Risks of neurodevelopmental disorders after prenatal exposure to antiepileptic drugs.
Abstract
Background
To determine whether in utero exposure to antiepileptic drugs (AEDs) is associated with autism spectrum disorder (ASD), attention-deficit/hyperactivity disorder (ADHD), and intellectual disability (ID) in offspring and to test robustness to familial confounding.
Methods
We conducted a nationwide population-based cohort of all live births between January 1, 2004, and December 31, 2016, with follow-up through 2021, linking birth, prescription, and health registries. Exposure was maternal dispensing of AEDs during pregnancy. We estimated hazard ratios with Cox models adjusted for parental demographics and psychiatric/medical comorbidities. To address shared familial factors, we performed sibling comparisons using stratified Cox models with mother as the stratum. We further conducted trimester-specific analyses, restricted to any neurodevelopmental disorder because of limited numbers for individual outcomes, by separately defining maternal AED dispensing during the first, second, and third trimesters to evaluate whether associations differed according to the timing of prenatal exposure.
Results
Among 2,196,156 births, 9,809 (0.45%) were exposed to AEDs during pregnancy. In adjusted population-wide analyses, carbamazepine, oxcarbazepine, valproic acid, and topiramate were associated with increased risk of any neurodevelopmental disorder, with some drug-specific associations also observed for ASD, ADHD, and ID. However, these associations were not observed in sibling comparison analyses. Trimester-specific sibling analyses showed no increased risk for first-trimester exposure, although several imprecise associations were observed for second- and third-trimester exposure. Sensitivity analyses were generally consistent with the main sibling comparison findings, except for an increased risk associated with topiramate when the exposure window was extended to 90 days before conception.
Conclusions
Population-wide associations between prenatal AED exposure and neurodevelopmental disorders were largely attenuated in sibling comparison analyses, suggesting that shared familial factors may explain much of the observed risk. These findings support careful individualized prescribing during pregnancy while highlighting the need for further research on specific AEDs and exposure windows.