ABSTRACT: Background There is limited evidence of long-term impact of exercise-based cardiac rehabilitation (CR) on clinical end points for patients with atrial fibrillation (AF). We therefore compared 18-month all-cause mortality, hospitalization, stroke, and heart failure in patients with AF and an electronic medical record of exercise-based CR to matched controls. Methods and Results This retrospective cohort study included patient data obtained on February 3, 2021 from a global federated health research network. Patients with AF undergoing exercise-based CR were propensity-score matched to patients with AF without exercise-based CR by age, sex, race, comorbidities, cardiovascular procedures, and cardiovascular medication. We ascertained 18-month incidence of all-cause mortality, hospitalization, stroke, and heart failure. Of 1 366 422 patients with AF, 11 947 patients had an electronic medical record of exercise-based CR within 6-months of incident AF who were propensity-score matched with 11 947 patients with AF without CR. Exercise-based CR was associated with 68% lower odds of all-cause mortality (odds ratio, 0.32; 95% CI, 0.29-0.35), 44% lower odds of rehospitalization (0.56; 95% CI, 0.53-0.59), and 16% lower odds of incident stroke (0.84; 95% CI, 0.72-0.99) compared with propensity-score matched controls. No significant associations were shown for incident heart failure (0.93; 95% CI, 0.84-1.04). The beneficial association of exercise-based CR on all-cause mortality was independent of sex, older age, comorbidities, and AF subtype. Conclusions Exercise-based CR among patients with incident AF was associated with lower odds of all-cause mortality, rehospitalization, and incident stroke at 18-month follow-up, supporting the provision of exercise-based CR for patients with AF.
