Venous thromboembolism (VTE) is a common complication of cancer, with the risk of developing VTE varying greatly depending on several patient and disease factors. In a study published in Lancet Haematology, researchers analyzed incidence of VTE in two large European cohorts to identify factors associated with an increased risk for VTE.
Ingrid Pabinger, MD, of the Medical University of Vienna in Austria, and co-authors determined that two variables, tumor site and D-dimer concentrations, predicted the risk of VTE. They then validated a risk-prediction model incorporating these variables, finding that “the [prediction] nomogram was able to discriminate between patients who did and did not develop VTE during six months of follow-up and was appropriately calibrated,” the authors reported.
Using this model could help clinicians select patients who would benefit from thromboprophylaxis (rather than a treatall or treat-none approach) “by reducing the risks of VTE and bleeding events caused by unnecessary thromboprophylaxis,” they continued.
To identify variables prognostic of VTE, the researchers first analyzed data from the prospective Vienna Cancer and Thrombosis Study (CATS), which included 1,423 patients with solid tumors or lymphoma. Then, the researchers validated the prognostic importance of these variables with data from 832 patients enrolled in the prospective Multinational Cohort Study to Identify Cancer Patients at High Risk of VTE (MICA).
The analysis excluded patients with multiple myeloma or high-grade glioma, but all patients had been treated as outpatients, “because about 75 percent of all cases of cancer-associated VTE occur within this population,” the authors added.
The primary outcomes for each cohort were as follows:
- CATS: symptomatic, independently assessed VTE (a composite of distal or proximal deep vein thrombosis [DVT] of the leg, upper-limb DVT, symptomatic splanchnic DVT, or pulmonary embolism [PE])
- MICA: a composite of symptomatic or incidental PE, distal or proximal DVT, non–catheter-related upper-limb DVT, or symptomatic catheter-related upper-limb DVT
The authors also categorized tumor sites as either low or intermediate risk for VTE, high risk for VTE, or very high risk for VTE.
In the CATS and MICA cohorts, 80 patients (6%) and 48 patients (6%) developed VTE, respectively, during a median follow-up of 180 days (range = 109-180 days). This translated to a cumulative risk of VTE at six months (primary endpoint) of 5.7 percent in CATS and 6.3 percent in MICA.
Eleven clinical prognostic factors and biomarkers emerged in the univariable model of cause-specific VTE hazards, but, on multivariable analysis, only two variables were significantly associated with an increased VTE risk:
- tumor site: Compared with patients with tumors located in a very–high-risk location, patients with tumors in a highrisk site were nearly twice as likely to develop VTE (hazard ratio [HR] = 1.96; 95% CI 1.41-2.72; p=0.0001). The same association was seen for patients with tumors located in a high-risk versus low- or intermediate-risk site.
- D-dimer concentrations: The risk for VTE increased by 32% as D-dimer concentration levels doubled (HR=1.32; 95% CI 1.12-1.56; p=0.001).
When the investigators created a two-variable risk model incorporating tumor site and D-dimer concentration levels, they found that the model predicted a six-month VTE risk of 5.7 percent in the CATS cohort.
Next, the researchers performed cross-validation analysis to calculate how the predicted incidence of objectively confirmed VTE at six months compared with the cumulative 6-month incidences observed in both cohorts (referred to as c-indices).
The cross-validated c-index of the risk-prediction model was 0.66 in CATS and 0.68 in MICA. The model also performed better than existing models; it correctly reclassified up to 31 percent of patients in CATS who were initially classified with the five-variable Khorana score, which accounts for tumor site, body mass index, and platelet, hemoglobin, and leukocyte counts.
The researchers noted that patients from the two cohorts were recruited from only academic centers, reducing the generalizability of the findings across the broader cancer population. In addition, considering that the primary outcome in the MICA cohort was a composite of DVT and PE, the authors questioned the validity of the model for predicting the two individual outcomes independently.
“Our clinical prediction model could outperform previous clinical prediction scores in predicting those patients at high risk of developing VTE,” the authors concluded. “Our simple clinical prediction model considerably improved prediction of cancer-associated VTE and could aid physicians in selection of those ambulatory patients with solid tumors who will most benefit from pharmacological thromboprophylaxis.”
The new prediction model is available as a paper-based nomogram, as well as an online risk calculator, the authors reported.
The authors reported no conflicts of interest.
Pabinger I, van Es N, Heinze G, et al. A clinical prediction model for cancer-associated venous thromboembolism: a development and validation study in two independent prospective cohorts. Lancet Haematol. 2018;5:e289-98.