Simplified predictive scores for thrombosis and bleeding complications in newly diagnosed acute leukemia patients

Abstract

BackgroundBleeding and thrombotic complications are the leading causes of death in acute leukemia patients. The Conventional International Society of Thrombosis and Haemostasis Disseminated Intravascular Coagulation (ISTH DIC) scoring system is utilized to assess DIC diagnoses in various conditions. Nevertheless, limited studies have tested the system’s accuracy in predicting thrombo-hemorrhagic events in individuals with acute leukemia. This study aimed to (1) validate the ISTH DIC scoring system and (2) propose a new Siriraj Acute Myeloid/Lymphoblastic Leukemia (SiAML) bleeding and thrombosis scoring system for thrombohemorrhagic risk assessment in acute leukemia.MethodsWe conducted a retro-prospective observational study of newly diagnosed acute leukemia patients between March 2014 and December 2019. We recorded thrombohemorrhagic episodes within 30 days postdiagnosis and DIC profiles, including prothrombin time, platelet level, D-dimer, and fibrinogen. The sensitivities, specificities, positive and negative predictive values, and areas under receiver operating characteristic curves for the ISTH DIC and SiAML scoring systems were calculated.ResultsIn all, 261 acute leukemia patients were identified: 64% with acute myeloid leukemia, 27% with acute lymphoblastic leukemia, and 9% with acute promyelocytic leukemia. Overall bleeding and thrombotic events were 16.8% and 6.1%, respectively. With a cutoff of 5 for the ISTH DIC score, the sensitivity and specificity for bleeding prediction were 43.5% and 74.4%, respectively, while the corresponding values for thrombotic prediction were 37.5% and 71.8%, respectively. D-dimer > 5000 µg FEU/L and fibrinogen ≤ 150 mg/dL were significantly associated with bleeding. A SiAML-bleeding score was calculated using these factors, with a sensitivity and specificity of 65.2% and 65.6%, respectively. Conversely, D-dimer > 7000 µg FEU/L, platelet > 40 × 109/L, and white blood cell level > 15 × 109/L were significant variables related to thrombosis. Using these variables, we established a SiAML-thrombosis score with a sensitivity and specificity of 93.8% and 66.1%, respectively.ConclusionsThe proposed SiAML scoring system might be valuable for prognosticating individuals at risk for bleeding and thrombotic complications. Prospective validation studies are needed to verify its usefulness.