B ra lute Keywords: Vitamin B1 Vitamin B2 HPLC Cereal products al o ece ed as b and biosynthesis of corticoids, among others (Illera Martín et al., 2000). These two compounds are unstable to heat, light and other factors, so the technological processes in cereal grains can extraction) and exogenous vitamins, should be given on the label of the food products. This is the reason why the majority of methods are targeting the determination of total vitamin content, and has led to a growing need for cheap, cost-effective, rapid and reliable analytical monitoring methods for the determination of nutrients and vitamins and other nutrients in food products, that can be easily applied in any laboratories for routine analysis (Heudi et al., 2005; Zafra-Gómez et al., 2006). Many analytical methods have been proposed for thiamine and riboflavine analysis. The choice of * Corresponding author. Tel./fax: þ34 91 394 17 99. E-mail addresses:
[email protected] (R. San José Rodriguez), vfernand@ farm.ucm.es (V. Fernández-Ruiz),
[email protected] (M. Cámara), cortesm@ Contents lists available at Journal of Cer journal homepage: www. Journal of Cereal Science 55 (2012) 293e299 farm.ucm.es (M.C. Sánchez-Mata). recent years, their consumption has declined due to new life styles. As the necessity of their recovery in the diet is recognized, some new cereal products have been developed by the industry, adapted to different uses and providing nutritional and functional benefits to the organism, including the intake of energy from complex carbohydrates, as well as the presence of dietary fibre and micro- nutrients as minerals and vitamins. Thiamine and riboflavin are two vitamins involved in different processes in the human body, such as glucose metabolism, nervous transmission, replication of genes, development of foetal tissues grain products is an important feature for the industry, in relation to final product quality. The recent regulations regarding food nutrition labelling, introduced by the governments and regulatory bodies around the world, require a guaranteed content of nutrients and other bioactive compounds in food products during all their shelf life. Particularly, “European Regulation (EC) No 1925/2006 of the European Parliament and of the Council of 20 December 2006, on the addition of vitamins and minerals and of certain other substances to foods”, establishes that not only the added fraction, but the total sum of endogenous (that could need enzymatic 1. Introduction Cereals have been traditionally us 0733-5210/$ e see front matter � 2012 Elsevier Ltd. doi:10.1016/j.jcs.2011.12.011 HPLC-UV methods have been described but they are less sensitive, and present difficulties due to interfering compounds, particularly in complex food matrixes, as grains and derivatives. A combination of extraction and separation systems, that allows enough sensitivity, precision and accuracy for the analysis of vitamin B1 and B2 in complex cereal food products, by isocratic UV-HPLC, in a single wavelength simultaneous separation is presented, with the advantage of using low-cost equipment requirements, simple sample pre-treatment and short time. The achievement of this goal has involved the optimization of compatible extraction and measurement protocols for cereal matrices, comparing seven different separation conditions and six extraction/clarification matrices analysis. The selected method was comparatively validated and compared to reference AOAC spectrofluorimetric methods, providing comparable linearity and accuracy, with better specificity and precision parameters, as well as practical applicability. � 2012 Elsevier Ltd. All rights reserved. asic foods; however in partially destroy these compounds. For that reason, many industrial cereal products are enriched in B group vitamins to recover their levels. The evaluation of nutritional or functional components in Accepted 15 December 2011 analytical methods that can be easily applied in any quality control laboratories for routine analysis. Spectrofluorimetric analysis of thiamine and riboflavin are sensitive, but need specific equipment. A few Received in revised form 11 December 2011 life of the product. For that reason, industry usually makes many efforts to develop simple and reliable Simultaneous determination of vitamin by reverse phase isocratic HPLC-UV R. San José Rodriguez, V. Fernández-Ruiz, M. Cáma Departamento de Nutrición y Bromatología II, Facultad de Farmacia, Universidad Comp a r t i c l e i n f o Article history: Received 20 July 2011 a b s t r a c t The evaluation of nutrition industry, especially when r All rights reserved. 