基于交互式列线图早期预测腹膜透析患者主要不良事件风险研究  

Early prediction of major adverse events inperitoneal dialysis patients based on interactive nomogram

作　　者：刘胜凤[1] 薄祥敏[1] 陈飞佳 张旋[1] LIU Shengfeng;BO Xiangmin;CHEN Feijia;ZHANG Xuan(Department of Nephrology,Jiangsu Province Hospital of Chinese Medicine Affiliated Hospital of Nanjing University of Chinese Medicine,Nanjing 210029,Jiangsu,China)

机构地区：[1]南京中医药大学附属江苏省中医院肾内科,南京210029

出　　处：《医学研究与战创伤救治》2025年第2期164-169,共6页Journal of Medical Research & Combat Trauma Care

基　　金：江苏省中医药科技发展计划项目(YB2020211)。

摘　　要：目的探讨腹膜透析(PD)患者发生主要不良事件(MAE)的危险因素并开发基于交互式列线图的预测模型。方法回顾性分析2014年5月至2022年8月在江苏省中医院肾内科腹透中心PD治疗的268例患者的临床资料。收集患者的基线特征、既往史和辅助检查结果。随访1年内发生全因死亡、主要心血管不良事件或转为血液透析任一事件,即判定为MAE。根据首次入院时间将总样本分为训练集(n=180)和验证集(n=88),采用最小绝对收缩和选择算子(LASSO)回归算法筛选MAE的最佳预测特征,利用多变量Logistic回归构建预测模型并在验证集中进行验证。通过受试者工作特征曲线下面积(AUC)、校准曲线和决策曲线分析(DCA)来评估模型的区分度、校准度和临床应用性。结果随访1年内训练集和验证集分别有32例(17.8%)和14例(15.9%)患者发生MAE。血清白蛋白、血红蛋白和血钠为PD患者MAE的最优预测因子。预测模型在训练集和验证集的AUC分别为0.92和0.88,并且均通过了校准度检验(P>0.05)。DCA显示模型具有临床应用价值,最后基于Shiny程序开发预测模型的网页交互式列线图。结论该预测模型可有效识别PD患者1年内MAE发生风险,交互式列线图可作为临床评估工具提供辅助决策建议。Objective To explore the risk factors for major adverse events(MAE)in peritoneal dialysis(PD)patients and to develop a predictive model based on interactive nomogram.Methods The clinical data of 268 patients treated with PD in the Peritoneal Dialysis Center of the Department of Nephrology,Jiangsu Provincial Hospital of Traditional Chinese Medicine from May 2014 to August 2022 were retrospectively analyzed.Baseline characteristics,medical history,and auxiliary examination results were collected.MAE was determined by the occurrence of any of all cause death,major adverse cardiovascular event or transition to hemodialysis within 1 year of follow-up.The total sample was divided into a training set(n=180)and a validation set(n=88)according to the time of first admission.The least absolute shrinkage and selection operator(LASSO)regression algorithm was employed to select optimal predictive features for MAEs.A multivariable logistic regression was used to construct the prediction model,which was then validated in the validation set.The discriminative ability,calibration,and clinical applicability of the model were evaluated through the area under the receiver operating characteristic curve(AUC),calibration curve,and decision curve analysis(DCA),respectively.Results Within 1 year of follow-up,MAE occurred in 32 patients(17.8%)in the training set and 14 patients(15.9%)in the validation set.Serum albumin,hemoglobinand serum sodium were identified as the optimal predictors for MAE in PD patients.The AUCs of the prediction model in the training and validation sets were 0.92 and 0.88,respectively,with both satisfying the calibration test(P>0.05).DCA demonstrated the clinical utility of the model,and a web based interactive nomogram was developed using the Shiny pro gram.Conclusion The prediction model developed in this study effectively identifies the risk of MAEs in PD patients within one year.The interactive nomogram can serve as a clinical assessment tool to provide auxiliary decision support.

关 键 词：腹膜透析 主要不良事件 预测模型 交互式列线图 

分 类 号：R692[医药卫生—泌尿科学]