24 Murillo, S. 624 Muris, D. M. J. 96 Murphy, A. T. 938 Murphy, N. P. 172 Murray, A. J. 24 Murray, R. 309 Musi, N. 74 Musil, P. 1239 Mussin, C. 864 Musso, C. 864 Muth, L. 1085 Mutlu, H. Huseyin. 449 Myakina, N. 996 Mylona, M. 631 Myroforidis, A. 603 N Nabi, H. 260 Nackiewicz, D. 60 Nagai, S. 1265 Nagai, Y. 876 Nagano, G. 694 Nagano, H. 502 Nagao, M. 1325, 708, 786 Nagashima, M. 871, 915 Nagashimada, M. 752 Nagata, M. 952 Nagata, N. 752 Nagel, F. 1030 Nagl, K. 1296 Nagy, Z. 746 Naidis, E. 985 Nair, K. 1263 Najmi, L. A. 363 Naka, Y. 634 Nakagami, T. 1158 Nakagawa, F. 779 Nakajima, K. 509 Nakajima, S. 952 Nakajima, Y. 786 Nakamura, A. 470, 874 Nakamura, K. 994 Nakamura, N. 1175, 1200, 189 Nakamura, T. 1247 Nakamura, T. 952 Nakamura, T. 952 Nakanishi, N. 1175 Nakano, Y. 493 Nakao, K. 779 Nakatsu, Y. 1277 Nakayama, H. 719 Nalla, A. 408 Nam, Y. 1193 Nandagopalan, S. 954 Nankervis, A. J. 1118 Nannipieri, M. 125 Naples, M. 1308, 921 Napoli, A. 1063, 1098 Nara, E. 542 Narayanan, R. P. 1029, 351 Narayanan, U. 1315 Narazaki, M. 967 Nardeli, T. R. 424 Naselli, G. 388 Näsman, P. 44, 45 Nasteska, D. 522, 535, 766 Nata, K. 392, 427 Natali, A. 166, 501, 527 Natalicchio, A. 1316, 299, 464 Natasha, K. 1106 Nauck, M. 37, 525, 815, 830, 846, 848 Nauck, M. A. 109 Nauck, M. A.. 928 Naujok, O. 390 Navarova, J. 1294 Navarro-Martínez, L. 553 Navis, G. 1219, 1232, 192 Nawrodt, B. 109 Nawroth, P. P. 1282 Nawroth, P. Paul. 1303 Nawroth, P. P. 1304 Nawroth, P. P.. 1313 Nawroth, P. P. 1314, 215 Nayak, A. U. 1088, 906 Neale, A. 388 Nédélec, A. 703 Negoro, T. 493 Negrato, C. A. 1119, 668 Negri, C. 511 Nelson, A. 1152 Nemcová, A. 1146 Nemcova, A. 1148 Nergård, B. 740 Neslusan, C. 817, 826 Neubauer, K. M. 979 Neumann, F. 419 Neumann, U. H. 33 Neunübel, K. 682 Ng, C. 465 Ng, M. C. Y. 1182 Nguewa, J.-L. 1227, 512 Nguyen, M. 1236, 660, 660 Nguyen, T.-V. 1170 Nguyen-Tu, M.-S. 97 Ni, Y. 752 Nicholas, L. M. 133 Nichols, G. A. 159 Nichols, L. 302 Nicolas, A. 89 Nicolay, C. 646 Nicolucci, A. 1063, 1066, 154, 890 Nieda, M. 914 Niedzwiecki, P. 1144, 222 Nielsen, A. A. 289 Nielsen, J. H. 408 Nielsen, J. Steen. 282 Nielsen, J. S. 321 Nielsen, K. E. 1025 Nielsen, L. B. 1250 Nielsen, R. 103 Nielsen, T. S. 103 Niemann, J. 1312, 727, 728 Niemi, T. 19 Nieminen, J. K. 1188 Nightingale, P. G. 303 Nigi, L. 231, 29 Nigro, C. 353 Nijpels, G. 1078, 14, 208, 258, 275, 342, 942 Nijpels, G. 1080 Niki, I. 422 Nikolajuk, A. 176, 211 Nikolic, D. 905 Nikonova, E. 1064, 841, 846 Nikonova, T. 441 Nillegoda, N. B. 1282 Nilsson, C. 1100 Nilsson, E. 352 Nilsson, P. M. 900 Nilsson, P. 345, 79 Nilsson, P. M. 336 Nilsson, S. K. 379 Nimri, R. 1000 Ning, G. 816 NIRAD Study Group, 451, 742 Nischal, A. 1047 Nishikawa, H. 437 Nishikawa, T. 542 Nishimura, F. 1277 Nishimura, R. 1260, 513 Nishimura, W. 26 Nishio, Y. 779 Niskanen, L. 974 Nislow, C. 715 Nitti, D. 729 Nitz, B. 82 Niwa, T. 509 Nixon, J. 1152 Njølstad, P. R. 363, 364, 370 Nocito, L. 175 Noda, M. 1215, 272 Noethlings, U. 772 Nogueira, T. C. 204 Nogueiras, R. 674 Noh, J. 1192 Noh, Y. 517 Nøhr, M. K. 700 Nohtomi, K. 915 Nolan, J. 548 Nomoto, H. 1265, 470, 874, 929 Nomura, H. 1215 Nonell, L. 687 Nordmann, T. 614 Norhammar, A. 44, 45 Norheim, F. 618 Norrbacka, K. 221, 839 Nørrelund, H. 789 Norris, J. M. 170 Nosek, L. 932, 934, 936 Nöthlings, U. 770 Noto, H. 1215 Notomi, K. 871 Novaes, F. S. 123 Novak, D. 912 Novakovic, A. 714 Novials, A. 200, 624, 706 Nóvoa Mogollón, F. 1050 Novo-Rodriguez, C. 587 Nowak, N. 360 Nowotny, B. 442 Nowotny, I. 953 Nowotny, P. 168, 557, 589 nPOD-V Consortium, 285 Ntemka, A. 7 Nuhoho, S. 1072 Nuijts, R. M. M. 1156 Nummenmaa, L. 177, 678, 83 Numminen, H. 678 Nuñez, I. 1205 Nurit, K. 425 Nürnberg, B. 714 NUTRIM, School of Nutrition, Toxicology and Metabo, l. 623 Nuutila, P. 121, 1283, 163, 177, 19, 20, 219, 23, 446, 678, 738, 83 Nyby, S. 922 Nylander, V. 248 Nystrom, F. H. 787 Nyumura, I. 1191 O Oarrak, T. 206 Obermayer-Pietsch, B. 124 Obeso, A. 562 O›Brien, K. T. 303 O›Brien, K. 150 Ocakoglu, I. 449 Ocampo Francisco, A. 964 Occhipinti, M. 373, 478 O›Connell, R. 1233 O´Connell, M. 995 Odawara, M. 272, 902 Odisho, T. 457 O›Donoghue, G. 548 Oelze, M. 819 Ofori, J. K. 400 Ogasawara, S. 1129 Ogata, M. 1264 Ogawa, W. 952 Ogbaa, I. 803 Ogino, J. 1194, 925 O›Gorman, D. 548 Ogunkolade, W. B. 236 Oh, B. 473 Oh, B.-C. 593 Oh, K.-W. 297 Oh, R. 878, 897 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 555 Oh, S. 1196 Oh, S.-H. 473 Oh, S. 473, 650 Oh, T. 1167 Ohara, N. 630 O›Hare, J. 669 Ohashi, Y. 1209 Ohata, Y. 598 Ohki, T. 719 Ohkura, H. 503 Ohkura, T. 503 Ohlsen, R. 309 Ohlsson, L. 518 Ohman, P. 118 Öhman, P. 244 Ohmi, S. 505 Ohn, J. 374 Ohno, H. 694 Ohta, A. 876 Ohta, M. 573 Oikarinen, M. 285 Oikarinen, S. 169 Oikawa, S. 1325, 708, 786 Oikawa, Y. 191 Oikonen, V. 20, 738 Oikonomou, D. 1314 Ojo, O. O. 495 Okada, H. 659, 966 Okada, H. 1175 Okada, S. 560 Okamoto, H. 392 Okamoto, M. 422 Okamoto, T. 115 Okerberg, C. 212 Oki, K. 694 Okisugi, M. 191 Okita, K. 626 Okubo, H. 1277 Okuno, Y. 952 Okuyama, K. 115 Oláh, Á. 444 Olaniru, O. E. 407 Olausson, J. 116 Olimpico, F. 373 Olivar, M. 1104 Oliveira, A. A. F. 1185 Oliveira, M. S. 123 Oliveira, T. C. 461 Oliveira, T. F. 1185 Oliver, N. 1044, 1153, 290 Oliver, N. S. 367 Oliyarnyk, O. 771 Olsovsky, J. 1169 Olsson, A. H. 234 Olszewska, M. 1144 Omar, B. 403, 491, 518, 55 Omar, S. Z. 1093 Omer, B. 307 Omura, M. 542 on behalf of the DIAGRAM Con- sortium, 104 on behalf of the DPV initiative, 1231 on behalf of the EDITION JP 1 study group, 975 on behalf of the EDITION JP 2 study group, 976 on behalf of the EPIC consortium, 770 on behalf of the Japan study group of Wolfram synd, r. 365 on behalf of the SUMMIT Investi- gators, 11 on behalf of the XC-Pleiotropy Group, 344 Onda, Y. 1260, 513 O›Neal, D. 843 O›Neal, D. N. 1001 O›Neill, C. 1092 Onfroy, L. 681 Ong, S. H. 813 Onishi, S. 1187 Ono, Y. 626 Onouchi, H. 812 Oohata, Y. 161 OpT2mise Study Group, 1002 O›Rahilly, S. 520 Oram, R. 510 Orban, P. C. 254 Orchard, T. 1255 Orho-Melander, M. 143, 377, 62, 79 Orhun, A. 449 ORIGIN Trial Investigators, 1270 Orime, K. 750 Orlando, M. R. 1316 Orme, M. E. 850 Ornstrup, M. J. 757 Orsi, E. 210, 305, 86 Ortega, F. 683, 684, 716 Ortega, F. J. 664, 687 Ortis, F. 424 Osmenda, G. 361 Osta, B. 227 Östenson, C.-G. 743 Ostensson, C. 355 Osterhoff, M. 765 Osterhoff, M. A. 682, 734, 82 Östgren, C. 787 Osthold, S. 419 Ota, H. 392 Ota, T. 752 Otabe, S. 719 Otonkoski, T. 172, 358 Otvos, J. D. 1309 Ou, H.