Laboratory diagnosis of Lyme neuroborreliosis is influenced by the test used: Comparison of two ELISAs, immunoblot and CXCL13 testing

Journal of the Neurological Sciences xxx (2014) xxx–xxx JNS-13442; No of Pages 8 Contents lists available at ScienceDirect Journal of the Neurological Sciences j ourna l homepage: www.e lsev ie r .com/ locate / jns Laboratory diagnosis of Lyme neuroborreliosis is influenced by the test used: Comparison of two ELISAs, immunoblot and CXCL13 testing Nora Wutte a,⁎, Juan Archelos b, Brian A. Crowe c, Werner Zenz d, Elisabeth Daghofer e, Franz Fazekas b, Elisabeth Aberer a a Department of Dermatology and Venerology, Division of Environmental Dermatology and Venerology, Medical University of Graz, Auenbrugger Platz 8, A-8036 Graz, Austria b Department of Neurology, Division of General Neurology, Medical University of Graz, Auenbrugger Platz 22, A-8036 Graz, Austria c Vaccines R&D, Baxter Bioscience, Orth an der Donau, Austria d Department of General Pediatrics, Medical University of Graz, Auenbrugger Platz 34/2, A-8036 Graz, Austria e Institute for Hygiene and Microbiology, Medical University of Graz, Universitätsplatz 4, 8010 Graz, Austria ⁎ Corresponding author at: Dept. of Dermatology, Auenbrugger Platz 8, A-8036 Graz, Austria. Tel.: +43 316 12466. E-mail addresses: [email protected] (N. Wut [email protected] (J. Archelos), Werner.zenz@ [email protected] (F. Fazekas), Elisabeth.abere http://dx.doi.org/10.1016/j.jns.2014.09.027 0022-510X/© 2014 Elsevier B.V. All rights reserved. Please cite this article as: Wutte N, et al, Labo immunoblot and CXCL13 testing, J Neurol Sc a b s t r a c t a r t i c l e i n f o Article history: Received 25 March 2014 Received in revised form 16 September 2014 Accepted 18 September 2014 Available online xxxx Keywords: CXCL13 Neuroborreliosis Diagnosis ELISA Immunoblot Purpose: To compare Borrelia-specific intrathecal antibodies by two different ELISAs, an immunoblot (IB) and CXCL13. Methods: Twenty-seven adults and 23 children with clinical symptoms compatible with NB were tested for Borrelia-specific intrathecal antibodies by flagellum ELISA-AI (flELISA), a recombinant ELISA-AI (rELISA) and by IB. Patients were classified according to the European Federation of Neurological Societies (EFNS) criteria as definite NB, possible NB, or non-NB. CSF CXCL13 levels were measured by ELISA. Results:Among 50 patients, definiteNBwas diagnosedwith the rELISA-AI in 29 (58%) patients, confirmed by IB in 19/29 patients, with flELISA-AI in 17 (34%) patients, confirmed by IB in 15/17 patients, and with IB in 20 (40%) patients. CXCL13waspositive in 22 (44%) patients. In 4 of 8 patientswith negative AI, IB showedmanydetectable bands both in the CSF and serum. Conclusions: The diagnosis of NB strongly relies on the used test method. The rELISA-AI test appears to be the most sensitive while the flELISA-AI is the least sensitive. However when the ELISA-AIs were confirmed by IB, different patients were identified as NB, while only 26% were identified by all performed test methods. There is a demand for standardized test methods with well-defined sensitivity and specificity to establish validated diagnostic criteria for NB including the use of the IB assay and CXCL13 as an additional non-Borrelia specific determinant in early NB. © 2014 Elsevier B.V. All rights reserved. 1. Introduction Acute neuroborreliosis (NB) is the most common systemic manifes- tation of Lyme borreliosis in North America and Europe, affecting 10–15% of patients infected with Borrelia burgdorferi (Bb) [6,24,25,32]. The main clinical presentations of NB are meningopolyradiculitis (Bannwarth's syndrome) and facial palsy, the latter especially in children [3]. Cerebrospinal fluid (CSF) analysis shows lymphocytic pleocytosis and disturbance of the blood–brain barrier. According to European consensus, the definite diagnosis of acute NB requires detec- tion of Borrelia burgdorferi sensu lato (Bbsl) by culture or PCR, or evi- dence of intrathecal production of specific antibodies by a positive Medical University of Graz, 385 30062; fax: +43 316 385 te), medunigraz.at (W. Zenz), [email protected] (E. Aberer). ratory diagnosis of Lyme neu i (2014), http://dx.doi.org/10 antibody index (AI) (ratio of specific antibodies in CSF versus serum) [14]. Serum antibodies to Bb are usually present, although they may be negative in early NB [10]. Pitfalls in diagnostic CSF procedures result from several disease- related and technical facts. PCR and culture for Bb are rarely positive [4,17,30], AI has a moderate sensitivity of 55–80% in early infection, and antibodies persist even after therapy [2,7,11]. Furthermore, there is a high seroprevalence of Bb antibodies even in healthy individuals in endemic regions [22]. Antibodies in CSF may also result from diffu- sion of serum antibodies through an otherwise impaired blood–brain barrier [15]. Borrelia serology is not yet standardized and test results from different laboratories may vary depending on the test kit used. Guidelines recommend that ELISA-AI tests employed for routine serological diagnosis of Lyme borreliosis (LB) be confirmed by immuno- blot [21,30], but at present this is not routinely done for specific Bb− antibodies in CSF. A few studies have tested CSF by immunoblot in the past [12,13,17, 20,29], showing IgG antibodies in 99% and IgM antibodies in 60%. IgM antibodies occur early after infection, whereas IgG predominates in roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 mailto:[email protected] mailto:[email protected] mailto:[email protected] mailto:[email protected] mailto:[email protected] http://dx.doi.org/10.1016/j.jns.2014.09.027 http://www.sciencedirect.com/science/journal/0022510X www.elsevier.com/locate/jns http://dx.doi.org/10.1016/j.jns.2014.09.027 2 N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx late NB, and the number of positive bands increaseswith the duration of infection [10,13]. Furthermore, it has been proposed that the immune response in the CSF directed to other antigens than those in serum re- flects intrathecal antibody production [12]. In recent years, CXCL13 has become an interesting new biomarker that has high predictive values for the diagnosis of