Amiodarone was used in 86 patients with ventricular tachycardia (VT) (67 patients) or ventricular fibrillation (19 patients) secondary to coronary artery disease. The mean +/- standard deviation left ventricular ejection fraction was 30 +/- 12% (range 8 to 65%). Prior trials with 4 +/- 1.2 alternate antiarrhythmic agents had been unsuccessful. Amiodarone was loaded at dosages of 1,200 to 1,800 mg/day, with maintenance dosages of 400 to 600 mg/day. Drug efficacy was evaluated by programmed stimulation at 10 to 14 days in 68 patients. In 38 patients sustained VT or ventricular fibrillation was inducible (group I), whereas 30 patients (group II) had either no inducible VT (8) or had nonsustained VT induced (22). Holter monitoring was used to assess drug efficacy in 18 patients (group III). All patients were evaluated at 3- to 6-month intervals with Holter monitors for efficacy and a standard protocol for toxicity. During a long-term follow-up of 18 +/- 16 months, sudden death occurred in 5 patients and nonfatal arrhythmia recurrences were detected in 16. The actuarial probability of freedom from fatal and nonfatal arrhythmia recurrences at 24 months was 0.52 for group I, 0.97 for group II and 0.68 for group III. The mode of induction, rate change or hemodynamic tolerance of the induced ventricular tachycardia did not predict arrhythmia recurrence. Among the clinical variables analyzed, only an ejection fraction of less than or equal to 30% was identified as a significant predictor of arrhythmia recurrence. Nonsudden cardiac death occurred in 21 patients, including 19 from heart failure and 2 from myocardial infarction. Noncardiac death occurred in 7 patients.(ABSTRACT TRUNCATED AT 250 WORDS)