-Y. 582 Ouarrak, T. 304 Šoupal, J. 568 Ouslim, A. 1083 Out, M. 795 Ouwens, D. M. 763 Ouwens, D. M.. 760 Oğuz, A. 941 Overgaard, R. V. 881 Overvad, K. 770, 772 Ovesen, P. 1121 Owada, Y. 766 Owolabi, B. O. 495 Oya, J. 1162 Oyadomari, S. 496 Øyen, O. 255 Ozaki, R. 1040, 1045, 1073 Ozasa, N. 781 Ozel, S. 1111 P Paavonen, K. 1299 Pácal, L. 1169, 568 Pach, F. P. 604 Pacini, G. 1112, 1272, 491, 518 Paczkowska, M. 356 Padhiar, A. 839 Page, K. A. 156 Page, M. M.. 385 Page, M. M. 705 Pagkalos, E. 900 Paiva, A. C. 1061 Paixão, A. 261 Pakseresht, M. 1037 Paldánius, P. M. 900 Paldánius, P. M. 893 Pallai, S. 131 Palming, J. 576 Palomer, X. 53 Pan, C. 1058 Pan, J. 1171 Pan, Q. 276 Pan, S.-F. 1267 Panagoulias, G. 1150 Panajotu, A. 80 Panduru, N. M. 87 Panse, M. 405 Pantalone, K. M. 323 Pantoja, A. 1104 Panzhinskiy, E. 458 Papadimitriou, A. 1164, 1166, 1173 Papadimitriou, A. 1086 Papadimitriou, Z. 782 Papakonstantinou, E. 561 Papanas, N. 1150 Papatheodorou, D. 782 Papavasileiou, S. 155 Papazafiropoulou, A. 1246 Papizan, J. 212 Pappas, A. 155 Paramalingam, N. 1256 Paramasivam, S. S. 1093 Pareja, J. C. 123 Parekh, W. A. 489 Parellada, C. B. 1117 Parikh, J. 627 Parikh, S. 798, 807, 848 Parikh, S. J. 800, 804 Parikh, S. J.. 805 Park, A. J. 531 Park, B. 35, 416 Park, C.-Y. 843 Park, C.-Y. 297 Park, H. 297 Park, H. 391 Park, J. 858 Park, J.-Y. 1290 Park, K.-J. 1087, 460 Park, K.-G. 1284, 238, 591, 867 Park, S. 297 Park, S. 1193 Park, S.-W. 297 Park, Y.-M. 10 Park, Y. 866, 972 Parker, A. 183, 887 Parkin, C. G. 1014, 545 Parkkola, R. 163, 219, 23 Paroni, F. 389, 744 Parra, P. P. 982 Parrillo, L. 353, 354 Parsons, L. 747, 748 Partenheimer, A. 732 Parthsarathy, V. 119 Partke, H.-J. 415 Partridge, H. 956 Parving, H.-H. 207 Pasch, A. 1286 Paschou, S. A. 603 Pasdar, N. Atabaki. 259 Pasquel, F. 1038, 114 Pastan, I. 354 Paster, A. 981 Patel, S. 1, 815, 851, 885, 894 Patel, T. 1139 Patel, V. 1079 Patenaude, V. 892 Patente, T. 1183 Patera, I. Patrizia. 1253 Paterson, A. 158 Paterson, A. D.. 283 Paterson, A. D. 51 Päth, G. 1096, 747, 748 Pathan, F. 883 Patrascioiu, I. 1205 Patrick, A. L. 955 Patsch, B. 1238 Patsch, W. 226 Patti, A. Maria. 905 Pattou, F. 233 Pattou, F. 609 Patzer, O.-M. 484 Pauli, J. R. 616 Pauli, L. S. S. 616 Paulsen, S. 875 Paulsen, S. J. 519 Pavlidis, G. 1300, 94 Pavo, I. 1299 Pawlowski, M. 1159 Peakman, M. 290 Pearson, E. 139, 339, 341, 510 Pearson, E. R. 257, 259 Pearson, G. L. 463 Pechlivangolou, P. 988 Pechtner, V. 38, 827 Pecoraro, G. 905 Pedersen, B. K. 178 Pedersen, H. K. 139, 339 Pedersen, J. 519, 57 Pedersen, L. 789 Pedersen, M. 352 Pedersen, N. L. 352 Pedersen, O. 103, 341 Pedersen, O. B. 339 Pedersen, S. B. 700 Pedersen, S. B.. 757 Pedersen-Bjergaard, U. 1329, 481, 485, 632, 917 Pedraza, E. 992 Pedro-Botet, J. 1224, 187 Peeters, P. H. M. 772 Peeters, S. A. 93 Peiretti, F. 575 Peiro, C. 1320 Peiró, C. 1319 Peixoto, E. Bouçada. Mauro Inácio. 1164 Peled, D. 453 Pelikanova, T. 735, 771 Pellegrini, F. 1235, 154, 890 Pellegrini, M. A. 1066 Peña, V. 1104 Penalba Martínez, M. 1035 Penfold, S. A. 1170 Penfornis, A. 900 Penfornis, A. 1071, 640, 832, 840 Peng, C.-H. 1081 Peng, L. 114 Pénicaud, L. 670 Penno, G. 210, 305, 59, 85, 86 Pepin, E. 802 Peppas, T. 1246 Perakakis, N. 1096, 747, 748 Peraldi-Roux, s. 756 Peräneva, L. 1188 Perego, C. 462, 612 Pereira, A. 261 Pereira, M. J. 576, 605 Pérez-Bravo, F. 332 Perez-Cadena, Z. 722 Perez-Gutthann, S. 890 Pérez-Luque, E. L. 664 Perez-Perez, A. 571 Perez-Sieira, S. 166 Perfilyev, A. 352 Perichon, R. 142 Perkins, B. A. 956 Perkins, C. M. 41 Permpikul, C. 556 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 556 Perrea, D. 1288 Perrin, L. 570 Perrini, S. 1316, 299, 464 Perron, P. 1094 Perry, C. R. 41, 838 Perry, R. 861 Persaud, S. 137 Persaud, S. J. 100, 406, 407, 99 Persson, F. 911 Pesl, P. 1044 Pesta, D. 861 Peter, A. 1096, 721 Peterkova, V. 371 Peterkova, V. A. 362 Peters, A. 760 Peters, J. L. 1068 Petersen, B. 1014 Peterson, G. 661 Petit, C. 648 Petit, M.-H. 642, 648 Peto, R. 274 Petropoulos, I. N. 1155 Petrucco, A. 18 Petruzelkova, L. 366 Petsiou, E. 561 Pettus, J. H. 1020 Pfeiffer, A. F.. H.. 525, 653 Pfeiffer, A. F. H. 682, 734, 764, 765, 82 Pfeilschifter, J. 472 Pfeilschifter, J. M. 230, 411 Pfohl, M. 943 Pham, I. 1236 Pham, T. 121 Phielix, E. 581 Philippe, J. 1103, 570 Philips, J.-C. 1143 Philis-Tsimikas, A. 145, 243 Phillip, M. 1251, 1257 Phillips, A. R. J. 1315 Philotheou, A. 964 Piaggesi, A. 1142, 1145 Piatti, P. 501 Pibernik-Okanovic, M. 1028 Piccinini, F. 123, 567 Pichaipillai, L. 1116 Pichard, B. 233 Pichotta, P. 937 Picon, M. J. 1108 Pieber, T. 1005 Pieber, T. R. 1019, 124, 147, 232 Pieber, T. R.. 979 Pieber, T. R. 989 Piechowiak, K. 1021, 622, 986 Pietiläinen, K. 36, 679, 680 Pietiläinen, K. H. 697 Pilacinski, S. 222 Pilemann-Lyberg, S. 809 Piljac, A. 1281 Pillon, N. J. 759 Pilorget, V. 110, 1251, 1257 Pilz, I. 748 Pilz, S. 275 Pimentel, F. 982, 983, 984 Pinchuk, A. 480 Pineau, M. 1016 Pingitore, A. 137 Piniés, J. A. 1222, 1268 Pinnick, K. E. 600 Pinnschmidt, H. 732 Piñol, J. L. 266 Pintaudi, B. 154 Pinto, J. B. 380 Pipeleers, D. 253 Pipeleers, D. G. 288, 452 Pipper, C. 1250 Pirola, L. 227 Pisani, D. 693 Pisani, G. D. 471 Pisarczyk-Wiza, D. 1022 Pi-Sunyer, X. 73, 833, 881, 903, 904 Pittman, D. L. 334 Pivackova, L. 1293, 1294 Pivovarova, O. 1331 Pivovarova, O. 525, 764, 765 Pizarro, P. P. 984 Pizarro Parra, P. 983 Pizzini, P. 1235 Pla, A. 387 Place, J. 1019 Pladevall, M. 890 Plank, J. 979 Platt, D. 545 Plence, P. 412 Pleus, S. 1017 Plicka, R. 995 Plomgaard, P. 178 Pochat, A. 642, 648 Poduri, M. 739 Pogue, J. 1270 Pohl, R. 150 Poirier, J.-Y. 1010 Poitout, V. 802 Pokrovskaya, M. 95 Polak, B. C. P. 208, 258 Polderman, J. A. W. 1053 Policardo, L. 1062 Policola, C. 620, 783 Polidori, D. 820, 855 Pollack, P. A. 186 Pollock, R. 221 Pollom, R. K. 146, 948, 969 Polonia, J. 1202 Polonsky, W. H. 1030 Polyzou, E. 1096 Poncina, N. 729 Pongratz, R. L. 394 Poon, W. 336, 345 Pop, L. A. 1036 Popa, S. 278 Pop-Busui, R. 127 Poplawska-Kita, A. 286, 337 Porcellati, F. 608 Porciello, G. 353 Porksen, N. 957 Pørksen, S. 1250 Porta, M. 51, 52 Porta, N. 628 Portal-Núñez, S. 544 Pospisilik, J. A.. 226 Postic, C. 681 Potasso, L. 1096 Poulain-Godefroy, O. 233 Poulou, A. 603 Poulsen, K. 1048, 1049 Poulsen, M. M. 757 Poulsen, P. L. 1124 Poulsen, P. 352, 78, 959 Poulsen, S. S. 519, 57 Pourteymour, S. 618 Powell, D. 803 Powell, D. R. 862 Powell, J. 520 Powell, N. 701 Powell, P. O. 594 Powell, R. 1068 Powers, M. 829 Pozzilli, P. 448, 453, 872 Pracht, C. 672 Pradhan-Nabzdyk, L. 251 Prados-Nodd, N. 571 Prahalathan, C. 1323 Prapaporn, J. 357 Prati, S. 1103 Prazny, M. 1305 Prázný, M. 568 Preckel, B. 1053 Preechasuk, L. 556 Prellberg, M. 709 Prevenzano, I. 354 Prevost, G. 554 Price, D. 954, 994 Price, D. A. 1020 Price, K. L. 1171 Price, L. 1139 Prietl, B. 232 Prikoszovich, T. 1112 Prince, M. J. 146, 948, 969 Probstfield, J. 1270 Profant, M. 369 Prokopenko, I. 106, 144, 344 Proks, P. 202 Protogerou, A. D. 96 Protopsaltis, J. 1295 Prudente, S. 357 Pruhova, S. 366, 372 Przemeck, G. 404 Ptaszynska, A. 6, 798, 800, 804, 822, 825 Puşcaşu, S. 1036 Pucci, L. 85 Puelles, J. 327 Pugliese, G. 1307, 210, 305, 86 Puigdecanet, E. 687 Puig-Domingo, M. 846 Pujol, A. 419 Pullen, T. J. 134 Pullen, T. J. 97 Punthakee, Z. 1270 Puntí, J. 1244 Pursiheimo, J.