In a single-blind randomized study, the efficacy and safety of intravenous propafenone (2 mg/kg body weight per 10 min) versus flecainide (2 mg/kg per 10 min) were assessed in 50 patients with atrial fibrillation or flutter. Treatment was considered successful if sinus rhythm occurred within 1 h. Conversion to sinus was achieved in 11 (55%) of 20 patients with atrial fibrillation treated with propafenone and in 18 (90%) of 20 with atrial fibrillation treated with flecainide (p less than 0.02). If atrial fibrillation was present less than or equal to 24 h, conversion to sinus rhythm was achieved in 8 (57%) of 14 patients in the propafenone group and 13 (93%) of 14 in the flecainide group (p less than 0.05). Atrial flutter was converted in two (40%) of five patients treated with propafenone and in one (20%) of five with flecainide (p = NS). Mean time to conversion was 16 +/- 10 min in the propafenone group versus 18 +/- 13 min in the flecainide group (p = NS). QRS lengthening (83 +/- 15 to 99 +/- 20 ms) was observed only in the patients treated with flecainide (p less than 0.001). Patients successfully treated with propafenone showed significantly higher plasma levels than those whose arrhythmia did not convert to sinus rhythm. Transient adverse effects were more frequent in the flecainide group (40%) than in the propafenone group (8%) (p less than 0.01). In conclusion, at a dose of 2 mg/kg in 10 min, flecainide is more effective than propafenone for conversion of paroxysmal atrial fibrillation to sinus rhythm. However, considering the propafenone plasma levels and very few adverse effects, the dose or infusion rate, or both, used in the propafenone group may not have been sufficient to achieve an optimal effect. Neither drug seems very effective in patients with atrial flutter.