(242) Association of prenatal antiseizure medication exposure with risk of neurodevelopmental disorders in children

Background: Valproate use during pregnancy is linked to poorer neurocognitive function in children, but information on the association of other anti-seizure medications (ASMs) with neurodevelopmental disorders (NDDs) is limited.

Objectives: To evaluate whether prenatal exposure to individual ASMs increases the risk of specific NDDs in children.

Methods: We identified a population-based cohort of publicly and commercially insured women with epilepsy linked to their children, using two US healthcare utilization databases, 2000-2021. Exposure to individual ASMs was defined based on ≥1 dispensing of the medication of interest from gestational week 19 to delivery, the period of synaptogenesis. Prenatally exposed children were compared to those born to mothers without any ASM dispensing from 3 months before pregnancy until delivery. Individual ASMs considered included carbamazepine, lacosamide, lamotrigine, levetiracetam, oxcarbazepine, phenobarbital, phenytoin, topiramate, valproate, and zonisamide. Specific NDDs – identified based on validated algorithms – included autism, attention-deficit/hyperactivity disorder, learning disorder, speech/language disorder, coordination disorder, intellectual disability (ID), and behavioral disorder. Children were followed until diagnosis of the specific NDD, disenrollment, death, or end of the study. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards models with propensity score overlap weighting to adjust for demographics, maternal mental health, neurologic conditions other than epilepsy, concomitant medications, proxies for lifestyle behavior, other comorbidities, and healthcare utilization.

Results: The cohort included 8,887 prenatally unexposed children. Exposed counts ranged from 219 for lacosamide to 5,261 for levetiracetam. After adjustment, phenobarbital, valproate and zonisamide showed associations with several NDDs (HR range: 1.52 to 4.50). Several ASMs were associated with ID, including lamotrigine (HR: 2.39; 95% CI: 1.20-4.73), oxcarbazepine (3.30; 1.29-8.41), phenobarbital (3.70; 0.85-16.11), topiramate (1.84; 0.75-4.51), valproate (1.72; 0.77-3.83), and zonisamide (4.50; 1.43-14.18), all driven by a small number of exposed cases. No associations were seen for other ASMs and NDDs. Results were robust across multiple sensitivity analyses.

Conclusions: Our findings strengthen the evidence of adverse effects of valproate and raise concern regarding risk increases for phenobarbital and zonisamide. Additional signals for some ASMs in relation to ID in the context of multiple testing and rare outcomes (resulting in imprecise estimates) require confirmation in follow-up studies as evidence continues to accumulate.