腹膜透析患者血管钙化的风险预测列线图模型的构建  

Construction of a nomogram model for risk prediction of vascular calcification in peritoneal dialysis patients

作　　者：杨建兵[1] 林武君 刘利 彭李[1] 陈德政[1] YANG Jianbing;LIN Wujun;LIU Li;PENG Li;CHEN Dezheng(Jianyang People's Hospital,Jianyang 641400,Sichuan,China)

机构地区：[1]简阳市人民医院,四川简阳641400

出　　处：《中南医学科学杂志》2025年第2期361-364,共4页Medical Science Journal of Central South China

基　　金：四川省卫生健康委科研课题(18PJ370)。

摘　　要：目的 分析影响腹膜透析患者发生血管钙化的可能因素,并构建列线图预测模型。方法 选择腹膜透析患者323例作为研究对象,按照7∶3的比例随机分为建模组(226例)和验证组(97例),建模组患者根据是否发生血管钙化,分为血管钙化组(98例)和非血管钙化组(128例)。基于Logistic回归模型分析结果构建预测腹膜透析患者发生血管钙化的列线图模型,并进行验证。结果 高水平甘油三酯、血清钙、碱性磷酸酶、成纤维细胞生长因子23、甲状旁腺激素是腹膜透析患者发生血管钙化的危险因素(P<0.05)。ROC曲线分析建模组和验证组在列线图模型中的曲线下面积分别为0.987、0.954,模型的区分度、诊断效能良好。Hosmer-Lemeshow检验模型的拟合优度,建模组的χ^(2)=5.066,P=0.751,验证组的χ^(2)=6.045,P=0.713。结论 利用筛选出的甘油三酯、血清钙、碱性磷酸酶、成纤维细胞生长因子23、甲状旁腺激素等因素构建的列线图模型能较准确地预测腹膜透析患者发生血管钙化的风险,辅助临床诊治。Aim To analyze the potential factors influencing vascular calcification in peritoneal dialysis patients and to construct a nomogram risk prediction model.Methods A total of 323 peritoneal dialysis patients were selected as the study subjects and randomly divided into a modeling group(226 cases) and a validation group(97 cases) at a 7∶3 ratio.Based on whether vascular calcification occurred in the modeling group,patients were further divided into a vascular calcification group(98 cases) and a non-vascular calcification group(128 cases).A nomogram model for predicting vascular calcification in peritoneal dialysis patients was constructed based on the Logistic regression analysis result and subsequently validated.Results High levels of triglycerides,serum calcium,alkaline phosphatase,fibroblast growth factor 23,and parathyroid hormone were identified as risk factors for vascular calcification in peritoneal dialysis patients(P<0.05).The area under the ROC curve for the modeling group and validation group in the nomogram model was 0.987 and 0.954,respectively,indicating good discrimination and diagnostic performance of the model.The Hosmer-Lemeshow tests showed good model fit,with χ~2=5.066,P=0.751 for the modeling group and χ~2=6.045,P=0.713 for the validation group.Conclusion The nomogram model constructed using the identified factors triglycerides,serum calcium,alkaline phosphatase,fibroblast growth factor 23,and parathyroid hormone demonstrates good diagnostic performance and can accurately predict the risk of vascular calcification in peritoneal dialysis patients,supporting clinical diagnosis and treatment.

关 键 词：腹膜透析 血管钙化 影响因素 列线图 

分 类 号：R692.5[医药卫生—泌尿科学]