BACKGROUND AND OBJECTIVE: Flow cytometry (FC) to identify platelet-associated (PA) immunoglobulin (Ig) is a potentially useful diagnostic test for idiopathic thrombocytopenic purpura (ITP). However, the restricted application of PAIg measurement to thrombocytopenic populations primarily comprised of ITP patients will artificially enhance the test's diagnostic specificity. For this reason, we performed a prospective study in which the results of a sensitive technique for detecting PAIg, as is FC, were correlated to the cause of the thrombocytopenia. DESIGN AND METHODS: A total of 118 patients with platelet counts <100 x 10(9)/L and 30 normal donors with a platelet count >200 x 10(9)/L were studied for PAIg employing a flow cytometer. Forty-two children and 20 adults were diagnosed as having immune thrombocytopenia and 27 children and 29 adults had nonimmune thrombocytopenia of different etiology. RESULTS: Raised levels of PAIg were found in 56/62 patients with immune thrombocytopenia and in 34/56 patients with non-immune thrombocytopenia. Diagnostic values of PAIg for the detection of immune thrombocytopenia were: sensitivity 90.3% and specificity 39. 3%. An enzyme-linked immunoabsorbant assay (ELISA) for the detection of autoantibodies to platelet glycoprotein (GP) complexes was used in adults, 9 with immune-related thrombocytopenia and 16 with non-immune thrombocytopenia, in order to determine the true non-specific nature of the positive PAIg test. By ELISA, 8/9 patients with immune thrombocytopenia and 7/16 with non-immune thrombocytopenic disorders showed autoantibodies to platelet GP complexes. INTERPRETATION AND CONCLUSIONS: PAIg detection by FC constitutes a sensitive but non-specific assay thus making it unnecessary and inappropriate for establishing the diagnosis of ITP.