Breastfeeding during infancy and neurocognitive function in adolescence: 16-year follow-up of the PROBIT cluster-randomized trial

Abstract

Background: Evidence on the long-term effect of breastfeeding on neurocognitive development is based almost exclusively on observational studies. In the 16-year follow-up study of a large, cluster-randomized trial of a breastfeeding promotion intervention, we evaluated the long-term persistence of the neurocognitive benefits of the breastfeeding promotion intervention previously observed at early school age.

Methods and findings: A total of 13,557 participants (79.5% of the 17,046 randomized) of the Promotion of Breastfeeding Intervention Trial (PROBIT) were followed up at age 16 from September 2012 to July 2015. At the follow-up, neurocognitive function was assessed in 7 verbal and nonverbal cognitive domains using a computerized, self-administered test battery among 13,427 participants. Using an intention-to-treat (ITT) analysis as our prespecified primary analysis, we estimated cluster- and baseline characteristic-adjusted mean differences between the intervention (prolonged and exclusive breastfeeding promotion modelled on the Baby-Friendly Hospital Initiative) and control (usual care) groups in 7 cognitive domains and a global cognitive score. In our prespecified secondary analysis, we estimated mean differences by instrumental variable (IV) analysis to account for noncompliance with the randomly assigned intervention and estimate causal effects of breastfeeding. The 16-year follow-up rates were similar in the intervention (79.7%) and control groups (79.3%), and baseline characteristics were comparable between the two. In the cluster-adjusted ITT analyses, children in the intervention group did not show statistically significant differences in the scores from children in the control group. Prespecified additional adjustment for baseline characteristics improved statistical precision and resulted in slightly higher scores among children in the intervention for verbal function (1.4 [95% CI 0.3-2.5]) and memory (1.2 [95% CI 0.01-2.4]). IV analysis showed that children who were exclusively breastfed for ≥3 (versus <3) months had a 3.5-point (95% CI 0.9-6.1) higher verbal function, but no differences were observed in other domains. While our computerized, self-administered cognitive testing reduced the cluster-level variability in the scores, it may have increased individual-level measurement errors in adolescents.

Conclusions: We observed no benefit of a breastfeeding promotion intervention on overall neurocognitive function. The only beneficial effect was on verbal function at age 16. The higher verbal ability is consistent with results observed at early school age; however, the effect size was substantially smaller in adolescence.

Probit trial registration: ClinicalTrials.gov NCT01561612.

Conflict of interest statement

I have read the journal's policy and the authors of this manuscript have the following competing interests: GD and SA are employees of NeuroTrax.

Figures

Fig 1. Flow diagram of clusters and participants of PROBIT recruitment follow-up phases at 12 mo, 6.5 y, 11.5 y, and 16 y.

a: During the 11.5-y follow-up, 6 deaths were reported in the intervention arm. Data-checking during the 16-y follow-up found one of these children had been incorrectly reported as deceased and data were amended. b: Of the 13,557 seen at the 16-y follow-up, 12,072 were seen at both 11.5-y and 16-y follow-ups, 274 were not seen at either 6.5-y or 11.5-y follow-ups, 449 were seen at 6.5 y but not seen at 11.5 y, and 762 were seen at 11.5 y but not seen at 6.5 y. Of the 3,489 children randomized but not followed up at 16 y, 267 attended the excluded site, 116 died after randomization, 2,674 were lost to follow-up, and 432 were unable or unwilling to come for their clinic visit.