华南预防医学 ›› 2026, Vol. 52 ›› Issue (4): 383-388.doi: 10.12183/j.scjpm.2026.0383

• 论著 • 上一篇    下一篇

2型糖尿病患者高尿酸血症的危险因素识别和患病风险预测模型研究

刘付亦国1,2, 吴志燊1, 谭嘉裕1, 吴跃千1, 唐慧1, 常义坤1, 陈津伟1, 李楠1, 张王剑1, 杜志成1   

  1. 1.中山大学公共卫生学院,广东 广州 510080;
    2.化州市中垌卫生院
  • 收稿日期:2025-04-30 出版日期:2026-04-20 发布日期:2026-05-08
  • 通讯作者: 杜志成,E-mail:duzhch5@mail.sysu.edu.cn
  • 作者简介:刘付亦国(1994—),男,硕士研究生,主治医师,主要从事疾病控制工作

A study on the identification of risk factors and the development of a risk prediction model for hyperuricemia in patients with type 2 diabetes mellitus

Liu Fuyiguo1,2, Wu Zhishen1, Tan Jiayu1, Wu yueqian1, Tang Hui1, Chang Yikun1, Chen Jinwei1, Li Nan1, Zhang Wangjian1, Du Zhicheng1   

  1. 1. School of Public Health, Sun Yat-sen University, Guangzhou, Guangdong 510080, China;
    2. Zhongdong Health Center, Huazhou
  • Received:2025-04-30 Online:2026-04-20 Published:2026-05-08

摘要: 目的 在2型糖尿病患者中,识别高尿酸血症的危险因素和构建风险预测模型,并根据模型识别2型糖尿病患者中患高尿酸血症的高危人群。方法 基于中国人民解放军总医院的回顾性数据。数据收集时间为2016年12月5日至2021年12月13日,共纳入2 329例2型糖尿病患者。首先采用单因素分析初步描述、探讨与高尿酸血症相关的因素;继而利用随机森林算法进行变量初筛和logistic回归分析确定具有统计学意义的独立影响因素,据此构建风险预测模型。最后,采用受试者工作特征(ROC)曲线等评估该模型的预测效能。结果 2型糖尿病患者中高尿酸血症的患病率为20.35%。基于随机森林的初次筛选,我们从所有变量中选出20个,并进一步通过logistic回归确定了5个关键变量:血尿素(BU)、甘油三酯(TG)、体质指数(BMI)、糖化血红蛋白(HbA1c)、总蛋白(TP)。最后构建logistic回归模型,得到的2型糖尿病患者中高尿酸血症患病的风险预测模型:logit(P)=-4.936-0.159×HbA1c(%)+0.062×BMI+0.196×TG(mmol/L)+0.155×BU(mmol/L)+0.023×TP(g/L)。训练集中ROC曲线下面积(AUC)为0.765(95% CI:0.739~0.791),测试集中AUC值为0.699(95% CI:0.632~0.766)。结论 在2型糖尿病患者中,较低的HbA1c水平以及较高的BMI、TG、BU和TP水平,均是高尿酸血症发生的危险因素。基于上述5个变量构建的预测模型,能够较为准确地识别出2型糖尿病患者中高尿酸血症的高危人群。

关键词: 2型糖尿病, 高尿酸血症, 风险预测模型

Abstract: Objective To identify the risk factors for hyperuricemia and construct a risk prediction model for patients with Type 2 Diabetes Mellitus (T2DM), thereby enabling the identification of high-risk individuals within this patient population. Methods This study was based on retrospective data from 2 329 patients with T2DM admitted to the Chinese PLA General Hospital between December 5, 2016, and December 13, 2021. Univariate analysis was initially employed to describe and explore factors associated with hyperuricemia. Subsequently, a random forest algorithm was utilized for preliminary variable screening, followed by logistic regression analysis to determine statistically significant independent predictors. These predictors were then used to construct the risk prediction model. The predictive efficacy of the model was evaluated using the Receiver Operating Characteristic (ROC) curve. Results The prevalence of hyperuricemia among the cohort of patients with T2DM was 20.35%. Following the initial screening with the random forest algorithm, 20 variables were selected from the initial dataset. Logistic regression analysis further identified five key independent variables: Blood Urea (BU), Triglycerides (TG), Body Mass Index (BMI), Glycated Hemoglobin (HbA1c), and Total Protein (TP). A logistic regression model was subsequently constructed, yielding the following risk prediction formula for hyperuricemia in patients with T2DM: logit(P)=-4.936-0.159×HbA1c(%)+0.062×BMI+0.196×TG(mmol/L)+0.155×BU(mmol/L)+0.023×TP(g/L). The Area Under the ROC Curve (AUC) was 0.765 (95% CI: 0.739-0.791) in the training set and 0.699 (95% CI: 0.632-0.766) in the testing set. Conclusion Lower levels of HbA1c and elevated levels of BMI, TG, BU, and TP are significant risk factors for the development of hyperuricemia in patients with T2DM. The prediction model, developed based on these five variables, demonstrates a competent capacity for accurately identifying individuals at high risk for hyperuricemia within the T2DM patient population.

Key words: Type 2 diabetes mellitus, Hyperuricemia, Risk prediction model

中图分类号: 

  • R587.1