Comparative effectiveness and safety of high‐dose rivaroxaban and apixaban for atrial fibrillation: A propensity score‐matched cohort study

Study Objective Observational studies assessing direct oral anticoagulant (DOACs) dosage in atrial fibrillation (AF) reported that a lower proportion of patients received high‐dose DOACs compared to those in randomized controlled trials (RCTs). Effectiveness and safety of high‐dose DOACs relative to apixaban in a real‐world AF population need to be addressed. The aim is to assess comparative effectiveness and safety of high‐dose rivaroxaban relative to apixaban. Design We conducted a cohort study. Setting We built a cohort of patients hospitalized and discharged in community with a primary or secondary AF diagnosis from 2011‐2017 using Quebec administrative databases (Med‐Echo and RAMQ). Patients Cohort entry was defined as the first OAC claim in new users of high‐dose rivaroxaban and apixaban, with no OAC claims in the prior year. Intervention To compare effectiveness and safety of high‐dose rivaroxaban to apixaban. Measurement We ascertained patient demographics, comorbidities, CHA2DS2‐VASc and HASBLED scores and Charlson score within 3 years prior to cohort entry. Primary effectiveness and safety were a composite of ischemic stroke/systemic thrombosis, death, myocardial infarction, and of intracranial bleeding (ICH), extracranial major bleeding, in the first year following drug initiation. We conducted propensity score matching and estimated hazard ratios (HRs) for outcomes using Cox proportional hazard models. All the analyses were conducted to account for competing risks. Main results The cohort consisted of 4,632 and 6,771 patients received high‐dose rivaroxaban and apixaban, respectively. High‐dose rivaroxaban users were younger with a mean age of 73.2 years, presented less associated comorbidities and had lower CHA2DS2‐VASc scores compared to apixaban. High‐dose rivaroxaban at the intention to treat was associated with a higher risk of stroke/SE/death (HR 1.21, 95% CI 1.04‐1.40) and worse composite effectiveness (HR 1.21: 1.05‐1.40); under treatment exposure, those values were at HR (1.66: 1.21‐2.29) and HR (1.58:1.19‐2.10), respectively. And, rivaroxaban presented a less favorable safety profile relative to apixaban. Conclusion In this study, composite effectiveness and safety varied between rivaroxaban and apixaban. High‐dose apixaban was observed to have a better effectiveness and safety.