Effect of long-chain polyunsaturated fatty acids in infant formula on long-term cognitive function in childhood: A systematic review and meta-analysis of randomised controlled trials

Abstract

Lack of preformed long-chain polyunsaturated fatty acids (LCPUFA) in infant formula has been hypothesised as contributing to cognitive differences between breast-fed and formula-fed infants. Previous systematic reviews found no cognitive differences between infants fed formula with LCPUFA and those fed formula without, but focused on early developmental measures, such as Bayley Scales of Infant Development, which are poorly differentiating and not predictive of cognitive ability in childhood. This systematic review examined the effect of randomising infants to formula supplemented with LCUFA vs unsupplemented formula on cognitive function ≥ age 2.5 years. We searched Medline, Embase the Cochrane Central Register of Controlled Trials without date limit, following a pre-published protocol according to PRISMA guidelines. We conducted random effects meta-analyses in RevMan v5.4 and followed GRADE and Cochrane Guidelines to evaluate strength of evidence and potential for bias. We included 8 trial cohorts which randomised participants between 1993 and 2004 and analyse 6 previously unpublished outcomes provided by various trialists. Age at the last available cognitive test ranged from 3.3 to 16 years. The pooled mean difference in Wechsler Preschool and Primary Scale of Intelligence-Revised from four trials in term-born children showed no benefit of LCPUFA: -0.04 points (95% confidence interval -5.94 to 5.85, 95% prediction interval -14.17 to 14.25). The pooled mean difference in Wechsler Abbreviated Scale of Intelligence score from two trials in preterm-born children also showed no benefit of LCPUFA: -7.71 (95% CI -24.63 to 9.22, 95% PI -97.80 to 82.38). Overall quality of evidence was low, due to substantial heterogeneity, low rates of follow-up, and indications of selective publication. The long-term effect of LCPUFA supplementation in term and preterm-born infants on cognition is highly uncertain and includes potential for large benefit as well as large harm. Based on our findings, LCPUFA supplementation of infant formula is not recommended until further robust evidence excludes long-term harm. Study registration PROSPERO registration numbers CRD42018105196 and CRD42018088868.