This prospective study tested the hypothesis that abnormal signal-averaged electrocardiograms (ECGs) and inducible ventricular arrhythmias identify patients awaiting cardiac transplantation who are prone to sustained ventricular tachycardia (VT) or ventricular fibrillation (VF). Thirty-seven patients with advanced symptoms of heart failure and a mean left ventricular ejection fraction of 20 +/- 7 were studied. In response to programmed ventricular stimulation using up to 3 extrastimuli, sustained monomorphic VT was induced in 8 (22%) and polymorphic VT or VF was induced in 5 patients (13%). Patients with inducible arrhythmias underwent drug therapy guided by results of programmed ventricular stimulation or implantation of a defibrillator. Patients in whom ventricular arrhythmias could not be induced were not treated for arrhythmias. The signal-averaged ECG was abnormal and sustained VT or VF was induced in 10 patients (27%). Follow-up ranged from 1 to 33 months (mean 12). Four patients (11%) died suddenly and 4 (11%) had nonfatal sustained VT or VF. The positive predictive value for sudden death or nonfatal VT/VF was 27% for the signal-averaged ECG, 38% for programmed ventricular stimulation, and 50% if both tests were abnormal. The negative predictive values for these tests were 87, 88 and 88%, respectively. The actuarial incidence of arrhythmic events was significantly higher in patients with inducible ventricular arrhythmias (p = 0.017) and in patients in whom both the results of signal-averaged electrocardiographic analysis and the response to programmed ventricular stimulation were abnormal (p = 0.002). This study demonstrates that results of signal-averaged electrocardiographic analysis and the response to programmed ventricular stimulation improve risk stratification for sudden cardiac death in patients awaiting cardiac transplantation.