Vitamin supplementation for preventing miscarriage.

Systematic review and meta-analysis of randomized and cluster-randomized trials in women planning pregnancy or pregnant women, with supplementation started before 20 weeks' gestation. Trials were conducted in diverse settings worldwide and included both low- and high-risk populations.

This review evaluated oral vitamin supplementation started before conception, periconceptionally, or in early pregnancy, including vitamin C alone or with vitamin E or multivitamins, vitamin A or beta-carotene, folic acid with or without multivitamins and/or iron, multivitamins plus iron and folic acid, and antioxidant vitamin combinations. Doses, formulations, and duration varied widely across trials, from short daily regimens to supplementation continued until delivery or postpartum.

Overall, vitamin supplementation before conception or in early pregnancy did not prevent miscarriage. Vitamin C with vitamin E showed no effect on total fetal loss, early or late miscarriage, stillbirth, or adverse effects, and vitamin A, beta-carotene, vitamin C alone, folic acid, and most multivitamin comparisons also showed no clear benefit. The main exception was multivitamins plus iron and folic acid, which reduced stillbirth versus iron and folate alone (RR 0.92, 95% CI 0.85 to 0.99; 10 trials, 79,851 women), but did not reduce total fetal loss or early/late miscarriage. A few single-trial comparisons suggested lower total fetal loss with multivitamin without folic acid (RR 0.49, 95% CI 0.34 to 0.70) or multivitamin with or without vitamin A (RR 0.60, 95% CI 0.39 to 0.92), but these findings require cautious interpretation.

The evidence base was heterogeneous in vitamin type, dose, timing, and comparison groups, making indirect comparisons difficult. Several potentially positive findings came from single small trials, and adverse-event reporting was limited or absent for many comparisons. Some pooled analyses were imprecise or showed substantial heterogeneity, reducing confidence in subgroup-specific conclusions.