动脉瘤性蛛网膜下腔出血并发急性脑积水的危险因素分析及预测模型构建

Risk factor analysis and prediction model construction of aneurysmal subarachnoid hemorrhage complicated by acute hydrocephalus

目的评估动脉瘤性蛛网膜下腔出血(aneurysmal subarachnoid hemorrhage,aSAH)并发急性脑积水(acute hydrocephalus,AHC)的危险因素,并构建预测模型。方法回顾性分析2015年1月至2024年1月秦皇岛市第一医院神经外科收治的aSAH患者的临床资料。按照7∶3比例,将aSAH患者随机分为训练集和验证集,根据是否发生AHC分为AHC组和非AHC组,训练集用于构建AHC风险预测模型,验证集用于验证AHC风险预测模型,采用受试者工作特征曲线(receiver operating characteristic curve,ROC)及其曲线下面积(area under curve,AUC)、校准曲线和决策曲线验证风险预测模型的可靠性和稳定性。结果共纳入1062例aSAH患者,其中324例患者发生AHC,发生率为30.51%。训练集、验证集分别有患者744、318例。多因素Logistic回归显示年龄≥60岁[OR=3.067,95%CI(1.710,5.499)]、破入脑室[OR=7.039,95%CI(3.792,13.068)]、Fisher分级为IV级[OR=3.371,95%CI(1.335,8.514)]、Hunt-Hess分级为IV级[OR=6.198,95%CI(2.218,17.324)]和高水平神经元特异性烯醇化酶[OR=1.746,95%CI(1.581,1.928)]是aSAH患者发生AHC的独立危险因素(P<0.05),而动脉瘤位于前循环[OR=0.397,95%CI(0.199,0.790)]是aSAH患者发生AHC的的独立保护因素(P<0.05)。训练集和验证集的AUC(95%CI)分别为0.950(0.932,0.967)和0.969(0.955,0.982);校准曲线显示“预测AHC概率”和“实际AHC概率”吻合;决策曲线均提示AHC风险预测模型临床净获益高于“全部”临床净获益。结论临床工作中应重点关注年龄≥60岁、破入脑室、Fisher分级与Hunt-Hess分级为IV级和高水平神经元特异性烯醇化酶aSAH患者AHC的发生风险,本研究构建的AHC预测模型可为早期识别AHC提供便捷工具。

Objective To evaluate the risk factors for the development of acute hydrocephalus(AHC)after aneurysmal subarachnoid hemorrhage(aSAH)and to construct a prediction model.Methods The clinical data of patients with aSAH treated in the department of neurosurgery in the First Hospital of Qinhuangdao from January 2015 to January 2024 were retrospectively analyzed.The patients were randomly divided into the training set and the validation set in a 7:3 ratio and were also divided into the AHC group and non-AHC group according to whether they developed AHC or not.The training set was used to construct an AHC risk prediction model,while the validation set was used to validate the AHC risk prediction model.The reliability and stability of the risk prediction model were verified by the receiver operating characteristic curve(ROC),its area under the curve(AUC),calibration curve,and decision curve.Results A total of 1,062 patients with aSAH were included,among whom 324 patients developed AHC,with an incidence rate of 30.51%.The training set and validation set had 744 and 318 patients,respectively.Multivariate Logistic regression showed that age≥60 years[OR=3.067,95%CI(1.710,5.499)],entering the ventricles[OR=7.039,95%CI(3.792,13.068)],Fisher grade IV[OR=3.371,95%CI(1.335,8.514)],Hunt-Hess grade IV[OR=6.198,95%CI(2.218,17.324)]and high level of neuron-specific enolase[OR=1.746,95%CI(1.581,1.928)]were independent risk factors for aSAH patients developing AHC(P<0.05),while the aneurysm located at anterior circulation[OR=0.397,95%CI(0.199,0.790)]was an independent protective factor(P<0.05).The AUC(95%CI)for the training set and validation set were 0.950(0.932,0.967)and 0.969(0.955,0.982),respectively;the calibration curve showed that the predicted AHC probability and actual AHC probability were consistent;the decision curves both indicated that the AHC risk prediction model had a higher net clinical benefit than the all net clinical benefit.Conclusion Clinical attention should be focused on patients with aSAH aged 60 years or older,entering the ventricles,Fisher grade IV,Hunt-Hess grade IV,and high level of neuron-specific enolase,and the AHC prediction model constructed in this study can provide a convenient tool for early identification of AHC.