The Impact of Dropping the Requirements for Goblet Cells for the Diagnosis of Barrett's Esophagus (BE).
Maria Westerhoff, Lindsey Hovan, John Hart. University of Chicago, IL
Background: Currently the American Gastroenterological Association (AGA) requires two elements for the diagnosis of BE: endoscopic evidence of columnar mucosa within the esophagus, and the presence of goblet cells (GC) in biopsies from the columnar lined portion. In contrast, the British Society of Gastroenterology only requires that columnar mucosa be documented by biopsy, regardless of the presence of GC. The AGA is presently strongly considering dropping the requirement for GC.
Design: Our pathology database was searched from 1990-2008 for all cases in which esophageal biopsies were performed to identify BE mucosa in patients without a previous diagnosis of BE. Information from the clinical history, endoscopy reports, biopsy findings and any follow-up were collected.
Results: A total of 708 patients without a previous diagnosis of BE underwent biopsies of endoscopically evident columnar mucosa. In 276, GC were identified and BE was diagnosed, while in 379 no GC were identified (53 had columnar mucosa visualized but only squamous mucosa on biopsy and were excluded). The length of the endoscopically apparent columnar mucosa was greater in patients with GC (X=4.6 cm) than in those without (X=1.6 cm). Patients with GC had more biopsies taken (X=5) compared to those without (X=4). A total of 125 patients (33%) without GC underwent at least one additional endoscopy. In 14 (11%) of these patients, GC were subsequently identified, requiring a mean follow-up of 4.9 years and 2.07 more procedures in which a mean of 6.2 more biopsies were obtained. In the other 111, who underwent a mean of 2.8 procedures (average follow-up 5.4 years) and a mean of 7 more biopsies, GC were still not identified. The mean total number of biopsies from all procedures was 11 for this subset of patients. The mean length of the columnar lined segment in the 14 patients in whom GC were identified upon repeat endoscopy was 4.1 cm, compared to a mean of 1.6 cm in the 111 without GC. GC were identified in only 9% of all patients with columnar mucosa <2 cm or in whom an irregular Z-line suspicious for Barrett's esophagus was reported. No patient without GC developed dysplasia, followed over a range of 1-23 years (X=5.4).
Conclusions: In this study, dropping the requirement for the identification of GC would have increased the initial diagnosis of BE by 53%. Among patients with a short segment of columnar mucosa, subsequent endoscopy generally does not reveal GC, suggesting that the columnar mucosa may in fact represent proximal gastric mucosa. Dropping the requirement for GC may cause many of these patients to be inaccurately labeled as BE.
Tuesday, March 1, 2011 8:15 AM
Platform Session: Section E, Tuesday Morning