Intravenous lipid emulsions with or without fish oil for the prevention of liver disease in VLBW neonates: A network meta-analysis.

BACKGROUND

Parenteral nutrition-associated liver disease is a severe complication in very low birth weight (≤1500 g) preterm neonates. Although previous pairwise meta-analyses have compared individual intravenous lipid emulsions, their relative efficacy across all available formulations has not been established in this population.

METHODS

We conducted a systematic review and frequentist random-effects network meta-analysis, searching PubMed, Embase, and the Cochrane Central Register through April 2026. Eligible studies were randomized trials of any intravenous lipid emulsion in very low birth weight neonates reporting this outcome. Treatments formed seven nodes; odds ratios versus soybean oil were estimated, ranking used the surface under the cumulative ranking curve (Bayesian model), and certainty was assessed with the Confidence in Network Meta-Analysis framework.

RESULTS

A total of 15 trials (1876 neonates) were included. Fish oil/olive oil/soybean oil was the only formulation significantly associated with reduced disease versus soybean oil (odds ratio, 0.099; 95% confidence interval, 0.016-0.613), but derived from a single high-risk-of-bias study. A soybean oil, medium-chain triglyceride, olive oil, and fish oil emulsion showed a nonsignificant reduction (odds ratio, 0.695; 0.440-1.099; 94% Bayesian probability of benefit), most pronounced among extremely low birth weight infants (odds ratio, 0.557; 0.307-1.010). In the ranking, fish oil/olive oil/soybean oil ranked first and this multicomponent emulsion second. No heterogeneity or inconsistency was detected; certainty was low to very low.

CONCLUSION

Fish oil-containing emulsions show a consistent signal toward reduced disease in very low birth weight neonates, but evidence remains insufficient for definitive recommendations. Adequately powered trials with standardized cholestasis definitions are needed.