BACKGROUND: The optimal pharmacologic means to restore and maintain sinus rhythm in patients with atrial fibrillation remains controversial. METHODS: In this double-blind, placebo-controlled trial, we randomly assigned 665 patients who were receiving anticoagulants and had persistent atrial fibrillation to receive amiodarone (267 patients), sotalol (261 patients), or placebo (137 patients) and monitored them for 1 to 4.5 years. The primary end point was the time to recurrence of atrial fibrillation beginning on day 28, determined by means of weekly transtelephonic monitoring. RESULTS: Spontaneous conversion occurred in 27.1 percent of the amiodarone group, 24.2 percent of the sotalol group, and 0.8 percent of the placebo group, and direct-current cardioversion failed in 27.7 percent, 26.5 percent, and 32.1 percent, respectively. The median times to a recurrence of atrial fibrillation were 487 days in the amiodarone group, 74 days in the sotalol group, and 6 days in the placebo group according to intention to treat and 809, 209, and 13 days, respectively, according to treatment received. Amiodarone was superior to sotalol (P<0.001) and to placebo (P<0.001), and sotalol was superior to placebo (P<0.001). In patients with ischemic heart disease, the median time to a recurrence of atrial fibrillation was 569 days with amiodarone therapy and 428 days with sotalol therapy (P=0.53). Restoration and maintenance of sinus rhythm significantly improved the quality of life and exercise capacity. There were no significant differences in major adverse events among the three groups. CONCLUSIONS: Amiodarone and sotalol are equally efficacious in converting atrial fibrillation to sinus rhythm. Amiodarone is superior for maintaining sinus rhythm, but both drugs have similar efficacy in patients with ischemic heart disease. Sustained sinus rhythm is associated with an improved quality of life and improved exercise performance.