https://orcid.org/0000-0002-9340-8771
Key Points
A high VTE incidence in asparaginase-treated ALL, despite thromboprophylaxis, indicates a need for novel approaches.
We derived and externally validated a VTE risk prediction model based on D-dimer and hemoglobin levels at ALL diagnosis.
The incidence of venous thromboembolism (VTE) in patients with acute lymphoblastic leukemia (ALL) receiving asparaginase-based induction is high despite primary thromboprophylaxis. Our aim was to derive and externally validate a VTE risk prediction model in patients with ALL receiving asparaginase-based induction. We conducted a multicenter retrospective cohort study of patients (≥18yrs) with newly-diagnosed ALL receiving asparaginase-based induction. The derivation and external validation cohorts included 306 and 94 patients respectively. Primary outcome was VTE at any site. Cause-specific Cox proportional hazards model, stratified by thromboprophylaxis and center, was performed to identify VTE risk factors in the derivation cohort. A risk prediction model for VTE at 30-days was derived using variables with p-value<0.05 in the multivariable model and was tested in the validation cohort. VTE risk factors on multivariable analysis in the derivation cohort included D-dimer ≥ 1 µg FEU/mL (Hazard ratio [HR] 2.64; 95% CI 1.07-6.5) and hemoglobin (HR for each 1 g/dL increment: 1.19; 95% CI 1.06-1.34) at ALL diagnosis. A VTE risk score based on these variables distinguished between a 4% (95% CI 0.72-12%) and 20% (95% CI 14-27%) 30-day cumulative incidence of VTE in the derivation cohort, with similar findings in the validation cohort (AUC 0.56). The negative predictive value for VTE at 30 days was 96% and 93% in the derivation and validation cohorts respectively, and the positive predictive value was 20% in both. We derived and validated a model using D-dimer and hemoglobin, which stratifies VTE risk in patients with ALL receiving asparaginase-based induction.