-P. 169 Püschel, G. P. 583 Puskas, J. D.. 114 Pussard, E. 648 Püttgen, S. 249 Pynes, M. 1016 Q Qian, K. 235 Qin, G. 1179, 1180 Qiu, S. 1057 Qiu, Y. 808, 811, 898 Quaranta, V. 299 Queniat, G. 609 Quimbo, R. A. 111 Quinlan, D. 303 R Rabbani, N. 1301, 1306, 558, 68 Rabbone, I. 1253 Rabe, B. 60 Rabinovitz, H. 776 Rabøl, R. 644 Racault, A.-S. 1009 Raccah, D. 1064 Raciti, G. A. 353, 354 Raddatz, K. 720 Rader, D. 65 Radican, L. 322, 898 Radin, M. S. 1089 Radkowski, P. 368 Radtke, F. 250 Radtke, M. 439 Ragot, S. 9 Rahou, A. 1083 Raiko, J. 177, 19, 20 Raiko, J. R. H. 23 Raimo, K. 1201 Raimondo, A. 359 raitakari, O. 312 Rajander, J. 446 Rajkovic, N. 331 Rajpathak, S. 898 Rakhimova, N. 1204 Rakyan, V. K. 236 Rama Chandran, S. 338 Ramachandran, A. 1270, 481 Ramanathan, S. 228, 691 Rami-Merhar, B. 1296, 232, 939 Ramírez García, O. 1101 Ramm, G. 1170 Rämö, J. 36, 679 Ramond, E. 396 Ramos-Álvarez, I. 544 Ramos-Leví, A. María. 506 Ramracheya, R. 613 Ramtoola, S. 1116 Ranasinghe, P. 265 Ranganath, L. 541 Ranson, A. 932, 934 Rao, S. 863 Raoux, M. 991, 992 Raposo, J. F. 1061 Rappou, E. 679, 697 Raptis, S. 1295 Raptis, S. A.. 561 Raptis, S. A. 564, 565 Rasenack, R. 1096 Rasheed, A. 65 Raskin, J. 1026 Raskin, P. 145 Rasmussen, S. 974 Rasouli, B. 173, 174 Rathmann, W. 1214, 760, 796 Rath-Wolfson, L. 801 Ratnasingam, J. 1093 Rauh, S. P. 208, 258, 275, 326 Ravaine, V. 992 Raverdy, V. 233 Ravussin, E. 507 Rawshani, A. 306, 318, 320 Ray, K. K.. 183, 641 Rayman, G. 993 Raymond, R. H. 500, 529 Raynal, P. 703 Rayner, C. K. 217, 530, 538, 774 Rayner, N. W. 104, 107 Rayner, W. N. 90 Raz, I. 183, 186, 453, 483, 641, 658, 725, 877, 887, 930 Razak, R. Abdul. 1093 Razak Hady, H. 741 Re, R. 1012 Reaney, M. 646, 839 Rebelos, E. 501 Rebuffat, S. 699 Reddy, M. 1044 Reddy, P. 150 Reed, S. D. 273 Rees, C. 1014 Reeves, N. D. 1131, 129, 130 Régent, A.-S. 233 Reghina, A. 790 Rehfeld, J. F. 532, 533 Rehman, T. M. 324 Rehorova, J. 1169 Reichenbach, A. 672 Reichetzeder, C. 1286, 1330 Reid, T. 1070 Reihill, G. 645 Reimann, F. 202, 520, 531 Reimer, A. 1015 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 557 Reimer, A. 1041 Rein, P. 1211, 1212, 1217, 1218, 1220, 1223, 1238, 1243, 762 Reinbeck, A. 415 Reinhold, P. E. 91 Reinholt, F. 255 Ren, G. 212 Ren, W. 546 Renard, E. 843 Renard, E. 1019 Renard, E. M. 1000 Renaud, S. 991 Rendell, M. S. 879, 889 Rendell, M. 830 Renehan, A. 246, 294 Renner, S. 918 Renstrom, E. 467 Renstrom, F. 65 Repas, C. 746 Repetto, E. M.. 657 Repina, E. 443 Resta, O. 299 Resuli, M. 909 Rett, K. G. 126 Rewers, A. 1252 Rewers, M. 1252, 170 Reyes-Garcia, R. 587 Rezaninova, L. 1148 Rezende, L. F. 424 Reznik, Y. 1002, 1009, 1012, 999 Rhee, B. 1192 Rhee, E.-J. 297 Rhee, N. A. 1042, 519, 57 Rhee, S. Youl. 280 Rhinehart, A. 1070 Riant, E. 225 Ribaux, P. 30 Ribeiro, M. J. 562 Ribeiro, R. T. 1061 Ribel-Madsen, R. 352 Ribitsch, A. 232 Riboni, L. 465 Ricart, W. 664, 683, 684, 687, 716 Ricci, C. 1307 Richards, P. 520 Richardsen, K. R. 293 Richardson, S. J. 285 Richelsen, B. 757 Richter, E. A. 922 Richter, O. 157 Richter, S. 829 Ricordeau, P. 151 Ridderstråle, M. 1065, 2, 289, 818 Riddle, M. C. 118, 1270, 148, 946, 947, 961, 980 Rietman, A. 775 Rigda, R. S. 530 Rijken, M. 1078, 14 Rinaudo, C. 905 Ringel, B. 249 Ringholm, L. 1105, 1117, 408 Riphagen, I. J. 1219 Riphagen, I. J.. 1237 Riphagen, I. J. 192 Risold, P.-Y. 670 Rissanen, A. 36, 680, 697 Ritterband, L. M. 13 Rittig, K. 768 Ritzel, R. A. 963 Ritzmann, M. 918 Rius, F. 190 Riva, M. 492 Rivas, M. 105 Riveline, J.-P. 1227, 396, 512 Rivero-Arias, O. 273 Rizell, M. 576, 605 Rizza, S. 343 Rizzo, M. 905 Roalfe, A. 302 Robb, P. M. 1170 Robert, J. 325 Robert, M. 227 Roberts, G. A. 303 Roberts, H. 1029 Roberts, L. D. 24 Robertson, N. 341 Robertson, R. P. 500, 529 Robinson, A. 993 Robinson, T. E. 1213 Rocío, S. 657 Rodbard, D. 949 Rodbard, H. 37 Rodbard, H. W. 147, 78, 835 Roden, M. 1172, 168, 249, 415, 442, 557, 581, 589, 592, 760, 790, 806 Røder, M. E. 809 Rodrigues, B. A. 616 Rodrigues, L. S. 1190 Rodrigues, T. 261 Rodríguez, L. 1205 Rodriguez-Calvo, T. 285 Rodríguez-Chacón, M. 438 Rodríguez Cordero, J. 1050 Rodriguez-Murua, J. L. 1222 Rodriguez-Murua, J. L. 1268 Rodríguez-Revuelto, A. 1244 Rogers, E. J. 1140 Rohde, K. 662 Rohde, U. 532 Rohleder, S. 271 Rohm, S. 157 Rohwedder, K. 807, 848, 852 Rojeski, M. 148 Rojo-Martínez, G. 438 Rokitta, I. 1172 Rola-Pleszczynski, M. 691 Roldán, M. 1168 Rolin, B. 1311 Roma, G. 1098 Roma, L. P. 469 Romacho, T. 1320, 692 Romacho, T. 1319 Roman, R. J. 824 Romano, M. 1084 Romero, O. 190 Romito, I. 620 Rondinini, L. 478 Rönn, T. 352, 468 Ropelle, E. R. 616 Ropelle, E. C. C. 616 Rorive, M. 1143 Rorsman, P. 202, 421, 613 Ros, M. 736 Rosa, P. 612 Rosales, M. Aparecida. Brunini. 1164 Rose-John, S. 60 Rosenbauer, J. 1258, 442 Rosenberg, R. 184 Rosenfeld, P. 1153 Rosengren, A. H. 859 Rosengren, A. 306, 318 Rosenkilde, M. M. 923 Rosenlund, S. 198 Rosenstock, J. 146, 241, 242, 244, 37, 4, 42, 793, 926, 948 Rosenthal, T. 801 Rosenzweig, A. 1332 Ross, S. 849 Ross, S. A. 894 Rossi, M. Chiara. 