1 and B2 in complex cereal foods, , M.C. Sánchez-Mata* nse de Madrid, Pza. Ramón y Cajal s/n, E-28040 Madrid, Spain r functional components in grain products is an important feature for the nt regulations require a correct nutrition labelling, valid during all the shelf SciVerse ScienceDirect eal Science elsevier .com/locate/ jcs l of one method usually depends on the accuracy and sensitivity required and the interferences encountered in the sample matrix. Both vitamin B1 and B2 can be easily determined in pharmaceutical products using HPLC-UV methods; but for foods, more specific techniques of analysis should be applied, due to the presence of a high number of interfering compounds, and different food matrixes (Lynch and Young, 2000). For this purpose, chromato- graphic techniques with ion-exchange or silica columns are of interest to separate vitamins prior to analysis by reverse-phase HPLC or other methods. Most authors use reverse-phase HPLC, with a C18 column and methanol/water as the eluent, although some authors also have reported the use of ion-exchange columns (Hilker and Clifford, 1982) or amide-based columns (Viñas et al., 2003). The presence of buffers is often useful to adjust pH in the solvent and get more reproducible conditions, and the addition of an ion-pair reagent is also applied to improve resolution of the peaks (Ayi et al., 1985; Finglas and Faulks, 1984; Ndaw et al., 2000; Van Niekerk, 1988). Spectrofluorimetric determination, either directly (AOAC, 2011) or coupled to HPLC separation (Arella et al., 1996; Kyritsi et al., 2011; Ndaw et al., 2000; Reyes and Subryan, 1989; Watada and Tran, 1985) is one of the most recommended methods for the determi- nation of these vitamins in foodstufs. Flavines as riboflavin, FAD and FMN present native fluorescence, either in a neutral isoxazoline or oxidated form. However, the reduced forms do not show fluores- cence and their binding to proteins also reduces the fluorescent signal. The oxidized derivative of thiamine (thiocrome) is fluores- cent in alkaline medium, so for the HPLC-fluorimetric method, either a pre- or post-column derivatization is needed (Lynch and Young, 2000; Ohta et al., 1993). Spectrofluorimetric analysis of vitamin B1 and B2 has the advantage of being extremely sensitive, but requires specific equipment (spectrofluorimeter or spectrofluorimetric detector for HPLC) that is not always available for many laboratories (Mondragón-Portocarrero et al., 2011). Some HPLC-UV methods have been described for food products (Albalá-Hurtado et al., 1997; Kamman et al., 1980; Nicholas and Pfender, 1990; Vidal-Valverde and Reche, 1990), but often have the disadvantages of being less sensitive, and with a limited application to various food samples, due to the presence of many interfering compounds absorbing light in the UV range. To solve these problems, careful hydrolysis and purification procedures have to be applied and also the chromatographic conditions have to be improved. In some cases, this induces very long retention times in the chromatograms (40min ormore), which can be accompanied by poor sensitivity and characteristics of the peaks. HPLC-UV analysis of vitamins B1 and B2 requires high amount of sample to make possible the detection of both compounds. In other cases, UV detection is accompanied by a more sensitive and specific technique, as the use of gradients, different wavelengths programs, Diode Array, Coulometric or Mass Spectrometry Detec- tion (Albalá-Hurtado et al., 1997; Aranda and Morlock, 2006; Engel et al., 2010; Heudi et al., 2005; Mandal et al., 2009; Marsza11 et al., 2005; Viñas et al., 2003; Zafra-Gómez et al., 2006), or other tech- niques such as capillary electrophoresis, biosensors or chem- iluminescence (Bai et al., 2008; Gao et al., 2008), which often makes necessary the availability of highly complex equipment. Grains and their derivatives are food matrixes that show special difficulties for this kind of analysis, mainly due to two reasons: first, they are starchy products and often gel when using high amount of sample and hot extraction; and second, they have a lot of inter- fering compounds, especially when analysing industrial complex cereal mixtures including many different ingredients. For that reason, only a few methods use HPLC-UV, and they often require R. San José Rodriguez et al. / Journa294 additional strategies to avoid the problems mentioned above. Due to the significant challenges that must be faced when a specific equipment is not available for the analysis of vitamins B1 and B2 in cereal food products, the contribution of this work has been the establishment of a protocol that combines compatible extraction and separation systems, to get enough sensitivity, precision and accuracy for the analysis of vitamin B1 and B2 in a very complex food matrix, with clean, well-resolved analyte peaks, in a single wavelength simultaneous separation, using very simple and easily available equipment. 2. Experimental 2.1. Reagents and samples Standards of thiamine and riboflavin, as well as enzymes (diastase, alpha-amylase, beta-amylase, acid phosphatase, papain and pepsin) and other common enzymes were purchased from Sigma (St. Louis, USA). HPLC grade solvents were purchased from Symta (Madrid, Spain). Commercial complex cereal products, con- sisting of breakfast cereal food (containing rice, oat, whole wheat, wheat flour and wheat bran) with enrichment levels of 2e2.7 mg/ 100 g of vitamins B1 and B2, were purchased from local markets. 