The aim of treating epilepsy is to control or at least decrease seizures without producing unacceptable adverse effects that impair quality of life. Antiepileptic drugs (AEDs) have been considered amongst the drugs most frequently associated with fatal suspected adverse drug reactions. Physicians must therefore be as familiar with safety and tolerability data of AEDs as they are with the expected therapeutic effects. AEDs may cause dose-related adverse effects (i.e. drowsiness, fatigue, dizziness, blurry vision and incoordination) that, in most cases, may be obviated by lowering the dosage, reducing the number of drugs or switching to a better tolerated AED. AEDs also have the potential of precipitating idiosyncratic adverse effects (i.e. serious cutaneous, haematological and hepatic events), which are more common in children and usually require withdrawal of the AED. Although occurrence of idiosyncratic adverse effects can only rarely be predicted or prevented, there are known risk factors that can help in identifying patients at high risk. Occurrence of an idiosyncratic event in a close relative, a concomitant autoimmune disease, co-treatment with specific drugs, history of a previous allergic drug reaction, starting treatment with high doses and rapid titration have all been associated with a higher risk of idiosyncratic adverse effects. New AEDs have been developed in the last two decades with the aim of improving the benefit-risk balance of AED therapy. Available evidence suggests that the newer AEDs are no more effective but may be somewhat better tolerated than older molecules. We performed a literature review with the aim of evaluating safety and tolerability of second- and third-generation AEDs in children. A PubMed search was conducted with the purpose of identifying English-language studies published between 1 January 1989 and 1 January 2011 that reported any adverse event having occurred in children with epilepsy in whom second- and third-generation AEDs were administered.