1063 Rossi, M. C. 1066, 154 Rossi, M. 1084 Rossing, P. 1249, 182, 198, 207, 336, 48, 911 Rossmeislova, L. 686 Rosso, C. 237 Rostoka, E. 1321 Rotella, C. M. 1228 Roth, J. 301 Roudaut, M. 850 Round, R. A. 303 Roussel, B. 693 Roussel, R. 1183, 799, 846, 89, 9, 963 Roustit, M. M. 577 Roux, F. 806 Roy-Duval, C. 75, 926 Rozova, E. 1331 Ruan, Y. 960 Rubio-Martín, E. 438 Rudovich, N. 764, 765 Rudovich, N. N. 525 Ruetten, H. 529 Ruetter, R. 442 Rugale, C. 1317 Ruige, J. 269, 288, 629 Ruige, J. B. 763 Ruilope, L. 205 Ruiter, R. 247 Ruiz, M. L. 1055 Ruiz de Adana, M. S. 438 Rukh, G. 143, 79 Ružman, L. 1075 Rune, I. 1311 Rungby, J. 224 Rungby, J. 282, 321 Runkle, I. 1168 Runzis, S. 999 Ruozi, G. 432 Rurali, E. 90 Rusavy, Z. 1132 Rusina, R. 602 Russell, G. 49 Russell, M. A. 172, 358 Russo, E. 85 Russo, G. L. 758 Russo, G. T. 1063 Russo, I. 913 Russo, L. 1084 Russo, R. 85 Rustenbeck, I. 135, 423 Rutten, G. 14 Rutten, G. E.. H.. 1074 Rutten, G. E. H. 1078, 944 Rutter, G. A. 102, 134, 199, 349, 97 Rutters, F. 208, 258, 275 Rutti, S. 431, 432 Ryan, L. 439 Ryan, R. 302 Rydén, L. 44, 45 Rydén, L. E. 1270 Ryhänen, S. J. 330 Ryu, J. 295 Rzajeva, A. 335 S Saada, A. 658, 725 Saari, T. 19, 20, 23 Sabater, M. 664, 687 Sabbion, A. 1261 Sabrautzki, S. 404 Sacchi, F. V. 612 Sacco, M. 154 Sacramento, J. F. 562, 778 Sada, Y. 876 Saely, C. H.. 1211, 1212, 1217, 1218, 1220, 1223, 1238, 1243, 762 Saemann, A. 945 Säemann, M. D. 1272 Safai, N. 1065 Sagarra, R. 266 Sagia, C. 1027 Sagun, G. 449 Sainsbury, A. 777 Saint-Laurent, C. 703 Saisho, Y. 376 Saito, J. 542 Saito, K. 298 Saito, M. 1189 Saito, T. 560 Sakagami, H. 526 Sakaguchi, K. 846, 952 Sakaguchi, S. 437 Sakamoto, K. 1187, 868 Sakamoto, K. 272 Sakamoto, Y. 300 Sakane, N. 635 Sakoda, H. 1277 Sakuma, Y. 925 Sakurada, B. 323 Sakuramoto-Tsuchida, S. 392 Saleh, F. 572 Saleheen, D. 65 Salehi, A. 613 Sales, S. 682, 734 Salic, K. 264 Salminen, P. 121, 678, 738, 83 Salo, H. 284 Salomaa, V. 158 Salonen, K. M. 330 Salonsaari, R.-T. 171 Salsali, A. 806 Salunkhe, V. A. 204 Salunkhe, V. A.. 400 Salunkhe, V. A. 426 Salvati, A. 1201 Salvemini, L. 1235 Salzsieder, E. 1051 Samarasinghe, D. G. 1153 Samczuk, P. 741 Sammeth, M. 204 Samos, M. 43 Sampaio, J. L. 682, 734 Sampedro-Núñez, M. Antonio. 506 Sampol, G. 190 Samuel, V. 861 Sancak, S. 941 Sánchez-Ferrer, C. F.. 1319 Sanchez-Ferrer, C. F. 1320 Sánchez Hernández, R. 1050 Sanchez-Jimenez, F. 571 Sánchez-López, I. 506 Sanchez-Margalet, V. 571 Sancho Bornez, V. 59, 85 Sandbæk, A. 1221 Sandboge, S. 177 Sander, S. 282 Sanders, P. E. 11 Sandholm, N. 107, 51, 87, 90 Sands, A. 803, 862 Sanghvi, A. 820 Sang-Man, J. 1130 Sango, K. 1129 Sano, H. 812 Sano, H. 1260 Sano, H. 437 Sano, M. 902 Sano, Y. 635 Sanoyama, D. 786 Santini, E. 504 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 558 Santos, A. F. 461 Santos, G. J. 424 Sanz, R. 544 Saponara, C. 620 Saponaro, C. 237 Saraheimo, M. 1299, 87 Saravanan, P. 1079, 1095 Sarem, Z. 653 Sargin, G. 655 Sargin, M. 931 Sarma, A. 127 Sarma, A. V.. 128 Sarri, K. 155 Sartorius, N. 1040 Sartorius, T. 714 Sasaki, K. 522 Sasaki, T. 659 Sasson, S. 425 Sastre, J. 1104 Satman, I. 1111, 307 Sato, S. 376 Sato, T. 659 Sato, Y. 834 Satoh, H. 580 Satoh, S. 1269 Saudek, F. 479 Saulnier, P. 9 Saulnier, P.-J. 89 Saunavaara, V. 163, 23 Saunders, R. 221 Saunders, W. 828 Sauque-Reyna, L. 846 Sauter, N. S. 751 Savisto, N. 19, 20 Sawicki, P. T. 1034 Sayers, S. 97 Scanlon, P. H. 1160, 49 Scapoli, C. 344 Scaramuzza, A. 1253 Scaranna, C. 197 Scardapane, M. 1066, 154, 890 Scardo, J. A. 1118 Scatena, F. 373 Schalkwijk, C. 713 Schalkwijk, C. G. 1234, 141, 753, 93 Schall, S. 1041 Schaller, H.-E. 768 Schalling, M. 743 Schanze, N. 583 Schaper, N. C. 900 Schaper, N. C. 96 Scharfmann, R. 381 Schatz, D. 939 Schaupp, L. 979 Scheele, C. 352 Scheel-Thomsen, J. 224 Scheen, A. J. 1143 Schellekens, H. 968 Schellevis, F. G. 1078 Scheltema, M. 533 Scherbaum, W. 940 Scherer, T. 669 Schernthaner, G.-H. 1296, 162 Schernthaner, G. 162, 842, 885 Schiel, R. 1259, 958 Schinner, S. 574 Schleicher, E. 1304, 215, 665 Schleinitz, D. 157, 662, 709 Schliess, F. 934 Schloot, N. 646 Schloot, N. C. 442 Schlosser, M. 334 Schmelzeisen-Redeker, G. 997 Schmid, C. 1017, 785 Schmidt, A. 970 Schmidt, K. 13 Schmidt, L. 1303 Schmidt, R. 413 Schmidt, U. 768 Schmieder, R. E.. 206, 304 Schmit, S. 1043 Schmitt, A. 1023, 1041 Schmitt, H. 1026 Schmitz-Peiffer, C. 720 Schnecke, V. 1245 Schneeberger, M. 674 Schneiderman, N. 504 Schoemaker, M. H. 692 Schoemaker, M. 997 Schoeppe, T. 413 Scholz, M. 144, 347 Schön, M. R. 662 Schouten, J. S. A. 1156 Schram, M. T. 1156, 96 Schrauwen, P. 623 Schrauwen-Hinderling, V. B. 623 Schreiner, B. 768 Schreiweis, H. 126 Schröder, S. 1312 Schroeder, M. 817 Schuling, J. 16 Schulman, K. A. 273 Schultes, B. 124 Schultz, J. 726, 727 Schulz, E. 819 Schulze, F. 614 Schulze, M. B. 270 Schumacher, K. 135, 423 Schürmann, C. 252 Schuster, D. 1026 Schuster-Klein, C. 263 Schütt, M. 646 Schutte, E. 8 Schwartz, F. 1003, 649 Schwarz, J. 775 Schweizer, A. 879, 886, 888, 889, 891 Scirica, B. M.. 183 Scirica, B. M. 186 Scirica, B. M.. 483, 641, 877, 887, 930 Scott, D. 182 Scott, D. A. 1060 Scott, R. A. 830 Scott, R. 1233 Scotti, L. 890 Sealls, W. 40 Seaquist, E. R. 638 Sebastian, D. 674 Sebastiani, G. 231, 29 Sebestova, B. 995 Secher, A. L. 1105, 1117 Secher, N. H. 178 Sechko, E. 371 Sechko, E. A. 362 Seeley, R. 820 Seeman, T. 372 Seferovic, J. 331 Segal-Lieberman, G. 981 Segel, S. 882 Seger, M. E. 935 Seghieri, G. 1062 Segin, V. B. 1125 Segura, J. 205 Seiça, R. 261 Seino, S. 418, 920, 952 Sejling, A.-S. 1329, 632 Sekerija, M. 1028 Sekine, O. 779 Selander, L. 379 Sell, D. R. 92 Sell, H. 1319 Seltmann, A. C. 734 Seltmann, A.-C. 682 Selvarajah, D. 1134, 1136, 131, 132 Sena, C. 261 Sendela, J. 622 Sener, A. 335, 609 Senesi, P. 731, 865 Senmaru, T. 1175, 1200 Seo, C. 513 Seo, J. 295 Seoane, A. 187 Sep, S. J. 96 Seppanen, M. 172 Serban, A. I. 1178 Sereda, S. 201 Sereda, S. B. 175 Sergentanis, T. 1295 Serhiyenko, A. A. 1125 Serhiyenko, V. A. 1125 Serrabulho, L. 1061 Serrano, M. 684, 716 Sert, M. 941 Servitja, J. M. 706 Sestakauskas, K. 947 Seufert, J. 1096, 747, 748 Sewon, O. 1130 Sgadari, A. 620 Shafie, A. 1301 Shaginian, R. 827, 839 Shagiwal, S. 1318 Shah, B. R. 488 Shah, F. Z. M. 1093 Shah, M. 1256 Shah, V. N. 987 Shakeel, N. 293 Shaltout, I. 883 Shamanna, P. 838 Shamkhalova, M. 182, 480 Shankar, A. 310 Shankar, S. S. 500, 529 Shankarappa, M. 1263 Shao, Q. 886, 888, 891 Shao, S. 1322, 26 Shao, W. 919 Sharif, O. 226 Sharipova, J. 1321 Sharma, A. 26 Sharma, S. 1037 Sharoyko, V. V. 398 Sharp, A. 984 Sharp, A. 982, 983 Shaw, J. 1266, 842 Shcherbina, L. 399, 523 She, J.