急性脑梗死患者溶栓后预测预后模型的构建

Construction of a model for predicting the prognosis of patients with acute cerebral infarction after thrombolysis

目的:建立急性脑梗死患者溶栓后预测预后的模型。方法:回顾性分析115例纳入研究范围的溶栓患者,其中预后良好组77例,预后不良组38例。收集患者信息:年龄、吸烟史、BMI、入院至开始溶栓时间(DNT)、糖化血红蛋白、相对表观扩散系数(rADC)、溶栓后24h的美国国立卫生研究院卒中量表(NIHSS)评分、发病3个月后Barthel指数评分、居住地及文化水平等。应用逻辑回归并根据赤池信息准则(AIC),找出进入模型的变量,并获得预测模型方程。采用受试者工作特征曲线(ROC)及曲线下面积(AUC)来评价模型的预测效能。结果:预后良好组77例(66.95%);预后不良组38例(33.05%)。rADC,年龄,DNT进入预测模型,并获得预测模型方程=-6.942×rADC+1.576×年龄+1.181×DNT。模型ROC曲线下面积为0.897(95%CI:0.826~0.946),敏感度及特异度分别为92.11%和76.62%,约登指数为0.6873。Hosmer and Lemeshow拟合优度检验P=0.187。结论:本研究所得预测模型可以科学、有效地预测急性脑梗死患者溶栓治疗的预后,且指标简单易得,方便指导临床工作。

Objective:To establish a model for predicting the prognosis of patients with acute cerebral infarction after thrombolysis.Methods:One hundred and fifteen patients with intravenous thrombolysis were retrospectively analyzed,including 77patients in the good prognosis group and 38patients in the bad prognosis group.Patient’s information was collected:age,smoking history,body mass index(BMI),door to needle time(DNT),relative apparent diffusion coefficient(rADC),glycosylated hemoglobin,The National Institutes of Health Stroke Scale(NIHSS)score of 24hours after thrombolysis,Barthel index three months after the onset of the disease,residential address,educational level.By means of Logistic regression and Akaike Information Criterion(AIC),the variables entering the model were found out.The predictive model equation was obtained.The predictive effectiveness of the model was evaluated by the Receiver Operating Characteristic curve(ROC)and area under the curve(AUC).Results:rADC,age,and DNT entered the prediction model,and the prediction model equation was obtained=-6.942×rADC+1.576×age+1.181×DNT.The area under the ROC curve of the model is 0.897,(95%CI:0.826-0.946),the sensitivity and specificity are 92.11%and 76.62%,respectively,and the Youden index is 0.6873.Hosmer and Lemeshow goodness-of-fit test P=0.187.Conclusion:The prediction model obtained in this study can scientifically and effectively predict the prognosis of patients with acute cerebral infarction after thrombolysis,and all the indicators are simple and easy to be obtained,which can guide the next clinical treatment.