Analysis of Discrepancy between Expected and Actual Donor-Recipient Matching in Heart Transplantation

Abstract

Purpose We hypothesised that for any donor, there is a significant discrepancy, driven by logistical variables, between the expected and actual recipient of a heart transplant (HTx). Methods 51 deidentified donor information sheets and waiting lists (comprising 1488 entries and 93 patients) from August 2017-2018 at St Vincent's Hospital Sydney (SVHS) were given to HTx surgeons and physicians, who ranked their top 10 recipients for transplant and variables relating to HTx by importance. Two Cardiac Allocation Scores (CASs) were used to rank recipients, filtered by ABO group, then weight/predicted heart mass, and adjusted for immunological cross-matching. Three ranker groups gave “expected” matches and rankings - CASs, surgeons and physicians. Inter-rater reliability on variable and waitlisted candidate rankings was analysed with Krippendorff's α, and pairwise inter-rater agreement on candidate rankings with Rank Biased Overlap (RBO). Logistical factors were correlated to a ranker's top candidate being transplanted with univariate and stepwise multivariate logistic and linear regression. Characteristics of top-ranked and actual recipients were compared with the t, Wilcoxon rank sum and χ2 tests. Results Across all rankers, expected and actual recipients matched in 25.66% of cases on average (31.38% for surgeons and physicians). Regarding rankings of HTx candidates and variables of importance in HTx, inter-rater reliability and agreement between rankers, ranged from low to moderate (αvariables=0.22-0.69, RBO=0.23-0.61; 1=total agreement) within and between groups. Flight as mode of transport was consistently significantly correlated with concordance of expected and actual recipient across multiple rankers (p= Conclusion At SVHS, discrepancy exists between actual and expected donor-recipient matching in HTx for doctors and CASs. Transplant logistics and individual practice variations contribute to this.