2.2. Instrumentation A PerkineElmer LS-3 Fluorescence Spectrophotometer (Massa- chusetts, USA) was used for this study. The HPLC apparatus con- sisted of a PU II isocratic pumping system; a Jasco (Tokyo, Japan) AS-1555 autosampler; an ERC-Gecko-2000 (Riemerling, Germany) column heater and a Spectra Series UV100 UVeVis detector (Micron Analitica, S.A., Madrid, Spain). For data processing and analysis, Biocrom 2000 3.0 version software (Micron Analitica, S.A., Madrid, Spain) was used. The analytical column was a Purospher�- STAR RP-18e (250 � 4 mm), 5 mm pore size, purchased from Merck (Darmstad, Germany). 2.3. Sample preparation Vitamins B1 and B2 are usually bound to macromolecules as proteins and carbohydrates, as well as in phosphorylated forms, less frequent in vegetal origin tissues (Ball, 1994). For this reason, acidic hydrolysis (with HCl or H2SO4 together with heating at 60e150 �C for 10e30 min), enzymatic hydrolysis (with at least diastase and protease, and in some cases phosphatase, at about 37 �C overnight), or both, has to be applied to the samples in order to extract these vitamins (Arella et al., 1996; Augustin et al., 1985; Matallana González et al., 1998; Ollilainen et al., 1993). According to these previous studies, the most suitable extrac- tion procedure for vitamin B1 and B2 in this kind of samples were chosen through different acid and enzymatic hydrolysis (with overnight incubation at 37 �C) assays in order to separate both vitamins from other compounds in the seeds (mainly proteins and carbohydrates) (Table 1). Some extracts were concentrated after incubation, using a Savant Speed Vac PD121P concentrator. 2.4. Chromatographic conditions According to the literature, different solvent systems with methanol, water and acetate buffer have been used for the sepa- ration of both vitamins in an endcapped-C18 column, with UV detection (268 nm) (Ndaw et al., 2000; Ottaway, 1993; Vidal- Valverde and Reche, 1990). The use of ion-pair reagents (sodium hexanesulfonate or heptanesulfonate) were tried for the separation Cereal Science 55 (2012) 293e299 of both compounds from other interfering substances in the Protocol Acid hydrolysis Heat pH a/ Ac Ac Ac Ac a/ a/ sulp l of 2.5. Reference spectrofluorimetric methods extracts (Albalá-Hurtado et al., 1997; Arella et al., 1996). The conditions assayed are presented in Table 1. number 1 0.1 M HCl 100 �C, 30 min 2.5 M AcN 2 e e 4.5 (5 mM 3 e e 4.5 (5 mM 4 50 mM HCl e a) 2.5 M b) 25 M 5 0.1 N H2SO4 e 2.5 M AcN 6 0.1 N H2SO4 e 2.5 M AcN Mobile phase assays Solvent system Methanol:H2O Na hexane-sulphonate Na heptane- A 50:50 5 mM e B 50:50 e 5 mM C 25:75 e 2.5 mM D 25:75 e 1 mM E 25:75 e 2.5 mM F 25:75 e 2.5 mM G 25:75 e 2.5 mM tR ¼ time of retention. Table 1 Different conditions assayed for the optimization of the analytical procedure. Extraction assays R. San José Rodriguez et al. / Journa The developed HPLC method was compared to the spectro- fluorimetry, as AOAC reference methods (Horwitz and Latimer, 2005), using the same extraction conditions: These methods were as follows: - Vitamin B1: 5 mL of filtered extract was added with 3 mL of 0.3 mg/mL K3Fe(CN)6, in 15% NaOH and 4 mL of 25% KCl for the alkalineoxidationof thiamine to itsfluorescent form (thiocrom). This compound was extracted with isobutanol, and the fluo- rescent signal in the alcoholic mediumwas quickly measured at 365/435nm.Foreach sample, a blankassaywas carriedout,with the addition of 3 mL of 15% NaOH instead of K3Fe(CN)6 solution. The 100% fluorescent signal was adjusted with a quinine sulphate solution, and calibration curves were made with solutions of thiamine chlorhydrate diluted in an alcoholeacidic medium, as reported by AOAC method number 953.17. - Vitamin B2: 10 mL of filtered extract of the samples were oxidized with 0.5 mL of KMnO4, in the presence of glacial acetic acid. After 2 min, the excess of KMnO4 was reduced with H2O2 until discolouration. Fluorescence was measured at 440/ 565 nm, and extracts in which riboflavin was reduced with 20mg de Na2S2O4 were used as blanks. Calibration curves were made with solutions of riboflavin in 0.02 N acetic acid, as reported by AOAC (method number 970.65). 3. Results and discussion Different extraction and HPLC measurement protocols were assayed in order to get the best extraction efficiency and separation conditions in the complex samples analysed. 