Chinese General Practice ›› 2023, Vol. 26 ›› Issue (26): 3259-3268.DOI: 10.12114/j.issn.1007-9572.2023.0002

• Original Research·Monographic Research·Type 2 Diabetic • Previous Articles     Next Articles

Development and Validation of a Risk Prediction Model for the Progression from Microalbuminuria to Macroalbuminuria in Patients with Type 2 Diabetes Mellitus

  

  1. 1. Department of Endocrinology, the First Affiliated Hospital of Air Force Medical University, Xi'an 710032, China
    2. Department of Internal Medicine, the Traditional Chinese Medicine Hospital of Lintong District, Xi'an 710600, China
  • Received:2022-11-08 Revised:2023-01-15 Published:2023-09-15 Online:2023-02-16
  • Contact: LI Xiaomiao

2型糖尿病患者微量白蛋白尿进展至大量白蛋白尿的风险预测模型研究

  

  1. 1.710032 陕西省西安市,空军军医大学第一附属医院内分泌科
    2.710600 陕西省西安市临潼区中医医院内科
  • 通讯作者: 李晓苗
  • 作者简介:
    作者贡献:卢作维、曹宏伟、刘涛、陈艳艳、李晓苗进行方案策划,论文的修订;卢作维、曹宏伟、刘涛、李晓苗开展研究与调查,负责文章的质量控制及审校;卢作维、曹宏伟、陈艳艳、史勤利、赖敬波进行数据策划、形式分析;卢作维、张娜娜、陈艳艳、史勤利、王琼、赖敬波进行统计学处理;刘向阳、王琼、李晓苗进行形式分析;卢作维、曹宏伟、刘涛、陈艳艳撰写论文初稿;卢作维、曹宏伟、刘涛、陈艳艳、李晓苗对文章整体负责,监督管理。

Abstract:

Background

The incidence of diabetic kidney disease (DKD) and the proportion of its related end-stage renal disease in dialysis patients in China are increasing. So it is urgent to take measures to prevent and control DKD. Intensified multifactorial interventions may prevent or delay the progression of DKD. Therefore, developing a personalized risk prediction model can effectively delay or even prevent the progression of DKD and be useful for the prevention and treatment of DKD.

Objective

The purpose of this study was to develop and validate a nomogram for the risk prediction of the progression from microalbuminuria (MAU) to macroalbuminuria (CAU) in type 2 diabetes mellitus (T2DM) patients.

Methods

A total of 1 263 T2DM patients with albuminuria who were hospitalized in Department of Endocrinology, the First Affiliated Hospital of Air Force Medical University from October 2016 to March 2020 were retrospectively recruited and divided into a development cohort of 906 cases and a validation cohort of 357 cases, according to the admission time. LASSO regression was used to screen the optimized variables measured at baseline for CAU. A Nomogram was constructed based on selected predictive factors identified by the multivariate logistic regression model of the development sub-cohort. The receiver operating characteristic (ROC) curve, calibration curve and Hosmer-Lemeshow (H-L) test were employed to assess the calibration and discrimination of the model. Decision curve analysis (DCA) was performed to evaluate the net clinical benefit of the Nomogram.

Results

The diabetes duration, systolic blood pressure (SBP), glycosylated hemoglobin A1c (HbA1c), low-density lipoprotein cholesterol (LDL-C), cystatin C (Cys-C), estimated glomerular filtration rate (eGFR), and diabetic retinopathy (DR) were screened as predictive factors for progression from MAU to CAU by LASSO penalty regression. Multivariable Logistic regression analysis using these factors indicated that seven of those potential predictors were present in the final model, diabetes duration≥10 years, SBP≥140 mmHg, HbA1c≥7.0 mmol/L, LDL-C≥1.8 mmol/L, Cys-C>1.09 mg/L, and DR were risk factors for the progression from MAU to CAU in T2DM patients (P<0.05), while eGFR showed no statistically significant association with the progression in stratified analysis (P>0.05). External and internal validations of the nomogram indicated a good predictive performance. The AUC of the model was 0.814〔95%CI (0.782, 0.846) 〕 in the development cohort, and was 0.768〔95%CI (0.713, 0.823) 〕 in the validation cohort. The model was well fit according to the calibration curve and the H-L goodness of fit test (internal validation: P=0.065; external validation: P=0.451). DCA curve showed that the Nomogram's net benefit was higher than both extreme curves when the threshold probability set between 0.08 and 0.74 in the development cohort, and between 0.14 and 0.70 in the external validation cohort, suggesting potential clinical benefits provided by this Nomogram.

