To evaluate the efficacy of propafenone for suppression of recurrent paroxysmal symptomatic atrial fibrillation (AF), patients with frequent episodes of AF entered an open-label, dose-ranging study to determine the maximal tolerated dose of propafenone and were subsequently randomized to alternate between propafenone and placebo every month for 4 months. Patients recorded each episode of AF in a diary and recorded a simultaneous electrocardiographic rhythm strip by means of a transtelephonic recorder and transmitter to validate the presence of AF. Eighteen patients were eligible for study. During dose ranging, 4 patients withdrew due to inadequate drug efficacy or poor compliance, 2 withdrew due to intolerable side effects and 1 died. The mean dose of propafenone at the end of dose ranging was 644 +/- 189 mg/day. During the crossover study, the percentage of days with an attack of AF was significantly reduced by propafenone compared with placebo (27 +/- 34 vs 51 +/- 34%, p less than 0.01). The rate of early crossover or withdrawal from the crossover study was 13.6% with propafenone and 45% with placebo (p = 0.056). Five patients went on to receive long-term propafenone and 4 continued treatment with suppression of AF for 12 to 21 months. During the crossover study there were 29 reported minor side effects with propafenone and 11 with placebo.