Clinical impact of switching or continuation of apixaban or rivaroxaban among patients with non-valvular atrial fibrillation

Abstract Background Numerous studies have demonstrated varied clinical outcomes with initiation of different direct oral anticoagulants (DOACs) among patients with non-valvular atrial fibrillation (NVAF). Although switching of DOACs either for medical or non-medical reasons occurs among some NVAF patients in real-world clinical practice, there has been a lack of evidence on the clinical outcomes of those switchers. Purpose Assess the risk of stroke/systemic embolism (SE) and major bleeding (MB) among NVAF patients who switched from apixaban therapy to rivaroxaban or switched from rivaroxaban therapy to apixaban. Methods This retrospective study identified NVAF patients who initiated the most commonly used DOACs (apixaban or rivaroxaban) between 01 Jan 2013 and 31 December 2021 using a large US administrative claims database (Optum’s de-identified Clinformatics® Data Mart Database) representing patients in commercial or Medicare Advantage insurance plans. Patients who switched from one DOAC to the other within 30 days before or 90 days after discontinuation of initial DOAC and patients who continued receiving the initial DOAC were included. For switchers, the switch date was the index date. For continuers, a hypothetic index date was randomly assigned based on the distribution of the time from initial DOAC prescription to the switch date in the switchers group. Two final cohorts (switchers and continuers) within apixaban initiators and rivaroxaban initiators were further propensity score matched (PSM) based on pre-index characteristics up to a 1:5 ratio. The risks of stroke/SE and MB after the index date were compared between patients who continued the initial DOAC and patients who switched to the other DOAC. Results There were 167,868 apixaban initiators and 65,888 rivaroxaban initiators. After PSM, 2,884 patients who switched from apixaban to rivaroxaban were matched to 14,296 patients who continued apixaban, and 2,853 patients who switched from rivaroxaban to apixaban were matched to 13,330 patients who continued rivaroxaban. Patient characteristics were balanced between the continuers and switchers for both apixaban and rivaroxaban initiators. Patients who switched from apixaban to rivaroxaban were associated with a higher risk of stroke/SE (Hazard Ratio (HR):1.54, 95% Confidence Interval (CI): 1.06-2.24) and MB (HR: 1.73, 95% CI: 1.40-2.14) than those who continued apixaban. Patients who switched from rivaroxaban to apixaban were associated with a simi