Conclusion

This study finally constructed a prediction model with seven indicators containing diabetes duration, SBP, HbA1c, LDL-C, Cys-C, eGFR, and DR, and will be a useful clinical predictive tool for the risk of progression from MAU to CAU in T2DM patients.

Key words: Diabetes mellitus, type 2, Albuminuria, Microalbuminuria, Macroalbuminuria, Nomograms

摘要:

背景

我国糖尿病肾病(DKD)的发病率及其所致终末期肾病(ESRD)在透析患者中所占比例不断攀升,DKD防控已刻不容缓,强化的多因素干预措施可延缓或者阻止DKD的病程进展,因而通过建立个性化的风险预测模型,可延缓或者阻止DKD病程进展,从而实现DKD的有效防治。

目的

开发和验证一个基于诺模图(Nomogram)的列线图模型,根据预测变量预测2型糖尿病(T2DM)患者从微量白蛋白尿(MAU)进展为大量白蛋白尿(CAU)的风险。

方法

回顾性收集2016年10月—2020年3月于空军军医大学第一附属医院内分泌科住院的2型糖尿病合并白蛋白尿患者1 263例,并根据入院时间将入组病例分为开发队列906例和验证队列357例。应用LASSO回归从收集的基线数值中筛选预测变量,并根据筛选的预测变量构建多因素Logistic回归模型,绘制模型的Nomogram。模型的验证和评估主要基于受试者工作特征(ROC)曲线、校准曲线和Hosmer-Lemeshow检验(H-L检验),并根据决策曲线分析法(DCA)评价该模型的实际临床净收益。

结果

基于LASSO回归惩罚收缩方法筛选出糖尿病病程、收缩压(SBP)、糖化血红蛋白(HbA1c)、低密度脂蛋白胆固醇(LDL-C)、胱抑素C(Cys-C)、估算肾小球滤过率(eGFR)及糖尿病视网膜病变(DR)7个预测变量,依据这些预测变量构建的多因素Logistic回归模型显示,糖尿病病程≥10年、SBP≥140 mmHg、HbA1c≥7.0 mmol/L、LDL-C≥1.8mmol/L、Cys-C>1.09 mg/L及合并DR是T2DM患者MAU进展为CAU的危险因素(P<0.05),而eGFR分层处理后无统计学意义(P>0.05)。预测模型的内外部验证显示该模型预测效能较优,开发队列的ROC曲线下面积(AUC)为0.814〔95%CI(0.782,0.846)〕,验证队列的AUC为0.768〔95%CI(0.713,0.823)〕,校准曲线和H-L检验(开发队列P=0.065;验证队列P=0.451)均显示该模型具有较好的一致性和拟合度,另外DCA结果显示当开发队列和验证队列的阈值概率分别为0.08~0.74和0.14~0.70时,该模型临床有效性较高。

结论

本研究开发了一个包含糖尿病病程、SBP、HbA1c、LDL-C、Cys-C、eGFR及是否合并DR 7个变量的列线图模型,可用于预测T2DM患者MAU进展为CAU的临床风险。

关键词: 糖尿病,2型, 白蛋白尿, 微量白蛋白尿, 大量白蛋白尿, 列线图