The management of women with epilepsy may involve difficult choices during pregnancies. Inadequate seizure control during pregnancy may harm both mother and foetus but antiepileptic drugs may also harm the foetus. A study in Liverpool and Manchester (N Adab and colleagues Journal of Neurology, Neurosurgery, and Psychiatry2004;75:1575–83OpenUrlAbstract/FREE Full Text) has pointed to treatment with sodium valproate during pregnancy as a particular risk factor for developmental and cognitive problems in the child.

A total of 547 mothers were identified through epilepsy and antenatal clinics, and 219 (40%) agreed to take part in the study. The mothers were interviewed and their 375 children aged 6 months to 16 years were examined. Two hundred and forty nine of 256 children aged over 5 years had WISC IQ tests. Among these 249, 80 had not been exposed in utero to antiepileptic drugs, 120 had been exposed to monotherapy (41 with valproate, 52 with carbamazepine, 21 with phenytoin, and six with other drugs), and 49 had been exposed to polytherapy (28 including valproate). Full scale and performance IQs did not vary significantly according to drug exposure in utero. Verbal IQ scores, however, were significantly lower (by 7 points on average) in children exposed to valproate monotherapy compared with children exposed to carbamazepine, phenytoin monotherapy, or unexposed. Among children whose mothers took valproate monotherapy in pregnancy 42% had a verbal IQ score of less than 80 (22%, <70). Exposure to valproate monotherapy in utero was associated with a 3.5 fold increase in the risk of a verbal IQ under 70 compared with no exposure to antiepileptic drugs. Three factors (low maternal IQ, valproate exposure, and frequent tonic-clonic seizures (>4) in pregnancy) were independently predictive of low verbal IQ in children aged 6 or more years. Children exposed to valproate were more likely to have additional educational needs. The authors of this paper call for an increase in the provision of information and counselling about anti-epileptic drug treatment during pregnancy for all women with epilepsy in the child bearing age group before they become pregnant. There was a significant correlation between increasing dose of valproate in the first trimester and lower verbal IQ with no significant effect at doses of less than 800 mg a day. Among children under the age of 6 years exposure to valproate in utero was associated with delays in the interaction and hearing and language domains of the Schedule of Growing Skills II test. The prevalence of moderate or severe dysmorphic features was 44% (valproate), 9% (carbamazepine), and 2% (unexposed), and there was an association between dysmorphic features and low verbal IQ.