The purpose of this study was to investigate whether in highly sensitized patients (HSPs) the acceptable HLA-A and -B mismatches (AMs) can be predicted on the basis of patients' HLA-phenotype. To this affect, 1000 historical serum samples obtained from 50 HSPs (PRA > 60%), panel reactive antibodies (PRA) value and the specificity of class I anti-HLA-antibodies were detected by two techniques in parallel: An anti-human globulin augmented cytotoxicity (AHG-CDC) and an Elisa technique. Thereafter, class I HLA-antigens of the nonreactive cells in the screening panel and class I HLA-antigens of the patients were assigned to CREGs. The AMs for each one of the patients were detected using a separate cell panel, which was prepared in a way to include almost all the HLA-antigens belonging to the CREGs of the patients as well as to those of the nonreactive cells in the screening panel. It was found that the AMs in HSPs, detected with this protocol were more, compared to those we usually detect using only the HLA-antigens of the nonreactive cells in the screening panel (up to 8 versus 2-5). Both, the definitively detected AMs, and the HLA-specificities of the nonreactive cells of the screening panel belonged to the same CREGs. These CREGs were equivalent to the CREGs of class I HLA-phenotypes of each patient. The data presented in this paper introduce a new, rapid and easier way for the detection of AMs in HSPs. According to this proposed protocol, the assignment of patients' standard class I private HLA-phenotypes in CREGs, not only greatly facilitates the detection of AMs, but the detected AMs are also in fact significantly more than those determined by the conventional methodology. We have also confirmed that the majority of antibodies induced by HLA alloimmunization are directed against mismatched shared or public group epitopes CREGs. Moreover, we have confirmed that prospective matching for major CREGs would be feasible on a national level and would not significantly prolong waiting time, which could result in a significant augmentation of the potential donor pool.