Standardisation of HPLC techniques for the determination of naturally-occurring folates in food

The aim of this work was to evaluate current in-house HPLC procedures for the determination of naturally-occurring folates in food, and to identify problem areas for further improvement. Five intercomparison studies were completed over the period 1990-1997 in which nine participants from six countries took part. Through careful validations and detailed discussions held at evaluation meetings, possible biases and sources of systematic error have been identified and reduced. The use of ascorbic acid and nitrogen flushing during extraction, sample clean-up using strong anion exchange columns, spectrophometrically calibrated standards and fluorescence detection are all recommended. Both in-house hog kidney and human plasma deconjugase enzymes gave similar results to the circulated common hog kidney enzyme which was prepared from fresh pig's kidneys. The most consistently reported values were for 5-CH3H4-PteGlu, and to a lesser extent, for H(4)PteGlu. Four candidate reference materials (CRM 121, wholemeal flour; CRM 421, milk powder; CRM 485, lyophilised mixed vegetables, and CRM 487, lyophilised pig's liver) have been proposed with both indicative values (mean +/- uncertainty) for 5-CH3H4-PteGlu in CRM 421 (0.25; +/- 0.02 mg/kg) and CRM 485 (2.14; +/- 0.42 mg/kg), and information values (mean; range) for 5-CH3H4-PteGlu in CRh4 121 (0.04; 0.03-0.08 mg/kg) and CRM 487 (2.6; 1.9-3.8 mg/kg), Certified values are also given for total folate by microbiological assay: CRM 121 (0.50; +/- 0.07 mg/kg), CRM 421 (1.42; +/- 0.14 mg/kg), CRM 485 (3.15; 0.28 mg/kg), and CRM 487 (13.4; 1.3 mg/kg). Average recovery of 5-CH3H4-PteGlu, added prior to extraction and deconjugation, was 91% (84-95%) for the four CRMs. The average within- and between-laboratory variations were 6 and 15% for the determination of 5-CH3H4-PteGlu by HPLC, and 9 and 18% for the determination of total folate by microbiological assay. These CRMs will be used for quality control of folate measurements for nutritional labelling, and validation of new techniques. Further methodology work is required for the HPLC analyses of folate forms other than 5-CH3H4-PteGlu. (C) 1998 Elsevier Science Ltd. All rights reserved.