3.1. Extraction procedure selection For extractions assays, the proportion of sample weight and extraction medium were about 15 g/100 mL, although these proportions were adjusted to get enough amount of vitamins, and improve the handling of the samples. Enzymes Time, temperature 4.5 Diastase 500 mg 18 h, 37 �C Na buffer) Diastase 500 mg 18 h, 37 �C Papain 500 mg Na buffer) Acid phosphatase 10 mg 18 h, 37 �C Na/ 2 Na/ 4.5 a) Pepsin 100 mg b) Acid phosphatase 10 mg a-amylase 40 mg b-amylase 0.75 mg a) 4 h, 37� b) 18 h, 37 �C 4.5 Diastase 500 mg 18 h, 37 �C a-amylase 500 mg 4.5 Diastase 500 mg 18 h, 37 �C a-amylase 500 mg Papain 500 mg honate Na acetate tR (min) vitamin B1 tR (min) vitamin B2 e 3.20 3.37 e 5.56 4.13 e 14.14 23.13 e 22.44 25.06 12.5 mM 8.25 18.67 25 mM 7.26 19.92 50 mM 6.56 20.64 Cereal Science 55 (2012) 293e299 295 From all the extraction assays (Table 1), protocols number 1 (HCl þ heat þ diastase) and number 2 (diastase þ papain) did not provide a complete extraction of vitamin B1, only being detected the vitamin B2. Besides, in protocol number 1, the use of heat in this kind of sample gave rise to great difficulties for the extraction process, since starch formed a gel, and for this reason, it was not possible to use a high amount of sample, as would be required by the low sensitivity of UV detection. As has been previously repor- ted, heat extraction can also partially destroy thiamine, especially in its free form and for this reason, Ball (1998) recommended a low temperature acidic hydrolysis in vegetal products rich in free thiamine. For all these reasons, heat treatment was discarded in this study. Protocols 3 and 4 (using an enzyme cocktail with acid phosphatase, alpha- and beta-amylase and a protease) provided a high amount of interference in the chromatograms, due to all the enzymes used. Hasselmann et al. (1989) obtained a complete dephosphorylation of vitamins B1 and B2 using takadyastase þ beta-amylase, even better than using a phos- phatase. According to Ndaw et al. (2000), there is no need of phosphatase for vegetal origin products, since phosphorylated forms of these vitamins are not abundant (5% of total vitamins in wheat flour) and they usually have some natural phosphatase activity; besides, the use of some amylases can provide also some residual protease and phosphatase activity (Bognår and Ollilainen, 1997). These authors also suggested that acidic hydrolysis can be avoided for thiamine and riboflavin extraction in food samples when an enzymatic cocktail with enough dia- stase and protease activity is used, which was also demon- strated by Ndaw et al. (2000). On the basis of those previous findings, phosphatase was dis- carded from the enzyme cocktail, and only diastases and proteases were used. Besides, low amount of samples and the avoidance of heat are desirable for the vitamin B1 and B2 extraction in starchy samples such as cereals and derivatives. The concentration of the extracts after the extraction avoids using high amount of samples, However, it was not a good alternative, because high concentra- tions of interfering substances also appeared. Someauthors (VanNiekerk,1988) recommendedtheuseofH2SO4 for this kindof analysis. Protocols5and6appliedacidhydrolysiswith H2SO4 þ diastase and alpha-amylase, with better results for the extraction procedure and detection of vitamins. The use of different types of diastases has led to good results in thiamine and riboflavin extraction of different foodstuffs (Ndaw et al., 2000). These authors recommended acid extraction only for starchy samples, due to the improvement in the filtration process, and indicated the possible substitution of acid hydrolysis by amylase activity for this purpose. These authors also suggested that the presence of protease activity in some diastase enzymes is often enough for a total extraction of vitamins, evenwithout acid hydrolysis. As there can be great differences regarding the activity of different diastases, in the present study, the addition of papain to the extraction medium (H2SO4 þ diastase and alpha-amylase) was considered useful and provided a better clean-up of some inter- fering substances, in agreement with Russel (1996) who recom- mended the use of papain for thiamine analysis, due to its high protease activity. The combination papain þ diastase has been successfully applied in previous studies, for vitamin B group hydrolysis in other types of vegetal products (Sánchez-Mata et al., 2003). In these conditions, some variations were introduced to avoid interferences of some protein compounds in the extracts, based on Nicholas and Pfender (1990): addition of Na acetate, change of pH to 1.7e2 and after to 4e4.2, and purification through Sep Pack C18 cartridge. These strategies did not improve the extraction and resolution of the vitamins studied. 3.2. Optimization of HPLC method From all the HPLC conditions assayed (Table 1), the A solvent system (methanol/water 50/50 with sodium hexanesulphonate) resulted in coelution of thiamine and riboflavin. However the change of the ion-pair reagent to sodium heptanesulphonate (B solvent system number) permitted the separation of both vita- mins. Increasing the water content in the mobile phase resulted in much longer retention time for riboflavin and better separation between both compounds and the interferences. The less the amountof sodiumheptanesulphonate, the longer the retention time of both vitamins (solvent systems BeD), and its concentration was fixed at 2.5 mM, as it provides better resolution between the peaks. Some problems related to poor reproducibility of retention time of thiamine were detected using these conditions. As has been previously reported by Vidal-Valverde and Reche (1990) and Nicholas and Pfender (1990), thiamine is very sensitive to pH changes in the mobile phase. In the assays performed in this study, the inclusion of sodium acetate in the eluent provided a fixed pH which improved the reproducibility of the retention time of vitamin B1. When increasing the concentration of sodium acetate in the mobile phase between 12.5 and 50 mM (solvent systems EeG), eal ase R. San José Rodriguez et al. / Journal of Cereal Science 55 (2012) 293e299296 Fig. 1. Chromatograms of a) standards of thiamine and riboflavin; b) commercial cer conditions: Purospher�STAR RP-18e (250 � 4 mm), 5 mm pore size (Merck); mobile ph acetate; ldetection ¼ 268 nm. mixture sample; and c) chromatographic parameters of the peaks. Chromatographic methanol/water (25:75 v/v) þ 2.5 mM sodium heptanosulfonate and 12.5 mM sodium close using both methods, for either vitamin B1 or B2. The RSD of r r2 Response factors (X � SD) RSD (%) 0.9978 99.57% 4642.0 � 159.2 3.43 0.9986 99.73% 966.1 � 42.0 4.35 0.9981 99.63% 10800.9 � 516.8 4.78 0.9985 99.71% 156.7 � 10.9 6.93 evia l of Cereal Science 55 (2012) 293e299 297 shorter retention times for thiamine and longer for riboflavin were obtained. The best conditions were those of solvent system E, which provided retention times of 8 � 0.3 min for thiamine and 18 � 0.3 min for riboflavin (Fig. 1), and therefore good resolution and retention times long enough to separate vitamins from other interferences with total runs no longer than 20 min. Therefore, the mobile phase selected for analysis was: 12.5 mM sodium acetate in a mixture of methanol/water 25/75 þ 2.5 mM sodium heptane- sulphonate, at a flow rate of 0.9 mL/min. In these conditions, the suitability of both peaks was evaluated by the calculation of the following chromatographic parameters (Hsu and Chien, 1994): N ¼ 16ðtR=wÞ2 (1) Hp ¼ L=N (2) k0 ¼ �tR=tRsolv: �� 1 (3) T ¼ W0:05=2f (4) R ¼ 2�tR2 � tR1 ��� W1=2ð2Þ �W1=2ð1Þ � (5) whereN is the number of theoretical plates;Hp is the height of each plate; tR, tR1 and tR2 are the retention times of each compound; tRsolv: is the dead time; w is the width of the peak at the baseline;W0.05 is the width of the peak at 5% of the peak height;W1/2 (1) andW1/2 (2) are the width of the peak at half-height for each compound; L is the length of the column; k0 is the retention factor; T is the tailing factor; f is the distance between the perpendicular dropped from the peak maximum and the leading edge of the peak at 5% of the peak height; and R is the resolution factor. In the applied conditions, good characteristics for the chro- matographic peaks were obtained (Fig. 1), with high column performance, calculated as the apparent number of theoretical plates and low value of Hp. Good retention factors (between 1 and 5), were obtained for both thiamine and riboflavin, which showed symmetric peaks (tailing factor between 0.8 and 1.5), as recom- mended by Hsu and Chien (1994), Skoog et al. (2001) and British Table 2 Linearity of HPLC and spectrofluorimetric methods. Intercept (X � SD) Slope (X � SD) Vitamin B1 HPLC �8633.9 � 3323.0 4872.1 � 120.7 Spectrofluorimetry �6.5 � 12.3 909.5 � 16.8 Vitamin B2 HPLC �7856.4 � 4533.1 11404.0 � 261.3 Spectrofluorimetry �2.4 � 1.3 166.1 � 4.4 X ¼ Mean value (n ¼ 3); SD ¼ Standard deviation (n � 1); RSD ¼ relative standard d R. San José Rodriguez et al. / Journa Pharmacopoeia (2004). The resolution factor was above the minimum 1.5 recommended by these authors, due to the long separation between the peaks. 3.3. Final analytical procedure Considering all previous assay results, the final selected extraction method was as follows: 15 g of cereal sample was mixed with 90 mL 0.1 N H2SO4, adjusted to pH 4.5 with 2.5 M Na acetate and added with papain, diastase and alpha-amylase (500 mg each). Samples were incubated at 37 �C overnight and made up to 100 mL with distilled water, filtered through Whatman 40 paper (Maid- stone, Kent, England) and through 0.45 mm polyvinylidene fluoride the slopes were always below 2.4%, and those of the response factors were not higher than 6.9%. 3.4.2. Precision Precision of the instrumental technique was evaluated, ana- lysing 6 standards of both vitamins in the same day (repeatability) and in different days (reproducibility), in amounts of 0.05e0.1 mg/ mL for spectrofluorimetry, and 20e40 mg/mL for HPLC. For the whole procedure including sample extraction and instrumental analysis, 6 equal cereal samples were analysed by both methods, also in the same and in different days. (PVDF) membrane (Millipore, Bedford, USA). This extract was injected in the chromatographic system, using a Purospher�- STAR RP-18e (250 � 4 mm, 5 mm) column (Merck) and 12.5 mM sodium acetate in a mixture of methanol/water 25/75 þ 2.5 mM sodium heptanesulphonate, as eluant at a flow rate of 0.9 mL/min and ldetection ¼ 268 nm. The identity of the peaks obtained was confirmed by the addi- tion of standards to the extracts, and excellent results were obtained using these extraction conditions, with no interference with the chromatographic peaks and the quantification of the expected amount of vitamins. 