In patients suffering from ankylosing spondylitis (AS) HLA-B27 determination by means of the microlymphocytotoxicity test (MLCT) sometimes gives equivocal or false-negative results even though it has been performed with meticulous care. These failures of the test did not arise when the isolated mononuclear cells (MNC) were incubated in lymphocyte culture medium at 37 degrees C under sterile conditions for 24 h. To objectify these observations two methods of HLA class I typing were implemented before and after incubation of the test MNC in culture medium: a bioluminescence method based on the loss of adenosine triphosphate (ATP) in lysed cells in a modification of the usual MLCT and a flow cytometric (FC) test using direct immunofluorescence with an anti-HLA-B27 monoclonal antibody (MAB). In this study 50 patients with AS and 12 healthy volunteers were typed by the usual MLCT according to the NIH standard method and with both of the quantitative methods. In most of the AS patients the discrimination between positive and negative typing results became more distinct after 24 h incubation in culture medium. In the entire group of AS patients tested three false-negative typing results were prevented by this method. Although the MAB against HLA-B27 is cross-reactive with HLA-B7 and HLA-B22, errors in the FC analysis could be avoided by calibration of the flow cytometer with standard calibration beads. Possible explanations for masking of the HLA-B27 in AS patients are discussed.