2型糖尿病肾病风险列线图预测模型的建立  

Construction of a predictive nomogram model for the risk of type 2 diabetic nephropathy

作　　者：马瑾 马芳琪 董小英[3] Ma Jin;Ma Fangqi;Dong Xiaoying(School of Clinical Medicine,General Hospital of Ningxia Medical University,Yinchuan 750003,Ningxia Hui Autonomous Region,China;Department of Nephrology,Wuzhong People's Hospital,Wuzhong 751199,Ningxia Hui Autonomous Region,China;Department of Endocrinology,General Hospital of Ningxia Medical University,Yinchuan 750003,Ningxia Hui Autonomous Region,China)

机构地区：[1]宁夏医科大学总医院临床医学院,银川750003 [2]吴忠市人民医院肾病内科,吴忠751199 [3]宁夏医科大学总医院内分泌科,银川750003

出　　处：《中国基层医药》2024年第5期734-741,共8页Chinese Journal of Primary Medicine and Pharmacy

摘　　要：目的分析2型糖尿病(type 2 diabetes mellitus,T2DM)患者并发糖尿病肾病(diabetic kidney disease,DKD)的危险因素,构建DKD的风险列线图预测模型并进行验证,以期为疾病的早期筛查及预防提供科学依据。方法回顾性收集2018年1月至2022年12月在宁夏医科大学总医院内分泌科住院的T2DM患者1223例的临床资料,用RStudio 4.2.1,通过LASSO回归及十折交叉验证筛选最优特征变量,利用多因素logistic回归分析确定最终预测因子并构建风险预测模型后绘制列线图。最后,分别采用C指数、受试者工作特征曲线(ROC曲线)、校准曲线和Hosmer-Lemeshow拟合优度检验验证评价模型的区分度和准确性,决策曲线分析评估模型的临床效能。结果多因素logistic回归分析结果显示,糖尿病病程、糖化血红蛋白(HbA1c)(OR=1.14,95%CI:1.05~1.24)、血肌酐(Scr)(OR=1.02,95%CI:1.01~1.04)、25-羟维生素D[25-(OH)-D](OR=0.97,95%CI:0.95~1.00)均是T2DM患者发生DKD的最佳预测因素(均P<0.05)。依据预测变量绘制列线图,构建预测模型。预测模型预测训练集T2DM患者发生DKD的C指数为0.762(95%CI:0.734~0.790),预测验证集T2DM患者发生DKD的C指数为0.742(95%CI:0.689~0.790),Hosmer-Lemeshow拟合优度检验显示模型拟合度较好(P=0.108),DCA结果显示预测模型在临床上是有益的。结论风险列线图预测模型的建立可早期筛查和防治DKD,为临床工作者提供了更为便捷、科学的方法和手段。Objective To explore the risk factors of diabetic nephropathy in type 2 diabetes patients,establish a risk prediction model for diabetic nephropathy and provide scientific reference for the prevention and screening of diabetic nephropathy.Methods Clinical data of 1223 patients admitted at Department of Endocrinology,General Hospital of Ningxia Medical University from January 2018 to December 2022 were retrospectively collected.In the training set,LASSO regression analysis and 10-fold cross-validation were used to screen the optimal feature variables by RStudio 4.2.1 software,and then multivariate logistic regression analysis was used to determine the final predictors selected from LASSO regression to construct the risk prediction model and draw the nomogram diagram.The receiver operating characteristic curve,C-index,calibration curve,and Hosmer-Lemeshow test were used to assess the discrimination and accuracy of the model;and the decision curve analysis was used to assess the clinical validity of the model.Results The multivariate logistic regression analysis showed that the duration of diabetes,glycosylated hemoglobin[odds ratio(OR)=1.14,95%confidence interval(CI):1.05-1.24],serum creatinine(OR=1.02,95%CI:1.01-1.04),25-(OH)-D(OR=0.97,95%CI:0.95-1.00)were the best predictors of diabetic nephropathy in patients with type 2 diabetes(P<0.05).The predictive model was constructed by plotting the nomogram graph based on the predictor variables.In the training cohort,the diabetic nephropathy risk model displayed medium predictive power with a C-index of 0.762(95%CI:0.734-0.790).Meanwhile,the risk model was also well validated in the validation set,where the C-index was 0.742(95%CI:0.689-0.790).Hosmer-Lemeshow test showed excellent degree of fit(P=0.108),and the results of the decision curve analysis showed that the prediction model could be clinically beneficial.Conclusion The establishment of the risk prediction nomogram model provides clinicians with a more convenient and scientific method for early screening and prev

关 键 词：糖尿病 2型 糖尿病肾病 列线图 LOGISTIC模型 ROC曲线 血红蛋白A 糖基化 

分 类 号：R587.2[医药卫生—内分泌] R692.9[医药卫生—内科学]