3.4. Validation of HPLC method The developed HPLC method was validated for enriched cereal products. Linearity, precision, accuracy and limit of detection (LOD) for thiamine and riboflavin were evaluated compared to the spec- trofluorimetric standard method (Tables 2e5). 3.4.1. Linearity Triplicate calibration curves of thiamine and riboflavin were performed from standards (in 10 different concentrations between 0.05 and 75 mg/mL), and linear calibration curves were obtained for a concentration range including those of the samples analysed. The linearity of the method was confirmed by regression statistics. As can be deduced from the calibration parameters (Table 2), determination coefficients were always above 99.57%, being very tion. All the RSD obtained in this study (Table 3) were below the limits of 11%, considered as maximum for substances around 1 mg/mL according to AOAC (1993). Slight variations in HPLC repeatability are Table 3 Precision assays (% RSD). HPLC Spectrofluorimetry Vitamin B1 Vitamin B2 Vitamin B1 Vitamin B2 Standards repeatability 2.84 4.46 4.86 6.44 Samples repeatability 4.78 4.88 3.29 1.85 Standards reproducibility 0.353 4.6 7.43 2.38 Samples reproducibility 10.17 2.67 5.16 7.36 Table 4 Assays for recovery percents. Vitamin B1 Vitamin B2 X � SD (%) X � SD (%) HPLC CL 1 95.82 � 9.18 100.33 � 10.58 CL 2 85.96 � 17.27 109.37 � 11.22 Mean 93.32 � 12.8 104.85 � 11.6 Spectrofluorimetry CL 1 101.62 � 15.50 111.06 � 18.67 CL 2 94.13 � 13.12 107.04 � 17.23 Mean 97.87 � 19.3 109.45 � 15.6 X ¼mean value (n ¼ 3); SD ¼ standard deviation (n � 1); CL ¼ Concentration level. R. San José Rodriguez et al. / Journal of298 due to small volume injection or integration variations, and were comparable to those obtained using spectrofluorimetry. 3.4.3. Accuracy assays Accuracy and matrix effects of both methods were evaluated by the recovery percents. Two addition levels (between 15 and 200 mg for vitamin B1 and between 30 and 400 mg for vitamin B2) of vitamin B1 and B2 standards were added to complex cereal samples (Table 4). Triplicate analyses were performed for each addition level, as well as for samples without addition, using both methods. Mean recovery percents for the HPLC method ranged between 85.9% and 109.4%. All these values are in the interval accepted by AOAC (1993) for substances around 10 ppm (80e110%), as it is the case for these vitamins. Non-statistical differences (p < 0.05) were detected for the recovery values using both analytical methods, but RSD were lower for recovery percents in HPLC than the spectro- fluorimetric method. No statistically significant differences (p < 0.05) were found from the comparison of the results obtained for analysis of vitamin B1 and B2 in complex cereal commercial samples (p < 0.05) using both methods (Table 5), so HPLC can be used for vitamin B1 and B2 analysis in complex cereal samples, with no loss of accuracy with respect to the AOAC recommended method, and lower RSD for recovery percents in HPLC. Both, the results obtained from HPLC and spectrofluorimetry were coincident, but different from the level of enrichment indicated in labelling. In the case of vitamin B1, the levels found by both methods were half of the added level, which could be indicative of the instability of thiamine during the shelf life of the product. In the case of vitamin B2, the analytical levels found were coincident or higher than the addition level, meaning that the endogenous content of riboflavin may not have been taken into account in the labelling (especially in sample number 2, which included oat and wheat bran). These facts confirm the necessity of accessible methodology to assess the final vitamin contents in enriched food products, for labelling purposes, in agreement with official regulations. 3.4.4. Limits of detection The sensitivity of the method was evaluated as the limit of detection estimated following ICH Guidelines (1997). For HPLC Table 5 Analysis of cereal samples using HPLC and spectrofluorimteric methods. Enrichment level (mg/100 g) HPLC X � SD (mg/100 g) Spectrofluorimetry X � SD (mg/100 g) Vitamin B1 Complex cereal mixture 1 2.3 1.14 � 0.20 1.26 � 0.06 Complex cereal mixture 2 2.3 1.30 � 0.29 0.90 � 0.22 Vitamin B2 Complex cereal mixture 1 2.7 2.75 � 0.07 2.85 � 0.20 Complex cereal mixture 2 2.0 3.57 � 0.10 3.48 � 0.10 X ¼ mean value (n ¼ 3); SD ¼ standard deviation (n � 1). analytical methods, the limit of detection (LOD) can be defined as the concentration resulting from a signal/noise ratio of 3. The LOD obtained for the HPLC method were 62.8 ng for vitamin B1 and 21.9 ng for vitamin B2, corresponding to 0.628 mg/100 g and 0.219 mg/100 g, respectively in the cereal samples. For spectrofluorimetry, the LOD can be calculated as: LD ¼ 3S=m (6) where S is the standard deviation of the intercept, and m is the slope of the calibration curve. Using this equation, the LOD for spectrofluorimetry is 