Maternal gestational vitamin D supplementation and offspring bone health (MAVIDOS): a multicentre, double-blind, randomised placebo-controlled trial.

UK multicentre, double-blind, randomised, placebo-controlled trial at three sites (Southampton, Oxford, Sheffield). Pregnant women >18 years with singleton pregnancies recruited before 17 weeks' gestation, baseline 25(OH)D 25–100 nmol/L at 10–17 weeks, randomised 1:1 to vitamin D supplementation or placebo until delivery. Primary outcome: neonatal whole-body BMC by DXA within 2 weeks of birth; analyses by intention-to-treat. 1134 randomised (569 placebo; 565 cholecalciferol); 965 delivered; 370 placebo and 367 cholecalciferol had usable DXA scans for the primary endpoint.

Cholecalciferol 1000 IU/day, orally, from 14 weeks' gestation until delivery; taken in addition to standard antenatal multivitamins containing up to 400 IU/day vitamin D.

Primary outcome: neonatal whole-body BMC at birth did not differ significantly between groups (61.6 g vs 60.5 g; p=0.21). A prespecified season‑of‑birth interaction showed greater benefit for winter births with cholecalciferol (mean difference 5.5 g; 95% CI 1.8–9.1; p=0.004); winter births also showed differences in bone area (11.5 cm2; p=0.05), BMD (0.01 g/cm2; p=0.04), and BMC adjusted for length (3.7 g; p=0.03) as well as fat mass (113.6 g; p=0.008). No significant differences in birthweight, length, spine indices, or lean mass. Safety: no treatment-related adverse events; postpartum haemorrhage was more frequent in placebo (17% vs 12%; p=0.01). Conclusions: supplementation did not increase overall neonatal bone mass but maintained maternal vitamin D status and was safe. A winter‑delivery interaction suggests possible season-specific benefits; replication in other populations is needed before changing clinical guidance. Ongoing MAVIDOS follow-up will clarify longer-term outcomes.

Limitations include exclusion of participants with baseline 25(OH)D <25 or >100 nmol/L; underrepresentation of nonwhite participants; neonatal DXA measurement challenges with movement artefacts leading to data loss (usable DXA data for primary endpoint: 370 placebo, 367 cholecalciferol); secondary analyses were underpowered and susceptible to false positives due to multiple comparisons; findings by season require replication in other populations.