External validation of a prognostic model predicting renal graft function one year after brain-dead donor kidney transplantation

Background. A large German transplant center recently published a prognostic model predicting renal graft function one year after deceased donor kidney transplantation relying solely on pre-transplant variables. The aim of this study is to externally validate this prognostic model. Methods. In this external validation cohort, we retrospectively analyzed clinical data from deceased donor kidney transplant recipients undergoing kidney transplantation between January 2007 and December 2023 at University Hospital RWTH Aachen. Receiver operating characteristics (ROC) curves were analyzed to validate the abovementioned prognostic model based on donor age, donor serum creatinine, recipient body mass index, re-transplantation > 2 ^nd kidney transplant, and cold ischemia time. Glomerular filtration rate levels were categorized using the Kidney Disease: Improving Global Outcomes (KDIGO) stages G1 – G5 with KDIGO G1 as reference category. Results. In the observation period, a total of 494 kidney transplants were performed at our institution, 350 (70.9%) thereof from donation after brain death (DBD). The median one-year estimated glomerular filtration (eGFR) was 42 [12-94] mL/min/1.73 m ² . The areas under the receiver operating characteristics curve for KDIGO categories G2, G3a, G3b, G4, and G5 were 0.8664, 0.7167, 0.7442, 0.7315, and 0.7331, respectively, thus indicating a high sensitivity and specificity of prediction for all stages of renal graft function. Conclusions. Using the proposed prognostic model, we could successfully predict each eGFR category in our external cohort of deceased donor kidney transplant recipients one year after transplantation. Thus, we consider the prognostic model as highly valuable. It could be used in clinical routine to facilitate kidney allocation, optimize donor-recipient matching, and enhance long-term graft survival.