Faculty Bibliography

BACKGROUND:A number of deceased donor kidney scoring systems have been developed to predict post-transplant graft failure. However, studies comparing the predictive ability of these scoring systems to each other are lacking. METHODS:We used single-center histopathologic and UNOS data from 140 marginal deceased donor kidneys and transplant recipients to compare the predictive accuracy of the Maryland Aggregate Pathology Index (MAPI), Kidney Donor Risk Index (KDRI), Remuzzi, and Nyberg scoring systems for 2-year graft survival using time-dependent receiver operating curves and Kaplan-Meier analysis. RESULTS:MAPI had the highest predictive accuracy (area under curve [AUC] = 0.81) compared to KDRI (AUC = 0.45), Remuzzi (AUC = 0.59), and Nyberg (AUC = 0.63) for 2-year graft survival. Furthermore, when analyzing each score according to its pre-defined risk strata, MAPI was the only scoring system for which 2-year graft survival was significantly different across strata (84.3% for low risk, 56.5% for intermediate risk, and 50% for high risk, P < .001). Additionally, MAPI was the only risk score significantly associated with 2-year graft survival (hazard ratio per point: 1.12, 95% confidence interval [CI]: 1.01-1.23, P = .03). CONCLUSIONS:In a single-center cohort of biopsied marginal kidneys used for transplantation, MAPI had the best predictive ability of these four scoring systems. When biopsy data are available for kidneys considered for transplantation, the MAPI score may provide additional information that could be used to better identify kidneys likely to have longer graft survival.