Abstract
OBJECTIVE: Residents of nursing homes are usually excluded from clinical trials, including trials to assess treatments for common conditions such as nonvalvular atrial fibrillation (NVAF). We aimed to quantify the real-world comparative safety and effectiveness of direct-acting oral anticoagulants (DOACs) vs warfarin among nursing home residents with NVAF. DESIGN: Retrospective cohort study using 100% national Minimum Data Set and linked Medicare claims from January 2011 through December 2018. SETTING AND PARTICIPANTS: Long-term care nursing home residents aged ≥66 years enrolled in fee-for-service Medicare. We included individuals diagnosed with NVAF newly initiating oral anticoagulants. METHODS: We identified exposure to DOACs (apixaban, dabigatran, rivaroxaban, and edoxaban) vs warfarin. Outcomes were hospitalization for ischemic stroke/systemic embolism, major bleeding, pneumonia (negative control outcome), and all-cause death. We used inverse probability of treatment weighting competing risk regression models for clinical outcomes and Cox proportional hazards regression for all-cause death. RESULTS: Of 38,983 individuals newly initiating anticoagulants, 19,366 (49.7%) initiated DOACs and 19,617 (50.3%) initiated warfarin. In the inverse probability of treatment weighting analysis, compared with warfarin, there was no statistically significant association between DOAC use and ischemic stroke/systemic embolism [4.5 vs 4.7 events per 100 person-years; adjusted hazard ratio (aHR), 0.94; 95% CI, 0.84-1.05] or major bleeding (12.6 vs 12.4 events per 100 person-years; aHR, 1.03; 95% CI, 0.96-1.10). DOACs use was associated with a modest but statistically significant lower risk of all-cause death (48.1 vs 49.0 events per 100 person-years; IPTW analysis aHR, 0.95; 95% CI, 0.91-0.98). CONCLUSIONS AND IMPLICATIONS: Among nursing home residents with NVAF, DOACs and warfarin were associated with a similar risk of ischemic stroke/systemic embolism and major bleeding. However, the use of DOACs was associated with a slightly reduced risk of all-cause mortality.