Prenatal Vitamin D Supplementation and Child Respiratory Health: A Randomised Controlled Trial

Randomized controlled trial in pregnant women attending St Mary's Hospital in London, United Kingdom, enrolled from April to November 2007. The trial was stratified by four ethnic groups in a 1:1:1 design and followed offspring to age 3 years for wheeze, eczema, atopy, respiratory health, and related immune outcomes.

Pregnant women were randomized at 27 weeks' gestation to either oral ergocalciferol 800 IU daily until delivery or a single oral bolus of 200,000 IU cholecalciferol, compared with no vitamin D supplementation. Participants were instructed to avoid other multivitamin supplements containing vitamin D.

Prenatal vitamin D supplementation in late pregnancy did not reduce childhood wheeze, eczema, atopy, lung function abnormalities, or inflammatory markers at age 3 years. For wheeze ever, daily vitamin D was 11/56 (20) versus 14/50 (28) in control (RR 0.70, 95% CI 0.35 to 1.40; P = 0.31), while bolus vitamin D was 15/52 (29) versus 14/50 (28) (RR 1.03, 95% CI 0.56 to 1.91; P = 0.93). Most outcomes were null, but the bolus group had higher bronchodilator use (14/48 [29] vs 4/49 [8]; RR 3.57, 95% CI 1.27 to 10.09; P = 0.008) and more lower respiratory tract infection events (17/47 [36] vs 11/50 [22]; aOR 2.87, 95% CI 1.03 to 8.03; P = 0.05), though these signals were not considered robust after multiple-testing adjustment. Cord blood vitamin D increased modestly, but this did not translate into clear clinical benefit.

The active-arm sample sizes were small and several outcomes had substantial missing data or incomplete objective measurements, reducing precision and increasing the risk of chance findings. Follow-up to age 3 years may have been too short to detect later respiratory or allergic effects, and the two vitamin D regimens were heterogeneous, making the overall intervention effect harder to interpret. The population was a relatively vitamin D deficient, ethnically mixed urban cohort, which may limit generalizability.