Effects of nutrition interventions during pregnancy on low birth weight: an overview of systematic reviews

Overview of systematic reviews of randomized, quasi-randomized, and cluster-randomized trials in pregnant women, mostly with singleton pregnancies, from low-, middle-, and high-income settings. Populations included both low-risk and high-risk women, and some trials enrolled adolescents or women with specific risks such as preeclampsia, anemia, hypertension, or infections.

Across the underlying reviews, active regimens included single vitamins and minerals (vitamin A, C, E, folic acid, calcium, iron, iodine, magnesium, zinc), multiple micronutrients, balanced energy/protein supplements, and marine oil or long-chain n-3 fatty acid supplementation during pregnancy. Doses and formulations varied widely across trials, including UNIMMAP-style multiple micronutrient products and differing iron contents.

Overall, improving women’s nutritional status during pregnancy was associated with lower risk of low birth weight, small-for-gestational-age birth, and preterm birth, with the strongest and most consistent support for multiple micronutrient supplementation and preventive antimalarial drugs. Specific effects varied by intervention: vitamin A added to iron plus folate reduced low birth weight by 33%, high-dose calcium (≥1 g/day) reduced preterm birth by 24%, multiple micronutrients reduced low birth weight by 11-14% and SGA by 10-17%, zinc in adolescent pregnancy reduced low birth weight by 61%, and long-chain n-3 fatty acids reduced preterm birth by 39%. Preventive antimalarial drugs reduced low birth weight by 27%, while high protein increased SGA by 58% and several single-nutrient supplements showed no clear benefit.

Many interventions were supported by only a small number of trials, and the evidence varied widely in dose, formulation, population risk, and setting. Several comparisons were null or inconsistent, some outcomes were not reported for every intervention, and not all reviews provided pooled sample sizes, limiting precision and generalizability.