Individual participant data (IPD)-level meta-analysis of randomised controlled trials with vitamin D-fortified foods to estimate Dietary Reference Values for vitamin D.

Abstract

Individual participant data-level meta-regression (IPD) analysis is superior to meta-regression based on aggregate data in determining Dietary Reference Values (DRV) for vitamin D. Using data from randomized controlled trials (RCTs) with vitamin D3-fortified foods, we undertook an IPD analysis of the response of winter serum 25-hydroxyvitamin (25(OH)D) to total vitamin D intake among children and adults and derived DRV for vitamin D. IPD analysis using data from 1429 participants (ages 2-89years) in 11 RCTs with vitamin D-fortified foods identified via a systematic review and predefined eligibility criteria. Outcome measures were vitamin D DRV estimates across a range of serum 25(OH)D thresholds using unadjusted and adjusted models. Our IPD-derived estimates of vitamin D intakes required to maintain 97.5% of winter 25(OH)D concentrations ≥ 25 and ≥ 30nmol/L are 6 and 12µg/day, respectively (unadjusted model). The intake estimates to maintain 90%, 95% and 97.5% of concentrations ≥ 50nmol/L are 33.4, 57.5 and 92.3µg/day, respectively (unadjusted) and 17.0, 28.1 and 43.6µg/day, respectively (adjusted for mean values for baseline serum 25(OH)D, age and BMI). IPD-derived vitamin D intakes required to maintain 90%, 95% and 97.5% of winter 25(OH)D concentrations ≥ 50nmol/L are much higher than those derived from standard meta-regression based on aggregate data, due to the inability of the latter to capture between person-variability. Our IPD provides further evidence that using food-based approaches to achieve an intake of 12µg/day could prevent vitamin D deficiency (i.e., serum 25(OH)D < 30nmol/L) in the general population.