acute neuroborreliosis [19,27,33,34]. In a previous study we found that the chemokine CXCL13 in CSF confirms the diagnosis of NB determined by the flagellum ELISA-AI, and also identifies patients with possible NB but antibodies still negative on ELISA-AI [33,34]. Although CXCL13 is not yet validated as a routine diagnostic tool, CSF CXCL13may be another option to increase sensitiv- ity and accuracy in diagnosing NB, with a negative predictive value of 100% [18] Comparative studieswith different ELISA-AIs and IBwere performed to determine the optimal sensitivity of test methods [12,17]. Analogous to the stepwise diagnostic procedure in serum, i.e. determining the ELISA-AI followed by immunoblot confirmation, our objective here and aim of the study was to compare intrathecally produced anti- bodies by two different ELISA-AIs and one immunoblot method in both adults and for the first time in pediatric patients with clinically confirmed or suspected NB and controls. In addition, we compared CXCL13 levels in CSF with test results of the two ELISA-AIs, the immunoblot, and specific CSF changes (lymphocytic pleocytosis). Comparing these tests should allow us to define the most appropriate test for the diagnosis of NB. 2. Patients, materials & methods 2.1. Patients Over a 9-month period we recruited fifty patients (27 adults and 23 children; 28 males and 22 females; age range 6 to 69 years [median, 11 years]) for this study at the Department of Neurology or the Depart- ment of General Pediatrics, Medical University of Graz (Table 1). In a previous paper the same patient groupwas analyzed and classified clin- ically as NB (n = 22) or as neurologic control patients with suspected viral meningitis, or facial palsy (n = 31) [33,34]. The median age of the 27 adults was 34 years (range 23–69) (Table 1). Thirty-seven per- cent were female, median disease duration was 4 days and 77% recalled a tick bite 2 weeks to 4 months before onset of neurologic symptoms. Evidence for an EM was found in 37% of patients. The median age of the 23 children was 10 years (range 6–17 years) and 48% percent were female; 34.7% of the children had previous EM. Fifty-two percent reported a tick bite in the preceding 2 weeks to 4 months and the me- dian duration of their neurologic symptoms upon initial examination was 3 days (range 0–5 days). NB can be diagnosed clinically by an experencied neurologist because of a typical clinical symptomatic combined with historical data such as previous tick bite or EM. Our study, hovewer, aimed to assess how many of the 50 patients could be identified as suffering from NB using laboratory diagnosis only. According to the guidelines of the European Federation of Neurological Societies (EFNS), patients were classified into 3 groups: 1) definite NB, 2) possible NB and 3) neu- rological controls with no serum and CSF antibodies. The requirements for the diagnosis of definite NBwere: a) clinical symptoms of meningitis/facial palsy/meningoradiculitis, b) CSF pleocytosis ≥4 cells/μl and c) intrathecal production of Borrelia- specific antibodies [14]. In possible NB, according to the EFNS Table 1 Demographics. Patients Age (mean/SD) Female (%) Duration o Total (n = 50) 11 (6–69) 42% 3 days (0– Adults (n = 27) 34 (23–69) 37% 4 days (1– Children (n = 23) 10 (6–17) 48% 3 days (0– Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 criteria, 2 of these 3 criteria were fulfilled. We defined definite NB if patients had both CSF pleiocytosis and intrathecal antibody pro- duction by ELISA. We defined possible NB I when one CSF test was positive without pleocytosis and possible NB II when pleocytosis was present and CSF tests were negative. Since we performed 3 different Borrelia-specific tests, a positive CSF test was defined when at least one test (recombinant ELISA, flagellum ELISA or immunboblot) was positive. Additionally, levels of the B-cell cytokine CXCL13 in serum, were measured by ELISA. All patients and parents of the children that were included in the study gave their written informed consent. The study protocol was approved by the ethics committee of the Medical University of Graz. 2.2. Material & methods 2.2.1. CSF — serum antibody index Serum antibodies were determined and intrathecal antibody pro- duction was measured by three specific CSF tests (recombinant ELISA- AI, flagellum ELISA-AI, immunoblot) as well as test combinations (re- combinant ELISA confirmed by IB and flagellum ELISA confirmed by IB). Flagellum ELISA (flELISA): To prove intrathecal antibody production, CSF IgM/IgG antibodies were determined using the IDEIA Lyme Neuroborreliosis kit (OXOID, Cambridgeshire, UK) according to the manufacturer's instructions, with native Borrelia afzelii (Ba) strain DK1 flagellum as the antigen. The flagellum ELISA uses a capture meth- odology that corrects for differences in CSF and serum immunoglobulin concentrations. In the presence of intrathecal synthesis of Bb antibodies the proportion of specific antibodies is higher in CSF than in serum. An antibody index (AI) of OD CSF / OD serum× (OD CSF−OD serum) N1.5 was considered positive for intrathecal production of Bb antibodies [33, 34]. According to the manufacturer's instructions, technical sensitivity of this test is 76%, the specificity is unknown. Assuming that all patients with definite NB should have pleocytosis (n= 40) we calculated sensi- tivity in patientswith andwithout pleocytosis (n=10). Diagnostic sen- sitivity in our patient cohort is 43% at a specificity of 100% using the flELISA. Recombinant ELISA: we further employed a recombinant ELISA (rELISA) (recomWell IgG and IgM, Mikrogen Diagnostik, Neuried, Germany) according to the manufacturer's instructions for CSF and serum antibodies. This test is an indirect ELISA used to identify IgM and IgG antibodies against Bbsl in serum as well as in CSF. The rELISA employs OspC and VlsE of all genospecies pathogenic to humans (B. afzelii, several B. garinii subgroups and B. burgdorferi sensu stricto). Cut-off for IgM and IgG serum levels was 24 U/ml. To detect antibodies specific to Bb, a diluted serum or CSF sample is incubated in the well with Bb antigens. Antibodies are bound by anti-human immunoglobu- lin coupled with peroxidase. CSF and