Background: Warfarin prevents ischaemic stroke in patients with non-valvular atrial fibrillation, but dose adjustment, coagulation monitoring, and bleeding risk limit its use. The oral direct thrombin inhibitor ximelagatran represents a potential alternative. We aimed to establish whether ximelagatran is non-inferior to warfarin, within a margin of 2% per year, for prevention of stroke and systemic embolism.
Methods: We randomised 3410 patients with atrial fibrillation and one or more stroke risk factors to open-label warfarin (adjusted-dose, international normalised ratio [INR] 2.0-3.0) or ximelagatran (fixed-dose, 36 mg twice daily); patients were recruited from 259 hospitals, doctor's offices, or health-care clinics. Primary analysis was based on masked event assessment and was by intention to treat. Primary endpoint was stroke or systemic embolism.
Findings: During 4941 patient-years of exposure (mean 17.4 months, SD 4.1), 96 patients had primary events (56 in the warfarin group vs 40 in the ximelagatran group). The primary event rate by intention to treat was 2.3% per year with warfarin and 1.6% per year with ximelagatran (absolute risk reduction 0.7% [95% CI -0.1 to 1.4], p=0.10; relative risk reduction 29% [95% CI -6.5 to 52]). Rates of disabling or fatal stroke, mortality, and major bleeding were similar between groups, but combined minor and major haemorrhages were lower with ximelagatran than with warfarin (29.8% vs 25.8% per year; relative risk reduction 14% [4 to 22]; p=0.007). Raised serum alanine aminotransferase was more common with ximelagatran.
Interpretation: In high-risk patients with atrial fibrillation, fixed-dose oral ximelagatran was at least as effective as well-controlled warfarin for prevention of stroke and systemic embolism.