BACKGROUND: The potential adverse effects of antiepileptic drug (AED) exposure in pregnancy have been well recognised but the relative risks of specific antiepileptic drug exposures remain poorly understood. OBJECTIVES: To assess the adverse effects of commonly used antiepileptic drugs on maternal and fetal outcomes in pregnancy in women with epilepsy. Comparison of outcomes following specific antiepileptic drug exposures in utero to unexposed pregnancies in the general population or women with epilepsy are described. The current manuscript reports the first phase of this review which focuses upon neurodevelopmental outcomes in children exposed to antiepileptic drugs in utero. SEARCH STRATEGY: We searched MEDLINE, Pharmline, EMBASE, Reprotox and TERIS from 1966 to December 2003. Review articles and conference abstracts were also hand searched. SELECTION CRITERIA: All randomized controlled trials, prospective cohorts of children of pregnant women with and without epilepsy and case control studies (cases: developmental delay or impaired cognitive outcome, control: normal development) were included. DATA COLLECTION AND ANALYSIS: Methodological quality was assessed using an adapted version of the Newcastle-Ottawa Scale. The wide variety of outcome measures and methodological approaches made meta-analysis difficult and a descriptive analysis of the results is presented. MAIN RESULTS: PART A 1b - DEVELOPMENTAL OUTCOMES: The majority of studies were of limited quality. There was little evidence about which specific drugs carry more risk than others to the development of children exposed in utero. The results between studies are conflicting and while most failed to find a significant detrimental outcome with in utero exposure to monotherapy with carbamazepine, phenytoin or phenobarbitone, this should be interpreted cautiously. There were very few studies of exposure to sodium valproate. Polytherapy exposure in utero was more commonly associated with poorer outcomes, as was exposure to any AEDs when analysis did not take into account type of AED. The latter may reflect the large proportion of children included in these studies who were in fact exposed to polytherapy. REVIEWERS' CONCLUSIONS: PART A 1b - DEVELOPMENTAL OUTCOMES: Based on the best current available evidence it would seem advisable for women to continue medication during pregnancy using monotherapy at the lowest dose required to achieve seizure control. Polytherapy would seem best avoided where possible. More population based studies adequately powered to examine the effects of in utero exposure to specific monotherapies which are used in everyday practice are required.