关键词: 慢性硬膜下血肿； 手术； 减压性出血； 风险； 预测模型

Abstract: Objective To investigate the risk factors of decompression bleeding of CSDH patients after minimally invasive drainage and to construct prediction model to provide more reference for developing more targeted prevention and treatment strategies and improving prognosis.Methods 165 CSDH patients after minimally invasive drainage were retrospectively chosen in the period from January 2019 to January 2023 in our hospital and grouped according to whether decompression bleeding occurred after surgery or not into bleeding group (33 cases) and non-bleeding group (132 cases).The univariate and multivariate method were used to analyze the risk factors of decompression bleeding of CSDH patients after minimally invasive drainage.Construction of risk prediction model for decompression bleeding after CSDH minimally invasive drainage.Results The results of single factor analysis showed that both intracavitary pressure and midline recovery rate of hematoma may be related to decompression bleeding after CSDH minimally invasive drainage (P<0.05).The results of multifactor analysis (bleeding group =1,non-bleeding group =0) showed that intracavitary pressure and midline recovery speed were independent factors influencing the risk of decompression bleeding after CSDH minimally invasive drainage (P<0.05).Intracranial pressure,postoperative midline recovery speed and Logistic regression model P value were used to predict decompression bleeding after CSDH minimally invasive drainage.The area under ROC curve was 0.784,0.778 and 0.840,respectively,and the approximate den index was 50.76%,44.70% and 56.82%.Conclusion The incidence of decompression bleeding after CSDH minimally invasive drainage was related to the intracavitary pressure of hematoma and the recovery rate of postoperative midline and the prediction model based on the above factors can guide the identification of high risk groups of early decompression bleeding.

Key words: chronic subdural hematoma; surgery; decompression bleeding; risk; prediction model