Twenty two patients with heart failure were studied in a double blind crossover trial to compare amiodarone (200 mg/day) with placebo. Each agent was given for three months. Extrasystoles and complex ventricular arrhythmias were common during ambulatory electrocardiographic monitoring and during exercise testing at entry to the study. Breathlessness and tiredness as assessed by visual analogue scores and duration of treadmill exercise did not become worse during amiodarone treatment. During the placebo and amiodarone phases of the study left ventricular ejection fraction and cardiac index determined by first pass radionuclide ventriculography were similar, both at rest and during upright bicycle exercise. Exercise induced ventricular tachycardia was abolished and simple and complex ventricular arrhythmias observed on 24 hour ambulatory monitoring were greatly diminished during amiodarone treatment. Three patients died, all suddenly, during the placebo phase. In two patients amiodarone was withdrawn after a further myocardial infarction in one and a worsening of symptoms of ventricular arrhythmia in the other. In contrast with other antiarrhythmic agents amiodarone is effective in suppressing ventricular arrhythmias in heart failure without causing adverse haemodynamic effects. Because frequent ventricular arrhythmias are known to be associated with a poor prognosis in heart failure, these data suggest that amiodarone may improve the poor prognosis in patients with heart failure.