Risk factors and risk nomogram model of reoperation for hemorrhages after severe traumatic brain injury craniotomy

OBJECTIVE: This study aimed to explore the risk factors associated with reoperation for postoperative hemorrhages after severe traumatic brain injury (sTBI) craniotomy and establish a risk nomogram model. METHODS: A retrospective case‐control study was performed. Overall, 367 patients who were diagnosed with sTBI and fulfilled the inclusion criteria were enrolled from the Department of Neurosurgery of the Affiliated Hospital of Zunyi Medical University between January 2015 and December 2020. They were divided into a reoperation group and a non‐reoperation group according to whether they underwent reoperation for hemorrhages. Using univariate binary logistic regression analysis, the possible risk factors were screened. Subsequently, the independent risk factors of reoperation for postoperative hemorrhages were screened using the forward step method of multivariate binary logistic regression analysis, and a corresponding nomogram model was constructed. The receiver operative characteristic (ROC) curve was used to evaluate the reliability of the model. Finally, 30% of the data were randomly selected for internal verification of the model. RESULTS: The reoperation rate for hemorrhage after sTBI emergency craniotomy was 14.71% (54/367); multivariate logistic regression analysis showed that multiple hemorrhages (odds ratio [OR] = 4.38, 95% confidence interval [CI]: 1.815–10.587, p = 0.001), day or night surgery (OR = 0.26, 95% CI: 0.119–0.547, p < 0.001), operation duration (OR = 0.74, 95% CI: 0.119–0.547, p < 0.025), and abnormal intraoperative blood pressure fluctuation (OR = 4.15, 95% CI: 2.090–8.245, p < 0.001) were statistically significant. The sensitivity and specificity of the nomogram model were 0.815 and 0.661, respectively, and the area under ROC curve was 0.76 (95% CI: 0.705–0.833). Internal verification showed that the area under the ROC curve was 0.783 (95% CI: 0.683–0.883). CONCLUSIONS: Taken together, the results of our study reveal that multiple preoperative intracranial hemorrhages, day and night operation, operation duration, and abnormal fluctuation of intraoperative blood pressure were independent risk factors for postoperative bleeding and reoperation for sTBI. Through the analysis of the influencing factors, a prediction model for the risk of bleeding and reoperation after craniocerebral trauma was developed. Compared with other relevant studies, this prediction model has good prediction efficiency and can be used to predict the occurrence of bleeding and reoperation after sTBI in patients.