Venous thromboembolism in patients with urothelial carcinoma receiving systemic therapy

Patients with urothelial carcinoma (UC) undergoing systemic therapy face an increased risk of venous thromboembolism (VTE). While the Khorana Risk Score (KRS) is widely applied in oncology to predict VTE, its performance in UC remains uncertain. This study evaluated VTE incidence and clinical predictors in a real-world UC cohort.

A retrospective cohort study of 140 patients with bladder or upper-tract UC treated with systemic therapy (2008-2020) was conducted. Clinical and laboratory data, including KRS variables, treatment details, and venous catheter (VC) status, were analyzed. Logistic regression, ROC, and decision curve analyses (DCA) assessed predictive factors and model performance.

VTE occurred in 24 patients (17.1%) during a median follow-up of 22.4 months. Among KRS components, only erythropoietin (EPO) use showed a statistically significant association with VTE (OR 4.70; 95% CI: 0.96-22.82, p = 0.049). Independent predictors were regional lymph node (LN) metastases (OR 4.35; 95% CI: 1.63-11.94, p = 0.003) and long-term VC presence (OR 4.78; 95% CI: 1.41-22.64, p = 0.023). No other KRS components or traditional factors were significant. None of the patients on therapeutic anticoagulation developed VTE. A multivariable model including EPO, LN metastases, and VC achieved numerically higher predictive accuracy compared with KRS (AUC 0.73 vs. 0.65) and showed greater net clinical benefit on DCA at 5%-40% thresholds.

UC patients on systemic therapy face a high VTE risk not well captured by the KRS. Incorporating EPO use, LN status, and VC presence may improve prediction and support the need for UC-specific risk models to guide thromboprophylaxis.