The safety and efficacy of long-term amiodarone therapy were examined in 12 patients who had previously developed torsade de pointes as a complication of previous antiarrhythmic therapy. The QTc intervals were determined at the time of torsade de pointes (570 +/- 40 ms), after 7 days of amiodarone loading (490 +/- 70 ms), and after 3 months of chronic amiodarone administration (580 +/- 80 ms). Compared to a drug-free control period, QTc was significantly prolonged (P less than 0.05) at the time of torsade de pointes, after amiodarone loading, and after 3 months of amiodarone therapy. The QTc intervals at the time of torsade de pointes and after chronic amiodarone treatment were not significantly different. At 16 +/- 7 months of follow-up, all patients remained free of subsequent torsade de pointes, syncope, or sudden death. In addition, 5 of 6 patients with a history of sustained ventricular tachycardia remained free from arrhythmic recurrence despite persistence of inducible ventricular tachycardia during programmed stimulation studies done before discharge. We conclude that amiodarone can often be used safely and effectively in patients who have previously had an episode of drug-mediated torsade de pointes. Amiodarone-induced QTc prolongation, even when marked, does not predict recurrent torsade de pointes. These observations also suggest that the propensity for a drug to produce this arrhythmia is dependent on other electrophysiologic effects in addition to its ability to simply lengthen repolarization.