Development and validation of a prediction model for postoperative lower extremity deep vein thrombosis in patients with traumatic cervical spinal cord injuries

Yu Miao; Yin Hong; Jiang Tao; Zheng Wenjie; Liu Minghan

doi:10.3969/j.issn.1009-4237.2025.04.007

Abstract：Objective Deep venous thrombosis(DVT)is an early postoperative complication in patients with traumatic cervical spinal cord injuries(TCSCI).Early identification of patients at high risk of postoperative DVT is essential.This study aims to develop and validate a nomogram model to predict postoperative DVT in pa-tients with TCSCI.Methods Clinical data of 358 patients with TCSCI admitted to our hospital between Jan.2012 and Dec.2023 were retrospectively analyzed,and the patients were classified into DVT group(47 cases)and non-DVT group(311 cases)according to the DVT outcomes.Univariate analysis and multivariate logistic regression a-nalysis were performed to demonstrate the risk factors of postoperative DVT.The R language was adopted to draw the visual nomogram of the model,while the predictive performance of the model was evaluated through the receiver operating characteristic(ROC)curve,calibration curve,and decision curve analysis.Internal validation of the mod-el was achieved by using the Bootstrap.Results The incidence of postoperative DVT was 13.1％(95％CI:9.6％-16.6％).The results of univariate analysis and multivariate logistic regression analysis showed that multiple trauma(OR:2.758,95％CI:1.306-5.715),ASIA grading of A-B(OR:5.934,95％CI:2.707-12.970),operation time≥ 3 h(OR:1.931,95％CI:O.922-3.945),blood transfusion(OR:2.214,95％CI:0.974-4.853),and D-dimer levels ≥1.68 mg/L(OR:2.728,95％CI:1.134-7.695)were independent risk factors for postoperative DVT in TCSCI pa-tients.A risk prediction model for postoperative DVT of patients with TCSCI was constructed.The area under the ROC curve was 0.797(95％CI:0.705-0.874),the optimal threshold 0.067,sensitivity 0.830,and specificity 0.633,indicating good discrimination of this model.The calibration curve and Hosmer-Lemeshow goodness-of-fit test(x2=9.837,P=0.080)showed good accuracy of the model.The decision curve analysis indicated that the predic-tive model had good clinical applicability,and the internal validation of the model was effective.Conclusion The nomogram for DVT risk showed a good ability to predict postoperative DVT in patients with TCSCI,which may bene-fit the development of prevention and control strategies.