杨亮, 程润, 窦俊凯, 刘欢, 周志庆. 维持性血液透析病人衰弱发生风险列线图模型的构建[J]. 蚌埠医学院学报, 2023, 48(4): 538-543. DOI: 10.13898/j.cnki.issn.1000-2200.2023.04.028
    引用本文: 杨亮, 程润, 窦俊凯, 刘欢, 周志庆. 维持性血液透析病人衰弱发生风险列线图模型的构建[J]. 蚌埠医学院学报, 2023, 48(4): 538-543. DOI: 10.13898/j.cnki.issn.1000-2200.2023.04.028
    YANG Liang, CHENG Run, DOU Jun-kai, LIU Huan, ZHOU Zhi-qing. Establishment of nomogram model for the risk of frailty in maintenance hemodialysis patients[J]. Journal of Bengbu Medical College, 2023, 48(4): 538-543. DOI: 10.13898/j.cnki.issn.1000-2200.2023.04.028
    Citation: YANG Liang, CHENG Run, DOU Jun-kai, LIU Huan, ZHOU Zhi-qing. Establishment of nomogram model for the risk of frailty in maintenance hemodialysis patients[J]. Journal of Bengbu Medical College, 2023, 48(4): 538-543. DOI: 10.13898/j.cnki.issn.1000-2200.2023.04.028

    维持性血液透析病人衰弱发生风险列线图模型的构建

    Establishment of nomogram model for the risk of frailty in maintenance hemodialysis patients

    • 摘要:
      目的分析维持性血液透析病人衰弱发生的影响因素, 并建立衰弱发生风险的列线图模型。
      方法采用便利抽样法选取维持性血液透析病人222例, 根据是否发生衰弱分为衰弱组(n=85)和非衰弱组(n=137)。采用单因素与多因素logistic回归模型筛选影响维持性血液透析病人衰弱的危险因素。采用Bootstrap进行模型内部验证。使用受试者工作曲线、Homster-Lemeshow拟合度检验合并校准曲线图、临床决策曲线图分别对模型进行性能评价。
      结果222例维持性血液透析病人中发生衰弱共85例, 衰弱发生率为38.3%。logistic回归分析表明, 年龄≥ 60岁(OR=3.460, 95%CI: 1.775~6.747)、透析并发症个数(OR=1.644, 95%CI: 1.192~2.268)、白蛋白水平(OR=0.904, 95%CI: 0.838~0.976)、锻炼(OR=0.567, 95%CI: 0.430~0.748)、夜间睡眠时间(OR=0.488, 95%CI: 0.325~0.731)均是维持性血液透析病人衰弱发生的独立危险因素(P < 0.05~P < 0.01)。基于5项独立危险因素建立MHD病人衰弱发生的列线图预测模型。受试者工作曲线曲线下面积为0.829(95%CI: 0.777~0.882), Bootstrap重抽样法进行内部验证后, 模型的一致性指数为0.812。列线图模型拟合度较好(P>0.05);校准曲线图显示, 预测概率与实际概率发生率的相关性良好。临床决策曲线图表明模型阈值概率在0.03~0.83时, 此模型具有较好的临床实用性。
      结论基于影响维持性血液透析病人衰弱的危险因素建立的预测模型具有良好的区分度、一致性与临床实用性, 可为预防维持性血液透析病人衰弱的发生提供指导。

       

      Abstract:
      ObjectiveTo analyze the influencing factors of frailty in maintenance hemodialysis (MHD) patients and establish a nomogram model for frailty risk.
      MethodsA total of 222 patients with MHD were selected by convenient sampling method.According to whether the patients had frailty, they were divided into frailty group (n=85) and non-frailty group (n=137).Univariate and multivariate logistic regression model were used to screen the risk factors of frailty in patients with MHD.Bootstrap was used for internal verification of the model.The receiver operating characteristic (ROC) curve, Homster-Lemeshow goodness-of-fit test combined with calibration curve, and clinical decision curve analysis (DCA) were used to evaluate the performance of the model.
      ResultsThere were 85 cases of frailty in 222 patients with MHD, and the incidence of frailty was 38.3%.Logistic regression analysis showed that age ≥60 years old (OR=3.454, 95%CI: 1.775-6.719), the number of dialysis complications (OR=1.609, 95%CI: 1.173-2.207), albumin level (OR=0.901, 95%CI: 0.835-0.971), physical exercise (OR=0.684, 95%CI: 0.564-0.829) and night sleep time (OR=0.476, 95%CI: 0.316-0.716) were independent risk factors of frailty in patients with MHD (P < 0.05 to P < 0.01).Based on five independent risk factors, the nomogram prediction model of frailty in MHD patients was established.The area under ROC curve was 0.829 (95%CI: 0.777-0.882).After internal verification by Bootstrap resampling method, the consistency index of the model was 0.812.The fitting degree of nomograph model was good (P>0.05).The calibration curve showed that the correlation between the predicted probability and the actual probability incidence was good.The DCA showed that the model had good clinical practicability when the threshold probability was 0.03-0.83.
      ConclusionsThe prediction model based on the risk factors affecting the frailty of MHD patients has good differentiation, consistency and clinical practicability, and can provide guidance for the prevention of the frailty of MHD patients.

       

    /

    返回文章
    返回