Background: To show results of a prespecified pooled analysis of the studies SPORTIF III (open-label) and SPORTIF V (double-blind), to assess the homogeneity of the results and to explore subgroup analyses and adverse events.
Methods and results: 7,329 patients with atrial fibrillation (AF) and >or=1 additional stroke risk factor were randomized to warfarin (international normalized ratio 2.0-3.0) or ximelagatran (36 mg twice daily). Over 11,346 patient-years (mean 18.5 months/patient), 184 patients developed primary events of stroke and systemic embolism (ximelagatran 1.62 vs. warfarin 1.65%/year; p = 0.94). Heterogeneity between trials with respect to the primary event rate (study-by-treatment interaction p = 0.026) was found. This could not be explained statistically by baseline patient characteristics or by treatment (except perhaps by the better anticoagulation with warfarin in SPORTIF V) and was not evident for secondary end-points. There was no conclusive difference in major bleeding rates (ximelagatran 1.88 vs. warfarin 2.46%/year; p = 0.054), but combined minor plus major bleeding was lower with ximelagatran (31.7 vs. 38.7%/year; p < 0.0001). Elevation of liver enzymes occurred more frequently in patients taking ximelagatran (6.1% vs. warfarin 0.8%; p < 0.0001) and was reversible except in rare cases.
Conclusions: Fixed-dose oral ximelagatran without coagulation monitoring prevented stroke and systemic embolism as effectively as warfarin in patients with AF. Differences in the results of the two trials might relate to consistency of warfarin anticoagulation, different degree of blinding in the two trials, other concomitant therapies or chance. Further investigation is required to explore the long-term safety profile of ximelagatran.