Omega-3 Long-Chain Polyunsaturated Fatty Acids for Extremely Preterm Infants: A Systematic Review

Systematic review of trials in preterm infants, with the main population focused on extremely preterm infants born at less than 29 weeks' gestation and some analyses restricted to infants born at 32 weeks or less. Included studies evaluated omega-3 exposure through infant feeding, maternal supplementation, or intravenous lipid emulsions.

Omega-3 long-chain polyunsaturated fatty acid exposure was delivered through several regimens, including enteral formulas or human milk supplementation with DHA and ARA, maternal DHA supplementation affecting breast milk composition, and parenteral lipid emulsions containing fish oil or other omega-3 sources. Doses varied across trials, ranging from DHA 32 mg and ARA 31 mg per 100 mL human milk to formula and lipid emulsions enriched with DHA, EPA, and sometimes ARA.

Overall, omega-3 long-chain polyunsaturated fatty acid supplementation did not significantly reduce bronchopulmonary dysplasia in preterm infants. Across 12 RCTs, pooled BPD risk was RR 0.97, 95% CI 0.82-1.13, in 2809 infants; in the subgroup born at 32 weeks or less, BPD risk was RR 0.88, 95% CI 0.74-1.05, in 1156 infants. Necrotizing enterocolitis in that subgroup showed a non-significant trend toward benefit, RR 0.50, 95% CI 0.23-1.10, in 900 infants. Other key outcomes were also not significantly improved, including neonatal death RR 0.76, 95% CI 0.42-1.38, retinopathy of prematurity RR 1.10, 95% CI 0.91-1.33, and intraventricular hemorrhage RR 0.96, 95% CI 0.76-1.21. The review suggests possible benefit in the most immature infants, but definitive efficacy was not demonstrated.

Trials used heterogeneous omega-3 sources, doses, routes, and co-nutrients, making the intervention difficult to compare across studies. BPD definitions differed between trials, secondary outcomes were reported inconsistently, and several pooled estimates were imprecise with confidence intervals crossing no effect. Findings in the extremely preterm subgroup were suggestive but not statistically significant, so the evidence remains underpowered for definitive conclusions.