Adjusted-dose warfarin is effective for stroke prevention in patients with nonrheumatic atrial fibrillation (AF), but the risk of bleeding is high, especially among the elderly. Fixed minidose warfarin is effective in preventing venous thromboembolism with low risk of bleeding and no need for frequent clinical monitoring. Patients > 60 years with nonrheumatic AF were randomized in an open-labeled trial to receive fixed minidose warfarin (1.25 mg/day) or standard adjusted-dose warfarin (International Normalized Ratio [INR] between 2.0 and 3.0). Primary outcome events were ischemic stroke, peripheral or visceral embolism, cerebral or fatal bleeding, and vascular death. Secondary end points were major bleeding, myocardial infarction, and death. This study was discontinued before completion in light of publication of the Stroke Prevention in Atrial Fibrillation III trial, which indicated that low-intensity fixed-dose warfarin treatment (i.e., INR < 1.5) was insufficient for stroke prevention in high-risk patients with nonrheumatic AF. From a total of 1,209 considered patients, 303 were randomized to be studied (150 in the minidose group and 153 in the adjusted-dose group). Mean follow-up was 14.5 months. The rate of cumulative primary events was 11.1% (95% confidence intervals [CI] 4.0 to 18.2) in the fixed minidose group and 6.1% (95% CI 1.1 to 11.1) in the adjusted-dose group (p = 0.29). The rate of ischemic stroke was significantly higher in the minidose group (3.7% vs 0% per year, p = 0.025). Major bleedings were more frequent in standard treatment group (2.6% vs 1% per year, p = 0.19). Most thromboembolic complications occurred at INRs < 1.2, whereas the majority of hemorrhages occurred at INRs > 3.0. No significant difference in primary outcome events was observed in the abbreviated study. However, the significantly increased occurrence of ischemic stroke in the fixed minidose warfarin group suggests that this regimen does not protect patients with nonrheumatic AF.