Predictors and Timing of Venous Thromboembolism in Lymphoma
Asaad Trabolsi,Thomas Plate,Rachel S. Kronenfeld,Leticia E. Campoverde,Dan Morgenstern Kaplan,Alyssa Mercadel,Michael Caballero,Wei Zhao,Izidore S. Lossos,Craig H. Moskowitz,Gerald A. Soff
DOI: https://doi.org/10.1182/blood-2023-173903
IF: 20.3
2023-11-28
Blood
Abstract:Introduction Cancer associated thrombosis (CAT) is common in lymphoma, but the risk is not well characterized. While prophylactic anticoagulation is recommended in certain solid tumors, routine prophylaxis is not practiced in hematologic malignancies. The Kohrana CAT risk score considers lymphoma a single entity. We aimed to identify the incidence rate of venous thromboembolism (VTE) and risk factors in lymphoma patients at our institution. Methods Patients diagnosed with lymphoma between 2018 and 2022 were identified through our institution's tumor registry. Using automated text search from our electronic medical record (EPIC®) combined with natural language processing review, VTE events were identified from 90 days before and up to one year(y) after initiation of chemotherapy initiation, or diagnosis date if no chemotherapy was administered. VTE events included lower extremity deep vein thrombosis (DVT) or pulmonary embolism (PE). Events were reviewed and adjudicated manually by us. We report time to event analysis and survival analysis using Kaplan-Meier curve. The study was censored at one year. Univariate and multivariate analyses were performed using COX-regression model. We evaluated the effect of indolent versus aggressive lymphoma, stage, age, sex, race, ethnicity, white blood cell count, hemoglobin, and platelet count. Results We identified 879 patients(pts) with lymphoma; of those, 378 pts had indolent lymphoma (I-Lym) and 501 pts had aggressive lymphoma (A-Lym). Median age was 61y. 464 pts (52.8%) were above 60 y, 477 pts (54.3%) were males, and 325 pts (37.0%) were Hispanic. Race was distributed as follows: 89.5% white, 9.1% Black, and 1.4% other. 56.9% had advanced stage lymphoma. VTE occurred in 4.9% (n=43) of the overall population within 12 months. The incidence was higher in A-Lym (7.2%, n=36) compared with I-Lym (1.9%, n=7), p<0.001. For both A-Lym and I-Lym, 32.6% of the VTE events occurred prior to diagnosis or initiation of chemotherapy. Among patients who developed thrombosis after the chemotherapy start date, the median time to thrombosis was 39 days. In a univariate analysis (Table 1) for thrombosis-free survival, age>60, receiving chemotherapy and A-Lym, were all associated with higher risk of thrombosis (p<0.05). In the multivariate analysis, only, age>60, and A-Lym remained significant in predicting thrombosis (HR: 3.48, 95%CI: 1.41, 8.61, p=0.007). Hemoglobin <10, platelet count >350, white blood cell count >11, advanced stage, and chemotherapy administration were not significant risks of thrombosis. In multivariate analysis for overall survival (Table 2), development of VTE was associated with increased risk of mortality (HR: 2.59, 95%CI: 1.38, 4.87, p=0.003). This relationship of VTE and mortality was observed independently in both I-Lym and A-Lym. As expected, A-Lym, age>60, male sex, and advanced stage all predicted worse survival in our lymphoma population (p<0.05 for all). Of note, hemoglobin <10 and white blood cell count >11, did not predict VTE in lymphoma but did predict poor OS. Conclusion Lymphoma is associated with a clinically relevant rate of CAT, but incidence is significantly higher in A-Lym compared with I-Lym with a third of events occurring prior to chemotherapy. The hematologic parameters of the Khorana Score (WBC, Hgb, platelet count) did not correlate with incidence of CAT in lymphoma, neither did chemotherapy. Development of CAT during therapy, in multivariate analysis for both A-Lym and I-Lym, was associated with a marked reduction in OS. These points suggest that the thrombotic risk is an inherent property of lymphoma biology in at least some cases, leading to an increased risk of CAT as well as reduced OS. We do not propose that CAT-related mortality is driving the reduced OS, but more likely a marker of poorer prognosis. Limitations: This was a single center/ single dataset and these findings should be considered a derivation cohort, and should be verified with independent additional datasets. If our findings are verified, prospective studies to assess benefit of prophylactic anticoagulation in aggressive lymphoma may be warranted.
hematology