Complete Cochrane Risk of Bias 2 assessment.

BackgroundNutrition is vital for preterm infant development. Vitamins play key roles as cofactors and gene regulators for metabolic and immune functions and are common added components of preterm infant nutrition. However, information on how vitamins impact in-hospital and neurodevelopmental outcomes is sparse. We aimed to determine the effect of fat- and water-soluble vitamin supplementation on clinical outcomes during neonatal care and later neurodevelopment of very preterm (≤32 weeks’ gestation) and very low birth weight (≤1500 g) infants.Methods and findings4 databases and 3 clinical trial registries were systematically searched for randomised controlled trials (RCTs). Two reviewers independently extracted data and assessed quality using the Cochrane Risk of Bias tool. Meta-analyses were conducted using a random-effect model for each vitamin subgroup. Data are presented as risk ratios [95% confidence intervals]. Of 4074 references identified, 43 studies were included in the review. Only 2 reported neurodevelopment at 2 years, and only 4 were studies of water-soluble vitamins (vitamin C, 3 studies; B12 and folate, 1 study). Survival free from neurodisability was not affected by supplementation of vitamin A (0.89 [0.74–1.08], n = 538, very low certainty of evidence) or vitamin D (0.76 [0.46–1.27], n = 78, very low certainty of evidence). The incidence of bronchopulmonary dysplasia was decreased by vitamins D (0.58 [0.41–0.83]) and C (0.59 [0.37–0.93]), very low certainty of evidence), retinopathy of prematurity was decreased by vitamins A (0.77 [0.61–0.98]) and E (0.10 [0.01–0.80]), very low to low certainty of evidence) and intraventricular haemorrhage was decreased by vitamin E (0.70 [0.52–0.92], moderate certainty of evidence). Culture-proven sepsis was decreased by vitamin A (0.88 [0.77–0.99], moderate certainty of evidence).ConclusionsThere are few and inconclusive data on the effect of vitamin supplementation in preterm infants on later neurodevelopment. Evidence for shorter-term outcomes is mostly of low certainty. Together with substantial heterogeneity in trial design, it therefore is difficult to recommend a specific supplementation regimen.Registry and Registry Number: This systematic review was prospectively registered on PROSPERO, ID CRD42023418552, available from https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023418552