基于列线图的成人高尿酸血症发病风险预测模型的构建  

Construction of a Risk Prediction Model for Hyperuricemia in Adults based on Nomogram

作　　者：李娟 戴晓霞 刘冰[2] 余可斐[3] 刘翔[4] LI Juan;DAI Xiaoxia;LIU Bing;YU Kefei;LIU Xiang(Hubei University of Medicine,Shiyan,Hubei,442000,China;不详)

机构地区：[1]湖北医药学院护理学院,湖北十堰442000 [2]湖北医药学院卫生管理与卫生事发展研究中心,湖北十堰442000 [3]湖北医药学院附属十堰市太和医院,湖北十堰442000 [4]湖北医药学院附属十堰市人民医院,湖北十堰442000

出　　处：《中国社会医学杂志》2024年第6期747-752,共6页Chinese Journal of Social Medicine

基　　金：国家自然科学基金项目(71774049)。

摘　　要：目的构建高尿酸血症(hyperuricemia,HUA)发病风险的列线图,为HUA早期预防和干预提供依据。方法从湖北省十堰市某三甲医院健康管理中心收集1865名体检者的体检资料及生活方式资料,随机分为训练集和验证集,利用Lasso回归法筛选变量,构建HUA列线图模型并进行危险分层。列线图的预测性能采用区分度、校准度、临床决策曲线和Bootstrap法进行评价。结果Lasso回归筛选了9个变量(性别、年龄、大豆类摄入量、中心性肥胖、BMI、LDLC、HDLC、TG、ALT)构建HUA网页版列线图。然后将HUA发病风险分为低危(发生率<8.29%)、中危(发生率<17.56%)、高危(发生率<30.81%)、极高危(发生率≥30.81%)。预测模型在训练集和验证集的AUC分别为0.749和0.741。模型的内部验证采用Bootstrap法,得到AUC为0.791。校准曲线表现出较好的一致性。临床决策曲线(DCA)表明,当训练集和验证集的阈值概率分别约在5%~70%和5%~65%的范围内时,对HUA患者采取干预可产生净收益。结论构建的HUA发病风险预测模型较为准确,可有效识别HUA高危人群,为HUA早期预防和干预提供依据。Objective To construct a nomogram for predicting the risk of hyperuricemia(HUA)and to provide evidence for early prevention and intervention of hyperuricemia(HUA).Methods The physical examination data and lifestyle data of 1865 subjects were collected from the health management center of a tertiary hospital in Shiyan city,Hubei Province.The subjects were randomly divided into training set and validation set.Lasso regression method was used to select variables to construct a nomogram model for HUA risk stratification.The predictive performance of the nomogram was evaluated by discrimination,calibration,clinical decision curve and Bootstrap method.Results Nine variables(gender,age,soybean intake,central obesity,BMI,LDLC,HDLC,TG,ALT)were selected from Lasso regression to construct the web version of HUA nomogram.Then,the risk of HUA was divided into low risk(incidence<8.29%),moderate risk(incidence<17.56%),high risk(incidence<30.81%),and extremely high risk(incidence≥30.81%).The AUC of the prediction model in the training set and validation set were 0.749 and 0.741,respectively.The Bootstrap method was used for internal validation of the model,and the AUC was 0.791.The calibration curve showed good consistency.The clinical decision curve(DCA)showed that when the threshold probability of the training set and the validation set were in the range of 5%~70%and 5%~65%,respectively,the intervention of HUA patients could produce net benefits.Conclusion A more accurate risk prediction model for HUA was established in order to identify high-risk groups of HUA,providing a basis for early prevention and intervention of HUA.

关 键 词：高尿酸血症 列线图 预测模型 

分 类 号：R589.7[医药卫生—内分泌]