Introduction: The bioethical principles of autonomy, beneficence, non-maleficence and justice, complemented by ethics committee evaluation, define conduct of clinical trials of anti-epileptic medications (AEM). Background information: Increased teratogenicity for offspring of women with epilepsy is presented in both lay and scientific literature. Specific drugs-older generation: Teratogenicity of phenobarbitone (PB), phenytoin (PHT), valproate (VPA) and carbamazepine (CBZ) is acknowledged, with drugs, such as trimethadione, being removed from the market because of teratogenicity. Specific drugs-new generation: Insufficient data allowed definitive commentary concerning teratogenicity of newer AEM, such as lamotrigine (LTG), gabapentin (GBP), tiagabine (TGB) or levetiracetam (LEV). All those suggestions favour some over others with specific AEM combinations being.:questioned. Pregnancy registries: Lack of information sporn AEM-specific plus national and international birth registries which endorse VPA, CBZ and LTG dose related concerns. Conflict of interest: Competing influences of AEM and epilepsy-specific factors need delineation although appear more teratogenetic than does epilepsy alone. Clinical trials: Most trials focus upon refractory epilepsy with potentially enhanced risks. Informed consent demands discussion of possible teratogenicity and exclusion of women unwitting to practice adequate contraception. Discussion: Trials must respect legal and ethical dictates including the exclusion of women unwilling to practice reliable contraception. Automatic exclusion from a trial, subsequent to confirmed pregnancy, is unlikely to protect the foetus as potential for teratogenicity already has occurred. Autonomy should empower prospective parents to decide continued trial participation consequent to detailed informed consent without coercion. All options demand review, including responsibility to future AEM users. (C) 2006 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.