Background: Serum antibodies to tissue transglutaminase (tTGA) are reported to have high sensitivity and specificity for coeliac disease and to correlate closely with endomysial antibodies (EmA). We assessed their performance in a coeliac population with a high proportion of EmA-negative patients, who have been under-represented in previous studies.
Methods: We used a commercial ELISA kit to test for IgA class tTGA in sera from a population of 73 untreated coeliac patients with normal serum IgA and a high percentage (19%) EmA-negative, taking 58 patients with normal duodenal biopsies as controls. EmA was measured using indirect immunofluorescence.
Results: Forty-six (63%) patients with villous atrophy (VA) had both tTGA and EmA. However, when considered separately, sensitivities of tTGA and EmA for VA were similar (75% versus 81%) and both had high specificity (98% versus 97%). As 9 patients were tTGA-positive only and 13 had EmA only, selection of patients for biopsy on the presence of either antibody would have had a sensitivity of 93% (68 of 73), with 5 (7%) patients seronegative for both.
Conclusion: Although the ELISA tTGA assay is more convenient than EmA testing, it offers no advantages in sensitivity or specificity if used in isolation. However, incomplete concordance between EmA and tTGA positivity means that combination screening with both assays offers higher sensitivity, as almost a third of patients have only one antibody. As some coeliac patients with normal serum IgA are negative for both antibodies, biopsies should still be performed in seronegative individuals deemed at high risk for coeliac disease.