The foetal outcomes of 2,635 pregnancies recorded in the Australian Pregnancy Register were studied. In at least the initial 4 months of 515 pregnancies, there had been no intrauterine exposure to antiepileptic drugs, though the women involved in 264 of these pregnancies took antiepileptic drugs in later pregnancies. Compared with these 515 drug-unexposed pregnancies, foetal malformations risks were increased more than five-fold in association with valproate monotherapy, and more than doubled in association with carbamazepine monotherapy (p < 0.05). There were no statistically significant increases in malformation rates associated with other more commonly used antiepileptic drugs, while the malformation risk in relation to levetiracetam exposure was lower than that in the drug-unexposed pregnancies. The published literature has rather consistently shown raised malformation rates associated with carbamazepine monotherapy, though only once was it statistically significant. There now appears to be enough evidence to make it likely that carbamazepine possesses some teratogenic capacity. This makes it unwise to employ the malformation rate associated with carbamazepine monotherapy as a comparator when assessing the foetal hazards from exposure to newer antiepileptic drugs. Levetiracetam may prove a better comparator if adequate untreated control material is unobtainable.