What Is the Prevalence of Celiac Disease Among U.S. Patients with Autoimmune Hepatitis?

Abstract

Purpose: Autoimmune hepatitis (AIH) is an insidious disorder of unknown etiology, characterized by elevated serum aminotransferase and immunoglobulin G concentrations, circulating autoantibodies, lymphoplasmacytic portal and interface hepatitis and responsiveness to prednisone. Most affected European and North American AIH patients possess either the DR3 or DR4 haplotype. The hallmarks of celiac disease (CD) are gluten intolerance, and the presence of either the DR3 and/or DQ2 haplotypes. The estimated prevalence of CD in the general U.S. population is 1:100, whereas the prevalence of CD in European subjects with AIH has been estimated at between 1:16 and 1:36 in two separate studies. Therefore, we proposed to determine the prevalence of CD among patients with AIH in the U.S. Methods: We studied patients aged ≥ 18 years with a diagnosis of AIH conforming to International Autoimmune Hepatitis Group criteria, presenting at a single U.S. center between September 2003 and May 2008. Patients' sera were tested for IgA endomysial antibody (EMA) by indirect immunofluorescence; their serum IgA concentration was measured concomitantly. Patients with selective IgA deficiency (SIgAD) were tested for serum IgG gliadin antibody (AGA) by enzyme immunoassay. Results: The sample comprised 153 patients of whom 22 were male (14.4%), with a median age of 50 years (range 18–81). The ethnic distribution was: 115 (75.2%) Caucasian, 33 (21.6%) African American non-Hispanic, two (1.3%) native American, two (1.3%) Asian, and one (0.6%) Caucasian Hispanic. One patient had a positive EMA result (duodenal mucosal biopsies were consistent with CD). One man had selective IgA deficiency, but his AGA IgG titer was normal (5.1 units: normal 0–25). Five women underwent endoscopic duodenal mucosal biopsy despite negative EMA results (four had iron deficiency anemia, the other loose stools and weight loss). One patient demonstrated histological findings consistent with CD, whereas the other four demonstrated normal histology. Conclusion: Although the prevalence of CD in these U.S. patients with AIH (1:76 [1.3%]) is considerably less than reported among European AIH patients, it is slightly greater than in the general U.S. population. This suggests the possibility of increased risk for CD among U.S. patients with AIH, and therefore appropriate serologic screening. Conversely, the study sample may have been underpowered, thereby resulting in findings which suggest a difference that does not truly exist. In either case, it appears that EMA may not be the most appropriate screening test for CD in AIH patients: only one of the two patients identified with CD in this group yielded a positive antibody result.