HLAMatchmaker determines HLA compatibility at the level of polymorphic amino acid triplets in antibody-accessible sequence positions. Recent studies have shown that among HLA-DR-matched kidney transplants, the HLA-A,B antigen mismatches which are compatible at the triplet level have almost identical graft survival rates as the zero-HLA-A,B antigen mismatches. This finding provides the basis of a new strategy to identify HLA-mismatched organs that have similar success rates as the zero-HLA-antigen mismatches. This report describes how in conjunction with the Acceptable Mismatch program in Eurotransplant, HLAMatchmaker can expand the pool of potential donors for highly sensitized patients, for whom it is very difficult to find a compatible transplant. Sera from 35 highly sensitized kidney transplant candidates with an average PRA of 96% were screened by lymphocytotoxicity with HLAtyped panels that included cells that were selectively mismatched for one or two HLA antigens for each patient. Acceptable and unacceptable HLA-A,B antigen mismatches were determined from the serum reactivity with the cell panel. HLAMatchmaker analysis was applied to identify additional HLA class I antigens that were matched at the triplet level. For each patient, we calculated the probability of finding a donor (PFD) in the different match categories from HLA gene frequencies in the kidney donor population. The median PFD for a zero-antigen mismatch was 0.025%. Matching at the triplet level increased the median PFD to 0.037% (P = 0.008). The median PFD was 0.058% for a 0-1-triplet mismatch and 0.226% for a 0-2-triplet mismatch. Serum screening identified acceptable antigen mismatches for 28 of 35 highly sensitized patients, and the median PFD increased to 0.307% for a zero/ acceptable antigen mismatch. The application of HLAMatchmaker permitted for 33 patients (or 92%) the identification of additional antigens that were acceptable at the triplet level, and the median PFD for a zero/acceptable triplet mismatch went up to 0.425%. Inclusion of one-triplet mismatches increased the median PFD to 1.112%. Validation studies have shown that patient sera reacted with none of the zero-tripletmismatched antigens, 8-13% of the one-triplet mismatches, and 12-19% of the two-triplet mismatches. Although most antigens with one or two mismatched triplets appear acceptable to highly sensitized patients, a serum analysis must ascertain that the patient's antibodies do not recognize such mismatched triplets. HLAMatchmaker offers a useful strategy of identifying more 23 donors with acceptable HLA mismatches and could alleviate the problem of accumulation of highly sensitized patients on the transplant waiting list.