Exposure to prenatal infection and the development of internalizing and externalizing problems in children: a longitudinal population-based study.

Abstract

A large body of work has reported a link between prenatal exposure to infection and increased psychiatric risk in offspring. However, studies to date have focused primarily on exposure to severe prenatal infections and/or individual psychiatric diagnoses in clinical samples, typically measured at single time points, and without accounting for important genetic and environmental confounders. In this study, we investigated whether exposure to common infections during pregnancy is prospectively associated with repeatedly assessed child psychiatric symptoms in a large population-based study. Our study was embedded in a prospective pregnancy cohort (Generation R; n = 3,598 mother-child dyads). We constructed a comprehensive prenatal infection score comprising common infections for each trimester of pregnancy. Child total, internalizing, and externalizing problems were assessed repeatedly using the parent-rated Child Behavioral Checklist (average age: 1.5, 3, 6, 10, and 14 years). Linear mixed-effects models were run adjusting for a range of confounders, including child polygenic scores for psychopathology, maternal chronic illness, birth complications, and infections during childhood. We also investigated trimester-specific effects and child sex as a potential moderator. Prenatal exposure to infections was associated with higher child total, internalizing, and externalizing problems, showing temporally persistent effects, even after adjusting for important genetic and environmental confounders. We found no evidence that prenatal infections were associated with changes in child psychiatric symptoms over time. Moreover, in our trimester-specific analysis, we did not find evidence of significant timing effects of prenatal infection on child psychiatric symptoms. No interactions with child sex were identified. Our research adds to evidence that common prenatal infections may be a risk factor for psychiatric symptoms in children. We also extend previous findings by showing that these associations are present early on, and that rather than changing over time, they persist into adolescence. However, unmeasured confounding may still explain in part these associations. In the future, employing more advanced causal inference designs will be crucial to establishing the degree to which these effects are causal.