Risk factors for cancer-related venous thromboembolism in ambulatory patients.

Background: The awareness of venous thromboembolism (VTE) burden in ambulatory cancer patients is growing. Due to the heterogeneity in the individual risk profile, predictive risk scores have been proven to be useful in the identification of high-risk patients for targeting thromboprophylaxis. The first validated risk score, the Khorana score, identified four factors resulting in an increased risk for cancer patients to develop VTE. In our study, we tried to identify additional factors associated with VTE to improve the prediction of high-risk patients. Methods: We performed a prospective, observational study in 544 ambulatory patients (56 ± 12 SD y, BMI 27 ± 5 kg/m?, 382 female [70 %]) with solid tumors and ongoing antineoplastic treatments. Tumor sites were breast n=216 (40 %), colo-rectal n=115 (21 %), gyn/uro n=76 (14 %), pancreas n=45 (8 %), gastric n=22 (4 %), lung n=15 (3 %) and others tumors n=55 (10 %). In 264 (52 %) patients there was metastatic disease. We screened for VTE with ultrasound of the upper and lower extremities and correlated cancer-, treatment- and patient-related conditions with the incidence of VTE. Results: We found a VTE incidence of 10 % (n=54), over a median follow up of 10.1 months. According to Khorana score, 51 % of patients (n=247) belonged to the low-risk (score=0), 44 % (n=213) to intermediate-risk (score=1-2) and 6 % (n=27) to high-risk (score?3) class. Odds Ratio for intermediate risk was 1.18 (95% CI 0.65-2.15, p=.588) and 2.66 (95% CI 1.14-6.17, p=.030) for high-risk patients. In our analysis we identified previous VTE (p<.001), metastatic disease (p<.001), vascular/lymphatic compression by the tumor (p=.002), edema of extremities (p=.005), surgery in last 6 months (p=.003) and presence of central venous catheters (p=.013) to be additional significant risk factors for VTE. Conclusions: Khorana score showed a good predictive value for VTE in ambulatory cancer patients. We found other clinical conditions to be associated with VTE events. These data could implement the existing predictive scores with new clinical risk factors, and refine the decision making in the oncologic clinical practice.