Nonvalvular atrial fibrillation (AF) is an independent risk factor for stroke that becomes increasingly prevalent as populations age. More than half a dozen clinical trials have demonstrated that anticoagulation with the vitamin K antagonist warfarin is the most effective therapy for stroke prophylaxis in AF. The narrow therapeutic index of warfarin requires that the intensity of anticoagulation be maintained within the international normalized ratio (INR) range of 2.0 to 3.0 to optimize efficacy while minimizing bleeding risk. The pharmacokinetics of warfarin are subject to variability due to interactions with multiple drugs and foods, making maintenance of the INR within this range difficult to achieve in clinical practice without close coagulation monitoring and frequent dose adjustments. Current guidelines recommend oral anticoagulation for high-risk individuals with AF but these inherent limitations lead to substantial underprescribing, particularly in elderly patients at greatest risk. This has stimulated the development of new agents with improved benefit-risk profiles, such as ximelagatran, the first of the oral direct thrombin inhibitors, which has a wider therapeutic margin and low potential for drug interactions, allowing fixed dosing without anticoagulation monitoring. Ximelagatran has been evaluated for stroke prevention in AF in the Stroke Prevention using an Oral Direct Thrombin Inhibitor in Atrial Fibrillation (SPORTIF) program, the largest clinical trials of antithrombotic therapy for stroke prevention in AF to date. The phase III trials of ximelagatran in AF, SPORTIF III and V, found a fixed oral dose of ximelagatran (36 mg twice daily) comparable to dose-adjusted warfarin (INR 2.0 to 3.0) in preventing stroke and systemic thromboembolic complications among high-risk patients with AF. Results from the population of over 7000 patients in SPORTIF III and V demonstrate noninferiority of ximelagatran compared with warfarin. Data from SPORTIF III show an absolute reduction in stroke and systemic embolic events with ximelagatran compared with warfarin of 1.6% per year versus 2.3% per year, respectively ( P = 0.10). SPORTIF V further supports noninferiority between the two agents with an absolute risk reduction of 0.45%, well within the predefined noninferiority margin (95% confidence interval -0.13, 1.03; P = 0.13). Although event rates for major bleeding did not differ significantly with ximelagatran versus warfarin in either study, combined rates for major and minor bleeding were significantly reduced with ximelagatran. The overall net clinical benefit, taking into account effects on stroke or systemic embolic events, major bleeding, and death, was also greater with ximelagatran compared with warfarin in both studies. Elevated serum transaminase enzymes were observed in approximately 6% of patients given ximelagatran in these trials. These typically occurred 1 to 6 months after initiating treatment and usually abated without clinical sequelae whether or not treatment was continued. Given the consistency of response, the favorable overall benefit-risk ratio and the convenience of fixed oral dosing, ximelagatran may increase the number of patients with AF eligible for anticoagulation and amplify the potential for prophylaxis against stroke.