A lack of systematic pharmacoepidemiological studies investigating adverse drug reactions (ADRs) to anticonvulsants makes it difficult to assess accurately the incidence of anticonvulsant-related ADRs. Most of the available information in this regard stems from clinical trial experience, case reports and postmarketing surveillance, sources that are not, by any means, structured to provide precise data on adverse event epidemiology. For various ethical, statistical and logistical reasons, the organisation of structured clinical trials that are likely to provide substantial data on ADRs is extremely difficult. This review concentrates on current literature concerning serious and life-threatening ADRs. As with the older anticonvulsants, the majority of ADRs to newer anticonvulsants are CNS-related, although there are several that are apparently unique to some of these new drugs. Gabapentin has been reported to cause aggravation of seizures, movement disorders and psychiatric disturbances. Felbamate should only be prescribed under close medical supervision because of aplastic anaemia and hepatotoxicity. Lamotrigine causes hypersensitivity reactions that range from simple morbilliform rashes to multi-organ failure. Psychiatric ADRs and deterioration of seizure control have also been reported with lamotrigine treatment. Oxcarbazepine has a safety profile similar to that of carbamazepine. Hyponatraemia associated with oxcarbazepine is also a problem; however, it is less likely to cause rash than carbamazepine. Nonconvulsive status epilepticus has been reported frequently with tiagabine, although there are insufficient data at present to identify risk factors for this ADR. Topiramate frequently causes cognitive ADRs and, in addition, also appears to cause word-finding difficulties, renal calculi and bodyweight loss. Vigabatrin has been reported to cause seizure aggravation, especially in myoclonic seizures. There have been rare reports of other neurological ADRs to vigabatrin, such as encephalopathy, aphasia and motor disturbances. Vigabatrin-induced visual field constriction is the latest and most worrying ADR. Many questions regarding the nature of this potentially serious ADR remain unanswered, as no prospective controlled study examining the phenomenon has been published. Rare cases of behavioural ADRs and IgA and IgG2 deficiency associated with the use of zonisamide have been reported. However, relatively few patients so far have been exposed to this drug, and therefore more postmarketing information is required. The relatively late establishment of aplastic anaemia and hepatic failure as potentially fatal ADRs of felbamate, and of visual field constriction with vigabatrin, should serve as ample reminders that ADRs can appear at any time.