sera are further diluted. At an ex- tinction of 1.0 of both serum and CSF, the Reiber index is calculated, taking into account total CSF and serum albumin, total IgG and total IgM concentrations, respectively, according to a standardized formula. AI was considered positive with values ≥ 1.5. According to the manufacturer's instructions, this ELISA test method has a technical sen- sitivity of 100%, with a technical specificity ranging from 97% (IgM) to 99% (IgG). Diagnostic sensitivity in our patient cohort would be 70% at a specificity of 90% by rELISA. In children only IgG antibodies were tested by recombinant ELISA and immunoblot due to the small amount of stored CSF as these tests require determination of total serum- and CSF-IgM to calculate AI. f disease (median/min-max) History of any tick bite (%) EM (%) 7) 58% 60% 7) 77% 37% 5) 52% 35% roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 Table 2 Serum antibodies and CXCL13 levels. Serum Total (n = 50) Positive antibodies in IgM IgG Total Serum positive by flagellum (Siemens) ELISA 31 (62%) 34 (68%) 38 (76%) Serum positive by recombinant (Mikrogen) ELISA 31 (62%) 30 (60%) 38 (76%) Serum immunoblot positive 17 (34%) 38 (76%) 38 (76%) Serum CXCL13 N 100 pg/ml − − 28 (56%) 3N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx Immunoblot analysis from CSF and serum was performed using recomBlot Borrelia NB (recomBlot IgG and IgM, Mikrogen Diagnostik, Neuried, Germany) as directed by the manufacturer. CSF and serum are diluted to obtain equal concentrations of immunoglobulins in both, since specific IgM or IgG can be compared between serum and CSF only if the same concentration is reached. The recomBlot Borrelia NB detects antibodies against the three genospecies Bb sensu stricto, B. garinii and B.afzelii (VlsE from different genospecies, OspC from all genospecies and different B. garinii strains including OspA, OspC, VlsE, p100, p39, p18, p41 and p41i). Immunoblot bands that were stronger in CSF than in serum, or detectable in CSF only, were considered positive for intrathecal antibody production. Immunoblot bands were analyzed by a blinded investigator. Again, in children only IgG antibodies could be determined as this test requires results from rELISA. Diagnostic sensitivity in our patient cohort is 46% at a specificity of 90% by IB. CXCL13 in CSF and serum was measured by ELISA, as described previously [34]. All analyses were performed blinded to the clinical or other laboratory information. CXCL13 serum levels above 105 pg/ml were considered elevated. All samples were run in duplicate. Sensi- tivity and specificity of CXCL13 had been determined in previous tests [33]. 50 pa�ents Adults n=27 Children n=23 Serology Seroposi�ve N=37 Seronega�ve N=13 Definite NB: n=31 29 seropos. 2 seroneg. CSF Pos. n=31 CSF Neg. n=6 Possible NB I (=AI pos): n=2 2 seropos. CSF Pos. n=2 CSF Neg. n=11 Def.: „CSF Pos.“= At least one CSF ELISA test posi�ve (rEL „CSF Neg.“= No test posi�ve Fig. 1. Flow chart of patient classification. Fig. 1 describes how the 50 patients, both adults and c and “Controls” by taking into account CSF pleiocytosis and at least one positive Borrelia-specific Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 2.2.2. Statistical analysis The statistical software package SPSS, Version 17.0 (SPSS Inc., USA) was used for statistical analysis. Means and standard deviations or me- dian, minimum and maximum are shown, as appropriate, for continu- ous data. Absolute and relative frequencies are shown for categorical data. Data were tested for normality using the Kolmogorov–Smirnov test. Continuous variables were analyzed by the Kruskal–Wallis test; categorical variables were compared by the chi-square or Fisher's exact test. To check agreement between the tests, Cohen's kappa coeffi- cient was calculated. A p-value below 0.05 was considered statistically significant. Sensitivity and specificity of the tests could not be deter- mined as we did not use one test as a reference standard, but instead compared frequencies of positive results obtained with the different tests. 3. Results 3.1. Cell count Forty of the 50 patients showed CSF pleiocytosis. The CSF cell number was not higher in definite NB patients (mean 105.86 cells/μl, range 4–501) as compared to possible NB II (mean 176.38 cells/μl, range 5–624). 3.2. Serology Altogether, 37 patients had positive Lyme serologywhile 13 patients were seronegative. The results of the serum ELISA test and subclasses of serum antibodies are shown in Table 2. Serum CXCL13 levels were ele- vated in 28/50 patients, but the significance of elevated serum CXCL13 levels in Lyme borreliosis has not been proven, as demonstrated in a previous paper [33]. Pleiocytosis N=40 No pleiocytosis N=10 Possible NB II (=Pleiocytosis): n=9 1 seropos. 8 seroneg. Controls: n=8 4 seropos. 4 seroneg. 50 pa�ents Pleiocytosis Adults n=27 Children n=23 ISA OR flELISA) hildren, were classified into the categories of “Definite NB”, “Possible NB I”, “Possible NB II” CSF test. Lyme serology was tested in all patients, but did not influence the classification. roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 FlELISA 0 Immunoblot 0 rELISA CXCL13 0 13 7 4 1 1 12 n = 29 n = 20 n = 17 n = 22 1 1 Fig. 2. Number of patients with CSF pleiocytosis identified as definite NB using different test methods. Fig. 2 shows a detailed analysis of the used test methods alone or test com- binations. 13 patients were positive by all 4 test methods. Four patients were positive by a combined rELISA, immunoblot and CXCL13 analysis, 1 patient was positive by combined flELISA, immunoblot and CXCL13 testing. Furthermore, one patientwas positive by rELISA, flELISA and immunoblot, and one other patient was positive by rELISA, flELISA and CXCL13. Moreover, one patient was positive by rELISA and immunoblot. Two patients were positive by rELISA and CXCL13, and one patient by flELISA and CXCL13. Additional 7 patients were positive only by rELISA. No single patient was positive by flELISA, immu- noblot or CXCL13 alone. 