Diabetic Donor Kidneys: Five-Year Outcomes After Matching
Key Takeaways
- Among adult deceased-donor kidney recipients at a São Paulo transplant center, diabetic donors were older and more often hypertensive, and their recipients had lower unadjusted graft survival.
- Matched kidney-function and graft-survival differences by donor diabetes status were not statistically significant.
- Recipient diabetes showed no statistically significant interaction with donor diabetes, although jointly diabetic cases were limited.
- Selected biopsy reports showed predominantly vascular rather than diabetic glomerular chronic lesions, and no individual Banff chronicity measure independently predicted the specified outcomes.
At Hospital do Rim in São Paulo, Brazil, investigators retrospectively assessed 3,058 adults aged at least 18 years who received deceased-donor kidneys from January 2013 through December 2017, with up to five years of follow-up. Combined-organ recipients and those lost to follow-up in the first month were excluded. The cohort included 244 diabetic-donor recipients, whom investigators matched 1:2 to 488 nondiabetic-donor recipients by donor age, hypertension, cause of brain death, and human leukocyte antigen DR (HLA-DR) compatibility. Donor diabetes was distinct from recipient diabetes, which was examined in adjusted full-cohort interaction tests. Annually assessed estimated glomerular filtration rate (eGFR) through year five was the primary outcome; death-censored graft survival was secondary. A selected pathology analysis used 164 available preimplantation reports from 238 biopsied transplanted diabetic-donor kidneys to assess year-five eGFR below 30 mL/min/1.73 m² and death-censored graft loss.
Diabetic donors were older, more often hypertensive, and had higher kidney donor profile index (KDPI) values. Full-cohort five-year death-censored graft survival was 76.1% with diabetic donors versus 85% with nondiabetic donors (log-rank p < 0.001). After matching, diabetic-donor survival remained at that level, while nondiabetic-donor survival was 78.9%; the difference was not statistically significant. Matching balanced selected covariates and most baseline features, although KDPI and kidney donor risk index (KDRI) remained higher for diabetic donors.
In the full cohort, year-five mean eGFR was 30.3 versus 39.8 mL/min/1.73 m² for diabetic versus nondiabetic donors. The matched overall adjusted mean eGFR difference was −2.30 mL/min/1.73 m² (95% CI −5.37 to 0.78). Matched function, year-five function, and longitudinal slope did not differ significantly; declines from year one through five were similar even before matching. Adjusted donor-by-recipient diabetes tests detected no significant interaction for graft loss or eGFR. In selected biopsies, diabetic glomerular changes were uncommon, vascular lesions were more common, and no individual Banff chronicity measure independently predicted either specified five-year endpoint.
These retrospective, single-center comparisons concern accepted kidneys in Brazil, not a causal effect of donor diabetes or proof of equivalent outcomes. Residual confounding remains possible, particularly given the higher KDPI and KDRI after matching. Available biopsy reports represent selected transplanted kidneys, not all recovered or offered kidneys; those with advanced histologic injury were less likely to be transplanted. No stand-alone biopsy acceptance cutoff was validated. Limited jointly diabetic cases and biopsy outcome events constrain what nonsignificant interaction and pathology findings can exclude. Donor diabetes duration, metabolic markers, systematic posttransplant biopsies, and longitudinal glycemic measures were unavailable for assessing lesion evolution.
According to the authors, function and graft-survival gaps apparent before matching among kidneys accepted from diabetic donors at this center were no longer statistically significant after matching, without evidence of a steeper eGFR decline. Available biopsy lesions yielded no independent individual Banff prognostic signal for the specified outcomes.
Clinician Questions
How many graft losses occurred when both kidney donor and recipient had diabetes?
In the adjusted recipient-diabetes interaction analysis of the São Paulo deceased-donor kidney transplant cohort, 39 recipients had both donor and recipient diabetes, with six death-censored graft losses. The limited events constrain what the nonsignificant interaction can exclude about modification by recipient diabetes.
How was eGFR handled after graft failure, death, or loss to follow-up in the São Paulo kidney transplant cohort?
For five-year eGFR trajectories among deceased-donor kidney recipients at Hospital do Rim, measurements after graft failure were set to zero; missing values following death or loss to follow-up used the last observation carried forward.