Apixaban vs. Rivaroxaban for Atrial Fibrillation

Study Questions:

What is the comparative safety and effectiveness of apixaban versus rivaroxaban for patients with nonvalvular atrial fibrillation (AF)?


The authors performed a US nationwide commercial health claims analysis between 2013-2018. Propensity matching was performed using demographic and comorbidity data. Effectiveness was defined as ischemic stroke or systemic embolism. Safety was defined as intracranial hemorrhage or gastrointestinal bleeding. Cox proportional hazards models were used to estimate safety and effectiveness rates without any further adjustment after the propensity match.


The authors identified 59,172 patients newly prescribed apixaban and 40,707 patients newly prescribed rivaroxaban. The mean age was 69 years and 40% were women. Mean follow-up was 288 (apixaban) or 291 (rivaroxaban) days. After propensity matching, the analysis cohort included 39,351 patients each with apixaban or rivaroxaban therapy. Patients with a CHADS2 score of 1 consisted of 43% of the cohort, while patients with a CHADS2 score of ≥2 consisted of 45% of the cohort. Concurrent antiplatelet use was documented in 10% of the cohort. The incidence rate of ischemic stroke/systemic embolism was 6.6 versus 8.0/1,000 patient-years for apixaban- and rivaroxaban-treated patients, respectively (hazard ratio [HR], 0.82; 95% confidence interval [CI], 0.68-0.98). The rate of intracranial or gastrointestinal bleeding was 12.9 versus 21.9/1,000 patient-years for apixaban and rivaroxaban, respectively (HR, 0.58; 95% CI, 0.52-0.66).


The authors concluded that adult patients with AF prescribed apixaban had lower rates of both ischemic stroke or systemic embolism as well as bleeding compared to patients prescribed rivaroxaban.


This analysis is one of the largest nondirect comparisons between apixaban and rivaroxaban for stroke prevention in AF. Three important findings are reported. First, the overall rate of bleeding is modest (~1-2% per year) in the first year of therapy. These mostly represented gastrointestinal bleeds and not the more critical intracranial bleeds. However, other types of bleeding (e.g., epistaxis, retroperitoneal) were not measured. Second, the rate of bleeding appears to be lower with apixaban therapy as compared to rivaroxaban. This has been shown in other large observational studies. Third, the rate of ischemic stroke or embolism also was lower in the apixaban-treated patients. This has been shown in some (but not all) prior observational studies. However, a few important caveats must be considered. First, the follow-up period was relatively short (only ~9-10 months). Second, the outcomes relied on claims data, which are not as reliable as chart abstraction. Third, a large number of patients in each treatment group discontinued therapy. Fourth, the actual rate of aspirin use is unknown, as it is most commonly taken over-the-counter. And finally, as with all observational studies, unmeasured confounding cannot be addressed even with sophisticated propensity matching. Nonetheless, this analysis supports the findings from prior work in different populations that both apixaban and rivaroxaban have good overall safety and effectiveness profiles, with the possibility that apixaban therapy might be slightly safer and more efficacious.

Clinical Topics: Arrhythmias and Clinical EP, Prevention, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias

Keywords: Arrhythmias, Cardiac, Atrial Fibrillation, Brain Ischemia, Comorbidity, Embolism, Gastrointestinal Hemorrhage, Hemorrhage, Intracranial Hemorrhages, Secondary Prevention, Stroke, Vascular Diseases

< Back to Listings