To assess the risk of major congenital abnormalities associated with specific antiepileptic drug regimens, a large retrospective cohort study was performed. The study comprised 1,411 children born between 1972 and 1992 in four provinces in the Netherlands who were born to mothers with epilepsy and using antiepileptic drugs during the first trimester of pregnancy, and 2,000 nonepileptic matched controls. We found significantly increased risks of major congenital abnormalities for carbamazepine and valproate monotherapy, with evidence for a significant dose–response relationship for valproate. The risk of major congenital abnormalities was nonsignificantly increased for phenobarbital monotherapy when caffeine comedication was excluded, but a significant increase in risk was found when caffeine was included. Phenytoin monotherapy was not associated with an increased risk of major congenital abnormalities. Regarding polytherapy regimens, increased risks were found for several antiepileptic drug combinations. Clonazepam, in combination with other antiepileptic drugs, showed a significantly increased relative risk. Furthermore, there were significantly increased relative risks for the combination of carbamazepine and valproate and the combination of phenobarbital and caffeine with other antiepileptic drugs. This study shows that most antiepileptic drug regimens were associated with an increased risk of major congenital abnormalities in the offspring, in particular valproate (dose–response relationship) and carbamazepine monotherapy, benzodiazepines in polytherapy, and caffeine comedication in combinations with phenobarbital.