BACKGROUND--Arrhythmias are a common cause of morbidity after cardiac surgery. This study assessed the efficacy of prophylactic amiodarone in reducing the incidence of atrial fibrillation or flutter and ventricular arrhythmias after coronary artery surgery. METHODS--A double blind, randomised, placebo controlled trial. 60 patients received a 24 hour intravenous infusion of amiodarone (15 mg/kg started after removal of the aortic cross clamp) followed by 200 mg orally three times daily for 5 days, and 60 patients received placebo. RESULTS--6 patients (10%) in the amiodarone group and 14 (23%) in the placebo group needed treatment for arrhythmias (95% confidence interval (95% CI) for the difference between groups was 0 to 26%, p = 0.05). The incidence of supraventricular tachycardia detected clinically and requiring treatment was lower in the amiodarone group (8% amiodarone v 20% placebo, 95% CI 0 to 24%, p = 0.07). The incidence detected by 24 hour Holter monitoring was similar (17% amiodarone v 20% placebo). Untreated arrhythmias in the amiodarone group were either clinically benign and undetected (n = 3) or the ventricular response rate was slow (n = 2). Age > 60 years was a positive risk factor for the development of supraventricular tachycardia in the amiodarone group but not in the placebo group. Fewer patients had episodes of ventricular tachycardia or fibrillation recorded on Holter monitoring in the amiodarone group (15% amiodarone v 33% placebo, 95% CI 3 to 33%, p = 0.02). Bradycardia (78% amiodarone v 48% placebo, 95% CI 14% to 46%, p < 0.005) and pauses (7% amiodarone v 0% placebo) occurred in more amiodarone treated patients. Bradycardia warranted discontinuation of treatment in one patient treated with amiodarone. CONCLUSIONS--The incidence of clinically significant tachycardia was reduced by amiodarone. The ventricular response rate was slowed in supraventricular tachycardia, but the induction of bradycardia may preclude the routine use of amiodarone for prophylaxis.