In the multi-centre MAVIDOS trial, pregnancy vitamin D supplementation had a bene fi cial effect on neonatal bone mass amongst winter-born babies but not in other seasons (1) . We aimed to assess whether this effect on bone mineralisation is sustained into later childhood, with bone indices assessed at 4 years old in the Southampton participants. Demonstrating persistence of this effect into childhood would increase con fi dence in a long-term bene fi t of this intervention for life-long bone health. In Southampton, Oxford and Shef fi eld, in a double-blind design, 1123 pregnant women with a baseline 25-hydroxyvitamin D level 25 – 100nmol/l were randomised to 1000 IU/day cholecalciferol or matched placebo from 14 weeks ’ gestation to delivery. At age 4 years (Southampton participants only, n = 723 births), offspring assessments included anthropometry and whole-body dual-energy x-ray absorptiometry (DXA) [Hologic Horizon, yielding whole body less head (WBLH) bone mineral content (BMC), areal bone mineral density (aBMD), bone area (BA) and lean mass (LM)]. T-tests or Mann-Whitney U tests were used to compare bone DXA outcomes by maternal randomisation group. Linear regression was used to estimate the mean difference (represented by β ) in outcomes between the two randomisation arms, adjusted for sex, and age at DXA. Outcomes were standardised to a standard deviation scale, for ease of comparison. Full ethics and MHRA approvals were granted. Trial registration: ISRCTN82927713, registered 11/04/2008. 564/723 Maternal pregnancy to greater offspring indices compared irrespective of example (95% CI: g/cm2; (0.466,0.475) β =0.18 (0.00,0.35) trend role in the trial execution, data collection, analysis or manuscript preparation.