Neurocognitive Outcomes of In Utero Exposure to Antiseizure Medication: An Australian Cohort Study

Continuation of antiseizure medication (ASM) during pregnancy is essential for seizure control in most women with epilepsy, yet few studies have characterized the neurocognitive phenotypes associated with prenatal ASM exposure, with limited data on newer agents. The aim of this study was to determine whether in utero exposure to specific types of ASM monotherapies is associated with poorer neurocognitive outcomes compared with ASM-unexposed children.

This semiprospective cohort study included children aged 3-18 years born to women enrolled in the Raoul Wallenberg Australian Pregnancy Register of Antiepileptic Drugs. Children without major congenital anomalies who were exposed prenatally to carbamazepine (n = 20), lamotrigine (n = 23), levetiracetam (n = 20), valproate (n = 19), or topiramate (n = 11) monotherapy, or who were unexposed to ASMs (n = 19), were followed from birth and underwent standardized neuropsychological assessment by a neuropsychologist blinded to prenatal ASM exposure status. Primary outcomes included Full-Scale IQ (FSIQ) and index scores from the Wechsler Intelligence Scales; secondary outcomes assessed academic skills, memory, language, and executive functions. Generalized linear mixed-effects models were used to estimate associations between prenatal ASM exposure and cognitive outcomes, adjusting for maternal IQ, epilepsy characteristics, and relevant perinatal factors.

Of 110 children assessed (mean age 9.6 years; 51% female), 91 were exposed to ASM monotherapy in utero and 19 were unexposed. Children exposed to valproate, levetiracetam, topiramate, or carbamazepine scored lower across multiple cognitive domains, including processing speed, verbal comprehension, visuospatial reasoning, and academic skills (Cohen d range = -0.8 to -1.0). The largest FSIQ differences were observed for topiramate (b = -17.4, 95% CI -30.6 to -6.1) and levetiracetam (b = -13.9, 95% CI -24.5 to -3.8). Lamotrigine-exposed children showed performance comparable to unexposed peers. Cognitive outcomes did not vary by ASM dose, likely because of smaller sample sizes.

Prenatal exposure to topiramate, levetiracetam, valproate, and carbamazepine was associated with domain-specific cognitive vulnerabilities, particularly in IQ, processing speed, and academic functioning. Topiramate and levetiracetam exposure were linked to the largest IQ differences compared with unexposed children. While valproate is typically associated with the greatest risk, its impact seemed attenuated in this cohort. By contrast, cognitive performance in lamotrigine-exposed children was comparable to unexposed peers, suggesting that it may be comparatively safer with respect to cognitive outcomes.

This study provides Class II evidence that in utero exposure to valproate, levetiracetam, topiramate, and carbamazepine, but not to lamotrigine, is associated with worse neurocognitive outcomes in children compared with unexposed children.