中国血液净化

中国血液净化 ›› 2023, Vol. 22 ›› Issue (06): 426-431,437.doi: 10.3969/j.issn.1671-4091.2023.06.006

• 临床研究 • 上一篇    下一篇

列线图模型对维持性血液透析患者心脏瓣膜钙化发生风险的预测及评估

张嘉欣   唐文武   谢席胜   

  1. 646000 泸州，1西南医科大学临床医学院
    637001 南充，2川北医学院附属南充市中心医院肾内科
  • 收稿日期:2023-02-10 修回日期:2023-04-23 出版日期:2023-06-12 发布日期:2023-06-12
  • 通讯作者: 谢席胜 E-mail:xishengx@163.com
  • 基金资助:
    四川省中医药管理局中医药科研专项(2020JC079)；四川省科技厅专项基金(2021YFS0259)；南充市科技计划项目(22JCYJPT0005)

Prediction and evaluation of the risk of cardiac valve calcification in maintenance hemodialysis patients with a nomograph model

ZHANG Jia-xin, TANG Wen-wu, XIE Xi-sheng   

  1. 1School of Clinical Medicine, Southwest Medical University, Luzhou 646000, China; 2Department of Nephrology, Nanchong Central Hospital Affiliated to North Sichuan Medical College, Nanchong 637000, China
  • Received:2023-02-10 Revised:2023-04-23 Online:2023-06-12 Published:2023-06-12
  • Contact: 646000 泸州，1西南医科大学临床医学院；637001 南充，2川北医学院附属南充市中心医院肾内科 E-mail:xishengx@163.com

PDF

52

摘要： 目的 通过构建列线图预测维持性血液透析(maintenance hemodialysis,MHD)患者心脏瓣膜钙化(cardiac valve calcification,CVC)的发生风险。 方法  纳入2014～2022年在南充市中心医院接受MHD治疗患者，收集基础信息、疾病信息、实验室及影像学检查数据，利用单因素、多因素Logistic回归分析相关危险因素，使用R语言软件建立预测CVC发生风险的列线图模型。采用Bootstrap法进行验证。以受试者工作特征(ROC)曲线下面积大小评估模型的预测效能，运用校准曲线、决策曲线分析(decision curve analysis,DCA)评估模型的准确度及实用性。 结果 共纳入MHD患者173例，其中CVC患者63例(36.4%)，多因素Logistic回归分析结果显示年龄(OR=-1.046,   95% CI 1.016～1.077，P=0.002)、血钙(OR=5.181,95% CI 1.015～27.252,P=0.042)、血磷(OR=2.401,95% CI 1.177～4.898，P=0.038)、糖尿病(OR=2.943,95% CI 1.397～6.195，P=0.013)、继发性甲状旁腺功能亢进(OR=2.844,95% CI 1.003～8.082,P=0.041)是CVC的独立危险因素。列线图显示出较好的辨别力，训练集ROC曲线下面积为0.757(95% CI 0.735～0.763)，内部验证C-指数为0.732，测试集ROC曲线下面积为0.700(95% CI 0.695～0.714)。校准曲线结果表明实际发生CVC概率与预测概率之间良好的一致性。 结论 本研究构建的列线图可用于识别发生CVC的高风险人群，具有一定的临床效用。

关键词: 维持性血液透析, 心脏瓣膜钙化, 列线图

Abstract: Objectives   The risk of cardiac valve calcification (CVC) in patients with maintenance hemodialysis (MHD) was predicted by a nomogram.  Methods   Patients with end-stage kidney disease (ESKD) who received MHD in Nanchong Central Hospital from 2014 to 2022 were included in this study. Basic information, disease information, laboratory and imaging examination data were collected. The risk factors for CVC were analyzed by univariate and multivariate logistic regression analyses, and a nomograph model for predicting the risk of CVC was established using R language software. Bootstrap method was used for the verification. The prediction efficiency of the model was evaluated by the area under the ROC curve, and the accuracy and practicability of the model were evaluated by calibration curve and decision curve analysis (DCA).  Results   A total of 173 MHD patients were included, including 63 CVC patients (36.4%). Multivariate logistic regression analysis showed that age (OR:1.046, 95% CI: 1.016~1.077, P=0.002), serum calcium (OR: 5.181, 95% CI: 1.015~27.252, P=0.042), serum phosphorus (OR:2.401,95% CI:1.177~4.898, P=0.038), diabetes (OR: 2.943, 95% CI: 1.397~6.195, P=0.013) and secondary hyperparathyroidism (OR: 2.844, 95% CI: 1.003~8.082, P=0.041) were the independent risk factors for CVC. The line chart showed good differentiation. The area under ROC curve of the training set was 0.757 (95% CI: 0.735 ~0.763), the internal verification C-index was 0.732, and the area under ROC curve of the test set was 0.700 (95% CI: 0.695~0.714). The calibration curve results show that the actual probability of CVC occurrence was in better agreement with the predicted probability.  Conclusions  The column graph constructed in this study can be used to identify the patient group in high risk of CVC, which may be clinically useful.

Key words: Maintenance hemodialysis, Cardiac valve calcification, Nomogram

中图分类号: