Login / Signup

A kidney discard decision strategy based on zero-time histology analysis could lead to an unjustified increase in the organ turndown rate among ECD.

Yosu LuqueMatthieu JammeOlivier AubertArthur RouxFrank MartinezLucile AmroucheClaire TinelLouise GalmicheJean-Paul Duong Van HuyenFrançois AudenetChristophe LegendreDany AnglicheauMarion Rabant
Published in: Transplant international : official journal of the European Society for Organ Transplantation (2021)
The utility of zero-time kidney biopsies (KB) in deciding to accept expanded criteria donor (ECD) kidneys remains controversial. However, zero-time histology is one of the main causes for discarding kidneys in the United States. In a single-centre study, we examined the utility and impact on outcome of the use of frozen section zero-time KB among ECD. Ninety-two zero-time KB were analysed for accept/discard decision between 2005 and 2015 among ECD. 53% of kidneys were rejected after zero-time KB analysis; there was no difference in individual clinical and biological data between accepted/rejected groups. However, histology of rejected kidneys showed more sclerotic glomeruli (20% vs. 8%; P < 0.001), increased interstitial fibrosis (1.25 ± 0.12 vs. 0.47 ± 0.09; P < 0.0001), more arteriosclerosis (2.14 ± 0.17 vs. 1.71 ± 0.11; P = 0.0032) and arteriolar hyalinosis (2.15 ± 0.12 vs. 1.55 ± 0.11; P = 0.0006). Using propensity score matching, we generated a group of 42 kidney allograft recipients who received a transplant matched for donor zero-time histology and clinical characteristics with donors whose kidneys were rejected. Interestingly, their 1- and 5-year graft survival and function were similar to the global cohort of ECD recipients. In conclusion, when performed, zero-time KB was a decisive element for kidney discard decision. However, adverse zero-time histology was not associated with poorer graft survival and kidney function among ECD.
Keyphrases
  • kidney transplantation
  • decision making
  • free survival
  • electronic health record
  • big data
  • machine learning
  • deep learning
  • adverse drug