The issue of prescribing anticonvulsant drugs during lactation is clinically important, but also complex. Data for some drugs are completely lacking and for other drugs information is only available from single dose or short term studies or case reports. Moreover, limited knowledge exists about the practical impact of the drug concentrations found in breast milk and there are great methodological problems in the assessment of possible adverse drug reactions in infants. Nevertheless, based on current knowledge, some recommendations can be suggested. Treatment with carbamazepine, valproic acid (sodium valproate) and phenytoin is considered compatible with breastfeeding. Treatment with ethosuximide or phenobarbital (phenobarbitone)/primidone should most probably be regarded as potentially unsafe and close clinical monitoring of the infant is recommended if it is decided to continue breastfeeding. Occasional or short term treatment with benzodiazepines could be considered as compatible with breastfeeding, although maternal diazepam treatment has caused sedation in suckling infants after short term use. During long term use of benzodiazepines, infants should be observed for signs of sedation and poor suckling. Only very limited clinical data are available for the new generation anticonvulsant drugs and no clearcut recommendations can be made until further data are present. If it is decided to continue breast feeding during treatment with these drugs, the infant should be monitored for possible adverse effects. In general, the drug should be given in the lowest effective dose, guided by maternal serum or plasma drug concentration monitoring. If breast feeding is avoided at times of peak drug levels in milk, the exposure of the infant can be reduced to some extent. As breast milk has considerable advantages over formula milk, the benefits of continuing breast feeding should always be taken into consideration in the risk-benefit analysis.