基于实验室指标的颈动脉硬化模型的建立与评价

发布时间：2018-02-15 11:37

本文关键词： 颈动脉硬化实验室指标模型评分系统　出处：《重庆医科大学》2011年硕士论文　论文类型：学位论文

【摘要】：目的:探索联合应用多个实验室指标在颈动脉硬化预测中的价值,建立并评价颈动脉硬化的预测模型及评分系统。方法:247例健康体检人群被纳入研究对象,按多普勒超声测量的颈动脉内膜厚度将其分为颈动脉硬化组(62例)和非颈动脉硬化组(185例)。问卷调查收集相关资料(包括年龄、性别、身体质量指数、腰围、脉压、糖尿病、高血压、吸烟),采集静脉血进行相关生化指标检测(包括肝功能、肾功能、血脂和血糖等16个项目)、血常规分析和血液流变学分析。用SPSS 17.0软件统计分析所有数据,并联合利用组间比较及单因素logistic回归对所有计量资料指标进行筛选,再用筛选的指标作多因素逐步logistic回归建立颈动脉硬化实验室指标的预测模型,在此基础上探索建立颈动脉硬化的预测评分系统,并应用受试者工作特征曲线(ROC)评估预测模型和评分系统对颈动脉硬化的预测效率。结果:组间比较发现年龄、脉压、天门冬氨酸氨基转移酶、γ-谷氨酰转移酶、肌酐、血糖、甘油三酯、高敏C反应蛋白、同型半胱氨酸、单核细胞计数10项计量资料指标有统计学差异(P0.05),且均表现为颈动脉硬化组高于非颈动脉硬化组。对上述10项指标分别作单因素logistic回归,选出P0.1的指标7个(包括年龄、脉压、γ-谷氨酰转移酶、肌酐、血糖、高敏C反应蛋白、单核细胞计数)进行多因素逐步logistic回归,建立如下数学模型:Log(PI)=0.14×Age+0.009×γ-GT+ 0.152×Glu+2.829×Mono-10.959。该模型的ROC曲线下面积为0.814(P0.001,95%CI:0.756-0.872),当以PI=0.157为界点时,灵敏度为0.903,特异度为0.573。结合临床资料,选出年龄、性别、吸烟、糖尿病、脉压、γ-谷氨酰转移酶、肌酐、血糖、高敏C反应蛋白、单核细胞计数10项指标建立颈动脉硬化的预测评分系统,总分为10分,其ROC曲线下面积为0.802(P0.001,95%CI:0.720-0.884),以6分为界点时,灵敏度为0.632,特异度为0.869。结论:年龄、γ-谷氨酰转移酶、血糖、单核细胞计数是颈动脉硬化的独立危险因素;本研究成功地建立了基于实验室指标的颈动脉硬化预测模型和评分系统,有较高的预测价值,值得在扩大样本的基础上进一步研究。
[Abstract]:Objective: to explore the value of combined use of several laboratory indexes in the prediction of carotid atherosclerosis, and to establish and evaluate the prediction model and scoring system of carotid atherosclerosis. Methods: a total of 247 healthy people were included in the study. They were divided into two groups according to carotid intima thickness measured by Doppler ultrasound: 62 cases in carotid atherosclerosis group and 185 cases in non-carotid atherosclerosis group. Data were collected by questionnaire (including age). Sex, body mass index, waist circumference, pulse pressure, diabetes, hypertension, smoking, collecting venous blood for biochemical tests (including liver function, renal function, etc.). Blood routine analysis and hemorheology analysis. All the data were statistically analyzed by SPSS 17.0 software, and combined with inter-group comparison and univariate logistic regression to screen all measurement data. Using the selected indexes as multi-factor stepwise logistic regression, the prediction model of carotid atherosclerosis laboratory indexes was established, and the prediction scoring system of carotid atherosclerosis was established on the basis of this model. The predictive model and scoring system were used to evaluate the predictive efficiency of carotid atherosclerosis. Results: age, pulse pressure, aspartate aminotransferase, 纬 -glutamyl transferase, creatinine, blood glucose, triglyceride, Gao Min C-reactive protein, homocysteine were found. There was a statistical difference in the count of monocytes in 10 items of measurement data (P 0.05), and all of them showed that the carotid atherosclerosis group was higher than that of the non-carotid atherosclerosis group. The 10 indexes mentioned above were analyzed by single factor logistic regression, and 7 indexes (including age) of P0. 1 were selected. Pulse pressure, 纬 -glutamyl transferase, creatinine, blood glucose, Gao Min C-reactive protein, monocyte count) were analyzed by stepwise logistic regression. The following mathematical model was established as follows: Age 0.14 脳 Age 0.009 脳纬 -GT 0.152 脳 Glu 2.829 脳 Mono-10.959.The area under the ROC curve of the model was 0.814P0.001 / 95CIW 0.756-0.872n. The sensitivity was 0.903 and the specificity was 0.573.According to the clinical data, age, sex, smoking, diabetes mellitus, pulse pressure, 纬 -glutamyl transferase, creatinine, blood glucose, Gao Min C-reactive protein were selected. A system for predicting carotid atherosclerosis was established with 10 indexes of monocyte count. The total score was 10. The area under the ROC curve was 0.802P0.001C95CIU: 0.720-0.8840.The sensitivity was 0.632and the specificity was 0.869when divided into 6 points. Conclusion: age, 纬 -glutamyltransferase, blood glucose and monocyte count are independent risk factors for carotid atherosclerosis. It has high predictive value and is worthy of further study on the basis of expanding the sample.
【学位授予单位】：重庆医科大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：R-332

【相似文献】