Factors associated with postoperative bloodstream infections after kidney transplantation and the development of a quantitative risk prediction model

CUI Yong; WANG Xue-ying; TANG Fei-hu; ZHANG Zhen-tao; GAO Yuan

doi:10.13820/j.cnki.gdyx.20250852

Abstract：Objective To analyze the risk factors associated with bloodstream infections(BSIs)after kidney transplantation,and to establish and validate a quantitative risk prediction model for postoperative BSI.Methods A to-tal of 248 patients who underwent kidney transplantation between December 2020 and December 2023 were retrospectively analyzed.Patients were followed up for six months and divided into BSI and non-BSI groups according to postoperative infection status.Candidate variables were initially screened using the least absolute shrinkage and selection operator(LASSO)regression model,and significant predictors were further analyzed using multivariate logistic regression to identi-fy independent factors associated with BSIs.Sensitivity analyses were performed to evaluate model robustness.A risk pre-diction model was constructed and assessed by receiver operating characteristic(ROC)curve analysis,calibration curve,and decision curve analysis(DCA)to evaluate discrimination,calibration,and clinical utility.Results Postoperative BSIs occurred in 26 of 248 patients(10.97％).LASSO regression identified three variables with non-zero coefficients:diabetes mellitus,serum albumin(ALB),and platelet count(PLT).Multivariate logistic regression showed that diabetes mellitus(OR=2.504,95％CI:1.307-4.794,P＜0.05)was an independent risk factor for postoperative BSIs,while ALB(OR=0.475,95％CI:0.310-0.726,P＜0.05)and PLT(OR=0.609,95％CI:0.398-0.932,P＜0.05)were protective factors.Sensitivity analyses confirmed the stability of the model.The established logistic regression equa-tion was:logit=-11.259+2.504x1-0.475x2-0.609x3.The model demonstrated excellent predictive performance,with an AUC of 0.925(95％CI:0.884-0.955,Z=18.379,P＜0.001),sensitivity of 96.15％,and specificity of 87.20％.The Hosmer-Lemeshow test indicated good calibration(x2=0.641,P=0.196),and DCA showed positive net clinical benefit within the risk threshold range of0.11-0.81.Conclusion Diabetes mellitus,serum albumin,and platelet count are independent factors influencing the occurrence of bloodstream infections after kidney transplantation.The established quantitative prediction model demonstrates high accuracy,good calibration,and strong clinical applicability,providing a valuable tool for early risk assessment and individualized prevention of BSIs in kidney transplant recipients.