Editorial: Endomysial Antibodies for a No-Biopsy Diagnosis of Coeliac Disease—The Jury Is Still out. Authors' Reply

We appreciate the editorial by Drs Shiha and Penny [1] on our study [2] concerning endomysial antibodies and clinical risk stratification for biopsy-sparing coeliac disease diagnosis. We acknowledge that endomysial antibodies present some drawbacks, including higher cost and need for trained personnel, potentially limiting their use. Our study's excellent endomysial antibody results [2] reflect extensive experience with this assay and may not generalise to less experienced centres. However, endomysial antibodies offer some advantages over tissue transglutaminase antibodies. First, endomysial antibodies exhibited 99.7% specificity for coeliac disease in a meta-analysis of 34 studies [3]. Second, while requiring skilled interpretation, endomysial antibodies are an objective measure, unlike tissue transglutaminase antibodies, which show significant inter-assay variability [4].

We agree that managing potential coeliac disease within a biopsy-sparing framework remains an open question depending on patients' clinical features. A meta-analysis of 17 studies found that nearly 90% of potential coeliac disease patients improve on a gluten-free diet [5]. Thus, in symptomatic, endomysial antibody-positive patients, a gluten-free diet is warranted, so distinguishing symptomatic potential coeliac disease from conventional coeliac disease is of limited importance. For the minority of asymptomatic adult potential coeliac disease patients identified via screening, the decision for a gluten-free diet versus monitoring on a gluten-containing diet remains case-by-case. However, in our experience, most adopt a gluten-free diet due to associated autoimmune conditions or family environment. Patients also generally prefer a gluten-free diet over endoscopic follow-up, rarely maintaining a gluten-containing diet long-term.

The issue is further complicated by ultra-short coeliac disease, which requires bulb biopsies to differentiate from potential coeliac disease [6]. However, bulb biopsies are not routinely performed, and obtaining ≥ 4 biopsies can already be considered an achievement [7]. A meta-analysis estimated ultra-short coeliac disease prevalence at 9.2% among coeliac disease patients, increasing with the number of bulb biopsies taken [8]. A large portion of ultra-short coeliac disease patients may be misdiagnosed as potential coeliac disease if only the second portion of the duodenum is biopsied. Collectively, these render distinguishing between potential coeliac disease, ultra-short coeliac disease, and other coeliac disease forms of questionable clinical significance.

Our findings on the safety of a biopsy-sparing approach are perhaps most crucial and generalise beyond specific serological tests. The lack of major concomitant pathology in coeliac disease patients diagnosed in our study [2] is reassuring, supporting the overall safety of avoiding biopsy, consistent with previous studies [9, 10]. Conversely, the high prevalence of non-coeliac enteropathies in patients ≥ 45 years old with alarm symptoms underscores the necessity of continued endoscopic evaluation in these patients [2].

In conclusion, despite some limitations, endomysial antibodies' advantages over tissue transglutaminase antibodies make them a valuable tool for biopsy-sparing coeliac disease diagnosis. Importantly, clinical features are important in selecting patients for a biopsy-sparing strategy, irrespective of chosen serological tests or cut-offs. Identifying a patient's underlying condition is the goal, beyond whether or not the patient has coeliac disease. Future research on biopsy-sparing coeliac disease diagnosis should incorporate clinical features, and comparative studies evaluating both tissue transglutaminase and endomysial antibodies are needed to assess their diagnostic performance in this context.