K DANSIE1,2, P CLAYTON1,2,3
1Australia and New Zealand Dialysis and Transplant Registry (ANZDATA), Adelaide, South Australia; 2University of Adelaide, Adelaide, South Australia; 3Central Northern Adelaide Renal and Transplantation Service, Adelaide, South Australia
Aim: To evaluate the performance of the US and UK kidney donor risk indices (KDRIs) in predicting death-censored graft survival in the Australian and New Zealand (ANZ) kidney transplant population.
Background: Kidney donor risk indices have been developed in the US and UK to estimate graft survival, and the US KDRI is used in US kidney allocation. Prior to adopting such an index in the Australian or NZ kidney allocation system, it is essential to ensure that the index performs adequately in the Australian/NZ transplant context. Discrimination is the most important performance measure for scores designed to rank kidneys for allocation.
Methods: We included all adult DD kidney-only transplants performed in ANZ from 2006-2015 (n=5053). KDRI was calculated using both the US (kidney-only) and UK KDRI formulae. Three Cox models were constructed for each index (Model 1: KDRI only, Model 2: Model 1 + transplant characteristics, Model 3: Model 2 + recipient characteristics) and discrimination compared using Harrell’s C statistics.
Results: The US KDRI and UK KDRI were both strongly associated with death-censored graft survival (p<0.0001 in all models). In the KDRI only US and UK models, discrimination was moderately good with c-statistics of 0.63 and 0.60 respectively. Adjusting for transplant characteristics resulted in a marginal improvement of US KDRI to 0.66 and UK KDRI to 0.64. The addition of recipient characteristics again resulted in a marginal improvement of US KDRI to 0.70 and UK KDRI to 0.68.
Conclusions: The US and UK KDRI scores demonstrated moderate discrimination in the Australian/NZ DD transplant population, with the US score demonstrating slightly better performance in all models.