Predictive model of venous thromboembolism in endometrial cancer

Abstract

ObjectiveTo profile characteristics and survival of endometrial cancer patients who develop venous thromboembolism (VTE) and to establish a predictive model of VTE in endometrial cancer. MethodsCases were identified using an institutional database between 2000 and 2011. VTE was correlated to clinico-pathological information and survival outcomes. Frequency and odds ratio (OR) of VTE were examined in a predictive model based on combination patterns of independent risk factors for VTE. ResultsVTE was seen in 42 (8.1%, 95% CI 5.8–10.5) out of 516 cases subsequent to the diagnosis of endometrial cancer. Multivariate analysis identified 4 independent risk factors for VTE: elevated CA-125 (hazard ratio [HR] 5.38, p<0.001), extrauterine disease (HR 2.87, p=0.019), thrombocytosis (HR 2.11, p=0.04), and high risk histology (serous and clear cell, HR 2.09, p=0.049). VTE was the strongest variable for decreased progression-free survival (HR 4.28) and the second strongest variable for decreased overall survival (HR 5.65) in multivariate analysis. In a predictive model of VTE, the presence of multiple risk factors was associated with significantly increased risk of VTE: frequency of VTE, 1.4% if no risk factors, 0–9.3% (OR 1.0–4.2) if a single risk factor, 11.1–25.0% (OR 9.0–24.0) if two risk factors, and 42.9–46.2% (OR 54.0–61.7) if ≥3 risk factors. ConclusionVTE represents a surrogate for aggressive disease in endometrial cancer. Multiple risk factors of VTE in our predictive model demonstrated exceedingly high risk of VTE, suggesting that there may be a certain population of endometrial cancer patients who would benefit from long-term anti-coagulant prophylaxis to improve survival outcome.