-X. 170 Sheehan, P. 545 Shen, J. 1126, 1127 Shen, L. 1056 Shen, Y. S. 590 Sheridan, P. J. 633 Shestakova, E. A. 607 Shestakova, M. 183, 443, 480, 841, 95 Shestakova, M. V.. 1241 Shestakova, M. V. 607 Sheu, W. H. H. 483 Sheu, W. H.. H.. 930 Shi, Y. 136 Shiba, T. 272 Shibata, H. 626 Shibue, K. 522, 535, 766 Shields, B. 367, 510 Shields, B. M. 1068 Shillo, P. 1134 Shimabukuro, M. 496 Shimada, A. 191 Shimada, F. 914 Shimizu, S. 975, 976 Shimoda, Y. 560 Shimodaira, M. 509 Shimomura, I. 502, 559 Shimomura, K. 202 Shin, J. 47 Shin, J. 1002, 1009, 1012 Shin, J. 652 Shinmura, K. 493 Shinozaki, T. 1209 Shiochi, H. 503 Shirakawa, J. 402, 750 Shirihai, O. 201, 690 Shirihai, O. S.. 175 Shiwa, T. 694 Shoji, K. 503 Shoji, M. 1187 Shojima, N. 1032 Shore, A. C. 1157 Shuichi, O. 797 Shulman, G. 861 Shungin, D. 259 Shuto, Y. 1325 Shymanskyy, I. 1279 Shyshko, V. M. 311 Siddals, K. 351 Siddaramaiah, N. H. 1198 Sidibeh, C. O. 576, 605 Sidot, E. 756 Siebelink, E. 775 Siegel-Axel, D. I. 768 Siewko, K. 286, 337 Sijbrands, E. J. G. 1318 Siklova, M. 686 Siljander, H. T. A. 171 Silva, J. 1202 Silva, K. Cristina. 1164 Silva, M. A. 1128 Silva-Nunes, J. 813 Silveira, A. 223 Silveiro, S. P. 1190 Silvério, M. 261 Silverstein, J. 964 Silvestre, L. 241 Silvestre, M. 262 Silvis, K. 170 Sim, E. 1285 Simar, D. 248 Simell, O. 169, 170, 171, 329, 330 Simell, S. 171 Simes, J. 1233 Simmgen, M. 1139, 1151 Simó, R. 1163, 1244, 190, 53 Simon, C. 642, 648 Simon, M.-C. 442 Simonyi, G. 647 Simpson, P. 621 Simpson, R. W. 916 Singh, B. 993 Singh, H. 13 Singh, M. M. 1047 Singh-Manoux, A. 260 Sinha, V. 957 Sinnott, M. 142, 548 Siow, R. CM. M. 1171 Sipilä, H. 20 Sitkin, I. Ivanovich. 160 Sivaprasad, S. 49 Sivertsen, J. 922 Sjakste, N. 1321 Sjogren, L. 116 Sjöstrand, M. 641 Sjöström, D. 822 Skare, S. 793 Skarupelova, S. 570 Skibová, J. 1146 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 559 Skibova, J. 1148 Skjøth, T. V. 181, 39, 76 Skopkova, M. 369 Skov, L. 788 Skovsø, S. 1311 Skrivarhaug, T. 370 Skriver, M. V.. 1221 Skudelny, I. 126 Skupien, J. 1115, 368 Skyler, J. S. 149, 504 Skytt, D. 671 Slagter, S. N. 1031 Slagter, S. N.. 283 Slamova, Z. 372 Slee, A. 826 Slover, R. 1003, 1252, 649 Sluijs, I. 770, 772 Smahelova, A. 483, 877 Smart, M. 236 Smiley, D. 1038, 114 Smircic Duvnjak, L. 1281 Smirčić-Duvnjak, L. 1161 Smirnova, O. 441 Smith, D. M. 409 Smith, F. E. 585 Smith, G. 1332 Smith, J. 1306 Smith, M. 860 Smith, S. 309 Smith, U. 685, 712 Snell-Bergeon, J. K. 987 Snorgaard, O. 268, 809 So, A. 952 So, W. Yee. 1182 So, W. 1073 So, W.-Y. 279 So, W. 1045 Soares, A. F. 579 Soares, A. A. 1190 Sobngwi, E. 1227, 512 Soejima, E. 719 Soeters, M. R. 21, 698 Søgaard, M. 789 Sogayar, M. C. 461 Soininen, P. 1283 Soinio, M. 678, 738 Sojka, S. 748 Sokolovska, J. 1321 Solà, C. 1163 Solecka, I. 361 Soleymanlou, N. 956 Solin, O. 446 Solini, A. 1201, 210, 305, 86 Solis, C. 552 Soma, Y. 427 Sommer, C. 293 Sommer, C. 479 Somogyi, A. 444 Son, H. 858 Son, H.-S. 652 Son, H.-S. 547 Son, J. 47 Son, J. 933 Son, S. 1285 Son, S. 1193 Sone, H. 1215, 298, 630 Sonestedt, E. 62 Song, D.-K. 593 Song, I. 591 Song, K.-H. 10 Song, M. 416 Song, M. 35 Song, S. 100 Songini, M. 451 Sonne, D. P. 532, 533 Sookja, C. K. 1129 Sørensen, H. Toft. 282 Sørensen, H. T. 321 Soresi, M. 905 Sorhede-Winzell, M. 409 Sorice, G. Pio. 619 Sorice, G. P. 620, 783 Sorice, G. 515, 524 Soro-Paavonen, A. 1299 Sosa, F. 899 Soska, J. 1144 Sosner, P. 9 Sotiropoulos, A. 1246 Sotnikova, R. 1294 Souhami, E. 241, 829 Soula, G. 932, 934 Soula, O. 932, 934 Soula, R. 932, 934 Soupal, J. 1305 Søvik, O. 370 Spadaro, S. 1235 Spalova, I. 998 Spanoudi, F. 561, 564, 565 Sparre-Ulrich, A. H. 923 Sparwasser, A. 525 Spat, S. 979 Specht, L. 248 Speck, M. 60 Spégel, P. 350, 375, 393, 398, 399 Sperrin, M. 246, 294 Spiliotis, I. I. 202 Spinelli, R. 354 Spiropoulou, P. 1240 Spoletini, M. 451, 742 Spoto, B. 1235, 343 Spranger, J. 653 Spraul, M. 1154 Srankova, J. 1293, 1294 Srinisavan, D. K. 495 Srinivasan, P. 476 Sriwijitkamol, A. 556 Stabilini, A. 231 Stadler, M. 1007, 194, 989 Stadler, P. F. 347 Staels, B. 609 Stafylas, P. 988 Stager, W. 841 Stahl, D. 309 Ståhle, M. 474 Stahl-Pehe, A. 1258 Stahr, A. 415 Stahre, C. 483, 877, 887 Staiger, H. 721 Stamenkovic, J. 398 Stanarcic, J. 331 Stanca, L. 1178 Stangé, G. 253 Stanik, J. 369 Stark, R. 672 Starup-Linde, J. 224 Stasko, J. 43 Staszel, T. 360, 368 Staten, M. 529 Stechova, K. 366, 998 Steckel-Hamann, B. 1299 Stefan, N. 435, 721, 768 Stefanovski, D. 500, 529 Stefanowicz, M. 176, 211 Stefansson, B. 315 Steg, P. 186 Stehouwer, C. 713 Stehouwer, C. D. A. 1156, 1234, 141, 275, 63, 753, 792, 795, 93, 96 Steigleder-Schweiger, C. 232 Stender-Petersen, K. 644 Stenlöf, K. 5 Stepanova, E. 443 Stermann, T. 419 Steven, S. 819 Stevens, R. J. 1203 Stewart, J. 1070 Stewart, M. W. 41, 830, 831, 837, 838 Stewart, R. 316 Stich, V. 686 Stienstra, R. 755 Stier, C. 126 Stiles, L. 175 Stoico, V. 511 Stolarczyk, E. 701 Stolz, K. 749 Stoop, R. 1324 Storaas, T. 618 Storgaard, H. 856, 922 Storkholm, J. 540 Storm, P. 173, 336, 375, 401 Strachan, D. 106 Straczkowski, M. 176, 211 Straetener, J. 747 Strandberg, C. 1113 Stratmann, B. 1302, 69 Stratton, I. M. 1160, 303, 49 Strawbridge, R. J. 223 Strazhesko, I. 95 Strazhesko, I. D.. 1241 Streckel, E. 918 Strevens, H. 1100, 152 Stricker, B. H. 247 Strindberg, L. 116 Strinović, M. 1075 Strock, E. 1005, 829 Stroek, K. 218 Strojek, K. 887 Strom, A. 249 Strotmann, H.-J. 895 Strumph, P. 803 Stumvoll, M. 144, 347, 662, 709, 761, 888, 891 Sturkenboom, M. 890 Su, H. 643 Suenens, K. 253 Suga, T. 626 Suganami, H. 663 Sugano, T. 573 Sugawara, M. 634 Sugaya, Y. 580 Sugg, J. 805 Sugihara, H. 1325, 708, 786 Sugimoto, T. 1229 Sugino, Y. 1206 Sugoka, O. 1321 Suhai, F. I. 80 Suissa, S. 892 Suleiman, M. 373 Sullivan, D. 5 Sullivan, D. 1233 Sullivan, M. A. 594 Sulpice, T. 802 Sultan, A. 586 Sumi, K. 503 SUMMIT, 158 SUMMIT Consortium, 107 Sumnik, Z. 366 Sun, A. V. 515, 524 Sun, G. 97 Sun, H. 1135, 666, 72 Sun, H. 1184 Sun, J. 350 Sun, J.-H. 38 Sun, V. A. 620, 783 Sun, W. 92 Sun, X. 1177, 1297 Sun, Z. 1056, 1057, 1059, 1184 Sunagawa, S. 496 Sund, R. 245 Sundquist, K. 319 Sunkari, V. G.. 250 Suomalainen, A. 680 Supale, S. 1280 Suraci, C. 1063 Susano, J. 1061 Suto, C. 1158 Suuronen, E. 46 Suwansaksri, N. 556 Suzuki, A. 630 Suzuki, E. 454 Suzuki, K. 522, 535, 766 Suzuki, Y. 1194, 925 Svacina, S. 912 Svare, J. A. 1113 Sveen, K. 1216 Svehlikova, E. 124 Sveinsson, T. 63 Svendsen, C. B. 833 Svendsen, M. L. 789 Svensson, A.-M. 196, 245, 292, 306, 317, 318, 320 Svensson, A.-M. 1208 Svensson, E. 282, 321, 789 Svensson, J. 1250 Svensson, M. K. 352, 576, 605 Svensson, P.