Risk factors for coagulation dysfunction in traumatic brain injury patients: A single-center retrospective cohort...

This study aimed to analyze the risk factors for coagulation dysfunction in patients with traumatic brain injury (TBI) using a logistic regression risk prediction model. A single-center retrospective cohort study was conducted on 121 patients with TBI, including 76 patients in the modeling group and 45 patients in the validation group. The occurrence of coagulation dysfunction was analyzed, and a logistic regression risk prediction model was constructed. Variable selection was performed using univariate analysis and multivariate logistic regression. Variables exhibiting P < .1 in univariate analysis were preselected for inclusion in the multivariate logistic regression model. The model was evaluated using the receiver operating characteristic curve, calibration curve, and decision curve analysis. The incidence of coagulation dysfunction did not differ significantly between the modeling and validation groups. Patients in the occurrence group showed a significantly higher rate of blood transfusion within 24 hours, diffuse axonal injury, and long-term use of anticoagulants, while arterial blood pH was significantly lower, indicating an inverse relationship with coagulation dysfunction. Logistic regression analysis identified 4 independent predictors, and the prediction model was formulated. Receiver operating characteristic curve analysis demonstrated good predictive efficacy of the model (area under the curve = 0.798, 95% confidence interval: 0.702-0.894). The calibration curve and decision curve analysis showed high consistency and clinical applicability of the risk prediction model. Coagulation dysfunction in patients with TBI is influenced by multiple risk factors, including blood transfusion within 24 hours, diffuse axonal injury, arterial blood pH value, and long-term use of anticoagulant drugs. The logistic regression risk prediction model based on these factors showed good internal performance; however, it should be considered preliminary and requires external validation before clinical application.