神经内镜下清除脑内血肿术后再出血风险列线图模型的构建与验证

目的构建神经内镜下清除脑内血肿术后再出血风险列线图模型并进行验证。方法回顾性分析2019年1月—2023年7月在宿迁市分金亭医院治疗的162例原发性脑出血患者的临床资料。所有患者均采用神经内镜下清除脑内血肿，根据术后是否再出血分为术后再出血组（n=32）和术后未再出血组（n=130）。结果术后再出血组年龄大于术后未再出血组，高血压史例数少于术后未再出血组，发病至手术时间大于术后未再出血组，入院时格拉斯哥昏迷指数评分（GCS）、入院时血肿量、岛征阳性例数、术后颅内压均高于术后未再出血组，血肿清除率、仔细处理血肿周围的出血血管例数低于术后未再出血组，差异具有显著性（P＜0.05）；Logistic回归分析结果显示，高血压史、发病至手术时间、入院时GCS评分、入院时血肿量、岛征阳性、仔细处理血肿周围的出血血管、术后颅内压、血肿清除率是神经内镜下清除脑内血肿术后再出血的影响因素（P＜0.05）；列线图模型显示，发病至手术时间每增加3 h，相应得分增加10分；入院时GCS评分每增加5分，相应得分增加10分；入院时血肿量每增加10 mL，相应得分增加5分；血肿清除率每减少10%，相应得分增加8分，术后颅内压每增加4 mmHg，相应得分增加8分，仔细处理血肿周围的出血血管否（0）相应得分为0，是（1）相应得分为10分；岛征阳性否（0）相应得分为0，是（1）相应得分为8分；受试者工作特征（ROC）曲线分析结果显示，该列线图模型预测神经内镜下清除脑内血肿术后再出血的曲线下面积（AUC）为0.992（95%置信区间：0.983~1.000）。Hosmer-Lemeshow拟合优度检验结果显示，该列线图模型预测神经内镜下清除脑内血肿术后再出血与实际发生率比较，无统计学差异（χ2=8.147，P=0.879）。结论高血压史、发病至手术时间、入院时GCS评分、入院时血肿量、岛征阳性、仔细处理血肿周围的出血血管、术后颅内压是神经内镜下清除脑内血肿术后再出血的危险因素，基于上述危险因素构建的风险预测列线图模型，对神经内镜下清除脑内血肿术后再出血具有一定预测价值，可为临床制定有效预防措施提供依据。

Objective To construct a risk line chart model for postoperative rebleeding after endoscopic removal of intracerebral hematoma and validate it. Method The clinical data of 162 patients with primary cerebral hemorrhage admitted to Fenjinting Hospital of Suqian from January 2019 to July 2023 were analyzed retrospectively. All patients underwent endoscopic removal of intracerebral hematoma and were divided into postoperative rebleeding group(n=32) and postoperative non rebleeding group(n=130) based on whether postoperative rebleeding occurred. Results The age of the postoperative re bleeding group was higher than that of the postoperative non rebleeding group, the number of hypertensive cases was lower than that of the postoperative non rebleeding group, and the time from onset to surgery was longer than that of the postoperative non rebleeding group. The Glasgow coma scale(GCS) at admission, the amount of hematoma at admission, the number of positive island sign cases, and postoperative intracranial pressure were all higher than those of the postoperative non rebleeding group. The hematoma clearance rate and the number of bleeding vessels around the hematoma were significantly lower than those of the postoperative non rebleeding group(P<0.05). The results of logistic regression analysis showed that a history of hypertension, time from onset to surgery, GCS score at admission, hematoma volume at admission, positive island sign, careful handling of bleeding vessels around the hematoma, postoperative intracranial pressure, and hematoma clearance rate were the influencing factors of postoperative rebleeding after neuroendoscopic removal of intracerebral hematoma(P<0.05). The column chart model showed that for every 3 hours increased in the time from onset to surgery, the corresponding score increased by 10 points. For every 5 points increased in GCS score upon admission, the corresponding score increased by 10 points. For every 10 mL increased in hematoma volume upon admission, the corresponding score increased by 5 points. For every 10% decreased in hematoma clearance rate, the corresponding score increased by 8 points. For every 4 mmHg increased in postoperative intracranial pressure, the corresponding score increased by 8 points. Careful handling of bleeding vessels around the hematoma results in a corresponding score of 0(no), and 10 points(yes). Island sign negative(0) corresponded to a score of 0, positive (1)corresponded to a score of 8. The receiver operating characteristic(ROC) curve analysis results showed that the area under curve(AUC) of the column chart model predicting postoperative rebleeding after endoscopic removal of intracerebral hematoma was 0.992(95% confidence interval: 0.983-1.00). The results of the Hosmer Lemeshow goodness of fit test showed that there was no statistically significant difference between the prediction of postoperative rebleeding under neuroendoscopic clearance of intracerebral hematoma using the column chart model and the actual incidence rate(χ2=8.147,P=0.879). Conclusions A history of hypertension, time from onset to surgery, GCS score upon admission, hematoma volume upon admission, positive island sign, careful management of bleeding vessels around the hematoma, and postoperative intracranial pressure are risk factors for postoperative rebleeding after neuroendoscopic removal of intracerebral hematoma. A risk prediction column chart model constructed based on these risk factors has certain predictive value for postoperative rebleeding after neuroendoscopic removal of intracerebral hematoma, this can provide a basis for the development of effective preventive measures in clinical practice.