Abstract 11583: Length of Hospital Stay and All-Cause Mortality Among Adults With Atrial Fibrillation Hospitalized for Bleeding During Treatment With Warfarin, Dabigatran, or Rivaroxaban

Introduction: Little is known about the relative outcomes of bleeding complications during warfarin therapy compared to bleeding complications during treatment with dabigatran or rivaroxaban. Hypothesis: Because warfarin is more easily reversible, bleeding complications during warfarin therapy may be associated with better outcomes. Methods: We performed a retrospective cohort study using a US commercial claims database of 38 million members from 11/1/10 to 3/31/14. Participants were adults with atrial fibrillation hospitalized for bleeding after starting warfarin, dabigatran, or rivaroxaban. The outcomes were length of hospital stay, ICU admission, length of ICU stay, and all-cause mortality. The analyses used propensity scores as covariates in a multivariable Poisson regression. Subgroup analyses examined patients with chronic kidney disease, heart failure, more than 7 comorbidities, hemorrhagic stroke, major gastrointestinal bleed, those over 75, those without ICU admission, and those restarting anticoagulation after discharge. Results: Admissions for bleeding included 2446 warfarin, 442 dabigatran, and 256 rivaroxaban users. Warfarin users were older than dabigatran or rivaroxaban users (74 vs 69 vs 68 years) and had more comorbidities. After adjustment, warfarin was associated with a 2.2 mean day increase in hospital stay (CI 2.0 - 2.5) compared to dabigatran and a 3.1 mean day increase (CI 2.8 - 3.4) compared to rivaroxaban. There were no significant differences in the proportions of ICU admissions among the 3 groups. Warfarin was associated with a mean ICU stay increase of 2.3 days (CI 0.94 - 3.6) compared to dabigatran. After 3.4 years follow up, there was a 17% adjusted mortality in the warfarin cohort, 13% in the dabigatran cohort, and 6% in the rivaroxaban cohort. Warfarin was associated with longer hospitalization in every subgroup. Conclusions: Despite warfarin’s reversibility, bleeding complications during warfarin therapy are associated with longer hospitalization and higher adjusted mortality. However, because patients treated with newer oral anticoagulants are younger and healthier, randomized trials are necessary to better understand these unexpected differences in bleeding-related outcomes.