AML-035: Bleeding Outcomes in Thrombocytopenic Acute Leukemia Patients with Venous Thromboembolism at a Comprehensive Cancer Center

Abstract

Context Cancer-associated thrombosis (CAT) in leukemia patients with severe thrombocytopenia poses a treatment challenge due to competing risks of bleeding and recurrent thrombosis resulting in a narrow therapeutic index for antithrombotic therapies. Objective To determine rates of clinically relevant bleeding (CRB) during management of acute venous thromboembolic event (VTE) in thrombocytopenic acute leukemia patients. Design Retrospective analysis using chart review. Patients Patients (n=74) with acute leukemia admitted to the University of Texas M.D. Anderson Cancer Center, January 2002 - March 2016. Included patients were older than 17 years, with proximal lower extremity deep venous thrombosis and/or pulmonary embolism, and thrombocytopenia (platelet levels ≤50×109/L) on at least two consecutive measures. Patients were sub-grouped into three VTE-treatment interventions: anticoagulation (32.4%), inferior vena cava filter (29.7%), and observation (37.8%). Main outcome measures Rates of CRB and by VTE-treatment intervention group. Association between CRB risk and clinical and laboratory characteristics by multivariate model based on backward model selection. Results Twenty-nine patients (39.2%) suffered CRB during the study (9 had clinically relevant non-major bleeding, 19 had major bleeding, and 1 had both major and non-major bleeding). Patients with relapsed leukemia bled more than those with de novo leukemia (P=0.0294). The use of anticoagulation did not increase the occurrence of CRB (P=0.6035). Use of chemotherapy increased the odds of bleeding by six times (P=0.043). Patients with CRB received more transfusions of platelets (P=0.0012), packed red blood cells (P=0.0017), and fresh frozen plasma (P=0.0573) compared to those who did not bleed. Baseline platelet count was not a significant factor associated to CRB(P=0.35), but the rate of increased platelet counts over time correlated with decreased bleeding chance (P=0.01). Conclusions There was a significant correlation between the occurrence of CRB and quantity of overall blood product transfusions, use of chemotherapy, and relapsed leukemia presentation. There was no difference in rates of CRB between VTE-treatment sub-groups, regardless of platelet count at time of VTE diagnosis. It appears the platelet count trajectory and the factors associated to a slower recovery of it are the main determinants for hemorrhagic complications during VTE treatment in acute leukemia.