Maternal infections during pregnancy and offspring cognitive outcome: A nationwide full-sibling cohort study

Abstract

Background: Maternal infections are common during pregnancy, but it is unclear how they impact the cognitive outcome of the offspring, with many studies suggesting adverse effects. Using long-term follow-up of a nationwide sibling cohort in Denmark with information on maternal antimicrobial prescriptions in community pharmacies and in-patient hospitalizations for infection, we aimed to estimate the effect of maternal infections during pregnancy on offspring school grades and intelligence test results in adolescence.

Methods and findings: From population-based national registries we defined a cohort of all full-siblings, born from January 1, 1996 to December 31, 2,003 in Denmark, and linked them to maternal filled prescription for antimicrobial pharmaceuticals and maternal hospitalizations for infection during pregnancy. Standardized examination grades in language and mathematics at the final year of compulsory schooling, in addition to intelligence test scores (calculated as IQ) for a nested sub-cohort of full brothers, were used as outcomes. Among 274,166 children in the full-sibling cohort, 80,817 (29.5%) had a mother who during her pregnancy filled a prescription for a systemic antimicrobial, while 5,628 (2.1%) had a mother who during her pregnancy was hospitalized due to an infection. We found no consistent difference in school grades in language (z-score difference, 0.0, 95% confidence interval [CI] [-0.0,0.0]; p = 0.920) and mathematics (z-score difference, -0.0, 95% CI [-0.0,-0.0]; p = 0.042), and in IQ (IQ-difference, 0.3, 95% CI [-0.2,0.7]; p = 0.217), in children whose mother filled one antimicrobial prescription compared with children whose mother did not fill any, when taking shared family factors into account, while many associations were consistently significant when not taking shared family factors into account. Furthermore, we found no indication of an impact of maternal in-patient hospitalizations for infections during pregnancy on school grades (z-score difference for language, -0.0, 95% CI [-0.1,0.0]; p = 0.103; z-score difference for mathematics, 0.0, 95% CI [-0.0,0.0]; p = 0.809) or IQ (IQ-difference, 0.4, 95% CI [-0.8,1.6]; p = 0.545), when also taking shared family factors into account. Similar findings were found when considering infections in bi-weekly exposure periods during gestation. The main limitations of the study were lacking information on within hospital pharmaceutical prescriptions and the underlying pathogenic microorganisms.

Conclusions: Our study does not support major effects of common maternal infections during pregnancy on offspring cognitive outcomes, and support the safety of commonly prescribed antimicrobials during pregnancy with respect to the long-term cognitive outcomes of the offspring.

Fig 1. Difference in standardized grade (z-score) in language and mathematics for the full-sibling cohort and difference in IQ for the nested full-brother sub-cohort, by maternal exposure to infection during pregnancy compared with no exposure.

All analyses are adjusted for maternal and paternal age at childbirth, maternal and paternal educational level, and number of older siblings. Estimates marked by diamonds (◊) are not adjusted for shared family factors (only covariate-adjusted), while estimates marked by squares (■) are additionally adjusted for shared family factors. Whiskers denote 95% confidence intervals. The estimates are additionally presented in tabular form in S1 Data. Asterisk (*) denote antimicrobial subtypes with 100 or fewer children exposed in the full-brother cohort, yielding uninformative IQ point estimates.

Fig 2. Difference in standardized grade (z-score) in language and mathematics for the full-sibling cohort and difference in IQ for the nested full-brother sub-cohort, given by gestational age at

β-lactam exposure compared with no exposure. Analyses are adjusted for maternal and paternal age at childbirth, maternal and paternal educational level, number of older siblings, and shared family factors. Whiskers denote 95% confidence intervals. Lines denote natural cubic splines (modeled on the day of exposure) and the shaded areas denote the corresponding 95% confidence bands of these splines.

Fig 3. Difference in language and mathematics z-scores by gestational age at maternal in-patient hospitalization due to infection compared with no such exposure.

Analyses are adjusted for maternal and paternal age at childbirth, maternal and paternal educational level, number of older siblings, and shared family factors. Whiskers denote 95% confidence intervals. Lines denote natural cubic splines (modeled on the day of exposure) and the shaded areas denote the corresponding 95% confidence bands of these splines.