Abstract:Objective To screen independent predictors of deep vein thrombosis（DVT）after traumatic fracture surgery based on LASSO-logistic regression analysis，and to construct a risk prediction model. Methods The study subjects were selected retrospectively from 302 patients who underwent surgical treatment for traumatic fractures in the Affiliated Hospital of Yangzhou University from January 2021 to January 2023. They were divided into two groups based on whether they developed DVT after surgery，namely the DVT group（45 cases）and the non-DVT group（257 cases）.Clinical data from both groups were collected，and independent predictors of postoperative DVT were screened using univariate analysis and LASSO-logistic regression analysis. A risk prediction model was constructed，and the predictive value was validated using receiver operating characteristic curve（ROC）curve，calibration curve，and decision curve analysis（DCA）. In addition，103 patients with traumatic fractures from February 2023 to February 2024 were selected as external validation samples. Results There were significant differences in age，BMI，time from injury to operation，Caprini score，multiple fractures，coexisting coronary heart disease，perioperative blood transfusion，American Society of Anesthesiologists（ASA）classification，operation time，fibrinogen（FIB）and D-dimer（D-D）on postoperative day 1 （P<0.05）. Lasso-logistic regression analysis showed that age，BMI，time from injury to operation，Caprini score，multiple fractures，coexisting coronary heart disease，operation time，FIB and D-D on postoperative day 1 were independent risk factors for postoperative DVT of traumatic fracture（P < 0.05）. The area under ROC curve（AUC）of the postoperative DVT risk prediction model for traumatic fractures constructed based on the above predictors was 0.858（95%CI：0.795-0.920），with a sensitivity of 82.22% and a specificity of 73.93%.The calibration plot showed a calibration of 0.887 and a C-index of 0.851. The DCA curve indicated that the model had a positive net clinical benefit. External validation ROC curve showed an AUC of 0.844（95%CI：0.727-0.961），with sensitivity of 78.57% and specificity of 73.03%. Conclusion Age，BMI，time from injury to operation，Caprini score，multiple fractures，coexisting coronary heart disease，time of surgery，and FIB and D-D on postoperative day 1 are independent predictors of DVT after traumatic fracture. The risk prediction model built based on these predictors has high sensitivity and clinical efficacy，can provide guidance for prevention and treatment of DVT after traumatic fracture.