Xing Yuwei, Chai Xuejiao, Liu Kuanzhi, Cao Guang, Wei Geng
Department of Endocrinology, The Second Hospital of Shijiazhuang, No. 53, Huaxi Road, Shijiazhuang, 050000, People's Republic of China.
Department of Endocrinology, The Third Hospital of Hebei Medical University, Shijiazhuang, 050000, People's Republic of China.
Int Urol Nephrol. 2024 Apr;56(4):1439-1448. doi: 10.1007/s11255-023-03815-7. Epub 2023 Oct 9.
There are few studies on the establishment of diagnostic models for diabetic nephropathy (DN) in in type 2 diabetes mellitus (T2DM) patients based on biomarkers. This study was to establish a model for diagnosing DN in T2DM.
In this cross-sectional study, data were collected from the Second Hospital of Shijiazhuang between August 2018 to March 2021. Totally, 359 eligible participants were included. Clinical characteristics and laboratory data were collected. LASSO regression analysis was used to screen out diagnostic factors, and the selected factors were input into the decision tree for fivefold cross validation; then a diagnostic model was established. The performances of the diagnosis model were evaluated by the area under the receiver operator characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. The diagnostic performance of the model was also validated through risk stratifications.
Totally, 199 patients (55.43%) were diagnosed with DN. Age, diastolic blood pressure (DBP), fasting blood glucose, insulin treatment, mean corpuscular hemoglobin concentration (MCHC), platelet distribution width (PDW), uric acid (UA), serum creatinine (SCR), fibrinogen (FIB), international normalized ratio (INR), and low-density lipoprotein cholesterol (LDL-C) were the diagnostic factors for DN in T2DM. The diagnostic model presented good performances, with the sensitivity, specificity, PPV, NPV, AUC, and accuracy being 0.849, 0.969, 0.971, 0.838, 0.965, and 0.903, respectively. The diagnostic model based on the stratifications also showed excellent diagnostic performance for diagnosing DN in T2DM patients.
Our diagnostic model with simple and accessible factors provides a noninvasive method for the diagnosis of DN.
基于生物标志物建立2型糖尿病（T2DM）患者糖尿病肾病（DN）诊断模型的研究较少。本研究旨在建立T2DM患者DN的诊断模型。
在这项横断面研究中，收集了2018年8月至2021年3月期间石家庄市第二医院的数据。共纳入359名符合条件的参与者。收集临床特征和实验室数据。采用LASSO回归分析筛选诊断因素，并将所选因素输入决策树进行五折交叉验证；然后建立诊断模型。通过受试者操作特征曲线下面积（AUC）、灵敏度、特异度、阳性预测值（PPV）、阴性预测值（NPV）和准确度评估诊断模型的性能。该模型的诊断性能也通过风险分层进行了验证。
共有199例患者（55.43%）被诊断为DN。年龄、舒张压（DBP）、空腹血糖、胰岛素治疗、平均红细胞血红蛋白浓度（MCHC）、血小板分布宽度（PDW）、尿酸（UA）、血清肌酐（SCR）、纤维蛋白原（FIB）、国际标准化比值（INR）和低密度脂蛋白胆固醇（LDL-C）是T2DM患者DN的诊断因素。该诊断模型表现良好，灵敏度、特异度、PPV、NPV、AUC和准确度分别为0.849、0.969、0.971、0.838、0.965和0.903。基于分层的诊断模型在诊断T2DM患者DN方面也显示出优异的诊断性能。
我们的诊断模型具有简单易获取的因素，为DN的诊断提供了一种非侵入性方法。
