HIGH-2-LOW risk model to predict venous thromboembolism in allogeneic transplant patients after platelet engraftment
Kylee L. Martens, Wilson Luiz da Costa, Christopher I. Amos, Chris Davis, Madeline Kesten, Stephanie J. Lee, Neil A. Zakai, David García, Ang Li
Abstract
Venous thromboembolism (VTE) after allogeneic hematopoietic cell transplantation (HCT) is a significant treatment-associated complication, although optimal timing of thromboprophylaxis remains uncertain when weighing concurrent risks of bleeding. We aimed to derive and internally validate a risk assessment model (RAM) using patients who underwent first allogeneic HCT from 2006 through 2015 (n = 1703). Index date was defined as the 30th day after transplant, at which point we estimated >75% of patients would have achieved platelet engraftment >50 × 109/L. Stepwise logistic regression modeling was used for model development, and internal validation was achieved by fitting a logistic regression model with 1000 bootstrapped resamples to estimate the optimism-corrected c-statistic. The final RAM, "HIGH-2-LOW," included 7 predictors obtained at 30 days after transplant: History of catheter-related deep venous thrombosis (DVT), Inpatient at day 30, Graft-versus-host disease grade 3 to 4, History of pulmonary embolism or lower-extremity DVT, Lymphoma diagnosis, Obesity with body mass index ≥35 kg/m2, and White blood cell count ≥11 × 109/L. Approximately 16% of patients were stratified as high risk, with incident VTE rate of 10.3% at 100 days compared with 1.5% for those at low risk. VTE odds ratios at 100 days were 5.87 (95% confidence interval [CI], 2.98-11.57) and 2.71 (95% CI, 1.38-5.35) in the high- and intermediate-risk vs low-risk groups, respectively. HIGH-2-LOW model serves as a novel and potentially clinically meaningful tool to identify high-risk allogeneic HCT patients who may benefit from early thromboprophylaxis after platelet engraftment.