4 N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx 3.3. CSF testing and classification of patients based on different Borrelia-specific tests 3.3.1. Definite NB Of the37 seropositive patientswith pleocytosis, 31 showedevidence of intrathecal antibody production according to at least one CSF ELISA test and thus were diagnosed as definite NB. (Fig. 1). Twenty-nine of these patients were seropositive, 2 patients were seronegative (6.7%). Fig. 2 illustrates the frequency of the diagnosis “definite NB” by different test methods. In detail, recombinant ELISA Table 4 CSF antibodies and CXCL13 in all patient groups. CSF Adults (n = 27) Positive antibodies in IgM IgG Total AI positive by recombinant (Mikrogen) ELISA 12 (44%) 17 (63%) 19 (70%) AI positive by flagellum (Oxoid) ELISA 4 (15%) 4 (15%) 6 (22%) Positive CSF Immunoblot 7 (26%) 11 (41%) 10 (37%) CSF CXCL13 N 100 pg/ml – – 13 (48%) rELISA & immunoblot – – 10 (37%) flELISA & immunoblot – – 6 (22%) Abs = antibodies Table 3 Classification of all patients based on different tests. Re Total positives by CSF test 30 Definite NB (clinical presentation, elevated cell count, positive CSF test) 29 Possible NB I (clinical presentation, positive CSF test, no elevated cell count) 1 ( Possible NB II (clinical presentation, elevated cell count, no positive CSF test) 11 Negative CSF test, no elevated cell count 9 ( Total patient count 50 Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 identified 29 patients with definite NB (58%), flELISA alone diagnosed 17 patients (34%). Analogous to the stepwise diagnostic approach used in serodiagnostics, we aimed to confirm positive results of rELISA and flELISA by a confirmatory immunoblot. In 19 patients rELISA was confirmed by immunoblot (38%) while flELISA was confirmed by im- munoblot in 15 patients (30%). Twenty patients had definite NB when rELISA and CSF CXCL13would have been combined, 16 patients accord- ing to combined flELISA and CXCL13. Thirteen of 50 patients (26%)were positive in all 4 tests. Immunoblot identified 20 patients with definite NB (40%) when this test would have been used as an independent marker (Table 3, Fig. 2). 3.3.2. Possible NB Possible NB was diagnosed in 11 patients (22%) (Fig. 1). Of these, 2 seropositive patients had a positive CSF test (1 rELISA, 1 IB), but showed no CSF pleiocytosis and thus were classified as possible NB I. Nine pa- tients had a positive cell count but negative CSF test (7 seronegative, 2 seropositive) and were classified as possible NB II. No patient showed a single positive flELISA or CXCL13 test. The detailed test results for adults and children and the distribution of immunoglobulin subclasses are shown in Table 4. 3.3.3. Controls Only 8 patients showed no pleiocytosis and no sign of intrathecal Borrelia-specific antibody production and thus were classified as con- trols. Four of these patients were seropositive and all were children diagnosed with idiopathic facial palsy. 3.3.4. CXCL13 in CSF CSF CXCL13 was elevated above the cut-off in 22 patients with defi- nite NB (44%). No control patient or possible NB patient was positive by CXCL13 testing. Generally, agreement between the tests was only fair with an in- complete overlap of positive patients with both ELISA-AI tests as well as immunoblot analysis (Kappa 0.28). The frequency of confirmed intra- thecal antibody production differed among the tests; the number of patients identified by rELISA was 1.7-fold higher than with flELISA (p = 0.002). AI results were also more often positive with rELISA than with IB (p = 0.039) or compared to CXCL13 levels (p = 0.022). There were no significant differences when agreement between flELISA, IB and CXCL13 was calculated. Children (n = 23) Total (n = 50) IgM IgG Total IgM IgG Total – 11 (48%) 11 (48%) 12 (24%) 28 (56%) 30 (60%) 5 (22%) 10 (43%) 11 (48%) 9 (18%) 14 (28%) 17 (34%) – 11 (48%) 11 (48%) 8 (16%) 19 (38%) 21 (42%) – – 9 (39%) – – 22 (44%) – – 9 (23%) – – 19 (38%) – – 9 (23%) – – 15 (30%) combinant ELISA Flagellum ELISA Immunoblot CXCL13 (60%) 17 (34%) 21 (42%) 22 (44%) (58%) 17 (34%) 20 (40%) 22 (44%) 2%) 0 1 (2%) 0 (22%) 23 (46%) 20 (40%) 18 (36%) 18%) 10 (20%) 9 (18%) 10 (20%) 50 50 50 roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 Fig. 3. CSF/serum-IgG-Immunoblot of a 10-year-old patient with definite NB. This fig- ure shows an example of stronger p41 and VlsE IgG bands in CSF than in serum and serum dilution. The p41i band is new in CSF as compared to serum and serum dilution. DIL = dilution of serum to an equivalent albumin concentration in serum and CSF. 5N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx With immunoblot, the most frequent intrathecal antibodies were directed against p41- and VlsE antigens (IgM and IgG). Stronger or new bands in CSF compared with serum were found in 44% of patients and were directed against VlsE, p41, p100, p39, p18 and p100. New bands occurred in adult patients only and were directed against p41 and VlsE, except for one patient who showed a new IgM p100 band compared with serum. Children showed no new bands in CSF, only stronger bands. Notably, stronger bands in CSF than in serum were also identified in 1 child with cephalea without CSF pleiocytosis (p100 positive). Interestingly, 4 patients (1 with possible NB II, 3 controls), without a positive AI (chronic headache, viral meningitis) had many positive serum and CSF IgG bands (VlsE, p41, p18, p39, p100, OspC), indicating previous infection (Fig. 3, Fig. 4, Table 5). 