The clinical utility of a flow cytometric assay (FCA) for intracellular HIV p24 antigen was evaluated in a group of HIV-1-infected subjects. A previously described method, p24-FCA (1), was modified to minimize nonspecific staining and to include irrelevant isotype-matched control antibodies. Blood mononuclear cells from 32 HIV-1 seropositive subjects and 14 HIV-1 seronegative controls were examined. In HIV-1 seropositive individuals, the proportion of CD4+ lymphocytes that bound the p24 monoclonal and the isotype control antibodies were not different. The frequency of cells binding p24 antibodies increased with declining CD4 counts and was highest in patients with low CD4+ lymphocyte counts. Although results of p24-FCA do reflect disease progression, the high nonspecific binding of monoclonal antibodies in infected subjects obscures specific p24 binding and precludes its use as an accurate measure of infection within single cells.