Attempt to improve the diagnosis of immune thrombocytopenia by combined use of two different platelet autoantibodies assays (PAIgG and MACE).

BACKGROUND AND OBJECTIVES Despite an extensive search for a definitive diagnostic assay for platelet autoantibodies, the laboratory diagnosis of immune thrombocytopenia (ITP) still remains a clinical challenge. Data in the literature have so far demonstrated that measurement of platelet-associated IgG (PAIgG) is sensitive, especially when flow cytometry is employed, but lacks adequate specificity. Measuring specific autoantibodies by antigen capture techniques increases specificity, but a large part of patients escape autoantibodies detection by such means too. The aim of the present study was to compare the diagnostic value of PAIgG with a modified antigen capture ELISA (MACE) in patients with primary and secondary immune thrombocytopenia and in patients with non-immune thrombocytopenia. DESIGN AND METHODS One hundred and four patients with a platelet count lower than 100x109/L were studied. Forty-two patients had primary ITP (P-ITP), 23 patients had ITP secondary to other immune diseases (S-ITP) and 39 patients had thrombocytopenia due to decreased platelet production (non-immune; NITP). PAIgG was measured by immunofluorescent flow cytometry, whereas specific platelet-associated autoantibodies (against GP IIb/IIIa, Ib/IX, Ia/Ia) were measured by a commercially available modified antigen capture assay (MACE, GTI, USA). RESULTS The sensitivity of the PAIgG assay for ITP was 60%, the specificity was 77%, the positive predictive value was 81% and the negative predictive value was 54%. The sensitivity of MACE was 60%, specificity was 97%, the positive predictive value 97% and the negative predictive value 59%. We found a 73% concordance between PAIgG and MACE assays. Both PAIgG and MACE had significantly greater sensitivity in S-ITP than in P-ITP. INTERPRETATION AND CONCLUSIONS Forty percent of patients with clinically diagnosed immune thrombocytopenia had no detectable platelet autoantibodies, possibly because of intrinsic methodological detection problems, different stages of disease, or absence of a true immune etiology.