5721Comparison of stroke and major bleeding risk of treatment with apixaban vs. rivaroxaban and dabigatran among elderly nonvalvular atrial fibrillation patients in the United States

Abstract

Background: Nonvalvular atrial fibrillation (NVAF) is a significant risk factor for disabling or fatal ischemic stroke and systemic embolism (SE) in the elderly population. It is important to understand clinical outcomes of elderly patients with NVAF treated with different direct oral anticoagulants (DOACs) in routine clinical practice. But evidence is lacking to compare apixaban with other DOACs in elderly NVAF patients. Purpose: To evaluate risk of stroke/SE and major bleeding (MB) among elderly NVAF patients treated with apixaban vs. rivaroxaban and apixaban vs. dabigatran in the real-world setting in the US Methods: Patients with Medicare coverage initiating apixaban, rivaroxaban or dabigatran (index event) were identified from the Humana database (1/1/2013- 9/30/2015). Patients were required to be ≥65 years, have an NVAF diagnosis, and 12 months of continuous health plan enrollment prior to the index event date (baseline period). NVAF patients were grouped into cohorts depending on the drug initiated and patient characteristics were evaluated during the baseline period. Propensity score matching (PSM) was then conducted to control for differences in patient characteristics of study cohorts. The rates of any stroke/SE (ischemic, hemorrhagic, SE) and any MB (intracranial hemorrhage, gastrointestinal (GI) MB, other MBs) were evaluated in the follow-up periods. Cox regressions were used to compare the risk of stroke/SE and MBs of treatment with apixaban vs. rivaroxaban or apixaban vs. dabigatran. Results: After PSM, 13,620 elderly patients treated with apixaban and rivaroxaban were matched, with 6,810 in each cohort (mean ages: 77 years). Patients treated with apixaban vs. rivaroxaban had lower unadjusted annual rates for any stroke/SE (2.3% vs. 3.3%, p=0.02) and any MB (3.8% vs. 7.7%, p<0.001). Cox regression showed that the risks for any stroke/SE (hazard ratio (HR): 0.72, p=0.003), ischemic stroke (HR: 0.67, p=0.01), any MB (HR: 0.49, p<0.001), GI MB (HR: 0.43, p<0.001), and other MB (HR: 0.51, p<0.001) were significantly lower during the follow-up for patients treated with apixaban vs. rivaroxaban. After PSM, 4,654 elderly patients treated with apixaban and dabigatran were matched with 2,327 in each cohort (mean ages: 77 years). Cox regression showed that the risks for any stroke/SE (HR: 0.78, p=0.27) and any MB (HR: 0.82, p=0.23) of NVAF patients treated with apixaban vs. dabigatran were not significantly different during the follow-up. Conclusions: In the real-world setting, after controlling for differences in patient characteristics treatment with apixaban is associated with lower risk of stroke/SE and MBs than treatment with rivaroxaban, and similar outcomes compared with treatment with dabigatran among elderly NVAF patients in the US. Head-to-head trials are needed to fully understand comparative efficacy and safety of different DOACs.