First BCR-intercomparison on the determination of folates in food

Abstract

The first BCR intercomparison on the determination of folate in food was designed to study the state-of-the-art in folate analysis in a group of experienced vitamin laboratories in Europe. In all 15 participants from 8 countries took part using microbiological and HPLC procedures, enzyme protein-binding assays (EPBA) and commercial radioprotein-binding assay kits (RPBA). The participants were asked to quantify folate levels in a lyophilised Brussels sprout material, which had been specifically prepared as a candidate reference material for vitamin work, using their preferred method of analysis. Three types of deconjugation were also investigated, human plasma (HP), chicken pancreas (CP) and hog kidney (HK). Generally good agreement was obtained by participants using the microbiological procedures. Folate levels determined after CP deconjugate treatment (mean = 984 μg per 100 g dry matter, SD = 237, n = 6) were 19% higher than those levels found after HP deconjugation (mean = 824 μg per 100 g dry matter, SD = 147, n = 6). The use of autoclaving followed by deconjugation with either HP or CP enzymes gave lower (10–20%) folate levels (determined by microbiological assay) when compared to refluxing and deconjugation with the same enzymes. Hog kidney deconjugase enzyme and autoclaving/refluxing was not as effective. Although the HPLC results from the 2 participants who were able to complete the study agree reasonably well with the microbiological data, there were differences in the proportion of the individual folate forms measured. One participant found 5-CH 3THF (55%), THF (25%) and 5-CHOTHF (20%), whereas the other only initially reported 5-CH 3THF but later confirmed small amounts (10–15%) of THF and 5-CHOTHF forms. Despite the use of HPLC with fluorometric detection, there were some problems in peak identification and calibration. The use of HPLC with UV detection gave unsatisfactory results due to difficulties in resolution of folate compounds and these results were excluded. RPBA results were generally higher (50–60%) than both the microbiological and HPLC results but more variable. EPBA results also varied between the three laboratories using this format but the mean folate content (HP only) agreed favourably with both the HPLC and microbiological results. The major problem identified with these methods was the response of the individual folate forms to the binding-protein used. Careful control of assay pH and calibrant are required if these methods are to be applied to the determination of food folates. Future work will focus on improvements in methodology of each type of assay prior to further intercomparisons.