Tissue transglutaminase antibodies in celiac disease: Assessment of a commercial kit

OBJECTIVE: Tissue transglutaminase was identified as the autoantigen elicit ing endomysial antibody. A homemade enzyme-linked immunosorbent assay (ELIS A)-based test was recently developed to determine quantitative titers of Ig A antitissue transglutaminase antibody. Our objective in this study was to assess the suitability of a newly developed commercial kit for quantitative determination of antibody in patients with untreated celiac disease. MATERIALS: We tested serum samples from 79 untreated celiac patients, 42 he althy blood donors, and 18 patients with nonceliac intestinal disorders eva luated in two different centers. Samples were tested for antitissue transgl utaminase, and antiendomysial and antigliadin antibodies in the center wher e diagnosis was performed. To assess interlaboratory variability of methods , 24 samples randomly selected were blindly tested in both centers. Antitis sue transglutaminase antibodies were determined using a commercial kit (INO VA Diagnostics, Inc., San Diego, CA). RESULTS: Untreated celiac patients had significantly higher titers of antit issue transglutaminase than healthy and disease controls (p < 0.00001). Acc ording to the cut-off provided by the manufacturers (20 AU/mL), overall sen sitivity was 92% (85% for one center and 100% for the other) and specificit y was 98% (100% and 95%, respectively). Antiendomysial antibody was 86% sen sitive and 100% specific. Discordance between antitissue transglutaminase a nd antiendomysial antibodies was detected in 13% of patients. Although two antitissue transglutaminase-negative cases had a positive antiendomysial an tibody, the inverse situation was found in eight cases. A blind determinati on of antitissue transglutaminase on the same samples evidenced a good agre ement (kappa statistic: 0.66) between both centers when assessment was qual itative (based on the decision of positive or negative). Although correlati on of titers for both determinations was highly significant (r: 0.902, p < 0.00001), a very wide interlaboratory variability (median: 50%) was detecte d when absolute values were considered. CONCLUSIONS: The quantitative determination of antitissue transglutaminase using a commercial kit was highly sensitive and specific for detection of c eliac disease. We observed an incomplete overlapping with antiendomysial an tibody. The very high variability of values between laboratories still rema ins to be salved so as to propose the commercial ELISA assay for the screen ing of celiac disease. (Am J Gastroenterol 2000;95:2318-2322. (C) 2000 by A m. Cell. of Gastroenterology).