4-5期慢性肾脏病非透析病人瘦体组织下降的列线图模型建立及分析

Establishment and analysis of the nomogram model of lean tissue decline in Stage 4-5 patients with chronic kidney disease without dialysis

目的:建立4-5期慢性肾脏病(CKD)非透析病人瘦体组织下降的列线图预测模型,以期为临床早期识别与筛选有关危险因素提供参考。方法:选取2018年7月至2021年12月在武汉市第五医院就诊的CKD 4-5期非透析病人247例,根据瘦体组织指数(LTI)分为正常LTI组和低LTI组,收集临床资料,Logistic回归分析瘦体组织下降的危险因素,并通过R软件包(R3.5.3)构建列线图预测模型,绘制校正曲线及ROC曲线评价模型区分度和精准度。结果:247例CKD 4-5期病人中正常LTI组148例(59.92%)、低LTI组99例(40.08%),Logistic回归分析显示年龄(OR=1.441,P=0.012)、糖尿病肾脏病(OR=2.153,P=0.029)、体质量指数(OR=1.639,P=0.023)、小腿围(OR=3.247,P=0.011)、二氧化碳(OR=3.573,P=0.004)进入回归方程,基于以上5个危险因素建立列线图预测模型并验证,校正曲线均趋近于理想曲线,校准曲线斜率接近1(χ^(2)=3.549,P=0.637),ROC曲线曲线下面积为0.792(95%CI:0.714~0.897)。结论:基于年龄、糖尿病肾脏病、体质量指数、小腿围、二氧化碳5个项目建立的列线图预测CKD 4-5期非透析病人瘦体组织下降具有较好的区分度和精确度。

Objective:To establish a nomogram prediction model for lean tissue decline in non-dialysis patients with Stage 4-5 chronic kidney disease(CKD),in order to provide reference for early clinical identification and screening of related risk factors.Methods:A total of 247 patients with CKD stage 4-5 who were treated in Wuhan Fifth Hospital from July 2018 to December 2021 were selected and divided into normal LTI group and low LTI group according to lean body tissue index(LTI).Clinical data were collected.Logistic regression were used to analyze the risk factors of lean tissue decline.A nomogram prediction model was constructed by R software package(R3.5.3),and calibration curve and ROC curve were drawn to evaluate the model differentiation and accuracy.Results:Among 247 CKD stage 4-5 patients,148 cases(59.92%)were in normal LTI group and 99 cases(40.08%)were in low LTI group.Logistic regression analysis showed age(OR=1.441,P=0.012),diabetic kidney disease(OR=2.153,P=0.029),body mass index(OR=1.639,P=0.023),calf circumference(OR=3.247,P=0.011)and carbon dioxide(OR=3.573,P=0.004)entered the regression equation.Based on the above five risk factors,a nomogram prediction model was established and verified.The correction curve all approached the ideal curve,and the slope of the calibration curve was close to 1(χ^(2)=3.549,P=0.637).The AUC of ROC curve was 0.792(95%CI:0.714~0.897).Conclusions:The nomogram based on age,diabetic kidney disease,body mass index,calf circumference and carbon dioxide showed good differentiation and accuracy in predicting lean tissue decline in CKD stage 4-5 non-dialysis patients.