Purpose: Most pediatric kidney transplant (KT) recipients eventually require retransplantation within their lifetime. It is unclear whether primary transplantation from the candidate's living donor (LD) followed by an eventual deceased donor (DD) is advantageous over primary transplantation from a DD followed by the LD, potentially years later.
Method: The decision of the order of LD and DD transplantation was modeled in a patient-oriented Markov decision process model. Pediatric KT candidates can be in one of five states in the model; on the waitlist, transplanted from their LD, transplanted from their DD, deceased, or post 2nd graft-failure. The living donor can be in one of two states in the model: eligible or ineligible to donate. The model includes a period of high-risk of graft loss during the adolescent years, sensitization following primary transplant, and risk of the living donor developing conditions that preclude donation. Cox models of waitlist, graft, and patient survival were based on data from Scientific Registry of Transplant Recipients (SRTR). Progression of sensitization, as measured by panel reactive antibody (PRA), following primary transplantation was modeled through ordered logistic regression based also on SRTR data. The risk of the living donor dying or developing conditions precluding kidney donation was modeled exponentially from the Atherosclerosis Risk in Communities and Coronary Artery Risk Development in Young Adults cohort studies.
Result: The Markov model simulates the recipient's survival for the subsequent 20 years after their decision to receive transplantation from their LD first or wait for a DD first. If the recipient chooses LD as their primary transplant, they immediately receive the transplant and start in the model in the LD-transplanted state. If the recipient requires retransplantation, they re-enter the waitlist for a DD; if however, the recipient chooses DD as their primary transplant, they spend a specified amount of time on the waitlist prior to receiving their primary transplant from a DD; if the recipient then requires retransplantation, they can immediately be retransplanted from their LD, provided the LD is still eligible to donate. If the LD is ineligible, the recipient then waits for a second DD transplant.
Conclusion: Given the patient's and donor's characteristics, the Markov model provides valuable patient-specific guidance on when to best utilize one's living donor graft.
