• Shay Perek, Alaa Khatib, Niv Izhaki, Ali Sleman Khalaila, Benjamin Brenner, and Netanel A Horowitz.
• Department of Internal Medicine A, Rambam Health Care Campus, 8, Ha'Aliya Street, Haifa 3109601, Israel; Department of Emergency Medicine, Rambam Health Care Campus, 8, Ha'Aliya Street, Haifa 3109601, Israel; The Ruth and Bruce Rappaport Faculty of Medicine, Technion - Israel Institute of Technology, 1, Efron St, Haifa 3109601, Israel.
• Eur. J. Intern. Med. 2022 Jul 1; 101: 68-75.

BackgroundCatheter-related thrombosis (CRT) is a common complication in cancer patients, that may lead to chemotherapy deferral, elevated risk for systemic infections and pulmonary embolism. This study aimed to assess CRT incidence and risk factors in newly-diagnosed acute myeloid leukemia (AML) patients and create predictive models potentially allowing to decrease CRT occurrence in this population.MethodsThis retrospective single-center analysis included all AML patients treated at the Rambam Health Care Campus between 2006 and 2019. Patient clinical and laboratory data were collected to evaluate thrombosis occurrence and time from AML diagnosis to CRT development. Multivariate classification models were created using logistic regression (LR) and competing risk analyzes.ResultsThe final analysis included 632 newly-diagnosed AML patients (mean age 54 ± 15 years). CRT incidence was 10.1% [confidence interval (CI) 7.7-12.9%], median time from AML diagnosis to CRT was 12.5 days [interquartile range 6-30]. In an LR multivariate model, prior history of venous thromboembolism [adjusted odds ratio (AOR) 12.046, p < 0.0001], acute promyelocytic leukemia (APL) (AOR 2.824, p = 0.015), a high body mass index and initial platelet counts <100 × 10E9/L (AOR 1.059 and 0.546; p = 0.011 and 0.040, respectively) were significantly associated with high CRT risk. Analysis of 587 non-APL patients demonstrated comparable results, with CRT incidence of 9.3% (CI 7.0%-12.1%) and emergence of chronic obstructive pulmonary disease (COPD) as a novel significant co-factor (AOR 34.491, p = 0.004). In both models, the area under curve (AUC) was ≥70%.ConclusionsSignificant CRT risk factors defined using the created model could be used for identification of high-risk newly-diagnosed AML patients requiring CRT prophylaxis.Copyright © 2022. Published by Elsevier B.V.

18 keywords...

hide…