Antipsychotic use during pregnancy and risk of specific neurodevelopmental disorders and learning difficulties in children: a multinational cohort study

Abstract

Background: Antipsychotics are commonly prescribed to treat a range of psychiatric conditions in women of reproductive age and during pregnancy, including schizophrenia, bipolar disorder, anxiety, depression, autism spectrum disorder, and insomnia. This study aimed to evaluate whether children exposed to antipsychotic medication prenatally are at increased risk of specific neurodevelopmental disorders and learning difficulties.

Methods: Our population-based cohort study used nationwide register data (1 January 2000-31 December 2020) on pregnant women diagnosed with a psychiatric disorder and their live-born singletons from Denmark, Finland, Iceland, Norway, and Sweden. Cox proportional hazard regression yielded propensity score-weighted hazard ratios (aHRs) and 95% confidence intervals (CIs) for risk of intellectual-, speech or language-, learning-developmental disorders, and a composite outcome of the listed disorders. We defined poor performance as scoring within the lowest quartile on national school tests in mathematics and language arts. We estimated propensity score-weighted risk ratios (aRRs) using Poisson regression. We analysed data from Denmark separately and pooled results using random effects meta-analysis.

Findings: Among 213,302 children (median follow-up: 6.7 years), 11 626 (5.5%) were exposed to antipsychotics prenatally. Adjusted risk estimates did not suggest an increased risk of neurodevelopmental disorders: aHR of 1.06 (95% CI 0.94-1.20) for the composite outcome, or for poor academic performance: aRR of 1.04 (95% CI 0.91-1.18) in mathematics, and of 1.00 (95% CI 0.87-1.15) in language arts. Results were generally consistent across individual medications, trimesters of exposure, sibling- and sensitivity analyses.

Interpretation: The findings of this large multinational cohort study suggest there is little to no increased risk of child neurodevelopmental disorders or learning difficulties after prenatal exposure to antipsychotics. Our findings can assist clinicians and women managing mental illness during pregnancy.

Funding: This study was funded by the NordForsk Nordic Program on Health and Welfare (Nordic Pregnancy Drug Safety Studies, project No. 83539), by the Research Council of Norway (International Pregnancy Drug Safety Studies, project No. 273366) and by the Research Council of Norway through its Centres of Excellence funding scheme (project No. 262700), and UNSW Scientia Programme Awards (PS46019, PS46019-A).

Keywords: Academic performance; Antipsychotics; Children; Neurodevelopmental disorders; Nordic health registers; Pregnancy.

Fig. 1

Crude and propensity score-weighted cumulative incidence (with 95% confidence intervals) for composite neurodevelopmental outcome from age 3–19 years in children of mothers with psychiatric disorders born between 2000 and 2020 by prenatal antipsychotic exposure. A) Crude cumulative incidence (with 95% confidence intervals) in combined cohort (Finland, Iceland, Norway, Sweden) B) Crude cumulative incidence (with 95% confidence intervals) in Denmark C) Propensity score-weighted cumulative incidence (with 95% confidence intervals) in combined cohort (Finland, Iceland, Norway, Sweden) D) Propensity score-weighted cumulative incidence (with 95% confidence intervals) in Denmark. ^aAdjusted for birth year, sex of child, child’s country of birth, maternal birth country, age, parity, education, cohabitation status, BMI & smoking in early pregnancy, use of other medications during pregnancy, or known/suspected teratogens and comorbidity prior to pregnancy using propensity score overlap weights.

Fig. 2

Hazard ratios (with 95% confidence intervals) for child neurodevelopmental disorders after prenatal antipsychotic exposure by timing of exposure and monotherapy. BMI, body mass index; CI, confidence interval Monotherapy was assumed if a prescription was filled for only one type of antipsychotic medication during the exposure period. For numbers less than 5 (<5), numbers are suppressed due to country-/data-specific suppression policy. ^aAdjusted for birth year and sex of child, child's country of birth, maternal age, parity using outcome regression. ^bAdjusted for birth year, sex of child, child's country of birth, maternal age, parity, education, maternal birth country, cohabitation status, BMI & smoking in early pregnancy, use of other medications during pregnancy, or known/suspected teratogens and comorbidity prior to pregnancy using propensity score overlap weights. ^cSubstantially different estimates were observed from Denmark versus the combined Nordic data (I² = 92%). Denmark, HR 0.35 (95% CI 0.09–0.64); combined Nordic cohort, 1.44 (0.94–2.20). ^dResults presented only for the combined cohort (Finland, Iceland, Norway, and Sweden), owing to low number of exposure/outcomes in the Danish cohort.

Fig. 3

Risk ratios (with 95% confidence intervals) for poor academic performance after prenatal antipsychotic exposure by timing of exposure and monotherapy. BMI, body mass index; CI, confidence interval. Monotherapy was assumed if a prescription was filled for only one type of antipsychotic medication during the exposure period. For numbers less than 5 (<5), numbers are suppressed due – country-/data-specific suppression policy. Finland was not included in this analysis due to data availability. ^aAdjusted for birth year and sex of child, child’s country of birth, maternal age, parity using outcome regression. ^bAdjusted for birth year, sex of child, child’s country of birth, maternal age, parity, education, maternal birth country, cohabitation status, BMI and smoking in early pregnancy, use of other medications during pregnancy, or known/suspected teratogens and comorbidity prior to pregnancy using propensity score overlap weights. ^cResults presented only for the combined academic cohort (Iceland, Norway, and Sweden), owing to low number of exposure/outcomes in the Danish cohort.

Fig. 4

Hazard and risk ratios (with 95% confidence intervals) for child neurodevelopmental disorders and poor academic performance after prenatal antipsychotic exposure in sibling matched analysis where siblings are discordant on exposure and outcome. AP, antipsychotic; BMI, body mass index; CI, confidence interval; I², I² statistic ^a Adjusted for birth year and sex of child, child’s country of birth, maternal age, parity using outcome regression. ^b Adjusted for birth year, sex of child, child’s country of birth, maternal birth country, age, parity, education, cohabitation status, BMI & smoking in early pregnancy, use of other medications during pregnancy, or known/suspected teratogens and comorbidity prior to pregnancy using propensity score overlap weights. ^c Results presented only for the combined cohort from Finland, Iceland, Norway, and Sweden, owing to low number of exposure/outcomes in the Danish cohort.