Background: Warfarin has been the only anticoagulant used for decades to prevent strokes and systemic embolisms in nonvalvular atrial fibrillation (NVAF) patients. Compared with rivaroxaban, warfarin has a narrow therapeutic range and many genetic and food-drug interactions that could potentially prolong hospital length of stay (LOS).
Objective: To compare hospital LOS between NVAF patients who were administered rivaroxaban versus warfarin with and without pretreatment of parenteral anticoagulant agents in a population of rivaroxaban-treated patients.
Methods: A retrospective matched-cohort analysis was conducted using the Premier Perspective Comparative Hospital Database from November 2010 to September 2012. Adult patients were included in the study if they had a hospitalization for NVAF. Rivaroxaban users were matched with up to 4 warfarin users based on propensity score analyses. Patients with and without pretreatment of parenteral anticoagulant agents were evaluated separately. Hospital LOS was compared between treatment groups using generalized estimating equations.
Results: The matched cohorts' characteristics were well balanced. Among the matched rivaroxaban and warfarin users who were administered parenteral agents, the mean age of the cohorts was 70 years and 47% of patients were female, whereas in the sample of patients who were not administered parenteral agents, the mean age was 72 years and 50% of patients were female. In the sample of patients who were administered parenteral agents, rivaroxaban users had significantly shorter hospital LOS (LOS difference: 1.38 days, P < 0.001) compared with warfarin users among rivaroxaban-treated patients. No significant difference in LOS was found in the sample of patients who were not administered parenteral anticoagulant agents (P = 0.169).
Conclusion: In the study sample of NVAF patients who were administered parenteral anticoagulant agents, rivaroxaban was associated with a significantly shorter hospital LOS compared with warfarin. The difference in LOS was not statistically significant in the sample of patients who were not administered parenteral anticoagulant agents.