Sixty-two patients with recent-onset (less than or equal to 1 week) atrial fibrillation (New York Heart Association functional class 1 and 2) were randomized in a single-blind study to 1 of the following treatment groups: (1) flecainide (300 mg) as a single oral loading dose; or (2) amiodarone (5 mg/kg) as an intravenous bolus, followed by 1.8 g/day; or (3) placebo for the first 8 hours. Twenty-four-hour Holter recording was performed, and conversion to sinus rhythm at 3, 8, 12 and 24 hours was considered as the criterion of efficacy. Conversion to sinus rhythm was achieved within 8 hours (placebo-controlled period) in 20 of 22 patients (91%) treated with flecainide, 7 of 19 (37%) treated with amiodarone (p less than 0.001 vs flecainide), and 10 of 21 (48%) treated with placebo (p less than 0.01 vs flecainide). Resumption of sinus rhythm within 24 hours occurred in 21 of 22 patients (95%) with flecainide and in 17 of 19 (89%) with amiodarone (p = not significant). Mean conversion times were shorter for flecainide (190 +/- 147 minutes) than for amiodarone (705 +/- 418; p less than 0.001). No major side effects occurred. At Holter monitoring, a pause of 9.3 seconds was observed in 1 asymptomatic patient treated with flecainide. Phases of atrial flutter with a ventricular rate less than or equal to 150 beats/min were detected before sinus conversion in 1 patient receiving placebo and in 2 receiving flecainide.(ABSTRACT TRUNCATED AT 250 WORDS)