Coeliac Patients Are Undiagnosed at Routine Upper Endoscopy

BACKGROUND AND AIMS Two out of three patients with Coeliac Disease (CD) in Australia are undiagnosed. This prospective clinical audit aimed to determine how many CD patients would be undiagnosed if duodenal biopsy had only been performed if the mucosa looked abnormal or the patient presented with typical CD symptoms. METHODS All eligible patients presenting for upper gastrointestinal endoscopy (OGD) in a regional center from 2004-2009 underwent prospective analysis of presenting symptoms and duodenal biopsy. Clinical presentations were defined as either Major (diarrhea, weight loss, iron deficiency, CD family history or positive celiac antibodies- Ab) or Minor Clinical Indicators (CI) to duodenal biopsy (atypical symptoms). Newly diagnosed CD patients had follow up celiac antibody testing. RESULTS Thirty-five (1.4%) new cases of CD were identified in the 2,559 patients biopsied at upper endoscopy. Almost a quarter (23%) of cases presented with atypical symptoms. There was an inverse relationship between presentation with Major CI's and increasing age (<16, 16-59 and >60: 100%, 81% and 50% respectively, p = 0.03); 28% of newly diagnosed CD patients were aged over 60 years. Endoscopic appearance was a useful diagnostic tool in only 51% (18/35) of CD patients. Coeliac antibodies were positive in 34/35 CD patients (sensitivity 97%). CONCLUSIONS Almost one quarter of new cases of CD presented with atypical symptoms and half of the new cases had unremarkable duodenal mucosa. At least 10% of new cases of celiac disease are likely to be undiagnosed at routine upper endoscopy, particularly patients over 60 years who more commonly present atypically. All new CD patients could be identified in this study by performing pre-operative celiac antibody testing on all patients presenting for OGD and proceeding to biopsy only positive antibody patients and those presenting with either Major CI or abnormal duodenal mucosa for an estimated cost of AUS$4,629 and AUS$3,710 respectively.