Comparing the Safety and Efficacy of Apixaban and Rivaroxaban

Study hypothesis and design: The study hypothesizes that apixaban will be superior to rivaroxaban for safety (using the International Society on Thrombosis and Haemostasis \[ISTH\] definition of major bleeding) and at least non inferior for efficacy among patients with atrial fibrillation (AF) or atrial flutter (AFL), henceforth collectively "AF." This is a phase 4, pragmatic, point of care, parallel group, unblinded randomized trial embedded in VA clinical care. Enrollment and allocation: Approximately 10,000 Veterans ages 65 years or older with AF and CHA2DS2-Vasc of 3 or more will be randomized 1:1 to apixaban or rivaroxaban; Veterans aged 22-64 may also enroll and will be included in exploratory analyses. Randomization uses site specific permuted blocks with random block sizes via a centralized Interactive Web Response System (IWRS). Co Primary Objectives 1. Determine whether apixaban is superior to rivaroxaban for ISTH major bleeding (safety). 2. Determine whether apixaban is non inferior to rivaroxaban for the composite efficacy outcome (ischemic stroke, systemic embolism, or all cause death). Secondary Objectives 1. Test superiority of apixaban vs rivaroxaban for the composite efficacy outcome. 2. Assess impact on hospitalization for heart failure, myocardial infarction, or acute coronary syndrome/unstable angina. 3. Examine each component of the composite efficacy endpoint individually (ischemic stroke, systemic embolism, all cause mortality). Co Primary Endpoints 1. Time to first ISTH defined major bleeding (Superiority Hypothesis). 2. Time to first ischemic stroke, systemic embolism, or all cause mortality (Non Inferiority Hypothesis). Secondary Endpoints (hierarchically ranked) 1. Time to first ischemic stroke, systemic embolism, or all cause mortality (Superiority Hypothesis). 2. Time to first ischemic stroke. 3. Time to first hospitalization for heart failure, myocardial infarction, or acute coronary syndrome. 4. Time to first systemic embolism. 5. Time to all cause death. Sites and representation: About 100 VA Medical Centers across the United States will enroll participants, representing both tertiary urban centers and rural CBOCs, with a goal of broad representation across VISNs. Efforts will promote enrollment of women and racial/ethnic minority Veterans. Site selection will give strong consideration to prior experience with ambulatory cardiac monitors, particularly the 14 day Zio patch. The study will not include sites outside the U.S. Population and eligibility: The primary analysis comprises Veterans ages 65 years or older with AF/AFL and CHA2DS2-Vasc of 3 or more. Inclusion requires a diagnosis of AF/AFL and current use of either apixaban or rivaroxaban. Antiplatelet therapy is permitted. Key exclusions include inability to switch anticoagulants if needed, other indications for anticoagulation, contraindication to OAC, bleeding diathesis, pregnancy/lactation, allergy/intolerance to study drugs, eGFR \<30 mL/min/1.73 m², mechanical valve, moderate-severe mitral stenosis, prior left atrial occlusion/excision/ligation, certain interacting medications (e.g., systemic ritonavir, itraconazole, ketoconazole; topical ketoconazole allowed), recent cardiac/thoracic surgery, cognitive impairment precluding consent, or concurrent interventional trial participation without waiver. Interventions and dosing: * Apixaban: 5 mg BID; 2.5 milligrams (mg) BID if 2 or more of: age 80 years, body weight 60 kilograms (kg), serum creatinine 1.5 milligrams per deciliter (mg/dL). * Rivaroxaban: 20 mg QD if creatinine clearance is 50 mL/min or more; 15 mg QD if creatinine clearance 15-50 mL/min. (Note: eGFR \<30 is an exclusion.) Medications are dispensed via VA pharmacy/CMOP per usual practice, with routine co pay policies; switching instructions are provided to avoid double dosing. Operations, follow up, and data capture: Enrollment and initiation typically occur within \<1 month, and can extend to up to 90 days for participants randomized to switch who recently received a 90 day supply. After randomization, all clinical management is per participants' usual VA providers. Outcomes (bleeding, stroke, systemic embolism, hospitalizations, death) are ascertained remotely from the VA EHR/CDW and, for participants ages 65 or older, from Medicare data; no protocol mandated study visits are required. Adherence is assessed via VA pharmacy refill data. Timeline: Planned 3 years of enrollment plus 3 or more years of follow up after the last participant (total \~6 years of data collection), with \~1 year for completion of analyses (\~7 years total, 84 months).