4. Discussion The sensitivity of commercially available tests for the detection of Borrelia-specific intrathecal antibodies varies considerably thereby hav- ing a profound effect and influence on the diagnosis of NB. Comparing rELISA, flELISA, IB in the CSF, the probability of diagnosing definite NB was highest (29/50 patients) with rELISA followed by different test combinations as shown in Fig. 2. Using all 3 Borrelia-specific test methods, definite NB was diagnosed in 13/50 patients only. According to the number of positive findings, recombinant ELISA-AI, which iden- tifies intrathecal synthesis of antibodies to recombinant OspC and VlsE, was a better screening test than flagellum ELISA-AI. These findings are in line with previous studies that showed a lower sensitivity of the flELISA [5,9]. When a more specific ELISA like flagellum AI is used, an infection may be missed in clinical practice. Using the flagellum ELISA, immunoblot confirmed the diagnosis in 15/17 patients. However, an additional 12 patients were positive by recombinant ELISA-AI although IB confirmed this in only 19/29 patients. In 5 of 12 additional patients, definite NB was confirmed using a combined recombinant ELISA-AI and IB, being finally 20 patients (rELISA and/or flELISA confirmed by IB); two patients were positive by flELISA who were not confirmed by rELISA. In patients with positive ELISA AI, immunoblot testing from the CSF can be performed as a confirmatory test, although this is not a routine procedure. Immunoblot testing from CSF is only available in few laboratory centers. Taking either the IB or CXCL13 for confirmation of ELISAs, all 17 flELISA positive patients and 22/29 rELISA positive patients would have had been identified as definite NB (Fig. 2). Altogether 24 patients would have been classified as definite NB (flELISA and/or rELISA con- firmed by IB and/or CXCL13) (Fig. 2).With regard to the high sensitivity of rELISA, and as single positive test, patients may have been falsely classified as definite NB. However, as there is no gold standard for diagnosis and different tests are used by different laboratories, a positive result by any one of the tests was regarded as confirmation of a clinically suspected diagno- sis. FlELISA was more often positive in children than rELISA, IB or CXCL13, because in children, both IgM and IgG titres could only be determined in CSF with the flELISA test. CXCL13 is not an established test method for NB. However, many studies have illustrated the high diagnostic sensitivity of CXCL13 espe- cially in early NB before intrathecal antibody production can be mea- sured [18,19,33,34]. Rupprecht et al. could show a negative predictive value of 100% for CXCL13 in 197 investigated patients [18]. However, exact sensitivity and specificity of CSF CXCL13 remain unknown. As suggested by Tumani et al. [28], an elevated CSF cell count and IgM-class dominance in the humoral immune response combined with the Bb specific AI could increase sensitivity and specificity in the diagnosis of NB. In other studies, IgG antibodies have been found as the relevant immunoglobulin isotype for detecting intrathecal borrelia antibodies in NB. [23]. Pleiocytosis was present in all our patients with definite NB and possible NB II, and also in several seronegatives, Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 although the cell number was not higher in definite NB patients as compared to possible NB. In a previous study, Roux and et al. determined specific Bb antibodies by 2different ELISA-AIs (Vidas andDade-Behring) and immunoblot in 11 patients hospitalized for Lyme meningoradiculitis and 16 controls [17]. roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 Fig. 4. CSF/serum–IgM–immunoblot of a 35-year-old patient with definite NB. This figure shows an example of a new p100 IgM band in CSF compared to serum and a p41 band stronger in CSF than in serum. VlsE is equally strong in CSF and serum, but stronger in CSF than in serum dilution. DIL = dilution of serum to an equivalent albumin concentra- tion in serum and CSF. CTRL = control strip. 6 N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx Additionally, PCR and culturewere performed to identifyB. burgdorferi in the CSF. Independent of positive IgM serology (40% IgM-positive, 40% IgM negative), intrathecal IgG synthesis was detected, but 20% IgM seropositives showed no intrathecal antibody synthesis. So, even in the Table 5 Overview of CSF immunoblot bands. Patient groups New bands Bands stron Total (n = 50) IgM: p41, VlsE, p100 IgM: p41, V IgG: p41, VlsE, p100 IgG: VlsE, p Adults (n = 27) IgM: p41, VlsE, p100 IgM: p41, V IgG: p41, VlsE, p100 IgG: p41, V Children (n = 23) IgM: – IgM: – IgG: – IgG: VlsE, p NB = Lyme neuroborreliosis. Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 absence of serum IgM antibodies there can be intrathecal antibody syn- thesis. Intrathecal antibody production is often demonstrable before the detection of specific antibodies in serum. Kaiser's studies also showed that seronegative patients can have NB [10] and in our study, this was the case with 2 seronegative patients who had definite NB. However, as international criteria do not take serum antibodies into account, a positive CSF test was required for diagnosis. Similarly to Roux et al., we screened all patients at the neurology and pediatric de- partments with suspected NB and control individuals within one season [17]. Thismeant that the number of patientswas small,which can be also seen as a limitation of our study results. The interpretation of CSF immunoblot differs among these tests. According to the European guidelines, IB is positive in the CSF when 3 bands of any molecular weight are present [16]. This is in contrast to our procedure and as prescribed in our manufacturer's manual; only those bands were counted as positive that were stronger than in serum or when new bands were seen. To our knowledge, ours is the first study to investigate immunoblot in CSF in children. As IgM antibodies were only measured in adults, the diagnostic sensitivity of immunoblot analysismay be even higher in our patientswith possible NB because an early neural infection should show a predominantly IgM-mediated immune response. This is in contrast to serum, in which IgG antibodies, especially against B. garinii, predomi- nate in NB [8,31]. CSF immunoblot showed one patient with possible NB I with stron- ger p100 antigen in CSF that might have indicated a previous Borrelia infection. This patient was a child with severe headache without pleocytosis. Three children with chronic headache from the control group also showed intrathecal antibodies with many different positive bands, even though not stronger than in serum. We suspect that these children with chronic headache and numerous positive CSF and serum IgG bands had a mild reaction of the CNS without typical neurological clinical symptoms, as has been described by Arnez et al. in children with multiple EM; these children might also benefit from antibiotic therapy [1]. Thus, IB would have provided important additional information in some of our patients without specific