Background: Antibodies to actin, chromatin, soluble liver antigen/liver pancreas and liver cytosol type 1 have been ascribed prognostic value in autoimmune hepatitis.
Aim: Evaluate the performance parameters of these nonstandard autoantibodies and determine the critical battery for clinical application.
Methods: All antibodies were tested concurrently by enzyme immunoassay in 106 patients who had reached a treatment outcome. Tests were repeated in 149 serum samples obtained later to assess durability of the findings.
Results: Antibodies to chromatin and soluble liver antigen/liver pancreas were superior to the other markers in predicting relapse. Patients with antibodies to chromatin and/or soluble liver antigen/liver pancreas relapsed more frequently than patients without these markers (100 versus 79%, p < 0.0003). Maximum sensitivity and predictability for relapse required combined testing, and they were 54 and 60%, respectively. Antibody status remained stable in 60% of patients during 127 +/- 9 months of follow-up, and antibodies to soluble liver antigen/liver pancreas were less labile than antibodies to chromatin (frequency of status change, 4 versus 22%). None of the antibodies were associated with treatment failure, death from hepatic failure or requirement for liver transplantation.
Conclusions: Antibodies to chromatin and soluble liver antigen/liver pancreas are associated with relapse after corticosteroid withdrawal, and they may be useful prognostic markers. Combined testing improves but does not eliminate deficiencies in sensitivity, predictability and durability.