A Thoracic Surgery–Specific Risk Assessment Model to Prevent Perioperative Venous Thromboembolism

The risk of venous thromboembolism (VTE) in patients undergoing resection in the setting of lung cancer represents a major challenge to improving postoperative outcomes. The Caprini risk assessment model (RAM) has been validated in general surgery to determine a role for extended chemoprophylaxis to reduce VTE events. Our goal was to simplify this burdensome model for the unique needs of this thoracic surgical population to better guide anticoagulation decision-making. Patients who underwent a resection for lung malignancy at our institution between June 2005 and June 2013 with available 60-day postoperative follow-up data were evaluated. Exclusion criteria were chronic anticoagulation, inferior vena cava filter, missing data, and loss to follow-up. 12 selected risk factors were analyzed for each patient by logistic regression with stepwise inclusion to model 60-day VTE incidence. 225 patients were eligible for inclusion and a VTE incidence of 5.8% was observed. We generated a five-variable model with similar predictive ability for VTE occurrence as the Caprini RAM ( p = 0.29). Weighting of sex, age, history of VTE, surgical approach, and duration of procedure provides a low-risk or high-risk composite score with 56% sensitivity and 77% specificity. In this first effort to model VTE incidence on the basis of a limited set of clinical risk factors, we demonstrated efficacy of retrospectively scoring patients with just five data points in anticipating risk of postoperative VTE. These high-risk surgical patients can be readily identified in the preoperative period to benefit from extended postoperative prophylaxis.