摘要
目的 探讨大面积脑梗死(MCI)患者脑疝发生的危险因素,构建并验证风险模型的预测效果.方法 选取2020年2月至2021年8月内蒙古呼和浩特市第一医院收治的145例MCI患者的临床资料进行回顾性研究.根据患者是否发生脑疝分成脑疝组(n=44)、非脑疝组(n=101),通过比较两组临床资料分析脑疝发生的单因素,利用多因素Logistic回归分析MCI后脑疝发生的独立危险因素.绘制受试者工作特征(ROC)曲线验证该模型评估脑疝发生风险的价值,确定曲线下面积(AUC),利用拟合优度检验分析风险预测模型的评估效能.结果 单因素分析提示,脑疝组与非脑疝组的性别构成比、冠心病史、糖尿病史、高血压史、饮酒史、发病至就诊时间、舒张压、呕吐、头痛、失语、双眼凝视、意识障碍、出血性梗死、去骨瓣减压术、取栓治疗、溶栓治疗情况比较,差异均无统计学意义(P＞0.05);脑疝组年龄小于非脑疝组,吸烟史、脑实质出血占比及收缩压、梗死体积均高于非脑疝组,差异均有统计学意义(P＜0.05).多因素Logistic回归分析提示,年龄、吸烟史、脑实质出血及收缩压过高是MCI患者的脑疝发生危险因素,而年龄越大,脑疝发生风险反而降低(P＜0.05).ROC曲线验证,本次预测概率模型预测脑疝发生风险的AUC为0.874,敏感度88.64％,特异度81.19％.Hosmer-Lemeshow拟合优度检验(x2=4.429,自由度df=8,P=0.817).结论 MCI患者脑疝发生与年龄、吸烟史、收缩压过高、脑实质出血密切相关,该预测概率模型对脑疝风险的预测效能较好,可用于MCI后脑疝风险评估.
Abstract
Objective To investigate the risk factors of cerebral hernia in patients with massive cerebral infarction(MCI),and establish and verify the prediction effect of the risk model.Methods The clinical data of 145 patients with MCI admitted to The First Hospital of Hohhot from February 2020 to Agust 2021 were retrospectively studied.Patients were divided into the brain hernia group(n=44)and the non-brain hernia group(n=101)according to whether they had brain hernia.The single factor of brain hernia occurrence was analyzed by comparing the clinical data of the two groups,and the independent risk factors of brain hernia occurrence after MCI were analyzed by multi-factor Logistic re-gression.The receiver operating characteristic(ROC)curve was plotted to validate the value of the model in assessing the risk of brain hernia,the area under the curve(AUC)was determined,and the evaluation efficacy of the risk prediction model was analyzed by goodness of fit test.Results Univariate analysis showed that there were no significant differences in gender composition ratio,history of coronary heart disease,history of dia-betes,history of hypertension,history of drinking,time from onset to treatment,diastolic blood pressure,vomiting,headache,aphasia,binocular gaze,disturbance of consciousness,hemorrhagic infarction,decompressive craniectomy,thrombectomy and thrombolytic therapy between the brain hernia group and the non-brain hernia group(P＞0.05),the age of the brain hernia group was smaller than that of the non-brain hernia group,and the smoking history,proportion of cerebral parenchymal hemorrhage,systolic pressure and infarct volume were higher than those of the non-brain hernia group(P＜0.05).Multivariate Logistic regression analysis showed that ages,smoking history,parenchymal hemorrhage and high systolic pressure were risk factors for brain hernia in MCI patients,and the risk of brain hernia decreased with increasing age(P＜0.05).ROC curve verified that the AUC of this prediction probability model to predict the risk of brain hernia was 0.874,the sensitivity was 88.64％,and the specificity was 81.19％.Hosmer-Lemeshow goodness of fit test(x 2=4.429,DOF df=8,P=0.817).Conclusion The incidence of brain hernia in patients with MCI is closely related to age,smoking history,high systolic pressure,and parenchymal hemorrhage.The prediction probability model has a good predictive effect on the risk of brain hernia,and can be used to assess the risk of brain hernia after MCI.
维普
万方数据