HLA matching by amino acid/positions in double and triple combinations

1. Began with the 2006 UNOS dataset and then selected for Caucasian donor/patients (post-1995 transplant and first transplant kidney only with full ABDR typing data). This was then split into cadaver donor (n = 20,542) and living donor (n = 21,890) populations. 2. Looked at every theoretical mismatch between donor and patients in the living donor population. Overall, 167 single, 4601 double, and 83,655 triple position/amino acid mismatches were included in the analysis. Two tallies for each mismatch were generated: "good" and "bad" based on patient's associated graft survival. 3. Generated a list of "bad" single, double, and triple position/ amino acid mismatches. 4. Looked at associated long-term graft survivals of patient/donor pairs with additive single, double, and triple "bad" mismatches in the cadaver donor population. 5. Survival curves of transplant pairs with associated mismatches were generated and compared with traditional standards of 0ABDR mismatch, 1 or more ABDR mismatch, and full 6ABDR mismatch survival curves. 6. The greatest effect was seen in re-grafted male patients where up to 50 double mismatches were associated with a 10-year survival 27% lower than 0ABDR mismatch. 7. Many of the "bad" position/amino acid mismatches identified were the same as those identified by generation of antibodies against Class I and Class II epitopes. This is evidence that antibody-defined epitopes and those functioning as transplantation epitopes influencing long-term graft survival are the same. 8. Overall, we are disappointed that position/ amino acid mismatches were not associated with more markedly lower survivals. It appears that current immunosuppression can largely negate the effect of mismatching for immunogenic epitopes.