The results obtained in the analysis of 130 thrombocytopenic patients with a radioimmunoassay (RIA) for platelet-associated IgG (PA-IgG) and the platelet suspension immunofluorescence test (PIFT) were compared. The RIA was positive in 33 of 41 (82.9%) patients with idiopathic thrombocytopenia (ITP) and in 51 of 79 (64.4%) patients with secondary thrombocytopenia (STP). The PIFT was positive in 37 of the 41 (90.2%) ITP patients and in 57 of the 79 (72.2%) STP patients. Sensitivity and specificity for the diagnosis of ITP of both tests were comparable: 82.9 and 40.9% for the PA-IgG(RIA) and 90.2 and 36.7% for the PIFT. A significant positive correlation was observed between the mean amount of PA-IgG measured and the height of PIFT scores with anti-IgG. Of 38 discrepancies between PA-IgG(RIA) and PIFT with anti-IgG, 15 were due to borderline results, 17 were associated with abnormal platelet-size distribution and 20 were associated with occurrence of IgM antibodies. These results suggest influences of platelet fragments and/or aggregates on accurate measurement of PA-IgG. Both fragments and aggregates escape from accurate platelet counting, while their contribution to the total IgG content remains. Therefore, a falsely elevated PA-IgG (RIA) may be measured.