0.0406 mg/mL (vitamin B1) and 0.025 mg/mL (vitamin B2), corresponding to 0.203 mg/100 g and 0.125 mg/100 g, respectively in the cereal samples. As expected, spectrofluorimetry is more sensitive than HPLC, which however was sensitive enough to analyse this kind of cereal sample. According to the assays performed and the results found, the developed protocol was suitable for grains and derivatives analysis (whole grains, flours, complex mixtures) in terms of no significant matrices effects, with the only limitation of the LOD and LOQ. 4. Conclusion The proposed method provides comparable linearity and accu- racy as the reference spectrofluorimetric method, being less sensi- tive but more precise for vitamin B1 and B2 analysis on different complex cereal samples. It also has the advantages of the rapid and simultaneous determination of both vitamins in a single chro- matographic run (samewavelength), being less time consuming and avoiding theuse of asmany reagents as officialmethods,with a quite simple sample preparation, and easily available equipment (isocratic HPLC-UV). The election of the analytical method depends on the equipment available and the sensitivity requirements of the analysis and, in this sense, the developed HPLC method can be of great interest for the analysis of complex enriched cereal products over 0.63 mg/100 g in vitamin B1 and 0.22 mg/100 g in vitamin B2, which are subject to special labelling regulations. Acknowledgements This work has been funded by University Complutense of Madrid (PR78/02-11037), and a grant for R. San Jose Rodriguez was provided by the Spanish Government. References Albalá-Hurtado, S., Veciana-Nogués, M.T., Izquierdo-Pulido, M., Marine-Font, A., 1997. Determination of water-soluble vitamins in infant milk by high- performance liquid chromatography. Journal of Chromatography 778, 247e253. AOAC, 2011. In: Horwitz, W., Latimer Jr., G. (Eds.), Official methods of analysis of AOAC international, 18th ed, Revision 4. AOAC, Gaithersburg. AOAC, 1993. Peer Verified Methods Program. AOAC, Manual on Policies and Procedures, Nov. 1993. AOAC International, Arlington. Aranda, M., Morlock, G., 2006. Simultaneous determination of riboflavin, pyri- doxine, nicotinamide, caffeine and taurine in energy drinks by planar chromatography-multiple detection with confirmation by electrospray ioniza- tion mass spectrometry. Journal of Chromatography A 1131 (1e2), 253e260. Arella, F., Lahély, S., Bourgignon, J.B., Hasselmann, C., 1996. Liquid chromatographic determination of vitamins B1 and B2 in foods. A collaborative study. Food Chemistry 56 (1), 81e86. Augustin, J., Klein, B.P., Becker, D., Vernugopal, P.B., 1985. Methods of Vitamin Assay. John Willey& Sons, New York. Ayi, B.K., Yuhas, D.A., Moffett, K.S., Joyce, D.M., Deangelis, N.J., 1985. Liquid chro- matographic determination of thiamine in infant formula products by using ultraviolet detection. Journal of AOAC 68, 1087e1092. Bai, J., Hun, X., Liu, Q., Chen, J.C., Bai, L., 2008. Capillary electrophoresis detection scheme for vitamin B1 and vitamin B2 based on the potassium hex- acyanoferrate(III) chemiluminescence system. Microchimica Acta 160, 165e171. Cereal Science 55 (2012) 293e299 Ball, G.F.M., 1994. Water Soluble Vitamins Assays in Human Nutrition. Chapman and Hall, London. Ball, G.F.M., 1998. Bioavailability and Analysis of Vitamins in Foods. Chapman and Hall, London. Bognår, A., Ollilainen, V., 1997. Influence of extraction on the determination of vitamin B6 in food by HPLC. Z. Lebensmittel-Untersuchung und -Forschung A 204, 327e335. British Pharmacopoeia Commission, 2004. British Pharmacopoeia. The Stationery Office, London. Engel, R., Stefanovits-Bánvai, E., Abrankó, L., 2010. LC simultaneous determination of the free forms of B group vitamins and vitamin C in various fortified food products. Chromatographia 71, 1069e1074. Finglas, P.M., Faulks, R.M., 1984. The HPLC analysis of thiamine and riboflavin in potatoes. Food Chemistry 15 (1), 37e44. Gao, Y.L., Guo, F., Gokavi, S., Chow, A., Sheng, Q.H., Guo, M.R., 2008. Quantification of water-soluble vitamins in milk-based infant formulae using biosensor-based assays. Food Chemistry 110, 769e776. Hasselmann, C., Franck, D., Grimm, P., Diop, P.A., Soules, C., 1989. High-performance liquid chromatography analysis of thiamine and riboflavin in dietetic foods. Journal of Micronutrient Analysis 5, 269e279. Heudi, O., Kilinç, T., Fontannaz, P., 2005. Separation of water-soluble vitamins by reversed-phase high performance liquid chromatography with ultra-violet detection: application to polyvitaminated premixes. Journal of Chromatog- raphy A 1070 (1e2), 49e56. Hilker, D.M., Clifford, A.J., 1982. Thaimin analysis and separation of thiamine phosphate esters by high-performance liquid chromatography. Journal of Chromatography 231, 433. Horwitz, W., Latimer, G.W., 2005. Official Methods of Analysis of AOAC Interna- tional, eighteenth ed. AOAC, Gaithersburg. Hsu, H.-C., Chien, C.