JAMA Netw Open. 2026 Oct 1;9(10):e2637533. doi: 10.1001/jamanetworkopen.2026.37533.
ABSTRACT
IMPORTANCE: Long-term nursing home residents with atrial fibrillation (AF) and activities of daily living (ADL) disability are at high risk of bleeding and death, making the net benefit of anticoagulation uncertain.
OBJECTIVE: To compare the association of oral anticoagulant (OAC) use vs nonuse with ischemic stroke and major bleeding in long-term nursing home residents with AF and ADL disability.
DESIGN, SETTING, AND PARTICIPANTS: This cohort study used US Medicare claims linked with the Minimum Data Set (January 1, 2013, to December 31, 2022). Nursing home residents with nonvalvular AF and impairment in 3 or more ADLs were included. Three cohorts comparing use of apixaban, rivaroxaban, or warfarin with nonuse were constructed using incidence-density sampling.
EXPOSURES: Initiation of apixaban, rivaroxaban, or warfarin use.
MAIN OUTCOMES AND MEASURES: Primary outcomes were ischemic stroke, major bleeding, and their composite. Propensity score (PS) matching balanced 111 covariates. Adjusted rate ratios (ARRs) with 95% CIs were estimated. Tipping-point analyses assessed the relative weighting ischemic stroke vs major bleeding required to favor anticoagulation. Direct active-comparator analyses compared apixaban with rivaroxaban and warfarin.
RESULTS: The study comprised 26 940 nursing home residents (median [IQR] age, 82 [74,88] years; 17 504 [65.0%] female), including PS-matched cohorts of 7600 residents receiving apixaban, 2654 receiving rivaroxaban, and 3216 receiving warfarin. Compared with nonuse, use of all OACs was associated with lower rates of ischemic stroke (apixaban: ARR, 0.53; 95% CI, 0.38-0.73; rivaroxaban: ARR, 0.48; 95% CI, 0.27-0.87; warfarin: ARR, 0.47; 95% CI, 0.25-0.88) and higher rates of major bleeding (apixaban: ARR, 1.62; 95% CI, 1.29-2.03; rivaroxaban: ARR, 2.68; 95% CI, 1.84-3.93; warfarin: ARR, 3.02; 95% CI, 2.15-4.25). Only apixaban was not associated with a higher composite outcome (ARR, 1.08; 95% CI, 0.90-1.30), whereas rivaroxaban (ARR, 1.62; 95% CI, 1.20-2.18) and warfarin (ARR, 1.99; 95% CI, 1.50-2.63) were associated with higher rates. Findings were consistent across frailty strata, although stroke reduction with apixaban was attenuated among frail individuals. Tipping-point analyses indicated a more favorable benefit-risk balance for apixaban than for rivaroxaban or warfarin.
CONCLUSIONS AND RELEVANCE: Among nursing home residents with AF and ADL disability, OAC use was associated with lower ischemic stroke but higher major bleeding rates than nonuse. Apixaban demonstrated the most favorable benefit-risk profile among the OACs evaluated. These findings may inform individualized anticoagulation decisions.
PMID:42832198 | DOI:10.1001/jamanetworkopen.2026.37533