肾动脉硬化性狭窄的预测模型研究

Study of Predictive Model for Atherosclerotic Renal Artery Stenosis in Patients with Coronary Artery Disease

目的:探讨动脉硬化性肾动脉狭窄(ARAS)的独立危险因素及预测因子,建立在冠心病(CAD)患者中筛查ARAS的预测模型。方法:选择经肾动脉造影确诊的ARAS患者232例为ARAS组,263例无ARAS者为对照组,记录动脉硬化危险因素,测定血糖、血脂、肌酐等指标,计算内生肌酐清除率(Ccr)。采用Logistic回归分析ARAS的独立预测因子,构建筛查ARAS的预测模型,以接收者操作特征曲线(ROC)分析模型的预测价值。结果:老龄(Age,≥65岁)、糖尿病(DM)、难治性高血压(RH)、Ccr降低(≤90ml/min)是ARAS的独立预测因子,Logistic回归方程为P/(1-P)=EXP(-1.87+1.17Age+1.87RH+0.58DM+0.70Ccr),预测ARAS的最佳概率值是0.41(敏感度78.4%,特异度71.1%),ROC曲线下面积为0.80。结论:年龄≥65岁、DM、RH、Ccr降低是CAD患者并发ARAS的独立预测因子,由这些独立预测因子拟合的预测模型可用于在CAD患者中筛查ARAS。

Objective:To investigate the predictors of atherosterotic renal artery stenosis(ARAS) in patients with coronary artery disease(CAD).Method:Serum creatinine was measured in 495 patients with CAD,creatinine clearance(Ccr) was estimated by Cockroft-Gault formula.Clinical characteristics and risk factors for CAD were recorded for each patient.All patients were divided into 2 groups based upon angiographic findings.ARAS group comprised 232 patients with renal artery stenosis,and 263 patients with CAD only served as control group.Logistic regression analysis was used to determine the power of each variable for predicting ARAS in patients with CAD.The value of predictive model was calculated by constructing ROC curve.Results:Older age(≥65 year),DM,resistant hypertension(RH) and decreased Ccr(Ccr≤90 ml/min) were independent predictors for ARAS in patients with CAD.The logistic regression model for predicting ARAS in CAD patients was defined as:P/(1-P) =EXP(-1.87+1.17Age+1.87RH+0.58DM+0.70Ccr).The optimal probability value of regression model to predict ARAS in patients with CAD was 0.41(sensitivity 78.4%,specificity 71.1%),the area under the ROC curve was 0.80.Conclusion:Older age(≥65 year),DM,RH and decreased Ccr are independent predictors for ARAS in CAD patients.Logistic regression model which incoperated with these predictors can be used to screening for ARAS before catheterization in CAD patients.