A risk prediction model for evaluating thrombosis extension of muscle calf venous thrombosis after craniotomy

OBJECTIVE: To explore the risk factors of muscle calf venous thrombosis (MCVT) after craniotomy and construct a risk prediction model, so as to provide tool for evaluating the prognosis of MCVT after craniotomy. METHODS: Retrospective analysis was performed on the data of patients undergoing craniotomy complicated with MCVT from January 1, 2018 to December 31, 2020. A prediction model was established by Logistic regression, and the predictive efficacy of the model was tested by ROC curve. The accuracy of the risk model was evaluated by Hosmer-Lemeshow (H-L) test, and the model was verified internally by cross validation. RESULTS: Among the 446 patients who underwent craniotomy complicated with MCVT, 112 cases (25.11%) had thrombosis extension. D-dimer, Capirini scores, length of hospital stay, malignant tumor, fracture, use of dehydrating agents and hemostatic agents were independently related to thrombosis extension after craniotomy. The area under ROC curve (AUROC) of the prediction model was 0.918 (0.888, 0.942), and the sensitivity and specificity of the maximum Youden index were 85.3% and 78.2%, respectively. H-L test showed that the prediction model was accurate (χ(2) = 12.426, P = 0.133). The internal verification results of the prediction model showed that the AUROC value of the prediction model is 0.892. CONCLUSION: The prediction model has a good prediction efficacy on the prognosis of post-craniotomy patients complicated with MCVT, and can be used as a tool to evaluate the risk of thrombosis extension.