Many antipsychotic drugs produce QT interval prolongation on the electrocardiogram (ECG). Blockade of the human cardiac K(+) channel known as human ether-a-go-go-related gene (HERG) often underlies such clinical findings. In fact, HERG channel inhibition is now commonly used as a screen to predict the ability of a drug to prolong QT interval. However, the exact relationship between HERG channel blockade, target receptor binding affinity and clinical QT prolongation is not known. Using patch-clamp electrophysiology, we examined a series of seven antipsychotic drugs for their ability to block HERG, and determined their IC(50) values. We then compared these results to their binding affinities (K(i) values) for the dopamine D(2) receptor, the 5-HT(2A) receptor and, where available, to clinical QT prolongation data. We found that sertindole, pimozide and thioridazine displayed little (<10-fold) or no selectivity for dopamine D(2) or 5-HT(2A) receptors relative to their HERG channel affinities. This lack of selectivity likely underlies the significant QT interval prolongation observed with administration of these drugs. Of the other drugs tested (ziprasidone, quetiapine, risperidone and olanzapine), olanzapine displayed the greatest selectivity for dopamine D(2) and 5-HT(2A) receptor binding (100-1000-fold) compared to its HERG channel IC(50). We also compared these HERG channel IC(50) values to QT interval prolongation and plasma drug levels obtained in a recent clinical study. We found that the ratio of total plasma drug concentration to HERG IC(50) value was indicative of the degree of QT prolongation observed. Target receptor affinity and expected clinical plasma levels are important parameters to consider for the interpretation of HERG channel data.