Abstract 11900: Influence of Competing Risks on the Association Between Warfarin and Ischemic Stroke in Atrial Fibrillation: The Anticoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study
Introduction: Studies examining the association between warfarin therapy and incidence of ischemic stroke in adults with atrial fibrillation (AF) have not accounted for patients who die of non-stroke causes.
Hypothesis: Accounting for the competing risk of death may provide greater understanding of the “real world” impact of warfarin on stroke risk during multiyear follow-up in a large, diverse cohort of AF patients.
Methods: We assessed this association in the ATRIA community-based cohort of AF patients (n=13,559; study years 1996-2003), with thromboembolic (>90% ischemic stroke) events (TEE) being clinician-adjudicated. Extended Cox proportional hazards regression with time-varying warfarin exposure estimated the cause-specific hazard ratio (HR) for TEE while adjusting for stroke risk factors. Fine and Gray subdistribution regression was used to estimate this association while also accounting for competing death events.
Results: Patients using warfarin were younger, more likely to have had a prior stroke, and to have known diabetes, coronary disease, and heart failure, and also had higher mean CHA2DS2VASc scores (3.68 vs. 3.22). The death rate was much higher in the non-warfarin group (8.1 deaths/100 person-years; 2637 deaths vs. 5.5 deaths/100 person-years; 1777 deaths on warfarin). The cause-specific HR indicated a large reduction in TE with warfarin use (adjusted HR: 0.57, 95% CI: 0.50-0.65). In subdistribution hazard models accounting for competing death events over the full follow-up of 6 years, this association was substantially attenuated (adjusted HR: 0.87, 95% CI: 0.77-0.99). In analyses limited to 1-year follow-up with only 648 competing death events, the results without accounting for competing risks (adjusted cause-specific HR: 0.55, 95% CI: 0.43-0.71) were similar to the results that did account for competing risks (adjusted subdistribution HR: 0.59, 95% CI: 0.46-0.75).
Conclusions: By accounting for competing death events, our results reflect a more realistic estimate of the multi-year stroke prevention benefits of warfarin for patients with AF. Many old/frail individuals with AF will not live long enough to gain substantial benefit from warfarin.
Author Disclosures: J.M. Ashburner: None. A.S. Go: Research Grant; Modest; CSL Behring. Y. Chang: None. M.C. Fang: None. L. Fredman: None. K.M. Applebaum: None. D.E. Singer: Research Grant; Modest; Medtronic, Inc., Johnson and Johnson, Briston-Myers Squibb. Consultant/Advisory Board; Modest; Bayer Healthcare, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Johnson and Johnson, St. Jude Medical, Pfizer, CSL Behring.
- © 2015 by American Heart Association, Inc.