Endoscopic markers of villous atrophy are not useful for the detection of celiac disease in patients with dyspeptic symptoms

Background and study aims Celiac disease can manifest with nonspecific symptoms, including functional gastrointestinal disorders such as dyspepsia. The aim of our study was to assess the usefulness of duodenal endoscopic markers of villous atrophy for the selection of dyspeptic patients for histological assessment. Patients and methods Esophagogastroduodenoscopy was performed in dyspeptic patients, in patients considered to be at risk of having celiac disease, and in healthy controls. At least three duodenal biopsies were performed for histological assessment of villous atrophy in all patients and controls. We looked for the following four duodenal endoscopic markers of celiac disease: reduction in the number of folds, scalloping of folds, mosaic-pattern mucosa, and nodular mucosa. Results A total of 175 people were enrolled (75 patients with dyspepsia; 75 patients who were "at risk" of having celiac disease; and 25 healthy volunteers, or "controls"). Of the dyspeptic patients, four had endoscopic markers of celiac disease with no histologically confirmed villous atrophy, while one patient without endoscopic markers was found to have Marsh type I villous atrophy. Of the patients at risk of having celiac disease, 16 had at least one endoscopic marker and 10/16 were found to have histological villous atrophy. In this group, the sensitivity and specificity of the endoscopic markers were 100 % and 90.8 % respectively. "At-risk" patients with two or more endoscopic markers all had histologically confirmed villous atrophy. Neither endoscopic markers nor villous atrophy were found in any of the control patients. Conclusions Additional endoscopic markers are valuable for diagnosis in patients with clinical symptoms suggestive of celiac disease. In contrast, endoscopic markers of villous atrophy are not useful for selecting a subgroup of dyspeptic patients for screening for celiac disease by duodenal histological assessment. These patients should be screened using other protocols.