Accuracy and learning curve and of the water-immersion technique in assessing marked villous atrophy of the duodenum: A single centre 4-year experience

Background: Upper endoscopy is not routinely performed to directly detect duodenal marked villous atrophy (MVA). With a prospective study, we investigated the accuracy, reliability and learning curve of the water-immersion technique for the assessment of MVA during routine upper endoscopy. Methods: A total of 9 endoscopists (who were previously trained to perform the immersion technique), from January 2003 to March 2006, routinely investigated the duodenum of 4,756 consecutive patients who were undergoing upper endoscopy. Among these, duodenal biopsies were requested for various reasons in 732 subjects. Duodenal villi were scored as present (including subjects with a normal or partial villous pattern) or absent (MVA) (including subjects with total or subtotal villous atrophy). Duodenal biopsies were performed when requested and when villi were judged as MVA at endoscopy. Using the Marsh's classification, villi were histologically scored as present (including normal or partial atrophy) or absent (MVA) (including total or subtotal atrophy). Results: On the whole, 105 (2.2%) patients with duodenal MVA were identified (91 among the 730 subjects in whom duodenal biopsies were expressively requested). The sensitivity, specificity, and positive and negative predictive values of the immersion technique for detection of MVA, irrespective of the endoscopist who performed the exam, was always 100%. All subjects with MVA were diagnosed as celiacs. Conclusions: The immersion technique may be regularly performed during the endoscopic exploration of the duodenum. This procedure is simple, feasible and can reliably detect MVA. This technique may avoid the need for biopsy in patients with MVA and could, therefore, reduce costs.