neurological symptoms. On the other hand, a negative immunoblot has a high diagnostic value in ruling out borrelial infection. Lakos et al. studied 22- and 41-kD proteins in IgM, and the 22-, 29-, 35-, 44-, 47-, 49- and 93-kD bands in IgG IB, and proposed that a quali- tative difference between the immunoblot pattern of CSF and serum is necessary to diagnose NB [12]. In our study, the main bands occurring in CSF were those against p41 and VlsE antigens. The p100 antigen was found in three cases, and the p39 antigen in one case. This is consis- tentwith data reported byWilske et al., who found that the Bb− specific CSF immune response is mainly directed against the p41 protein [30]. In another study, IgM antibodies to OspC were frequently reported in serum as well as CSF in NB patients, and IgM antibodies to OspB and OspA antigens were found in CSF but not in serum [12]. These Osp bands were not detected in our patients. Roux et al. detected new bands in the CSF of 6 patients (against different antigens 22, 44, 51, 53, 50, 65, 58, 39, 62 KD) and the same bands in serum and CSF in 3 patients. In 1/16 control patients N3 bands were detected in the CSF [17]. ger than in serum Bands detected but not stronger than in serum lsE, p39 IgM: p18, OspC, VlsE, p41, p39 39, p41, p18, p100 IgG: p41, VlsE, p100, p18, p39, p18, OspC lsE, p39 IgM: p18, OspC, VlsE, p41, p39 lsE, p100, p39 IgG: VlsE, p41, p100, p39, p18, OspC IgM: – 41, p18, p100 IgG: p41, VlsE, p18, p39, p18, OspC roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://dx.doi.org/10.1016/j.jns.2014.09.027 7N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx Studies about CXCL13 in CSF are mainly conducted in Europe, by working groups in Germany, Slovenia, the Netherlands, Norway, Finland and Poland. CSF CXCL13 testing in NB has not yet been reported from other countries and to best of our knowledge from the US. By analyzing all the positive tests and test combinations, we realized that all patients classified as definite NB, and only those patients, also had elevated CXCL13 levels. However, classification of our patients was not influenced by CXCL13 levels at all. Based on our own data we have had positive experiences with this test [33,34]. Similarly to immu- noblot testing, both tests might prove to be helpful in the diagnosis of NB, although larger cohort studies should be conducted to confirm this. It was difficult to calculate sensitivity and specificity of our test methods since the control group was small and no golden standard for CSF diagnosis is available. Sensitivity can also be influenced by the genospecies used by the different test methods. Our recombinant ELISA and IB comprised all 3 genospecies, which should have also covered B. garinii in particular, as it has been found by PCR in 3 of 4 CSF isolates and B. afzelii in one case in the same geographic area by Stünzner et al. [26]. Eight patients were negative for all our CSF tests and were termed “controls”. Four patients were also seronegative and thus, a borrelial infection can only be ruled out in these patients. In the 4 seropositive control patients the neurologic disease seen could not be related to NB; however, it also could not be ruled out that these patients had active non-neurologic Lyme disease or had seroconverted due to a past infection. The main limitation of our study is that the attempt to establish relative sensitivity and specificity is problematic. Control patients were difficult to define since there is also a high Bb seroprevalence in our healthy population [22]. Depending on the used test, a diverse num- ber of patients were identified as definite NB. We are aware that our approach has limitations, but, in the absence of a gold standard, we tried to calculate relative sensitivities based on typical clinical presenta- tion and lymphocytic pleiocytosis. As suggested by Blanc et al., we need new and pragmatic diagnostic criteria in NB [2]. There is a demand for standardized test methods with well-defined sensitivity and specificity to establish validated di- agnostic criteria for NB. When devising a new classification system, the role of IB analysis will have to be included. In our opinion, IB could be an important supplement to recombinant ELISA-AI to rule out the diagnosis of NB in children and adults with headache or in un- clear clinical cases. In one study, IB was not usable as a screening assay for diagnosis because of its low sensitivity (67%) and specificity (60%) [28]. In view of positive CSF findings in children with possible NB with headache−which could point to NB in these patients – it may be con- cluded that CSF immunoblot possibly is the most sensitive of all tests. CXCL13 levels in CSF can be used as an additional non-Borrelia specific determinant in early NB. Conflict of interest Brian A. Crowe is employed by Baxter, a developer of vaccines against Lyme borreliosis, and has equity interests in the company. Brian A. Crowe is author of planned, pending or granted patents pertaining to vaccines against Lyme borreliosis. This does not alter our adherence to all the European Journal of Clinical Microbiology policies on sharing data and materials. The other authors declare that they have no conflict of interest. Acknowledgments We greatly acknowledge the laboratory work of Ingrid Krainberger and Monika Joch who performed ELISA and immunoblot testing. We thank Prof. Andrea Berghold for fruitful discussion concerning presenta- tion of statistic calculations. Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 References [1] Arnez M, Pleterski-Rigler D, Ahcan J, Ruzić-Sabljić E, Strle F. Demographic features, clinical characteristics and laboratory findings in children with multiple erythema migrans in Slovenia. Wien Klin Wochenschr 2011;113: 98–101. [2] Blanc F, Jaulhac B, Fleury M, de Seze J, de Martino SJ, et al. Relevance of the antibody index to diagnose Lyme neuroborreliosis among seropositive patients. Neurology 2007;69:953–8. [3] Broekhuijsen-van Henten DM, Braun KP,Wolfs TF. Clinical presentation of childhood neuroborreliosis; neurological examination may be normal. Arch Dis Child 2010;95: 910–4. [4] Cerar T, Ogrinc K, Cimperman J, Lotric-Furlan S, Strle F, et al. Validation of cultivation and PCR methods for diagnosis of Lyme neuroborreliosis. J Clin Microbiol 2008;46: 3375–9. [5] Cerar T, Ogrinc K, Strle F, Ruzić-Sabljić E. Humoral immune responses in patients with Lyme neuroborreliosis. Clin Vaccine Immunol 2010;17:645–50. [6] Halperin JJ. Nervous system Lyme disease: diagnosis and treatment. Curr Treat Options Neurol 2013;15:454–64. [7] Hansen K, Lebech AM. Lyme neuroborreliosis: a new sensitive diagnostic assay for intrathecal synthesis of Borrelia burgdorferi-specific immunoglobulin G, A, and M. Ann Neurol 1991;30:197–205. [8] Hauser U, Lehnert G, Lobentanzer R, Wilske B. Interpretation criteria for standard- ized Western blots for three European species of Borrelia burgdorferi sensu lato. J Clin Microbiol 1997;35:1433–44. [9] Henningsson AJ, Christiansson M, Tjernberg I, Löfgren S, Matussek A. Laboratory di- agnosis of Lyme neuroborreliosis: a comparison of three CSF anti-i antibody assays. Eur J Clin Microbiol Infect Dis 2013;33:797–803. [10] Kaiser R. False-negative serology in patients with neuroborreliosis and the value of employing of different borrelial strains in serological assays. J Med Microbiol 2000; 49:911–5. [11] Krüger H, Reuss K, Pulz M, Rohrbach E, Pflughaupt KW, et al. Meningoradiculitis and encephalomyelitis due to Borrelia burgdorferi: a follow-up study of 72 patients over 27 years. J Neurol 1989;236:322–8. [12] Lakos A, Ferenczi E, Komoly S, GranströmM. Different B-cell populations are respon- sible for the peripheral and intrathecal antibody production in neuroborreliosis. Int Immunol 2005;17:1631–7. [13] Murray N, Kristoferitsch W, Stanek G, Steck AJ. Specificity of CSF antibodies against components of Borrelia burgdorferi in patients with meningopolyneuritis Garin– Bujadoux–Bannwarth. J Neurol 1986;233:224–7. [14] Mygland A, Ljøstad U, Fingerle V, Rupprecht T, Schmutzhard E, et al. European Fed- eration of Neurological Societies. EFNS guidelines on the diagnosis andmanagement of European Lyme neuroborreliosis. Eur J Neurol 2010;17:8–16. [15] Pícha D, Moravcová L, Zdárský E, Benes J. Clinical comparison of immunoblot and antibody index for detection of intrathecal synthesis of specific antibodies in Lyme neuroborreliosis. Eur J Clin Microbiol Infect Dis 2000;19:805–6. [16] Robertson J, Guy E, Andrews N,Wilske B, Anda P, et al. A Europeanmulticenter study of immunoblotting in serodiagnosis of Lyme borreliosis. J Clin Microbiol 2000;38: 2097–102. [17] Roux F, Boyer E, Jaulhac B, Dernis E, Closs-Prophette F, et al. Lymemeningoradiculitis: a prospective evaluation of biological diagnosis methods. Eur J Clin Microbiol Infect Dis 2007;26:685–93. [18] Rupprecht TA, Lechner C, Tumani H, Fingerle V. CXCL13: a biomarker for acute Lyme neuroborreliosis: investigation of the predictive value in the clinical routine. Nervenarzt 2014;85:459–64. [19] Schmidt C, Plate A, Angele B, Pfister HW, Wick M, Koedel U, et al. A prospective study on the role of CXCL13 in Lyme neuroborreliosis. Neurology 2011;76: 1051–8. [20] Schutzer SE, Coyle PK, Krupp LB, Deng Z, Belman AL, et al. Simultaneous expression of Borrelia OspA and OspC and IgM response in cerebrospinal fluid in early neuro- logic Lyme disease. J Clin Invest 1997;100:763–7. [21] Stanek G, O'Connell S, Cimmino M, Aberer E, Kristoferitsch W, et al. European Union concerted action on risk assessment in Lyme borreliosis: clinical case definitions for Lyme borreliosis. Wien Klin Wochenschr 1996;108:741–7. [22] Stanek G, Cetin E, Sotoudeh M, Auer H. Paradigm Burgenland: risk of Borrelia burgdorferi sensu lato infection indicated by variable seroprevalence rates in hunters. Wien Klin Wochenschr 2006;118:677–81. [23] Stanek G, Lusa L, Ogrinc K, Markowicz M, Strle F. Intrathecally produced IgG and IgM antibodies to recombinant VlsE, VlsE peptide, recombinant OspC and whole cell extracts in the diagnosis of Lyme neuroborreliosis. Med Microbiol Immunol 2014;203:125–32. [24] Steere AC, Bartenhagen NH, Craft JE, Hutchinson GJ, Newman JH, et al. The early clinical manifestations of Lyme disease. Ann Intern Med 1983;99:76–82. [25] Steere AC, McHugh G, Damle N, Sikand VK. Prospective study of serologic tests for Lyme disease. Clin Infect Dis 2008;47:188–95. [26] Stünzner D, Hubálek Z, Halouzka J, Wendelin I, Sixl W, Marth E. Prevalence of Borrelia burgdorferi sensu lato in the tick Ixodes ricinus in the Styrian mountains of Austria. Wien Klin Wochenschr 2006;118:682–5. [27] Tjernberg I, Henningsson AJ, Eliasson I, Forsberg P, Ernerudh J. Diagnostic perfor- mance of cerebrospinal fluid chemokine CXCL13 and antibodies to the C6-peptide in Lyme neuroborreliosis. J Infect 2011;62:149–58. [28] Tumani H, Nölker G, Reiber H. Relevance of cerebrospinal fluid variables for early diagnosis of neuroborreliosis. Neurology 1995;45:1663–70. [29] Wilske B, Schierz G, Preac-Mursic V, Weber K, Pfister HW, et al. Serological diag- nosis of erythema migrans disease and related disorders. Infection 1984;12: 331–7. roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0005 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0005 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0005 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0005 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0010 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0010 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0010 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0015 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0015 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0015 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0020 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0020 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0020 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0025 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0025 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0030 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0030 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0035 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0035 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0035 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0040 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0040 