-S., 1994. Validation of analytical methods: a simple model for HPLC assay methods. Journal of Food and Drug Analysis 2 (3), 161e176. ICH Guidelines, 1997. Q2B Federal Registers, vol. 62, pp. 27463e27467. Illera Martín, M., Illera Del Portal, J., Illera Del Portal, J.C., 2000. Vitaminas y min- erales. In: Colección estudios Complutenses, 1aed. Ed. Complutense, S.A, Madrid. Kamman, J.F., Labuza, T.P., Warthesen, J.J., 1980. Thiamin and riboflavin analysis by high performance liquid chromatography. Journal of Food Science 45, 1497e1504. Kyritsi, A., Tzia, C., Karathanos, V.T., 2011. Vitamin fortified rice grain using spraying Matallana González, M.C., Cámara Hurtado, M., Martínez Tomé, M.J., 1998. Estudio del contenido de vitaminas hidrosolubles (B1, B2, B6, C) en lechuga (Lactuca sativa L.) procedente de cultivo ecológico. Alimentaria 293, 39e43. Mondragón-Portocarrero, A.C., Vázquez-Odériz, L., romero-rodríguez, M.A., 2011. Development and validation of an HPLC method for the determination of thiamine and riboflavin in green leafy vegetables using Clara-Diastase. Journal of Food Science 76 (4), C639eC643. Ndaw, S., Bergaentzlé, M., Aoudé-Werner, D., Hasselmann, C., 2000. Extraction procedures for the liquid chromatographic determination of thiamine, ribo- flavin and vitamin B6 in foodstuffs. Food Chemistry 71, 129e138. Nicholas, E.C., Pfender, K.A., 1990. Fast and simple liquid chromatography deter- mination of non phosphorylated thiamine in infant formula milk and other foods. Journal of AOAC 73 (5), 792e798. Ohta, H., Maeda, M., Nogata, Y., 1993. A simple determination of thiamine in rice (Oryza sativa L.) by high-performance liquid chromatography with postcolumn derivatization. Journal of Liquid Chromatography 16, 2617e2629. Ollilainen, V., Vahteristo, L., Uusi-Rauva, A., Varo, P., Koivistoinen, P., Huttunen, J., 1993. The HPLC determination of total thiamine (Vitamin B1) in foods. Journal of Food Composition and Analysis 6, 152e165. Ottaway, P.B., 1993. The Technology of Vitamins in Food. Blackie Academic and Professional, Glasgow. Reyes, E.S.P., Subryan, L., 1989. An improved method of simultaneous HPLC and thiamine in selected cereal products. Journal of Food Composition and Analysis 2 (1), 41e47. Russel, F., 1996. Water-soluble vitamins. In: Nollet, L.M.L. (Ed.), Handbook of Food Analysis. Marcel Dekker Inc., New York, pp. 649e713. Sánchez-Mata, M.C., Cámara-Hurtado, M., Díez-Marqués, C., 2003. Extending shelf- life and nutritive value of green beans (Phaseolus vulgaris L.), by controlled atmosphere storage: micronutrients. Food Chemistry 80, 317e322. Skoog, D.A., Holler, F.J., Nieman, T.A., 2001. Principios de análisis instrumental, 5a ed. Mc Graw Hill, Madrid. Van Niekerk, P.J., 1988. Determination of vitamins. In: Macrae, R. (Ed.), HPLC in Food Analysis. Academic Press, London, pp. 133e184. Vidal-Valverde, C., Reche, A., 1990. Reliable system for the analysis of riboflavin in foods by high performance liquid chromatography and UV detection. Journal of Liquid Chromatography 13 (10), 2089e2101. R. San José Rodriguez et al. / Journal of Cereal Science 55 (2012) 293e299 299 and soaking methods. LWT e Food Science and Technology 44, 312e320. Lynch, P.L.M., Young, I.S., 2000. Determination of thiamine by high-performance liquid chromatography. Journal of Chromatography A 881 (1e2), 267e284. Mandal, S.M., Mandal, M., Ghosh, A.K., Dey, S., 2009. Rapid determination of vitamin B-2 and B-12 in human urine by isocratic liquid chromatography. Analytica Chimica Acta 640, 110e113. Marsza11, M.L., Lebiedzi�nska, A., Czarnowski, W., Szefer, P., 2005. High-performance liquid chromatography method for the simultaneous determination of thiamine hydrochloride, pyridoxine hydrochloride and cyanocobalamin in pharmaceu- tical formulations using coulometric electrochemical and ultraviolet detection. Journal of Chromatography A 1094 (1e2), 91e98. Viñas, P., López-Erroz, C., Balsalobre, N., Hernández-Córdoba, M., 2003. Reversed- phase liquid chromatography on an amide stationary phase for the determi- nation of the B group vitamins in baby foods. Journal of Chromatography A 1007 (1e2), 77e84. Watada, A.E., Tran, T.T., 1985. A sensitive high-performance liquid chromatography method for analyzing riboflavin in fresh fruits and vegetables. Journal of Liquid Chromatography 8, 1651e1660. Zafra-Gómez, A., Garballo, A., Morales, J.C., García-Ayuso, L.E., 2006. Simultaneous determination of eight water-soluble vitamins in supplemented foods by liquid chromatography. Journal of Agricultural and Food Chemistry 54 (13), 4531e4536. Simultaneous determination of vitamin B1 and B2 in complex cereal foods, by reverse phase isocratic HPLC-UV 1. Introduction 2. Experimental 2.1. Reagents and samples 2.2. Instrumentation 2.3. Sample preparation 2.4. Chromatographic conditions 2.5. Reference spectrofluorimetric methods 3. Results and discussion 3.1. Extraction procedure selection 3.2. Optimization of HPLC method 3.3. Final analytical procedure 3.4. Validation of HPLC method 3.4.1. Linearity 3.4.2. Precision 3.4.3. Accuracy assays 3.4.4. Limits of detection 4. Conclusion Acknowledgements References