Postoperative venous thromboembolism risk-prediction in foot and ankle fracture surgery

Background

Venous thromboembolism (VTE) are rare but serious complications after foot and ankle fracture surgery. A consensus definition of a high-risk patient has not been reached, leading to significant variability in the use of pharmacologic agents for VTE prophylaxis. The aim of this study was to develop a model for predicting VTE risk in patients undergoing surgery for foot and ankle fractures that is usable and scalable in clinical practice.

Methods

A retrospective review of 15,342 patients, within the ACS-NSQIP database, who had undergone surgical repair of foot and ankle fractures from 2015 to 2019 was performed. Univariate analysis evaluated differences in demographics and comorbidities. Stepwise multivariate logistic regression was generated based on a 60 % development cohort to evaluate risk factors for VTE. A receiver operator curve based on the 40 % test cohort calculated area under the curve (AUC) to measure the accuracy of the model in predicting VTE within the 30-day postoperative period.

Results

Of the 15,342 patients, 1.2 % patients experienced VTE, and 98.8 % patients did not. Patients who experienced VTE were significantly older and had an overall higher comorbidity burden. Those who had VTE spent on average 10.5 more minutes in the operating room. In the final model, age over 65, diabetes, dyspnea, CHF, dialysis, wound infection and bleeding disorders were all found to be significant predictors of VTE after controlling for all other factors. The model generated an AUC of 0.731, indicating good predictive accuracy. The predictive model is publicly available at https://shinyapps.io/VTE_Prediction/.

Conclusions

In alignment with previous studies, we identified increased age and bleeding disorders as independent risk factors for VTE after foot and ankle fracture surgery. This is one of the first studies to generate and test a model for identifying patients at risk for VTE in this population. This evidence-based model may help surgeons prospectively identify high-risk patients who may benefit from pharmacologic VTE prophylaxis.