The use of antiepileptic drugs (AED) is common in women of childbearing age. One study 1 reported that prenatal AED exposure occurred in 2.19% of all pregnancies in the United States. In the past decade, findings have highlighted that it is no longer sufficient to only consider the association between firsttrimester exposure and the risk for structural teratogenicity, because prenatal AED exposure throughout the entire pregnancy can have longterm negative consequences on neurodevelopment. The published studies are limited in scope but have consistently reported that valproate prenatal exposure is associated with lower scores on neuropsychometric batteries during early childhood, 2,3 while prenatal exposure to carbamazepine and lamotrigine monotherapies is associated with relatively normal scores. Risks associated with most other AEDs have not been adequately studied, to my knowledge.Findings from the Danish, population-based cohort study by Elkjaer et al 4 are not just additive but rather synergistic with the current body of literature, because the article provides complementary data from a unique source and study design. Prior studies have been either retrospective, case-control, registry-based, or prospective and observational, with generally low numbers of participants (eg, 13 to 62 for cohorts exposed to valproate monotherapy). 2,3,5 In contrast, the design of this study is population-based and focused on school-aged children exposed prenatally to AEDs. Children born between 1997 and 2006 were identified using the Danish Medical Birth Registry. Standardized test scores in Danish and mathematics from 479 027 children were available for analysis, including scores from 1865 children who had been exposed to AEDs prenatally. These were standardized to z scores and adjusted for calendar year, sex, maternal education level, and parental income level. The size of the sample is impressive, with 253 exposed to valproate monotherapy and substantial numbers exposed to other AEDs. The primary outcomes are objective measures based on standardized tests administered to children in second through eighth grades, with the largest data set from sixth-grade tests (mean age, 12.9 years). 4 The children with prenatal valproate exposure had lower z scores on sixthgrade Danish and mathematics tests, as well as Danish tests in the fourth and eighth grades. Even the low-dose valproate group (whose mothers took <1000 mg/day during pregnancy) had similarly reduced scores. The impact of these lower test scores on academic potential cannot be refuted. Moreover, the outcomes in the lamotrigine monotherapy group were identical to those of the unexposed group.