Clinical and health care resource use burden of hospitalizations for oral factor Xa inhibitor-associated major bleeding: A real-world analysis of Medicare beneficiaries

Abstract

Objective: To characterize the burden of illness associated with oral factor Xa (FXa) inhibitor-related bleeding in the US Medicare population. Methods: This retrospective cohort study used the full 20% Medicare random sample claims database to identify patients who experienced their first hospitalization for an FXa inhibitor-related major bleed between October 2013 and September 2017. Bleeding types were classified as intracranial hemorrhage (ICH), gastrointestinal (GI), and other. Associations between risk factors and outcomes (in-hospital and 30-day mortality, 30-day readmission, and discharge to a location other than home) adjusted for patient demographic characteristics, baseline clinical conditions, index event characteristics, treatment with hemostatic/factor replacement agents or transfusion (ie, usual care prereversal agent availability), multicompartment ICH and neurosurgical procedures (ICH cohort), and endoscopy (GI cohort) were assessed using multivariable regression and reported as crude incidences and adjusted odds ratios (ORs) stratified by bleed type. Results: Of the 11,593 patients identified, 2737 (23.6%) had ICH, 8169 (70.5%) had GI bleeds, and 687 (5.9%) had other bleeds. The incidences of in-hospital mortality, 30-day mortality, need for postdischarge out-of-home care, and 30-day readmission were 15.7%, 29.1%, 78.3%, and 20.3% in the single-compartment ICH cohort, respectively; and 1.7%, 6.8%, 41.3%, and 18.8% in the GI bleeds cohort, respectively. Increased odds of both in-hospital mortality and 30-day mortality were significantly associated with: multicompartment ICH (reference, single compartment ICH; OR = 3.35 [95% confidence interval (CI): 2.41–4.66]; 2.18 [95% CI: 1.63–2.91]), loss of consciousness during index hospitalization (yes vs no; OR = 2.03 [95% CI: 1.38–2.97]; 1.49 [95% CI: 1.11–2.02]), receiving usual care (yes vs no; OR = 1.55 [95% CI: 1.22–1.98]; 1.33 [95% CI: 1.09–1.63]) during index hospitalization, and increasing number of Elixhauser comorbidities at baseline (OR = 1.07 [95% CI: 1.03–1.10]; 1.09 [95% CI: 1.06–1.12]) in the ICH cohort; intensive care unit admission (yes vs no; OR = 1.88 [95% CI: 1.32–2.67]; 1.51 [95% CI: 1.26–1.81]), increasing number of Elixhauser comorbidities at baseline (OR = 1.12 [95% CI: 1.07–1.18]; 1.15 [1.12–1.18]), and increasing age on index date (OR = 1.04 [95% CI: 1.02–1.07]; 1.05 [95% CI: 1.04–1.07]) in the GI bleeds cohort. Conclusions: In this large sample of Medicare patients, FXa inhibitor-related major bleeding was associated with substantial burden in terms of adverse clinical outcomes and health care resource use. Incidence of ICH was lower than GI bleeds; however, burden of illness was notably higher with ICH.