中国血液净化

中国血液净化 ›› 2026, Vol. 25 ›› Issue (01): 16-20.doi: 10.3969/j.issn.1671-4091.2026.01.004

• 临床研究 • 上一篇    下一篇

腹膜透析患者残余肾功能快速下降的列线图预测模型

郭小雨   付晓慧   姜泽仪   刘映红   

  1. 410011 长沙，1中南大学湘雅二医院肾内科
  • 收稿日期:2025-05-21 修回日期:2025-10-15 出版日期:2026-01-12 发布日期:2025-12-31
  • 通讯作者: 刘映红 E-mail:liuyingh2002@csu.edu.cn
  • 基金资助:
    湖南省自然科学基金项目(2021JJ30942)；长沙市自然科学基金项目(kq2014235)

Nomogram prediction model for rapid decline of residual renal function in peritoneal dialysis patients

GUO Xiao-yu, FU Xiao-hui, JIANG Ze-yi, LIU Ying-hong   

  1. Department of Nephrology, the Second Xiangya Hospital of Central South University, Changsha 410011, China
  • Received:2025-05-21 Revised:2025-10-15 Online:2026-01-12 Published:2025-12-31
  • Contact: 410011 长沙，1中南大学湘雅二医院肾内科 E-mail:liuyingh2002@csu.edu.cn

摘要： 目的  分析腹膜透析(peritoneal dialysis，PD)患者残余肾功能(residual renal function，RRF)快速下降的危险因素并构建预测模型。 方法   纳入2012年1月1日—2022年12月31日于中南大学湘雅二医院肾内科就诊的PD患者，以RRF下降速率的中位数将患者分为快速下降组和缓慢下降组。应用最小绝对收缩与选择算子(least absolute shrinkage and selection operator，LASSO)回归初步筛选变量，再通过多因素逐步Logistic回归分析确定RRF快速下降的独立危险因素，并构建列线图预测模型。受试者工作特征(receiver operator characteristic，ROC)曲线下面积(area under the curve，AUC)、校准曲线、Hosmer-Lemeshow检验和决策曲线分析(decision curve analysis，DCA)评估模型的效能，并使用Bootstrap法进行内部验证。 结果  共纳入710例患者，其中快速下降组354例，缓慢下降组356例。LASSO回归及多因素逐步Logistic回归分析提示体质量指数(body mass index，BMI)＞28 kg/m2(OR =2.640，95% CI：1.180～5.926，P=0.018)、基线RRF≥2(OR =3.140，95% CI：1.845～5.342，P＜0.001)及血镁(OR =0.341，95% CI：0.128～0.909，P=0.032)是PD患者RRF快速下降的独立危险因素。基于以上因素构建的列线图预测模型AUC为0.719(95% CI：0.682～0.756)。经1000次Bootstrap法重采样内部验证后的AUC为0.722(95% CI：0.685～0.759)，Hosmer-Lemeshow拟合优度检验结果为χ2=5.268(P=0.729)，DAC显示其临床适用性阈概率范围为0.01～0.91。 结论  BMI、基线RRF、血镁是PD患者RRF快速下降的独立危险因素，基于此构建的列线图预测模型具有良好的区分度、校准度和临床适用度。

关键词: 腹膜透析, 残余肾功能下降速率, 危险因素

Abstract: Objective To analyze the risk factors for rapid decline of residual renal function (RRF) in peritoneal dialysis (PD) patients and to establish a predictive model. Methods PD patients who visited the Department of Nephrology at the Second Xiangya Hospital of Central South University from January 1, 2012, to December 31, 2022, were enrolled. Based on the median rate of RRF decline, patients were divided into a rapid decline group and a slow decline group. The Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed for preliminary variable selection. Independent risk factors associated with rapid RRF decline were then determined via multivariable stepwise logistic regression analysis. A nomogram prediction model was constructed based on the final predictors. The model's performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), calibration curves, the Hosmer-Lemeshow test, and decision curve analysis (DCA). Internal validation of the model was performed using the Bootstrap method. Results A total of 710 patients were included, with 354 in the rapid decline group and 356 in the slow decline group. LASSO regression and multivariable stepwise logistic regression identified that body mass index (BMI)＞28 (OR=2.640，95% CI：1.180～5.926，P=0.018), baseline RRF≥2(OR=3.140，95%CI：1.845～5.342，P＜0.001), and serum magnesium (OR=0.341，95%CI：0.128～0.909，P=0.032) as independent risk factors for rapid RRF decline in PD patients. The nomogram prediction model constructed based on these factors exhibited an  AUC of 0.719 (95%CI：0.682～0.756), and internal validation via 1000 Bootstrap resamples yielded an AUC of 0.722 (95%CI：0.685～0.759).The Hosmer-Lemeshow goodness-of-fit test yielded χ²=5.268 (P=0.729), and DCA indicated clinical applicability across a threshold probability range of 0.01～0.91. Conclusion    BMI, baseline RRF, and serum magnesium are independent risk factors for rapid RRF decline in PD patients. The nomogram prediction model based on these factors demonstrates good discrimination, calibration, and clinical applicability.

Key words: Peritoneal dialysis, Residual renal function decline rate, Risk factors

中图分类号: