Although chronic amiodarone has been proven to be effective to suppress ventricular tachycardia (VT) and ventricular fibrillation (VF), how we predict the recurrence of VT/VF after chronic amiodarone remains unknown. This study evaluated the predictive value of the QT interval, spatial, and transmural dispersions of repolarization (SDR and TDR) for further arrhythmic events after chronic amiodarone. Eighty-seven leads body surface ECGs were recorded before (pre) and one month after (post) chronic oral amiodarone in 50 patients with sustained monomorphic VT associated with organic heart disease. The Q-Tend (QTe), the Q-Tpeak (QTp), and the interval between Tpeak and Tend (Tp-e) as an index of TDR were measured automatically from 87-lead ECG, corrected Bazett's method (QTce, QTcp, Tcp-e), and averaged among all 87 leads. As an index of SDR, the maximum (max) minus minimum (min) QTce (max-min QTce) and standard deviation of QTce (SD-QTce) was obtained among 87 leads. All patients were prospectively followed (15 +/- 10 months) after starting amiodarone, and 20 patients had arrhythmic events. The univariate analysis revealed that post max QTce, post SD-QTce, post max-min QTce, and post mean Tcp-e from 87-lead but not from 12-lead ECG were the significant predictors for further arrhythmic events. ROC analysis indicated the post max-min QTce > or = 106 ms as the best predictor of events (hazard ratio = 10.4, 95%, CI 2.7 to 40.5, P = 0.0008). Excessive QT prolongation associated with increased spatial and transmural dispersions of repolarization predict the recurrence of VT/VF after amiodarone treatment.