Validation of summer and winter ELISA measurements of serum 25-hydroxyvitamin D concentrations in Mongolia.

BACKGROUND AND OBJECTIVES Assay cost, quality, and availability pose challenges for vitamin D surveys in limited resource settings. This study aimed to validate an inexpensive vitamin D assay (ELISA) under real-world conditions in Mongolia, the northernmost developing country, to characterize the assay's usefulness and inform the design of epidemiologic studies in similar regions. METHODS AND STUDY DESIGN We collected paired summer and winter serum samples from 120 men and women (aged 20-57 years) in urban and rural Mongolia, analyzed each sample for 25(OH)D concentration using both Immunodiagnostic Systems ELISA and DiaSorin LIAISON 25(OH)D TOTAL, and compared the assays using multiple statistics. LIAISON was itself validated by participation in the DEQAS program. RESULTS Correlation and agreement between assays were higher in summer (Pearson's correlation=0.60, Spearman's rank correlation=0.67, Lin's concordance correlation=0.56) than winter (rP=0.37, rS=0.43, rC=0.33), although ELISA less accurately assigned subjects to sufficiency categories in summer (percent agreement=44%) than winter (58%), during the latter of which most subjects were deficient ([25(OH)D] categories used: >75 nmol/L (optimal), 50-75 nmol/L (adequate), 25-50 nmol/L (inadequate), <25 nmol/L (deficient)). Compared with LIAISON, ELISA tended to indicate higher vitamin D status in both seasons (mean paired difference: 7.0 nmol/L (95% CI: 3.5-10.5) in summer, 5.2 nmol/L (95% CI: 2.9-7.5) in winter). CONCLUSIONS ELISA proved useful for measuring and ranking subjects' vitamin D status in Mongolia during summer, but levels were too low in winter to sensitively discriminate between subjects, and ELISA overestimated status in both seasons. These findings have implications for the timing and interpretation, respectively, of vitamin D surveys in highly deficient populations.