-A. 576 Svietleisha, T. 345 Svojanovsky, J. 1169 Swisa, A. 97 Syed, F. 373 Syrén, A. 182 Szabat, M. 385, 705 Szabo, M. 746 Szamatowicz, J. 1090 Szamatowicz, M. 1090 Szatkowski, L. 302 Szelachowska, M. 286, 337 Szendroedi, J. 442, 581, 589 Szepes, M. 746 Szepessy, E. 253 Szeto, C. Chun. 1182 Szopa, M. 360, 361, 368 Szyllo, N. 1159 Szypowska, A. 1021, 622, 977, 986 T Tabak, A. G. 1085 Tabák, A. G. 61 Tabakov, K. 996 Tack, C. J. J. 447 Tacke, C. 764 Tacke, F. 185 Tadler, F. 124 Tagaya, Y. 560 Taghizadeh, F. 458 Taguchi, K. 659, 966 Taguchi, T. 505 Tahara, Y. 834 Taher, J. 1308, 921 Tahhan, R. 482 Taittonen, M. 19 Tajan, M. 703 Tajima, K. 1194 Tajima, K. 750 Tajima, N. 1260 Tajiri, Y. 719, 794 Takabe, M. 952 Takada, S. 626 Takagi, A. 724 Takagi, M. 1191 Takahashi, A. 108 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 560 Takahashi, H. 920 Takahashi, I. 427 Takahashi, K. 1158 Takahashi, K. 635 Takahashi, K. 967 Takahashi, M. 870 Takahashi, S. 427 Takao, T. 209 Takasawa, S. 392 Takata, H. 505 Takatani, T. 203 Takayama, C. 496 Takayama, M. 1206 Takayanagi, R. 766 Takebe, N. 967 Takeda, M. 392 Takeda, Y. 526 Takemoto, D. 626 Takemoto, M. 1187, 1215, 868, 88, 914 Takemura, S. 1191 Takiyama, Y. 1176, 526, 536 Takizawa, M. 1264 Tam, C. H. T. 1182 Tamanaha, T. 161, 598 Tamarit-Rodriguez, J. 671 Tamayo, T. 1214 Tambascia, M. 841 Tamborlane, W. 939 Tamer, G. 449 Tamer, I. 449 Tamir, M. 453 Tamura, K. 920 Tamura, Y. 1269 Tamura, Y. 1209 Tamura, Y. 578 Tamvakos, E. 1246 Tan, A. T. B. 1093 Tan, B. M. 98 Tan, P. C. 1093 Tanabe, K. 365 Tanaka, H. 496 Tanaka, K. 719 Tanaka, M. 1175 Tanaka, N. 1191 Tanaka, Y. 876 Tang, J. 808, 811, 898 Tang, J. 1056 Tang, J. 1327 Tang, M. 25 Tang, Y. 859 Taniguchi, S. 503 Tanimoto, k. 1274 Tanizawa, Y. 365 Tankova, T. 1122 Tapp, R. 312 Tarasov, A. I. 421 Tarnoki, A. D. 80 Tarnoki, D. L. 80 Tarnow, L. 1124, 632 Taskinen, M.-R. 1233 Tatsuoka, H. 884 Tauschmann, M. 1007, 960 Tavakoli, M. 1155 Tavares, I. F. 1128 Tavernier, J. 428 Taylor, H. S. 572 Taylor, R. 585, 627 Tega, C. 427 Teichert, L. 953 Teixeira, D. 225 Tejedor, N. V. 106 Telejko, B. 1090 Tellechea, A. 251 Tellez, N. 387 Templeman, N. M. 499, 704 Teng, G.-J. 188 Tengholm, A. 395, 420, 610 ten Kulve, J. S. 81 Tennagels, N. 970 Tentolouris, N. 1150, 1288 Tentolouris, N. 1064 Teoh, H. 877 Terabayashi, T. 422 Terada, Y. 505 Terasaki, J. 1274, 437, 611 Terasaki, M. 871 Terasaki, M. 915 Terauchi, Y. 1269, 272, 402, 470, 750, 976 Terkildsen, R. S. 1065 Terra, L. F. 430, 461 Terruzzi, I. 625, 731, 865 Teruyama, K. 418 Terzi, T. 603 Teschemaker, A. 817 Tesfaye, S. 1134, 1136, 131, 132 Tesic, D. S. 1150 Teslovich, T. 105 Teslovich, T. M. 104 Testa, A. 343 TEST-DM1 study group, 1224 Tetreault, J.-C. 1207 Teuho, J. 20 Te Velde, S. J. 275 Tez, D. K. 1198 Thabit, H. 1007, 194, 989 Thalange, N. 964 Thamer, C. 849 ‹t Hart, L. M. 139, 341, 342 ‘t Hart, L. M. 339 Thayer, S. 111 the CIMT trial group, Copenhagen university hospit, a. 1271 The DCCT/EDIC Research Group, 92 the Diabetes Medication Choice Cards Trial in Gree, c. 1077 the DIAGRAM consortium, 140 the DIRECT Consortium, 326 the FIELD Study, 1233 Theilade, S. 1249 the IMPROVE study group, 223 the Japanese Elderly Diabetes Intervention Trial S, t. 1209 the J-PREDICT study investigators, 272 Thelwall, P. 585 Theodorakis, M. J. 281 Theodorou, K. 988 Theodosis-Georgilas, A. 1240 the TEDDY Study Group, 170 The TURDEP Study Group, 307 Theurl, E. 1039 Thevenet, J. 609 Thiemann, S. 894 Thienel, C. 751 Thoma, C. 627 Thomas, M. K. 11 Thomopoulos, A. 782 Thompson, G. 817 Thompson, J. C. 839 Thomsen, R. W. 282, 321, 789 Thomsen, S. 359 Thomson, H. 1044 Thorand, B. 760 Thorel, F. 614 Thorén Fischer, A.-H. 490, 492 Thorens, B. 102, 349, 97 Thornalley, P. J. 1301, 1306, 558, 68 Thorsteinsson, B. 1042, 1329, 485, 632, 809, 917 Thorup, A. C. S. 784 Thourani, V. 114 Thrasher, J. 1003, 649 Thuan, J.-F. 1242 Thurnheer, M. 124 Thybo, T. 548 Tiberti, C. 451, 742 Tiedge, M. 1312, 413, 727, 728 Tiemann, J. 1302 Tiemann, T. 484 Tiengo, A. 197 Timlin, L. 221 Timmer, A. E. 1219 Timper, K. 751 Tinahones, F. 1108 Tinahones, F. J. 687 Tinnes, K. 168 Tkacheva, O. 95 Tkacheva, O. N.. 1241 Tobe, K. 272 Tochiya, M. 161 Toda, K. 1181 Todorov, V. T. 821 Toeller, M. 732 Tofeec, K. 1116 Tofe Povedano, S. 40 Togashi, Y. 750 Tognoni, G. 154 Tokubuchi, I. 794 Tokuyama, H. 1187, 868, 914 Tolvanen, T. 20 Tomasselli, F. 1261 Tomé, D. 775 Tome, M. 1108 Tomesova, J. 1132 Tomić, M. 1161 Tomic, M. 1281 Tomita, M. 191 Tommasi, E. 18 Tommasi, M. 1261 Tomoichiro, A. 1277 Tomono, J. 1133 Tønder, N. 1329 Tong, C. 847, 853, 855 Tong, P. C. Y. 1182 Toni, S. 1253 Tonietti, M. 864 Toniolo, A. 285 Tönjes, A. 144, 347, 761 Toorawa, R. 815 Top, W. M. C. 792 Topolcan, O. 771 Toppari, J. 329 Toppila, I. 51 Torlone, E. 1063 Tornoczky, J. 1085 Tornøe, K. 37 Török, A. 1254 Torsvik, J. 364 Tortosa, F. 464 Toscano, V. 1098 Toumazou, C. 1044 Toumi, T. 174 Tournier, M. 756 Tovar, S. 166 Tołwińska, J. 195 Townsend, R. 850 Toya, K. 1191 Toyoda, Y. 520 Toyonaga, A. 1191 Trahair, L. G. 530, 538 Traina, S. 826 Trautmann, M. 845, 866, 933, 972 Trautmann, M. E.. 844 Trautwein, C. 185 Traylor, L. 968 Treebak, J. T. 103 Treguer, K. 703 Treiber, G. 232 Tremoli, E. 223 Trenell, M. I. 621, 627 Trevisan, R. 197, 210, 305, 86 Triantafyllidi, H. 1300, 94 Trico, D. 527 Trifone, L. 864 Tripathy, D. 552, 722, 74 Tripepi, G. 343 Triplitt, C. 74 Trischitta, V. 1235, 343, 357 Trnovska, J. 71 Trombetta, M. 511 Troy-Fioramonti, S. 711 Trucco, M. 385, 705 Trudeau, K. 201, 690 Trustrup, U. 1065 Tryggvason, K. 88 Tsapaeva, N. L. 311 Tsapas, A. 1077 Tschirhart, H. 956 Tschoepe, D. 1302, 69 Tschöpe, D. 206, 304 Tsuboi, K. 1129 Tsuchida, C. 392 Tsui, S. K. W. 1182 Tsujii, S. 313 Tsujinaka, H. 392 Tsujino, D. 513 Tsuprykov, O. 1286, 1330 Tsutsui, H. 626 Tsutsumi, C. 1274, 437 Tsuzaki, K. 635 Tu, Y. 863 Tuca, A. 124 Tulipani, A. 527 Tumini, S. 1253, 964 Tummers, M. 679, 697 Tunceli, K. 322, 898 Tuomi, T. 173 Tuomilehto, J. 271, 307 Tuominen, L. 83 Tups, A. 672 Tura, A. 1112, 1272 Turatsinze, J.-V. 204 Turiel, D. 982, 983, 984 Turker, F. 1111 Tutuncu, Y. 307 Tuulari, J. J. 177, 678, 738, 83 Tuzcu, M. 1216 Twig, G. 690 Tyberghein, M. 1143 Tyedmers, J. 1303 Tykhomyrov, A. 1165 Tyrwhitt, J. 1270 U Übel, N. 918 Uçak, S. 1046 Uchigata, Y. 1158, 1191, 1264, 50 Łuczyński, W. 195 U Din, M. 20 Ueda, R. 496 Ugi, S. 724, 779 Ukropcova, B. 167, 180 Ukropec, J. 167, 180 Šulentić, P. 1075 Ulivi, A. 612 Ullrich, S. 404, 405, 435, 768 Uma, R. 1082 Umanath, K. 8 Umez-Eronini, A. 186 Umpierrez, G. 843 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 561 Umpierrez, G. 1038, 182 Umpierrez, G. E. 114 Undlien, D. E. 370 Ungaro, P. 758 University of Helsinki, 680 Unnikrishnan, R. 1082 Uno, S. 502 Uphues, I. 97 Urbanova, J. 810 Ursing, D. 1100, 152 Ursoleo, P. 656 Ushigome, E. 1200 Usiskin, K. 5 Usman, K. 1047 Ustunomiya, K. 513 Usui, T. 659, 966 Utsunomiya, K. 1260, 300 Uusitalo, U. 170 Uzu, T. 724 Uzum, A. Kubat. 