This retrospective single-center cohort included adult recipients of deceased-donor kidney transplants performed between 2013 and 2017. Overall, 3058 recipients were analyzed; 244 received kidneys from donors with diabetes mellitus. Compared with non-diabetic donors, donors with diabetes were older, had a higher prevalence of hypertension, and presented with a higher Kidney Donor Profile Index (KDPI). These baseline differences informed subsequent analyses designed to account for confounding.
In unadjusted analyses, recipients of kidneys from diabetic donors demonstrated lower longitudinal estimated glomerular filtration rate (eGFR) over the 5-year follow-up and worse death-censored graft survival compared with recipients of kidneys from non-diabetic donors. These unadjusted comparisons reflect the raw outcomes in the cohort prior to accounting for the many baseline differences between donor groups, including donor age, comorbidities, and KDPI.
To address imbalances in baseline donor and recipient characteristics, the investigators performed 1:2 propensity score matching. After matching, differences between groups were substantially reduced. In the matched sample, mean eGFR from years 1 to 5, the eGFR slope across follow-up, and death-censored graft survival did not differ significantly between recipients of kidneys from diabetic versus non-diabetic donors. These adjusted results indicate that the initial unadjusted disparities were largely attributable to baseline characteristic differences rather than donor diabetes per se.
The study also assessed whether recipient diabetes modified the association between donor diabetes and outcomes. The analysis did not find statistically significant effect modification: recipient diabetes did not meaningfully change the relationship between donor diabetes and either graft survival or longitudinal eGFR. The authors note that the subgroup with both donor and recipient diabetes (D+/R+) was small (n = 39), limiting precision for conclusions specific to that subgroup.
A secondary exploratory clinicopathological analysis evaluated 164 available pre-implantation biopsy reports from accepted diabetic donor kidneys. In these reports, lesions specifically attributable to donor diabetes in glomeruli were uncommon. Chronic histological changes were predominantly vascular in distribution and were graded as mild to moderate. The investigators applied Banff chronicity parameters to these pre-implantation biopsies; however, no single Banff parameter independently predicted a 5-year outcome defined either as eGFR <30 mL/min/1.73 m2 or death-censored graft loss. This finding was reported as part of an exploratory analysis and does not imply that histology lacks prognostic value in other settings, but in this dataset individual Banff features did not provide independent 5-year prognostic information.
In this single-center cohort, kidneys from deceased donors with diabetes were initially associated with lower 5-year eGFR and worse death-censored graft survival on unadjusted analyses. After balancing donor and recipient baseline characteristics using propensity score matching, donor diabetes was not associated with significantly lower 5-year kidney function or death-censored graft survival. The secondary pathological review of pre-implantation biopsies from diabetic donors revealed uncommon diabetes-specific glomerular lesions and primarily mild-to-moderate vascular chronicity; Banff chronicity parameters did not independently predict 5-year outcomes in this cohort.
Taken together, these findings suggest that donor diabetes alone—when considered in the context of other donor and recipient characteristics—may not independently determine 5-year transplant outcomes at this center. This could inform decision-making around the acceptance of kidneys from donors with diabetes, emphasizing individualized assessment and adjustment for other risk factors.
The analysis is retrospective and single-center, which limits generalizability. Although propensity score matching reduced measured confounding, unmeasured confounders may remain. The exploratory pathology analysis included 164 pre-implantation biopsy reports from diabetic donors; details beyond the summary provided in the source (for example, specific Banff scores and their distributions) were not reported in the source material excerpt. The small size of the D+/R+ subgroup (n = 39) reduces power to detect interactions by combined donor-recipient diabetes status. The source reports no conflicts of interest.
Overall, in this cohort of 3058 kidney transplant recipients from 2013–2017, donor diabetes was associated with worse unadjusted outcomes but was not independently associated with lower 5-year eGFR or death-censored graft survival after adjustment; pre-implantation Banff chronicity parameters did not independently predict 5-year outcomes in the examined biopsies.