Construction and validation of risk prediction model for vascular calcification in autogenous arteriovenous fistulas among patients undergoing maintenance hemodialysis
WANG Chun
LI Kui
ZHAO Hong
SONG Yan
YANG Jing
Abstract：Objective:To analyze the influencing factors of vascular calcification in autogenous arteriovenous fistulas among patients undergoing maintenance hemodialysis(MHD)and to construct decision tree model.Methods:A total of 233 patients undergoing MHD treated in our hospital were selected as study subjects from January 2022 to September 2023.And clinical data were collected.The patients were divided into modeling group(163 patients)and validation group(70 patients)at 7∶3 ratio.Logistic regression analysis was used to screen for influencing factors of vascular calcification in autogenous arteriovenous fistulas among patients undergoing MHD.The decision tree model was constructed.The model was validated using area under the curve(AUC)of receiver operator characteristic.Results:The incidence of vascular calcification in autogenous arteriovenous fistulas among patients undergoing MHD in the modeling group was 39.88%.The result of Logistic regression analysis showed that age,dialysis duration,serum phosphorus,high-sensitivity C-reactive protein,fibroblast growth factor-21,and bone morphogenetic protein-2 were all influencing factors of vascular calcification in autogenous arteriovenous fistulas among patients undergoing MHD(P<0.05).Decision tree model showed that bone morphogenetic protein-2 in the first layer.The AUC of the model in the modeling group was 0.815(95%CI 0.747-0.872),while that in the validation group was 0.830(95%CI 0.761-0.899).Conclusions:Age,dialysis duration,serum phosphorus,high-sensitivity C-reactive protein,fibroblast growth factor-21,and bone morphogenetic protein-2 were significantly associated with vascular calcification in autogenous arteriovenous fistulas among patients undergoing MHD.The constructed decision tree model of vascular calcification in autogenous arteriovenous fistulas among patients undergoing MHD had better predictive performance.
Keywords：maintenance hemodialysisautogenous arteriovenous fistulavascular calcificationinfluencing factorsdecision treemodel validation
Publication Date：2025-07-25
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 2450-2455 )