Abstract
BACKGROUND: Diabetes mellitus is increasingly common among deceased donors and may signal donor-derived kidney injury that affects post-transplant outcomes. OBJECTIVE: To evaluate whether donor proteinuria is associated with graft and patient outcomes after kidney transplantation from deceased donors with diabetes. DESIGN: Retrospective cohort study using a national transplant registry. SETTING: United States; Scientific Registry of Transplant Recipients (SRTR), February 28, 2013-2023. PATIENTS: 9486 kidney-alone transplant recipients from deceased donors with diabetes in whom a pre-implantation (procurement) biopsy was performed and donor proteinuria status was available. MEASUREMENTS: Primary outcome: death-censored graft failure (DCGF). Secondary outcomes: all-cause graft failure (ACGF), death with graft function (DWGF), and delayed graft function (DGF). Exposure: donor proteinuria (present vs absent). METHODS: Kaplan-Meier analyses and multivariable Cox models (a priori covariables from a directed acyclic graph) assessed associations between donor proteinuria and time-to-event outcomes. Because proportional hazards were violated for DCGF, analyses were performed in two periods: an "early" cohort up to 2.5 years post-transplant and a landmarked cohort of recipients with functioning grafts at 2.5 years. Logistic regression evaluated DGF. Sensitivity analyses adjusted for donor insulin dependence (proxy for diabetes severity) and recipient characteristics; exploratory effect modification by biopsy glomerulosclerosis (GS) was assessed. RESULTS: Donor proteinuria was present in 54.9% of cases. In adjusted Cox models, donor proteinuria was not associated with early DCGF (<2.5 years; HR 1.14, 95% CI: 0.99, 1.32) but was associated with increased risk of late DCGF >2.5 years post-transplant (HR 1.36, 95% CI: 1.15, 1.62), with similar findings for ACGF. No associations were observed with DWGF or DGF. Results were consistent after adjustment for donor insulin dependence as a proxy for severity and recipient factors including diabetes status. The association between proteinuria and late graft failure was more pronounced in kidneys with lower GS, suggesting proteinuria may reflect chronic injury not well-captured by biopsy. LIMITATIONS: Observational design with potential residual confounding. Because the cohort includes only kidneys that were actually transplanted, findings reflect outcomes among accepted organs and are not intended to guide offer acceptance or decline decisions. Donor proteinuria was recorded only as present or absent, without standardized measurement. This may have led to misclassification, prevented assessment of dose-response relationships, and likely made it harder to detect true associations. Registry constraints limited histologic detail beyond GS. CONCLUSIONS: Among diabetic deceased donors, the presence of proteinuria is a time-dependent marker of increased long-term graft-failure risk, complementing biopsy and clinical data. Standardized, quantitative proteinuria assessment may improve risk stratification and post-transplant management while supporting judicious utilization of diabetic donor kidneys.