Virtual Crossmatch Approach to Maximize Matching in Paired Kidney Donation

We developed and tested a new computer program to match maximal sets of incompatible live donor/recipient pairs from a national paired kidney donation (PKD) registry. Data of 32 incompatible pairs included ABO and 4 digit-high-resolution donor and recipient HLA antigens and recipient's HLA antibodies. Three test runs were compared, in which donors were excluded from matching to recipients with either donor-specific antibodies (DSA) > 8000MFI (mean fluorescent intensity) at low-resolution (Run 1) or > 8000MFI at high-resolution (Run 2) or > 2000MFI and high-resolution (Run 3). Run 1 identified 22 703 possible combinations, with 20 pairs in the top ranked, Run 2 identified 24 113 combinations, with 19 pairs in the top ranked and Run 3 identified 8843 combinations, with 17 pairs in the top ranked. Review of DSA in Run 1 revealed that six recipients had DSA 2000-8000MFI causing a possible positive crossmatch resulting in breakdown of two 3-way and three 2-way chains. In Run 2, four recipients had DSA 2000-8000MFI, also potentially causing breakdown of three 2-way chains. The more prudent approach of excluding from matching recipients with DSA with > 2000MFI reduces the probability of matched pairs having a positive crossmatch without significantly decreasing the number of possible transplants.