1111 V Vaag, A. 282, 321, 352, 974 Vaarala, O. 1188, 284 Vacinova, G. 602 Vadas, M. 165 Vaish, A. Kumar. 1047 Valabhji, J. 1153 Valensi, P. 1236, 1275, 508, 660, 901 Valente, L. 448 Valentine, V. 854 Valentine, W. 221 Valentine, W. J. 17 Valentini, M. 29 Valentini, U. 197 Valentinova, L. 369 Valet, P. 703 Valikovics, A. 1254 Vallejo, M. 378 Valle Morales, L. 1101 Vallon, V. 824 Vallova, S. 180 Valsamakis, G. 782 Valtat, B. 133 Valtat, B. 393 Van Baal, J. G.. 1154 van Beek, A. P. 1031 van Bloemendaal, L. 81 Van Brunt, K. 907 van Buuren, S. 1324 Van Dalem, A. 288, 452 van de Bunt, M. 359 Vandemeulebroucke, E. 288 Vandenhende, F. 1299 van den Hoek, A. M. 1324, 264 van der Berg, J. D. 63 van der Heijden, A. A. 14 van der Heijden, A. A. W. 1080, 208, 258 van der Heijden, A. A.. W. 942 van der Kallen, C. 96 van der Kallen, C. J. H. 1234, 141 van der Klauw, M. M. 1031 van der Klauw, M. M.. 283 van der Reest, J. 755 van der Schans, C. P. 795 van der Stelt, I. 558 Van de Velde, U. 452 van Dijk, P. R. 193, 314 van Eijck, C. H. J. 247 van Ewijk, P. A. 623 van Greevenbroek, M. M. J. 1234, 141 Vanhaecke, T. 763 van Hall, G. 122 van Hateren, K. J. J. 1237, 192 van Hoek, M. 1318 Vankova, M. 1091, 602 van Leeuwen, N. 342 Vannasaeng, S. 556 Van Netten, J. J. 1154 Van Nieuwenhove, Y. 763 Van Noord, M. 271 van Puffelen, A. L. 1078 van Schothorst, E. M. 558 van Sloten, T. T. 96 van Tol, E. A. F. 692 van ‹t Riet, E. 258 van Vliet-Ostaptchouk, J. 1031 van Vliet-Ostaptchouk, J. V.. 283 van Wegberg, J. J. G. 753 Van Zuydam, N. 90 Van Zuydam, N. R. 107, 158 Varadhan, L. 1088 Varadhan, L. 906 Varas-Lorenzo, C. 890 Varga, L. 369 Varnhorn, A. 109 Varol, N. 839 Varoudi, M. 1300, 94 Varughese, G. I. 906 Vas, P. 1151 Vas, P. R. J. 1139 Vasantharajah, L. 66 Vasgiourakis, V. 631 Vasileiou, V. 1086 Vasques, A. C. J. 123 Vassilaki, M. 155 Vassilatou, E. 564, 565 Vassileva, M. T. 500, 529 Vastolo, V. 758 Vasylkova, H. 345 Vaughan, J. M. 577 Vaughn, D. E. 149 Vaur, L. 640 Vaz, J. A. 883 Vázquez-Carrera, M. 53 Vcelak, J. 1091, 602 Vcelakova, J. 366 Vecsei, Z. 604 Vedovato, M. 197, 210, 305, 86 Vedtofte, L. 1113 Vega Guedes, B. 1101 Veganzones, M. 1104 Veglia, F. 223 Veijola, R. 169, 171, 329, 330 Vejlstrup, N. 1329 Vejlstrup, N. G. 917 Vejrazkova, D. 1091, 602 Velho, G. 1183, 89, 9 Veliky, M. 1279 Vella, A. 529, 567 Veltman, D. J. 81 Vendrell, J. 1244 Venn, K. 955 Ventriglia, G. 231, 29 Venturi, M. 1142 Veras, K. 469 Verberne, H. 218, 22 Verberne, H. J.. 21, 698 Verchere, C. B. 254 Vereşiu, I. A. 1036, 1137 Veret, J. 465 Vergès, B. 711 Verhaeghe, J. 1110 Verhaeghen, K. 452 Verhoeven, A. J. M. 1318 Verlekar, P. 1140 Vernhes, C. 970 Vernochet, C. 1332 Vernon, W. 409 Verras, C. 1027 Versteyhe, S. 235 Vestergaard, H. 809 Vestergaard, P. 224 Vethakkan, S. R. 1093 Vetter, M. 244 Vettor, R. 125 Veves, A. 251 Viana, R. 813 Viani, P. 465 Viberti, G. 66 Vician, M. 167, 180 Vicuña, A. 506 Vidal, H. 227, 570 Vidal, J. 1064, 1205 Vidal, M. 716 Vidal, M. 636 Vidal Flor, M. 1035 Vieira, E. 699 Vietri, J. 1072 Vigier-Simmore, N. 1242 Vijapurkar, U. 854 Vikman, J. 426 Vikman, P. 336 Vila, I. K. 32 Vila, L. 1224 Vilariño-García, T. 571 Vilaseca, M. 387 Villalobos, L. 1320 Villalobos, L. A.. 1319 Villard, M. 412 Villaró, M. 628 Villate, O. 204 Villena, J. E.. 887 Vilman, P. 57 Vilmann, P. 519 Vilsbøll, T. 1113, 243, 519, 532, 533, 539, 540, 57, 606, 788, 835, 836, 856, 922, 923 Vilsboll, T. L. 541 Vincent, M. 110, 113, 940, 943, 973 Vinet, L. 961, 963 Vinik, A. I. 1137 Vinik, E. J. 1137 Violante Ortiz, R. 73, 833, 881, 903, 904 Viriato, D. 813 Virtanen, K. 19, 23 Virtanen, K. A. 20, 219, 697 Virtanen, S. M. 170 Visa, M. 706 Vistisen, D. 61 Vistoli, F. 478 Vitai, M. 604 Vitali, A. 688 Vitali, F. 1245 Vitolo, E. 1201 Vittas, S. 1314 Vlachopoulou, E. 158 Vlachová, Z. 1121 Vlajnic, A. 1070, 968 Vlcek, M. 180 Vogt, L. 1051 Vojtek, J. 995 Volchuk, A. 457 Volk, C. B. 971 Volk, N. 1313, 1314 Volkov, P. 234, 468 Vomscheid, M. 703 Vonbank, A. 1211, 1212, 1217, 1218, 1220, 1223, 1238, 1243, 762 Vondrova, H. 479 von Düring, M. 296 von Eynatten, M. 823, 894 von Gall, C. 415 von Herrath, M. 285 von Loeffelholz, C. 765 Vons, C. 1275 von Scholten, B. J. 48 von Scholten, B. 911 von Websky, K. 1286, 1330 Voorham, J. 16 Vora, J. 541, 836, 845 Vos, R. C.. 1074 Vos, R. C. 944 Voshol, P. J. 60 Voska, L. 479 Vrakas, S. 1027 Vrang, N. 875 Vrkljan, M. 1075 Vryonidou, A. 603 Vrzalova, J. 771 Vucic Lovrencic, M. 1028 Vulesevic, B. 46 W Waage, C. W. 293 Waagepetersen, H. 671 Wada, N. 719 Wada, Y. 884 Wadden, T. 184 Wade, J. D. 520 Wagenknecht, D. 126 Wägner, A. M. 1101 Wägner, A. 1050 Wagner, O. 226 Wagner, R. S. 1014 Wahlgren, C. D. 519, 57 Wailemann, R. A. M. 461 Wainstein, J. 1069, 637, 776 Walder, K. 707 Waldron, S. 1251, 1257 Walji, S. 488 Walker, A. B. 906 Walker, A. W. 701 Walkey, H. 290 Walkinshaw, E. 1007, 194 Wallström, P. 62 Walraven, I. 208, 258, 942 Walshe, K. 1008 Walter, T. L. 70 Walters, S. N. 229, 98 Walus-Miarka, M. 1310 Wan, G. 1267 Wan, L. Li. 791 Wandall, E. P. 519 Wang, C.-C. 1197 Wang, C.-Y. 1210 Wang, C. 769 Wang, D. 1327 Wang, H. 543 Wang, H. 239 Wang, H. 486, 639 Wang, H. 827 Wang, J. 103 Wang, P. 1135, 666, 667, 676, 72 Wang, Q. 1179, 1180 Wang, R. 27, 28 Wang, S.-H. 675 Wang, S. 1135, 1184, 666, 667, 676, 72 Wang, W. 165, 816 Wang, W. 816 Wang, X. 383 Wang, X. 156 Wang, Y. 410, 643 Wang, Y. 1184 Wang, Y. 1182 Wanke, R. 918 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 562 Wan Mohamud, W. 355 Ward, G. M. 1001 Ward, T. 1067 Wardęcki, M. 961 Wardecki, M. 946 Warner, J. 939 Warren, F. J. 594 Washer, S. L. 91 Watada, H. 578 Watanabe, A. 25 Watanabe, T. 580 Watt, T. 1105 Wawrusiewicz-Kurylonek, N. 1090, 356 Weaver, J. 1038 Weber, M. A. 6, 825 Weber-Born, S. 815 Webers, C. A. B. 1156 Weets, I. 288, 452 Wegscheider, K. 732 Wei, C. 183, 483, 930 Wei, W. 1058, 111 Weickert, M. O.. 653 Weiderpass, E. 772 Weigert, C. 665, 689 Weill, A. 151 Weiner, R. 126 Weinstock, R. 1003 Weir, G. C. 529 Weiss, H. 413, 728 Weiss, R. 1003, 649 Weissman, P. N. 837 Weißmann, J. 1015 Weitz, E. 126 Weksler-Zangen, S. 658, 725 Welch, A. J. 1140 Wells, B. J. 323 Welschof, P. 819 Welsh, J. 1003 Welters, H. J. 384, 409 Welungoda, I. 916 Wendt, A. 400, 426 Weng, J. ping. 1123 Weng, J. 1040, 287, 546, 710, 919 Weng, W. 323 Wenzlau, J. M. 452 Werzowa, J. 1272 Wessells, H. 128 Wessels, H. 127 West, D. J. 621 Westenfelder, C. 389 Wetzels, S. M. W. 753 Whelan, M. 573 White, K. E. 1170 White, W. B. 878 Whiting, D. 308 Whiting, M. 1233 Widdop, R. E. 916 Wiebe, J. C. 1101 Wielinga, P. Y. 264 Wierup, N. 133, 375, 399, 490, 492, 523, 740 Wierusz-Wysocka, B. 1022, 1144 Wietzke, M. 728 Wijesuriya, M. 66 Wijnands, E. 753 Wikstrom, J. D. 690 Wilbrink, M. F. J. 141 Wild, S. 245 Wilding, J. 847 Wilding, J. P. H. 73, 833, 881, 903, 904 Wilinska, M. E. 194, 960 Wilk, G. 361 Wilke, T. 12 Wilkens, L. R. 270 Wilkinson, I. D. 1134, 1136, 131, 132 Willaing, I. 1048 Willenborg, M. 135 Williams, A. J. K. 334 Williams, A. J. K. 450 Williams, A. S. 1141 Williams, J. D. 982 Willis, J. 701 Willis, M. 817 Wilms, B. 124 Wilson, B. R. 150 Wilson, C. 878, 897 Wilson, I. V. 450 Wing, Y.