Racial disparities in the incidence and risk factors of major bleeding during extended anticoagulant therapy for venous thromboembolism.

Abstract

BACKGROUND: Guidelines recommend extended anticoagulation after a first unprovoked venous thromboembolism (VTE) for individuals at low risk of bleeding. However, racial disparities in bleeding risks during extended treatment remain understudied.

OBJECTIVES: To compare risks of anticoagulant-associated bleeding and performance of a risk assessment model by racial group during extended VTE treatment.

METHODS: We analyzed 2 prospective cohorts of patients (223 Black participants and 4314 White participants) with a first unprovoked/weakly provoked VTE who continued anticoagulation after =3 months of initial treatment. Primary outcome was adjudicated International Society on Thrombosis and Haemostasis-defined major bleeding. Secondary outcomes included intracranial hemorrhage, fatal bleeding, and clinically relevant nonmajor bleeding. We determined incidence and hazard ratios (HRs) by race, then adjusted for bleeding risk factors that included the Creatinine, Hemoglobin, Age, antiPlatelet model.

RESULTS: Black participants had higher prevalence of bleeding risk factors and a 1.9-fold higher risk of major bleeding (HR, 1.87; 95% CI, 1.04-3.36) compared with White participants. Adjustment attenuated racial difference for major bleeding but not intracranial hemorrhage (adjusted HR, 2.35; 95% CI, 1.23-4.48). Among those classified as low risk by Creatinine, Hemoglobin, Age, antiPlatelet model, Black participants had numerically higher major bleeding incidence than White participants (2.5 vs 1.1 per 100 person-years). We did not observe racial disparities in fatal bleeding or clinically relevant nonmajor bleeding.

CONCLUSION: Black individuals on extended anticoagulation have higher risk of major bleeding compared with White individuals. This effect appears to persist in those classified as low risk for bleeding. Risk assessment models for anticoagulant-associated bleeding that are generalizable to racialized populations are needed.