Abstracts

Rationale: Evidence regarding the potential negative effects of prenatal antiseizure medication (ASM) exposure on cognitive measures, such as school performance, is accumulating, but has so far primarily addressed risks associated with the most commonly used ASMs, such as valproate, lamotrigine, and carbamazepine. We conducted a Danish register-based cohort study, estimating the association of prenatal ASM exposure with school performance in children attending primary and lower secondary state schools, thereby extending a prior Danish investigation to also include associations with topiramate, gabapentin, and levetiracetam.

Methods: We conducted a register-based cohort study of all singletons born in Denmark between 1997 and 2009 with Danish born parents and who had at least one test result in either language (Danish) or mathematics as part of the Danish National School Test Program between 2010 and 2019. We obtained information on maternal prescription fill for ASM from the National Prescription Register. We used linear regression with robust standard errors to estimate the difference in test scores in language and mathematics (scale 1-100, higher scores indicate better results) between children born to mothers with and without ASM use during pregnancy, adjusting for relevant confounders. We further examined associations with specific ASM monotherapies.

Results: Among 555,479 children included (51.3% males), 2,474 (0.45%) were exposed to maternal ASM use during pregnancy. The mean score was 56.3 (95% confidence interval (CI): 56.3; 56.4) for all language tests, and 57.2 (95% CI: 57.2; 57.3) for all mathematics tests. The adjusted difference in test scores between children of mothers with and without ASM use during pregnancy was -1.9 (95% CI: -2.9; -1.0) in language and -1.3 (95% CI: -2.4; -0.2) in mathematics, indicating a decrease in school performance. For the newly included monotherapies, we found an adjusted difference of 2.9 (95% CI: -2.0; 7.8) in language and -1.0 (95% CI: -7.4; 5.3) in mathematics for topiramate (n=41), -4.2 (95% CI: -8.9; 0.5) in language and -1.6 (95% CI: -6.8; 3.5) in mathematics for gabapentin (n=57) and 5.3 (95% CI: -1.6; 12.2) in language and 5.4 (95% CI: -1.3; 12.1) in mathematics for levetiracetam (n=28). As in previous studies, we observed no association with the use of lamotrigine during pregnancy and school performance in the offspring, and confirmed the previously reported negative association with valproate.

Conclusions: This nation-wide cohort study confirmed the previously reported negative association with school performance in children prenatally exposed to ASM and in particular valproate. No statistically significant association were detected in children prenatally exposed to topiramate, gabapentin, and levetiracetam. The statistical power of these analyses was however limited, and findings should be interpreted with caution.

Funding: The study was supported by the Danish Epilepsy Association, the Central Denmark Region, the Novo Nordisk Foundation (NNF16OC0019126, NNF22OC0075033), and the Independent Research Fund Denmark (1133-00026B).