Atrial fibrillation, even in the absence of rheumatic valvular disease, predisposes patients to embolic complications, but the role of antithrombotic therapy in the prevention of such complications has not been fully clarified. We therefore performed a randomized, placebo-controlled trial to evaluate warfarin and aspirin individually as prophylaxis against ischemic stroke and systemic embolism (the primary events) in such patients. Patients eligible to receive warfarin (group 1) were assigned to warfarin (open label), aspirin (325 mg per day), or placebo (aspirin and placebo were given in a doubleblind fashion). Those who were not eligible for warfarin (group 2) received either aspirin or placebo in a double-blind fashion. The placebo arm of group 1 was recently terminated, when evidence emerged that each active agent was superior to placebo. In this paper we report preliminary data on active therapy (with either warfarin or aspirin) as compared with placebo in group 1, and on aspirin as compared with placebo in groups 1 and 2 combined. By November 1989, 1244 patients had been followed for a mean of 1.13 years. The event rates were 1.6 percent per year in the 393 patients who made up the two active treatment arms (warfarin and aspirin) of group 1, and 8.3 percent per year in the 195 patients who made up the placebo arm (P less than 0.00005) (risk reduction, 81 percent; 95 percent confidence interval, 56 to 91). In all 517 patients given aspirin, the rate of primary events (3.2 percent per year) was lower than that in the 528 patients given placebo (6.3 percent per year; P = 0.014) (risk reduction, 49 percent; 95 percent confidence interval, 15 to 69). However, we were unable to show a benefit of aspirin in patients over 75 years of age. These preliminary data indicate that antithrombotic therapy with warfarin or aspirin is effective in the short term in reducing the risk of stroke and systemic embolism in patients with atrial fibrillation due to causes other than rheumatic valvular disease. The relative benefits of aspirin and warfarin remain unclear, and the trial is continuing in order to address this issue.