进展性缺血性脑卒中危险因素筛选及其预测评分系统的构建

发布时间：2018-08-27 08:01

【摘要】：目的筛选进展性缺血性脑卒中(Progressive ischemic stroke,PIS)的相关危险因素,基于这些危险因素构建PIS的预测评分系统,以便临床应用,为其预防和治疗提供依据。方法收集2015年12月到2016年12月于唐山工人医院神经内科住院的急性缺血性脑卒中患者186例。在患者入院时和病情变化时采用美国国立研究院卒中量表(National Institutes of Health Stroke Scale,NIHSS)进行神经功能评分,根据发病6小时至发病7天内NIHSS评分是否增加2分或2分以上,分为进展性缺血性脑卒中(PIS)组和非进展缺血性脑卒中(Non progressive ischemic stroke,NPIS)组,其中PIS组86例,NPIS组100例。收集可能影响患者病情的相关指标,包括性别、年龄、既往病史、吸烟、饮酒史、发热、感染、肥胖、体重指数、血白细胞计数、空腹血糖、纤维蛋白原、血脂、同型半胱氨酸、超敏C反应蛋白、尿酸、入院时血压及血压变化、贫血、颈动脉斑块、责任血管狭窄程度、梗死部位、入院NIHSS评分、病情变化时NIHSS评分等。对上述资料进行单因素分析及多因素Logistic回归分析,筛选PIS的危险因素,建立Logistic回归方程;基于各因素的b系数,采用“计分法”构建PIS的预测评分系统。通过绘制ROC曲线及配对卡方检验和一致性检验等对PIS预测评分系统的辨别力和准确度进行评估。结果1单因素分析:饮酒史、发热、感染、体重指数、血白细胞计数、甘油三酯、超敏C反应蛋白、尿酸、入院收缩压、脑梗死后血压下降、责任血管中-重度狭窄、内囊后肢梗死、分水岭梗死在PIS组和NPIS组比较,差异有统计学意义(P0.05);性别、年龄、高血压病史、冠心病史、糖尿病史、房颤病史、脑卒中/TIA病史、吸烟史、肥胖、空腹血糖、纤维蛋白原、总胆固醇、低密度脂蛋白胆固醇、同型半胱氨酸、入院舒张压、贫血、颈动脉斑块、颈动脉不稳定斑块、OCSP各型、桥脑梗死、入院时NIHSS评分在两组间比较,差异无统计学意义(P0.05)。2多因素Logistic回归分析:体重指数、饮酒史、梗死后血压下降、责任血管中-重度狭窄、内囊后肢梗死、分水岭梗死、发热,7个因素纳入Logistic回归模型,是PIS的危险因素,各因素的优势(Odds ratio,OR)比分别为1.300、4.027、15.852、4.702、4.322、4.185、11.999。3构建预测PIS的Logistic回归模型:Logit P=-10.035+0.262X1+1.393X2+2.763X3+1.548X4+1.464X5+1.432X6+2.458X7(注:X1表示BMI;X2表示饮酒史;X3表示梗死后血压下降;X4表示责任血管中-重度狭窄;X5表示内囊后肢梗死;X6表示分水岭区梗死;X7表示发热)。4简化Logistic回归模型,构建了总分为10分、预测界值为4分的PIS预测评分系统,该评分系统ROC曲线下面积为0.911,评价效果如下:灵敏度0.860,特异度0.920,总符合率0.892,阳性似然比10.750,阴性似然比0.152,Kappa值0.783。结论1体重指数、饮酒史、脑梗死后血压下降、责任血管中-重度狭窄、内囊后肢梗死、分水岭梗死、发热是PIS的重要危险因素。2构建的PIS预测评分系统,评估效果较好,应用方便简捷,具有一定的临床实用性。
[Abstract]:Objective to screen the risk factors associated with progressive ischemic stroke (Progressive ischemic stroke,PIS) and to construct a predictive scoring system for PIS based on these risk factors so as to provide evidence for its prevention and treatment. Methods 186 patients with acute ischemic stroke were collected from December 2015 to December 2016 in Department of Neurology, Tangshan Workers Hospital. The neurological function was assessed by the National Institutes of America Stroke scale (National Institutes of Health Stroke Scale,NIHSS) on admission and at the time of disease change. According to whether the NIHSS score increased by 2 points or more within 6 hours to 7 days after the onset of the disease, the neurological function of the patients was assessed by the National Institute of Stroke scale (National Institutes of Health Stroke Scale,NIHSS). The patients were divided into progressive ischemic stroke (PIS) group and non progressive ischemic stroke (Non progressive ischemic stroke,NPIS) group. There were 86 cases in PIS group and 100 cases in PIS group. Collect relevant indicators that may affect the patient's condition, including sex, age, past medical history, smoking, alcohol consumption, fever, infection, obesity, body mass index, white blood cell count, fasting blood glucose, fibrinogen, blood lipids, Homocysteine, hypersensitive C-reactive protein, uric acid, blood pressure and blood pressure changes at admission, anemia, carotid plaque, degree of responsible vascular stenosis, infarct location, admission NIHSS score, NIHSS score at the time of disease