Sudden unexpected deaths have been reported with antipsychotic use since the early 1960s. In some cases the antipsychotic may be unrelated to death, but in others it appears to be a causal factor. Antipsychotics can cause sudden death by several mechanisms, but particular interest has centred on torsade de pointes (TdP), a polymorphic ventricular arrhythmia that can progress to ventricular fibrillation and sudden death. The QTc interval is a heart rate-corrected value that represents the time between the onset of electrical depolarisation of the ventricles and the end of repolarisation. Prolongation of the QTc interval is a surrogate marker for the ability of a drug to cause TdP. In individual patients an absolute QTc interval of >500 msec or an increase of 60 msec from baseline is regarded as indicating an increased risk of TdP. However, TdP can occur with lower QTc values or changes. Concern about a relationship between QTc prolongation, TdP and sudden death applies to a wide range of drugs and has led to the withdrawal or restricted labelling of several. Among antipsychotics available in the UK, sertindole was voluntarily suspended, droperidol was withdrawn, and restricted labelling introduced for thioridazine and pimozide. The degree of QTc prolongation is dose dependent and varies between antipsychotics reflecting their different capacity to block cardiac ion channels. Significant prolongation is not a class effect. Among currently available agents, thioridazine and ziprasidone are associated with the greatest QTc prolongation. Virtually all drugs known to cause TdP block the rapidly activating component of the delayed rectifier potassium current (I(kr)). Arrhythmias are more likely to occur if drug-induced QTc prolongation coexists with other risk factors, such as individual susceptibility, presence of congenital long QT syndromes, heart failure, bradycardia, electrolyte imbalance, overdose of a QTc prolonging drug, female sex, restraint, old age, hepatic or renal impairment, and slow metaboliser status. Pharmacodynamic and pharmacokinetic interactions can also increase the risk of arrhythmias. Further research is needed to quantify the risk of sudden death with antipsychotics. The risk should be viewed in the context of the overall risks and benefits of antipsychotic treatment. It seems prudent, where possible, to select antipsychotics that are not associated with marked QTc prolongation. If use of a QTc-prolonging drug is warranted, then measures to reduce the risk should be adopted.