Drug-induced QT interval prolongation with or without Torsade de Pointes has led to the withdrawal, or severe prescribing restrictions being placed on the use, of many drugs. Other drugs have been denied regulatory approval because of their 'QT liability'. This mechanism-based toxicity results principally from inhibition of IKr, the rapid component of delayed rectifier potassium current. The KCNH2 (hERG) gene encodes the physiologically germane alpha-subunits of the channels that conduct this current. Among the battery of non-dinical tests used to characterize a drug for its 'QT liability' are the hERG channel studies. Studies with the hERG channel have been used for early screening of lead compounds and making 'go-no-go' decisions. However, the predictive value of these studies is limited by inter-laboratory variations, a high false-positive rate and lack of a consensus on the definition of a negative study ('safety margins'). From a regulatory perspective, it is reassuring that clinical torsadogens have always been hERG positive with no false negatives. hERG channel studies are useful in guiding electrocardiographic safety monitoring in early human studies, evaluating the relative risks of metabolites and enantiomers of a drug and clarifying uncertain mechanisms of action. One emerging issue of concern is the effect of drugs on hERG trafficking. Classical hERG channel studies will not identify these drugs. For adequate risk assessment, hERG channel data should be integrated with all other non-clinical and clinical data; otherwise there is a risk of novel and valuable drugs being rejected from development and/or denied regulatory approval.