Inpatient Mortality and Length of Stay Among Direct-Acting Oral Anticoagulant (DOAC) and Warfarin Users Presenting with Major Hemorrhage

Background: Use of direct-acting oral anticoagulants (DOAC) is increasingly common among patients with atrial fibrillation and venous thromboembolic disease. Differences in the mechanisms of action as compared to warfarin could impact transfusion patterns and clinical outcomes in patients, especially for those presenting with major hemorrhage. The management of patients taking these newer medications and corresponding outcomes are relevant to optimizing clinical decision making in situations of major hemorrhage.

Methods: We tested the hypothesis that inpatient all-cause mortality among patients presenting with major hemorrhage differs based on the home-administered anticoagulant medication class (DOAC versus warfarin). A cohort of patients presenting to twelve US hospitals from 2013 to 2016 was identified using the Recipient Epidemiology and Donor Evaluation Study (REDS)-III Recipient Database. Primary ICD diagnosis codes, issued blood products, laboratory data, and early mortality events were used in the application of the International Society on Thrombosis and Hemostasis definition of major hemorrhage. Exposure status was defined as a record of home-administered DOAC (apixaban, dabigatran, edoxaban, or rivaroxaban; exposed) or warfarin (non-exposed). Patients with multiple encounters and those transferred into or out of network were excluded from the analysis. Proportional hazards regression was used to compare all-cause mortality and hospital length of stay. We then repeated the analysis using a cohort matched on propensity scores to account for confounding by age, gender, concurrent aspirin and anti-platelet use, liver and renal dysfunction, cancer, CHA₂DS₂-VASc score, traumatic injury, and hospital. We then repeated the propensity score matched analysis stratified by anatomic location of bleed and traumatic injury.

Results: More than 1.5 million hospitalizations were screened for eligibility. Exclusion of minors, outpatients, hospitalizations without a medication of interest, absence of major hemorrhage, multiple hospitalizations, and hospital transfers resulted in 3,731 patients available for the unadjusted analysis. Inpatient all-cause mortality was lower among DOAC users when the entire cohort was considered (HR = 0.60, 95%CI 0.45 - 0.80, p=0.0005). Implementation of propensity score matching to account for confounding abrogated this difference (HR=0.84, 95%CI 0.58 - 1.22, p=0.36). Time to hospital discharge was shorter for DOAC users (HR = 1.17, 95%CI 1.05 - 1.30, p=0.0034). Transfusion patterns were similar by medication, except for plasma transfusion occurring in 42% of warfarin encounters and 11% of DOAC encounters. Vitamin K was administered in 63% of warfarin encounters, whereas specific DOAC reversal agents were largely unavailable during the analysis period [used in 5 (1%) DOAC encounters]. There were no statistically significant differences in inpatient all-cause mortality in the stratified analysis (warfarin as reference): HR = 0.69 (95%CI 0.31 - 1.55) for traumatic head injuries; HR = 1.10 (95%CI 0.62 - 1.95) for non-traumatic head injuries; HR = 0.62 (95%CI 0.20 - 1.94) for traumatic, non-head injuries; and HR = 0.69 (95%CI 0.29 - 1.63) for non-traumatic, non-head injuries.

Conclusions: Analysis of a population taking oral anticoagulation and presenting with major hemorrhage showed that transfusion of plasma was more commonly employed to treat major hemorrhage among warfarin users than DOAC users. Inpatient all-cause mortality was lower among DOAC users in the overall cohort; however, accounting for potential confounding factors using propensity score matching abrogated this difference. Hospital length of stay was shorter for DOAC users compared to warfarin users. Stratification by location of bleed and traumatic injury did not alter these findings. Less plasma use and a shorter length of hospitalization in this study, combined with no observable difference in inpatient all-cause mortality, suggests that outcomes following major hemorrhage are at least no different for DOAC users as compared to warfarin users.