-K. 1040, 279 Winkler, H. 1220 Winner, H. 1039 Winship, S. 1198 Winter, W. E. 334 Wirfält, E. 62 Witek, P. 1115 Witte, D. R.. 61 Wlazlo, N. 141 Woerle, H. J. 1, 2, 806, 818, 823 Woerle, H. J.. 849 Woerle, H. J. 851, 956 Woerle, H.-J. 815, 885, 894 Wohland, T. 484 Wojtusciszyn, A. 477 Wolf, E. 918 Wolf, G. 1024, 945 Wolf, G. 301 Wolf, M. 665 Wolf, X. A. 1311 Wolfert, R. 142 Wolffenbuttel, B. H. R. 1031, 283 Wolffenbuttel, B. H.. R.. 8 Wolfrum, C. 167 Woliner-van der Weg, W. 447 Wolkow, P. 368 Wollheim, C. 399 Wollheim, C. B. 398 Woloschak, M. 816 Won, J. 1192 Wong, R. 1045 Wong, S. 450 Woo, C. 1290 Woo, J.-.Taek. 280 Woo, J.-T. 858 Woo, M. 382, 383 Woo, V. 42 Woo, V. C. 78, 835 Wood, J. 324 Worbes-Cerezo, M. 1072 Workman, C. 139, 339 Wosková, V. 1146 Woskova, V. 1148 Wouters, K. 753 Wozniakiewicz, E. 1310 Wright-Hughes, A. 1152 Wroblewski, V. J. 938 Wronkowitz, N. 1319 Wu, G. Y. 971 Wu, H.-T. 582 Wu, J.-S. 582 Wu, L. 1179 Wu, Q. 672 Wu, T. 217, 774 Wu, X. 1266, 382, 383 Wu, X. W. 149 Wulff, B. S. 534 Wulff, J. 142 Wurst, U. 761 Wybranska, I. 360, 368 Wygant, G. 850 Wyne, K. 590 Wysham, C. H. 244 X Xia, M. 277, 863 Xia, P. 165 Xia, Q. 348 Xia, W.-Q. 675 Xia, W. 1135, 666, 667, 676, 72 Xiang, A. H. 156 Xie, R.-R. 1267 Xifra, G. 683, 684, 687, 716 Xiong, J. 552 Xu, F. 546, 710, 919 Xu, G. 665, 689 Xu, H. 546, 710 Xu, J. 1315 Xu, L. 279 Xu, L. 752 Xu, M. 328 Xu, S. 614, 751 Xu, W. 1123, 546 Xu, Y. 1266, 382, 383 Xue, M. 328 Xue, M. 1301, 68 Xue, Y. Ming. 643 Xueyao, Y. 276 Y Yagihashi, S. 1129 Yagüe, I. 624 Yamada, E. 560 Yamada, K. 719, 794 Yamada, M. 560 Yamada, S. 1181 Yamada, S. 505 Yamada, T. 630 Yamada, T. 376 Yamada, T. 1032 Yamada-Fowler, N. 1306 Yamaga, M. 1187 Yamaguchi, T. 1229 Yamamoto, K. 503 Yamamoto, K. 1265 Yamamoto, N. 503 Yamane, S. 522, 535, 766 Yamaoka, M. 422 Yamashita, S. 870 Yamashita, S. 663 Yamauchi, A. 392 Yamauchi, M. 966 Yamazaki, M. 1175, 1200 Yamazaki, T. 272 Yan, H. 863 Yan, H.-M. 277 Yan, J. 287, 919 Yan, K. 88 Yan, P. 239 Yanagimachi, T. 1176, 526, 536 Yanagisawa, H. 209 Yang, B. 1184 Yang, D. 287 Yang, F. 41 Yang, G. 239, 543, 588, 702, 769 Yang, G. 136 Yang, G.-R. 1267 Yang, H. zhang. 1123 Yang, J. 170 Yang, L. 274 Yang, M. 816 Yang, S.-N. 136 Yang, W. 1040 Yang, X. bin. 1123 Yang, Y. 1322 Yang, Y. 1056, 1059 Yang, Y.-C. 582 Yang, Y.-S. 1081 Yang, Y. 458 Yang, Y. 1135, 666, 667, 676 Yao, B. 287 Yao, X. 863 Yaochite, J. U. 471 Yaqoob, M. M. 262 Yarborough, O. 394 Yart, A. 703 Yavin, Y. 798, 800 Ye, C. 955 Ye, J. 830, 831, 837, 908 Yee, J. 847 Yengo, L. 233 Yenidunya, G. 1111 Yeo, K. 965 Yeung, R. 1040, 1045 Yeung, R. O. 1073 Yi, D. 1285 Yilmaz, O. 1111 Yilmaz, T. 307 Yim, C.-B. 446 Yin, S. Nan. 410 Yki-Järvinen, H. 718, 946, 961, 963 Yki-Järvinen, H. 980 Yılmaz, M. 941 Yoganathan, P. 60 Yokota, T. 626 Yokote, K. 1187, 1215, 663, 868, 88, 914 Yoldi, C. 636 Yoneda, C. 1194 Yoneda, M. 1277, 694 Yoneda, S. 502 Yoo, H. 295 Yoo, S.-Y. 514, 549 Yoo, S. 47 Yoon, K. 858 Yoon, K.-H. 1196 Yoshida, E. 402 Yoshida, N. 1191 Yoshida, N. 1191 Yoshida, S. 925 Yoshikawa, A. 502 Yoshiki, M. 902 Yoshimura, K. 505 Young, A. 983 Young, A. A. 982 Young, B. 1147 Younis, F. M. 801 Yu, C. 323 Yu, C. 978 Yu, D. 276 Yu, J. Y. 1118 Yu, J. 295 Yu, J. 136 Yu, L. 136 Yu, M. 907 Yu, M. 328 Yu, Q. 420, 610 Yu, S. 12 Yuan, L. 702 Yuan, L. 1056, 1059 Yuan, M.-X. 1267 Yuan, S.-Y. 1267 Yuan, T. 429 Yuan, X. 719 Yuan, Y. 1184, 666, 667 Yubero, S. 562 Yuen, K. 118 Yun, J.-S. 10 Yusuf, S. 1270 Diabetologia (2014) 57:[Suppl1]S1–S564 1 C S 563 Z Zabeo, E. 125 Zacho, J. 145, 147, 931 Zaffani, S. 1261 Zafon, C. 190 Zagar, A. 1026 Zairi, M. 1086 Zakaroff, A. 225 Zalesakova, H. 995 Zambrowicz, B. 803, 862 Zammit, G. 184 Zammit, N. W. 229, 98 Zampetti, S. 451, 742 Zander, A. R. 389 Zander, M. 809 Zanolin, D. 1211, 1212, 1217, 1218, 1220, 1223, 1238, 1243 Zantidis, A. 7 Zaoui, P. 9 Zapala, B. 360, 368 Zapanti, E. 1086 Zardi, F. 511 Zarra, E. 197 Zatterale, F. 354 Zavaroni, I. 745 Zbucka-Kretowska, M. 1090 Zdunczyk, B. M. 1021 Zee, P. 4 Zelada, M. 506 Zeller, C. 2, 818 Zeng, H. 1127 Zeng, K. 546, 919 Zeng, M. 863 Zeniya, M. 300 Zerbini, G. 86 Zerva, K. 631 Zethelius, B. 1208, 306, 317, 320 Zhang, D. Dan. 791 Zhang, D. 861 Zhang, E. 467 Zhang, H. 382, 383 Zhang, J. 26 Zhang, L. 1328 Zhang, L. 457 Zhang, L. 1315, 601 Zhang, L. 863 Zhang, Q. 202, 613 Zhang, Q. 486, 639 Zhang, S. 276 Zhang, S. 1315, 601 Zhang, W. 723 Zhang, W. Z. 1328 Zhang, X. 1126 Zhang, X. 857, 880, 935 Zhang, X.-L. 1267 Zhang, Y. 287 Zhang, Y. 328 Zhang, Y. 723 Zhang, Y. 1180 Zhang, Y. 1040, 1073 Zhang, Z. 346 Zhao, C. 322 Zhao, C. 978 Zhao, F. 1056 Zhao, J. 1177 Zhao, J. 1297 Zhao, J. 1127 Zhao, J. 1127 Zhao, L. 1038 Zhao, M. 91 Zhao, T. 739 Zhao, W. 65 Zhao, X. 665 Zheng, H. 91 Zheng, T. Shan. 791 Zheng, X. 250 Zheng, X. 287 Zhou, J. 276 Zhou, J. X. 1328 Zhou, K. 257 Zhou, L. 723 Zhou, M. 77, 845 Zhou, R. 112, 962 Zhou, S. 1058 Zhou, T. 241 Zhou, W. 1070 Zhou, Y. 666, 667 Zhou, Y. 1179, 1180 Zhou, Y. 718 Zhu, L.-X. 1267 Zhu, T. 328 Zhu, W. 588 Zhu, Y. hua. 1123 Zhuge, F. 752 Ziegler, A. 170 Ziegler, D. 249 Ziegler, I. 249 Ziegler, O. 1043 Ziegler, R. 1008, 1014 Zielinska, A. 286, 337 Zielinski, E. 669 Zielonka, J. S. 969 Ziemek, D. 107 Ziemen, M. 148, 946, 947, 949, 961, 980 Zierath, J. R. 235, 248 Zilberman, L. 371 Zilberman, L. I. 362 Zimmerman, R. S. 323 Zimmermann, A. G. 38 Zingaretti, M. C. 688 Zinman, B. 855, 956 Ziol, M. 1275 Zisser, H. 1000 Zoccali, C. 343 Zóka, A. 444 Zoppini, G. 210 Zorzano, A. 674 Zoso, A. 625 Zou, D. 723 Zozulinska-Ziolkiewicz, D. 1022, 1144, 222 Zucchini, S. 1253 Zulewski, H. 17 Zuurmond, A.-M. 1324 Diabetologia (2014) 57:[Suppl1]S1–S564 S1_S6_suppl-2014.pdf S7_S563_suppl-2014.pdf Blank Page