A quantitative antiglobulin consumption technique has been used to measure antiplatelet antibody (platelet-bound IgG) in 25 patients with idiopathic thrombocytopenic purpura (ITP). The assay confirmed the presumptive clinical diagnosis in all patients. Additional observations however, have demonstrated that if thrombocytopenic patients have platelet counts greater than 75 x 10(9)l, false-negative results may be obtained. Antiplatelet antibody levels closely mirrored the clinical progress of patients receiving therapy, but the level of antibody at the time of diagnosis was not of value in predicting the response to corticosteroid treatment. In a proportion of patients receiving corticosteroids, platelet counts returned to normal before antibody levels, implying a reduction in the sequestration of antibody-sensitized platelets within the reticulo-endothelial system. Because this response was observed in proportionately more patients in whom drug therapy was successful than others, this steroid effect may in some patients be an important factor in determining the successful outcome of medical treatment. Four out of 10 patients in clinical remission had evidence of disease activity based on platelet antibody and platelet survival measurements. The pattern of results obtained demonstrates that disease activity may go unrecognized in some patients unless both measurements are performed.