UCSF

Background

Extrapolation of clinical trial results comparing warfarin and direct-acting oral anticoagulant (DOAC) users experiencing major hemorrhage to clinical care is challenging due to differences seen among non-randomized oral anticoagulant users, bleed location, and etiology. We hypothesized that inpatient all-cause-mortality among patients presenting with major hemorrhage differed based on the home-administered anticoagulant medication class, DOAC versus warfarin.

Methods

More than 1.5 million hospitalizations were screened and 3731 patients with major hemorrhage were identified in the REDS-III Recipient Database. Propensity score matching and stratification were used to account for potentially confounding factors.

Results

Inpatient all-cause-mortality was lower for DOAC (HR = 0.60, 95%CI 0.45-0.80, p = 0.0005) before accounting for confounding and competing events. Inpatient all-cause-mortality for 1266 propensity-score-matched patients compared using proportional hazards regression did not differ (HR = 0.84, 95%CI 0.58-1.22, p = 0.36). Inpatient all-cause-mortality in stratified analyses (warfarin as reference) produced: 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. Mean time to discharge was shorter for DOAC (HR = 1.17, 95%CI 1.05-1.30, p = 0.0034) in the propensity score matched analysis. Plasma transfusion occurred in 42% of warfarin hospitalizations and 11% of DOAC hospitalizations. Vitamin K was administered in 63% of warfarin hospitalizations.

Conclusions

After accounting for differences in patient characteristics, location of bleed, and traumatic injury, inpatient survival was no different in patients presenting with major hemorrhage while on DOAC or warfarin.