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0040 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0165 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0165 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0165 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0045 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0045 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0045 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0050 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0050 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0050 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0055 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0055 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0055 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0060 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0060 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0060 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0065 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0065 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0065 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0070 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0070 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0070 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0075 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0075 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0075 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0080 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0080 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0080 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0085 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0085 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0085 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0090 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0090 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0090 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0095 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0095 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0095 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0100 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0100 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0100 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0105 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0105 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0105 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0170 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0170 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0170 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0170 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0110 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0110 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0115 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0115 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0120 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0120 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0120 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0125 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0125 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0125 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0130 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0130 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0135 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0135 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0135 http://dx.doi.org/10.1016/j.jns.2014.09.027 8 N. Wutte et al. / Journal of the Neurological Sciences xxx (2014) xxx–xxx [30] Wilske B, Schierz G, Preac-Mursic V, von Busch K, Kühbeck R, et al. Intrathecal production of specific antibodies against Borrelia burgdorferi in patients with lymphocytic meningoradiculitis (Bannwarth's syndrome). J Infect Dis 1986; 153:304–14. [31] Wormser GP. Clinical practice. Early Lyme disease. N Engl J Med 2006;354: 2794–801. [32] Wormser GP, Dattwyler RJ, Shapiro ED, Halperin JJ, Steere AC, et al. The clinical assessment, treatment, and prevention of Lyme disease, human granulocytic Please cite this article as: Wutte N, et al, Laboratory diagnosis of Lyme neu immunoblot and CXCL13 testing, J Neurol Sci (2014), http://dx.doi.org/10 anaplasmosis, and babesiosis: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis 2006;43:1089–134. [33] Wutte N, Berghold A, Krainberger I, Aberer E. Serum CXCL13 chemokine is not a marker for active Lyme borreliosis. Acta Derm Venereol 2011;2011:724–5. [34] Wutte N, Berghold A, Löffler S, Zenz W, Daghofer E, et al. CXCL13 chemokine in pe- diatric and adult neuroborreliosis. Acta Neurol Scand 2011;124:321–8. roborreliosis is influenced by the test used: Comparison of two ELISAs, .1016/j.jns.2014.09.027 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0140 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0140 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0140 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0140 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0145 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0145 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0150 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0150 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0150 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0150 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0155 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0155 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0160 http://refhub.elsevier.com/S0022-510X(14)00625-X/rf0160 http://dx.doi.org/10.1016/j.jns.2014.09.027 Laboratory diagnosis of Lyme neuroborreliosis is influenced by the test used: Comparison of two ELISAs, immunoblot and CXCL... 1. Introduction 2. Patients, materials & methods 2.1. Patients 2.2. Material & methods 2.2.1. CSF — serum antibody index 2.2.2. Statistical analysis 3. Results 3.1. Cell count 3.2. Serology 3.3. CSF testing and classification of patients based on different Borrelia-specific tests 3.3.1. Definite NB 3.3.2. Possible NB 3.3.3. Controls 3.3.4. CXCL13 in CSF 4. Discussion Conflict of interest Acknowledgments References