Prediction Model Of Delayed Graft Function Based On Clinical Characteristics Combined With Donor Serum IL-2 Level

Background:Kidney transplantation is an effective treatment for end stage renal disease(ESRD).Delayed graft function(DGF)is a common complication after kidney transplantation.Now,there are many prediction models to predict the occurrence of DGF after kidney transplantation,but there are few models combined with serum cytokines.This study established a DGF prediction model based on clinical features and serum cytokines.Methods:The discovery group data were obtained from 71 patients who underwent the first kidney transplantation in the First Affiliated Hospital of Nanchang University from October 2020 to October 2021.Logistic regression was used to analyze the correlation of each factors with DGF.A model for prediction of DGF after transplantation based on risk factors was established.In addition,the validation group data were obtained from 32 patients who underwent first kidney transplantation at the First Affiliated Hospital of Nanchang University from October 2021 to February2022.This model was verified by internal validation and external validation(1000bootstrap,Hosmer-Lemeshow statistic,Brier score,R~2).And,this model is compared with the DGF scoring system(DGFS,Chapal et al.,J Kidney Int.2014;86(6):1130-113))and Kidney Donor Risk Index(KDRI,Rao et al.,J Transplantation.2009;88(2):231-236)by area under receiver operating characteristic curve(AUC).Results:The DGF prediction model was composed of Cold ischemia time(CIT)(OR:1.445(1.046-1.995),P=0.026),Donor history of diabetes mellitus(OR :5.982(1.191-30.048),P=0.030),Donor interleukin-2(IL-2)level(OR :1.048(1.017-1.080),P=0.003)and Donor terminal creatinine(OR:2.115(1.054-4.245),P=0.035).In the internal validation,the areas under the receiver’s operating characteristic curve(AUC)of the prediction model was 0.894(95%CI=0.798-0.955),which higher than KDRI(0.764(95%CI=0.649-0.857))and DGFS(0.720(95%CI=0.601-0.820)).The good calibration of the model was confirmed(Brier score=0.116;R~2=55.8%).Also,in the external validation,AUC of the prediction model was 0.879(95%CI=0.715-0.967),which higher than KDRI(0.829(95%CI=0.655-0.938))and DGFS(0.667(95%CI=0.479-0.823)).The good calibration of the model was confirmed again(P=0.4117,Hosmer-Lemeshow statistic;Brier score=0.138;R~2= 53.9%).Conclusions:Donor serum IL-2 levels were associated with the DGF.A high accuracy DGF predictive model which included donor history of diabetes mellitus,donor serum IL-2,Donor terminal creatinine and CIT was established.