Disparate Stroke Rates on Warfarin Among Contemporaneous Cohorts With Atrial Fibrillation Potential Insights Into Risk From a Comparative Analysis of SPORTIF III Versus SPORTIF V

Background and Purpose-The rate of stroke among warfarin-treated patients in SPORTIF V was approximately half that of patients enrolled in SPORTIF III (1.16%/year versus 2.30%/year). SPORTIF III was an open-label trial comparing ximelagatran with warfarin for stroke prevention in atrial fibrillation. SPORTIF V was a double-blind trial performed in North America. The trial design was otherwise identical. We sought to determine if differences in baseline characteristics, use of potentially risk-modifying medications, or anticoagulation control help to explain the lower risk of stroke among warfarin-treated patients in SPORTIF V. Methods-Cox regression with stepwise model selection was used to define the covariates independently associated with stroke. Secondary analyses identified covariates with the strongest influence on the study factor (V/III). These covariates were then added to the primary model. Cox regression was used to determine the degree of confounding exerted by these covariates that might help to explain the differences between the trials. Results-Independent risk factors for stroke on warfarin included prior stroke/transient ischemic attack, coronary artery disease, international normalized ratio, weight, and study. Patients in SPORTIF V were at half the risk as those in SPORTIF III. We found that lower international normalized ratio variability, a higher proportion of prevalent warfarin use, lower systolic blood pressure, high-density lipoprotein, and a greater proportion of statin use among patients in SPORTIF V collectively conferred a lower risk of stroke. Conclusion-Differences in blood pressure control, international normalized ratio variability, proportion of prevalent warfarin users, statin exposure, and high-density lipoprotein collectively conferred a lower risk of stroke to patients in SPORTIF V. These findings suggest that the different event rates were not due to chance and provide potential insights into stroke risk among warfarin-treated patients with atrial fibrillation. (Stroke. 2008; 39: 3009-3014.)