change, etc. Univariate analysis and multivariate Logistic regression analysis were carried out to screen the risk factors of PIS and establish the Logistic regression equation. Based on the b coefficients of each factor, the prediction scoring system of PIS was constructed by "scoring method". The discriminative power and accuracy of PIS prediction scoring system are evaluated by drawing ROC curve, paired chi-square test and consistency test. Results 1 single factor analysis: alcohol history, fever, infection, body mass index, white blood cell count, triglyceride, hypersensitive C-reactive protein, uric acid, systolic blood pressure, blood pressure after cerebral infarction, moderate to severe stenosis of responsible blood vessel. There were significant differences between PIS group and NPIS group (P0.05); gender, age, history of hypertension, coronary heart disease, diabetes mellitus, atrial fibrillation, stroke / TIA history, smoking history, obesity, fasting blood glucose, Fibrinogen, total cholesterol, low density lipoprotein cholesterol, homocysteine, diastolic blood pressure, anemia, carotid plaque, carotid artery unstable plaque, pontine infarction, NIHSS scores at admission were compared between the two groups. There was no significant difference (P0.05) in multivariate Logistic regression analysis: body mass index (BMI), alcohol consumption history, blood pressure after infarction, moderate to severe stenosis of responsible vessels, internal capsule hind limb infarction, watershed infarction, fever, and 7 factors were included in Logistic regression model. Is a risk factor for PIS, The odds ratio of each factor (Odds ratio,OR) was 1.300 / 4.02715.852n 4.702n 4.3224.1854.1855 / 11.999.3 to construct a Logistic regression model for predicting PIS: logit Pn-10.035 0.262X1 1.393X2 2.763X3 1.548X4 1.464X5 1.432X6 2.458X7 (note: BMI;X2 indicates drinking history; X3 means blood pressure after infarction; X4 indicates moderate to severe stenosis of the responsible vessel; X4 means the internal of the responsible vessels with moderate to severe stenoses; X4 indicates that the blood pressure after infarction is decreased; X4 indicates that the responsible vessels have moderate to severe stenosis. X6 for watershed area infarction X7 for fever) .4 simplified Logistic regression model. A PIS prediction scoring system with a total score of 10 points and a prediction threshold of 4 points was constructed. The area under the ROC curve was 0.911. The evaluation results were as follows: sensitivity 0.860, specificity 0.920, total coincidence rate 0.889 2, positive likelihood ratio 10.750, negative likelihood ratio 0.152% and Kappa value 0.783. Conclusion 1 body mass index (BMI), alcohol consumption history, blood pressure decrease after cerebral infarction, moderate to severe stenosis of responsible vessels, posterior limb infarction, watershed infarction and fever are the important risk factors of PIS. 2. The application is simple and convenient, and has certain clinical practicability.
【学位授予单位】：华北理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R743.3

【参考文献】