ABSTRACT
BACKGROUND: The relationship between cancer and thrombosis has been studied for years, but reliable guidelines for thromboprophylaxis in that situation are still unclear. METHODS: We retrospectively reviewed the files of 3159 consecutive patients with newly diagnosed solid tumors at Jules Bordet Institute from January 2008 to December 2011. Among them, 99 developed a symptomatic thromboembolic episode and were matched with 2 controls (nested case control). The aim was to identify risk factors of thromboembolic events and to validate in our setting the Khorana score. RESULTS: In the cohort study, nodal status ≥ 2, presence of metastases, and primary tumor site were found to be the most significant predictive factors of a thromboembolic event (n = 99; 3.1%) in the multivariate analysis. In the nested study (n = 265), hemoglobin < 13 g/dL or treatment with a red cell growth factor, CRP ≥ 31.6 mg/L, creatinine level > 0.96 mg/dL, chronic inflammatory disease, and personal or familial history of thromboembolic events were found to be the most significant predictive factors of a thromboembolic event in the multivariate analysis. In our population, the sensitivity, specificity, positive predictive value, and negative predictive value of the Khorana score were respectively 29%, 93%, 15%, and 96%. CONCLUSION: We confirm the value of the risk factors identified in the literature with the additional presence of nodal involvement, elevated CRP, and creatinine levels, which may be helpful for patient risk stratification and should be considered in future clinical trials. Our results also suggest that the Khorana score might help to identify patients who can safely be spared of thromboprophylaxis.
