Performance Parameters of the Conventional Serological Markers for Autoimmune Hepatitis

Abstract

Autoimmune hepatitis is defined by a conventional battery of autoantibodies that may also be present in other liver diseases. To define the performance parameters of the conventional autoantibodies for autoimmune hepatitis, determine the diagnostic implications of simultaneous autoantibody production, and assess the performances of serological assays based on indirect immunofluorescence and enzyme immunoassay. In this study, 265 adults satisfying codified criteria for autoimmune hepatitis and 342 adults who satisfied conventional diagnostic criteria for another chronic liver disease were each assessed for the conventional autoantibodies. Antinuclear antibodies, smooth muscle antibodies, and antibodies to liver kidney type 1 had sensitivities of only 32, 16, and 1%, respectively, for the diagnosis of autoimmune hepatitis, and their diagnostic accuracy ranged from 56 to 61%. The combination of antinuclear antibodies and smooth muscle antibodies at presentation had superior sensitivity (43%), specificity (99%), positive (97%) and negative (69%) predictabilities, and diagnostic accuracy (74%) than each marker alone. The occurrence of multiple autoantibodies was lower in other chronic liver diseases than in autoimmune hepatitis (8% versus 51%, p < 0.000001). The enzyme immunoassay for antinuclear antibodies had an accuracy that was not discriminative for autoimmune hepatitis (p = 0.06). Isolated autoantibodies have a low diagnostic accuracy for autoimmune hepatitis. Concurrent antinuclear antibodies and smooth muscle antibodies have superior performance parameters. The enzyme immunoassay for antinuclear antibodies performs less well than the assay based on indirect immunofluorescence.