In 1969, a study by Patel and Terasaki persuaded the renal transplant community that a pre-transplant cross-match should always be performed between donor and recipient to detect HLA antibodies and prevent hyperacute allograft rejection. Although the role of the cross-match among nonsensitized patients is controversial, its importance among sensitized recipients is undeniable. Over the past 30 years, more sensitive techniques, such as the flow cytometric cross-match (FCXM), were developed to identify low levels of antibodies undetectable by other approaches. The clinical relevance of a positive FCXM, however, has been hotly disputed, with some investigators maintaining that the FCXM is 'too sensitive' and rules out acceptable donor-recipient combinations. An alternative explanation is that the FCXM is non-specific, and, at least in certain situations, identifies non-HLA antibodies that are clinically irrelevant. Recently, a solid phase immunoassay utilizing purified HLA Class I or Class II molecules bound to microparticles (FlowPRA) was developed. Ideally, use of the FlowPRA for the identification of HLA antibodies in recipient sera would help ascertain whether a positive FCXM with donor cells was truly the result of an HLA-specific antibody. As shown here, this may not always be true. In this study, two unexpected serum patterns were observed. Pattern 1: FlowPRA beads were positive (with an associated HLA Class I specificity) and the FCXM with cells expressing the HLA antigen(s) to which the antibody was directed, was negative. Sequence analysis of the HLA antigens reactive with this unexpected antibody suggests that the epitope recognized resides on the floor of the groove, a site generally not expected to generate antibody activity. Pattern 2: FlowPRA beads were negative yet the FCXM was T and B cell positive. Further analysis of the FlowPRA negative/FCXM positive sera using a flow cytometric cell-based panel reactive antibody (PRA) approach revealed those sera to have specific anti-HLA Class I activity. We suspect that both types of antibodies described above have clinical relevance. Thus, a negative or positive FCXM (when the FlowPRA against donor antigens is positive or negative, respectively) is not always a straightforward interpretation.