Comparison of warfarin, dabigatran, rivaroxaban and apixaban for effectiveness and safety among elderly patients with atrial fibrillation: a nationwide cohort study

Abstract

Abstract Introduction Age is a strong independent risk factor for stroke and systemic embolism (SE) in patients with atrial fibrillation (AF). Reducing risk of stroke/SE with oral anticoagulation (OAC) in elderly patients involves a correspondingly greater risk of bleeding than in younger patients. Non-vitamin K antagonist oral anticoagulants (NOACs) are associated with a net clinical benefit over vitamin K antagonists in the elderly population, but knowledge is lacking about the comparative effectiveness and safety between specific oral anticoagulants in these patients. Purpose The aim of this study was to compare the rates of stroke/SE and major bleeding between new users of warfarin, dabigatran, rivaroxaban, and apixaban, in a nationwide cohort of AF patients over 75 years. Methods From Norwegian national registries we identified all anticoagulant naïve initiators of warfarin, dabigatran, rivaroxaban and apixaban over 75 years of age between January 2013 and December 2017. During follow-up, patients were censored upon switching OAC, discontinuation of OAC, death, or end of study period. Multivariate competing risk regression was used to evaluate association between treatment and the outcomes stroke/se and major bleeding, treating death as a competing risk. Results A total of 30 401 patients were identified; 6 650 starting warfarin, 3 857 starting dabigatran, 6 108 starting rivaroxaban, and 13 786 starting apixaban. The median age was 82 years. Dabigatran-users had less comorbidity than all other OAC-users; the greatest difference was seen in the proportion of patients with chronic kidney disease (4.3% in the dabigatran-group versus 7.0%, 10.5%, and 16.5% in the rivaroxaban, apixaban, and warfarin groups, respectively). The median follow-up time was 15 months, during which time 1 386 (4.6%) patients suffered a stroke/SE; 1 277 (4.2%) patients had a major bleeding episode; and 3 270 (10.8%) died. Adjusted subhazard ratios for stroke/SE and major bleeding are presented in the figure. Conclusion Comparing NOACs with warfarin, we found no significant differences in risk of stroke/SE, while apixaban was associated with lower risk of major bleeding than warfarin. Comparing NOACs with each other; dabigatran was associated with a significantly lower risk of stroke/SE compared with rivaroxaban and apixaban, while both dabigatran and apixaban were associated with significantly lower risks of major bleeding compared with rivaroxaban. Incidence rates and subhazard ratios Funding Acknowledgement Type of funding source: Public hospital(s). Main funding source(s): South-Eastern Norway regional Health Authority