云南省安宁市老年人群高血压合并高尿酸血症列线图预测模型构建及验证  

Construction and verification of a line graph prediction model for hypertension complicated with hyperuricemia in elderly in Anning City,Yannan Province

作　　者：卢双艳 张梅 张茹 普惠婕 李梦启 李贵涛 刁翌 毛勇 LU Shuangyan;ZHANG Mei;ZHANG Ru;PU Huijie;LI Mengqi;LI Guitao;DIAO Yi;MAO Yong(School of Public Health,Kunming Medical University,Kunming,Yunnan 650500,China)

机构地区：[1]昆明医科大学公共卫生学院,云南昆明650500

出　　处：《中国预防医学杂志》2024年第9期1117-1123,共7页Chinese Preventive Medicine

基　　金：昆明医科大学科技创新团队建设项目(CXTD202103)。

摘　　要：目的 基于Lasso-logistic回归模型探讨≥60岁人群高血压合并高尿酸血症(hyperuricemia,HUA)的影响因素,为高血压防控提供依据。方法 2021年5—11月在云南省安宁市开展≥60岁老年人慢性病调查,收集研究对象的一般资料及临床生化指标。采用Lasso回归模型进行特征筛选,在此基础上利用多因素logistic回归模型进一步筛选潜在危险因素,最后建立列线图预测模型。采用受试者工作特征(receiver operating characteristic,ROC)曲线下面积(area under the curve,AUC)评估列线图预测模型的预测能力。利用bootstrap自抽样法对模型进行内部验证,利用一致性指数(C-index)、Hosmer-Lemeshow拟合优度检验、校准曲线评估模型的区分度和校准度。采用临床决策曲线(clinical decision curve,DCA)、临床影响曲线(clinical impact curve,CIC)对模型进行临床有效性分析。结果 现场招募≥60岁的老年人11 397例,高血压合并HUA者1 726例,Lasso-logistic回归模型筛选出12个独立危险因素,分别为年龄、性别、中心性肥胖、24.0 kg/m^(2)≤体质量指数(body mass index,BMI)<28.0 m^(2)、BMI≥28.0 kg/m^(2)以及血肌酐、三酰甘油、非高密度脂蛋白胆固醇(non-high density lipoprotein cholesterol,non-HDL-C)、丙氨酸氨基转移酶和天冬氨酸氨基转移酶比值(ALT/AST)、有高血压既往史、高血压合并2型糖尿病、HUA合并高ALT(P<0.05)。构建的高血压合并HUA的列线图预测模型ROC的AUC为0.81 (95%CI:0.79~0.84);内部验证结果提示,该模型的一致性指数为0.81 (95%CI:0.79~0.84)。该列线图模型预测高血压合并HUA的风险概率与实际概率基本一致,具有良好的校准度(χ^(2)=13.118,P>0.05)。DCA曲线和CIC曲线显示,该列线图预测模型具有一定的临床实用价值。结论 本研究基于Lasso-logistic回归构建的列线图模型具有较好的区分度和校准度,预测的风险概率与实际概率基本一致,对于高血压合并HUA高风险人群的�Objective This study aimed to investigate the influencing factors of hypertension combined with hyperuricemia (HUA) in individuals aged 60 and above,to provide basis for the prevention and control of hypertension.Methods From May to November 2021,a survey of chronic diseases in the elderly was carried out in Anning City,Yunnan Province.General information and clinical biochemical indicators of the subjects were collected.The Lasso regression model was used for feature screening,followed by multivariable logistic regression to identify potential risk factors,and finally,a line graph prediction model was established.The predictive ability of the model was evaluated using the area under the receiver operating characteristic (ROC)area under the curve (AUC).Internal validation was conducted using bootstrap resampling,and the discrimination and calibration of the model were assessed using the C-index,Hosmer-Lemeshow goodnessof-fit test,and calibration curve.Clinical decision curve (DCA) and clinical impact curve (CIC) were used to analyze the clinical effectiveness of the model.Results 11 397 participants aged 60 and above were recruited There were 1 726 patients with hypertension combined with HUA.The Lasso-logistic regression model screened 12 independent risk factors,including age,gender,central obesity,24.0 kg/m^(2)≤BMI<28.0 kg/m^(2),BMI≥28.0 kg/m^(2),serum creatinine,triacylglycerol,non-high-density lipoprotein cholesterol (non-HDL-C),aspartate aminotransferase to alanine aminotransferase ratio (ALT/AST),history of hypertension,hypertension combined with type 2 diabetes,and HUA combined with high alanine aminotransferase (ALT)(P<0.05).The line graph prediction model for hypertension complicated with HUA had an area under the ROC curve (AUC)of 0.81 (95%CI:0.79-0.84).Internal validation indicated good consistency with a C-index of 0.81 (95%CI:0.79-0.84).The model showed good calibration,with the predicted risk probabilities aligning well with the actual probabilities (χ^(2)=13.118,P>0.05).DCA curve and CIC curv

关 键 词：高血压 高尿酸血症 Lasso-logistic回归 列线图模型 

分 类 号：R587[医药卫生—内分泌]