A randomized trial was conducted to assess the efficacy of amiodarone versus metoprolol or no antiarrhythmic treatment to suppress asymptomatic ectopic activity and improve survival in patients who have had myocardial infarction with a left ventricular ejection fraction of 20 to 45% and > or = 3 ventricular premature complexes per hour (pairs or runs). Patients (n = 368) were randomly assigned to receive amiodarone 200 mg/day (n = 115) 10 to 60 days after the acute episode, and metoprolol 100 to 200 mg/day (n = 130) or no antiarrhythmic therapy (n = 123). After a median follow-up of 2.8 years, mortality in the amiodarone-treated patients (3.5 +/- 2% SEM) did not differ significantly from that of untreated control subjects (7.7 +/- 2.5%, p = 0.19), but was lower than that in the metoprolol group (15.4 +/- 3.5%, p = 0.006). Patients treated with metoprolol had twice the mortality seen in control subjects, even though the differences were not statistically significant. Holter studies performed at 1, 6 and 12 months showed that both amiodarone and metoprolol were equally effective in reducing heart rate, whereas only amiodarone significantly reduced ectopic activity (p < 0.0001). Thus, long-term treatment with amiodarone was clearly safe in patients with an ejection fraction of 20 to 45%, was effective in suppressing arrhythmias, and was associated with a lower mortality than metoprolol; corroboration is required in a larger trial.