Recent data suggest that risk for venous thromboembolism is associated with biologic aggressiveness of cancer. Findings in the Vienna Cancer and Thrombosis Study, recently reported by Ahlbrecht and colleagues in Journal of Clinical Oncology, indicate that patients with higher-grade tumors are at significantly increased risk of venous thromboembolism.1 Further, tumor grade was found to be associated with changes in levels of D-dimer, leukocytes, and hemoglobin—biomarkers that have also been linked with venous thromboembolism risk.
Study Details
In the study, 747 patients with newly diagnosed solid tumors or progression of disease after remission were followed for a median of 526 days for occurrence of symptomatic and objectively confirmed venous thromboembolism. Patients were not routinely screened for the disease. Accidentally detected thrombotic events were counted as venous thromboembolism events if they were considered clinically significant by the study adjudication committee.
Patients had a median age of 63 years, and 54% were male. Sites of cancer included lung in 20.3%, breast in 19.1%, colorectal in 17.4%, prostate in 17.4%, pancreas in 8.8%, stomach in 6.7%, kidney in 4.6%, and “other” in 5.8%.
For tumor grade, a three-level grading system (grades 1 to 3) was used for breast and prostate cancer, and a four-level system was used for other cancers. Overall, 62.7% of patients had low-grade tumors (7.0% grade 1 and 55.7% grade 2) and 37.3% had high-grade tumors (35.7% grade 3 and 1.6% grade 4). Disease was localized in 49.0% and consisted of distant metastasis in 50.5%.
Histology consisted of adenocarcinoma in 83.5% of patients and nonadenocarcinoma in 16.1%, primarily squamous cell carcinoma of various sites, renal clear cell carcinoma, and small cell lung cancer. During the study period, 46.9% of patients received surgery, 66.1% received chemotherapy, and 44.8% received radiotherapy.
During up to 2 years of follow-up, 282 patients (37.8%) died without clear evidence of venous thromboembolism. Objectively confirmed venous thromboembolism occurred in 52 patients (7.0%), including asymptomatic disease in 11. The cumulative probability of venous thromboembolism in the entire population was 5.5% after 6 months and 6.2% after 1 year.
Of all venous thromboembolism events, most were either isolated deep-vein thrombosis of the lower extremity (36.5%) or isolated pulmonary embolism (36.5%). By cancer site, 26.9% of events occurred in patients with pancreas cancer, 23.1% in those with colorectal cancer, 17.3% in those with stomach cancer, 15.4% in those with lung cancer, and 5.8% each in those with prostate, breast, and other cancer sites.
High-grade Tumors Predict Increased Risk
Overall, 51.9% of venous thromboembolism events occurred in patients with low-grade tumors (5.8% in those with grade 1 and 46.2% in those with grade 2) and 48.1% occurred in those with high-grade tumors (46.2% in those with grade 3 and 1.9% in those with grade 4). On univariate analysis, patients with high-grade tumors had a significant 80% increase in risk for venous thromboembolism (hazard ratio [HR] = 1.8, 95% confidence interval [CI] = 1.0–3.1, P = .04).
A multivariate model adjusted for distant metastasis, sex, age, and interaction between histology (adenocarcinoma vs nonadenocarcinoma) and tumor sites (high-risk vs low-risk); sites considered to pose high risk for venous thromboembolism were the pancreas, gastrointestinal system, lung, and kidney, and low-risk sites were the breast, prostate, and “other” sites. On this analysis, high-grade tumors were associated with a significant twofold increased risk of venous thromboembolism (HR = 2.0, 95% CI = 1.1–3.5, P = .015).
A similar outcome was found in a second multivariate model including distant metastasis, interaction between histology and tumor sites, and treatment (surgery, chemotherapy, and radiotherapy), with high-grade tumors being associated with a 90% increased risk of venous thromboembolism (HR = 1.9, 95% CI = 1.1–3.3, P = .023). On Kaplan-Meier analysis, the cumulative probability of venous thromboembolism after 6 months was 8.2% in patients with high-grade tumors vs 4.0% in those with low-grade tumors (P = .037).
Tumor Grade and Biomarkers
High-grade tumors were associated with differences in some of the biomarkers previously reported to be predictive of venous thromboembolism. Patients with high-grade tumors had significantly higher median D-dimer levels (0.78 vs 0.64 µg/mL, P = .008) and leukocyte counts (7.5 vs 6.8 × 109/L, P < .001) and significantly lower median hemoglobin levels (12.9 vs 13.2 g/dL, P = .008). There were no differences between patients with high-grade tumors and those with low-grade tumors with regard to other biomarkers, including prothrombin fragment 1 + 2, sP-selectin, clotting factor VIII activity, and platelet count.
A third multivariate model included the independent variables from the first two models (ie, tumor grade, histology, and tumor sites) and the biomarkers D-dimer, hemoglobin, platelet count, and leukocyte count—all routinely available clinical or laboratory variables. On this model, the adjusted hazard ratio for venous thromboembolism for high-grade tumors was 1.9 (95% CI = 1.1–3.3, P = .023). Also independently associated with venous thromboembolism were high-risk tumor sites (HR = 4.3, 95% CI = 2.1–9.0, P < .001), adenocarcinoma histology (HR = 2.7, 95% CI = 1.1–6.9, P = .038), and D-dimer level greater than 1.32 µg/mL (75% percentile of the population; HR = 2.2, 95% CI = 1.3–3.9, P = .005).
The findings in this study suggest that tumor differentiation may play an important role in the development of cancer-related venous thromboembolism. As stated by the investigators, “Tumor grade is a routine parameter available for almost all solid tumor types and thus represents an easy tool for risk assessment of cancer-associated [venous thromboembolism]. Because thromboprophylaxis … is still a major challenge, it may be useful to incorporate the tumor grade in risk assessment models…. This could enable improved assessment of the [venous thromboembolism] risk in the heterogeneous population of patients with cancer.” ■
Reference
1. Ahlbrecht J, Dickmann B, Ay C, et al: Tumor grade is associated with venous thromboembolism in patients with cancer: Results from the Vienna Cancer and Thrombosis Study. J Clin Oncol 30:3870-3875, 2012.
