Two International Round-Robin Studies Showed Good Comparability of 5-Methyltetrahydrofolate but Poor Comparability of Folic Acid Measured in Serum by Different High-Performance Liquid Chromatography-Tandem Mass Spectrometry Methods

Background: Serum folate methods produce different results. The comparability of HPLC-mass spectrometry (MS)/MS methods is not well documented.Objective: We conducted an international "round-robin" investigation to assess the comparability, precision, and accuracy of serum folate HPLC-MS/MS methods.Methods: The CDC laboratory, 7 laboratories with independently developed methods (group 1), and 6 laboratories with an adapted CDC method (group 2) analyzed folate forms in 6 serum pools and 6 calibrators from the CDC (duplicate analysis over 2 d) and in two 3-level reference materials (duplicate analysis).Results: All laboratories measured 5-methyltetrahydrofolate (5-methylTHF) and folic acid; some measured additional folate forms. The geometric mean (range) concentrations (nanomoles per liter) for 5-methylTHF in the 6 serum pools were 18.3 nmol/L (CDC), 13.8-28.9 nmol/L (group 1), and 16.8-18.6 nmol/L (group 2); for folic acid the concentrations were 3.42 nmol/L (CDC), 1.09-4.74 nmol/L (group 1), and 1.74-2.90 nmol/L (group 2). The median imprecision (CV) for 5-methylTHF was 4.1% (CDC), 4.6-11% (group 1), and 1.7-6.0% (group 2); for folic acid it was 6.9% (CDC), 4.9-20% (group 1), and 3.9-23% (group 2). The mean ± SD (range) recovery of 5-methylTHF spiked into serum was 98% ± 27% (59-138%) for group 1 and 98% ± 10% (82-111%) for group 2; for folic acid it was 93% ± 29% (67-198%) for group 1 and 81% ± 16% (64-102%) for group 2. The mean relative bias for 5-methylTHF compared with the reference material certificate value was 12% (CDC), -24% to 30% (group 1), and -0.6% to 16% (group 2); for folic acid it was 73% (CDC), -47% to 578% (group 1), and -3.3% to 67% (group 2).Conclusions: For 5-methylTHF, group 2 laboratories demonstrated better agreement and precision, less variable spiking recovery, and less bias by using a reference material. Laboratory performance for folic acid was highly variable and needs improvement. Certified reference materials for serum folate forms and total folate are needed to improve method accuracy.