The length of columnar-lined esophagus and the frequency of intestinal metaplasia

PURPOSE: The evolving definition of Barrett's esophagus (BE) requires the histologic finding of intestinal metaplasia rather than the endoscopic length of columnar-lined esophagus. We sought to determine the correlation between the probability of biopsying intestinal metaplasia (IM) in patients with endoscopic evidence of columnar-lined esophagus (CLE). METHODS: We reviewed our endoscopy and histology data from 1992 to 1997. In consecutive patients undergoing EGD for any indication, 4 quadrant biopsies were taken just below the squamocolumnar junction and every two centimeters within any CLE. The extent of CLE above the gastric folds and the worst pathologic diagnosis were recorded. Patients undergoing Barrett's surveillance were counted only once (longest CLE and worst pathology recorded). RESULTS: 349 patients (mean age 62, 45% female) had complete data. Twenty-four percent had strictures and 22% had evidence of esophagitis at the EGD recorded. The length and pathology correlation was as follows: Length (cm) n %IM dysplasia 1 83 26.5 0 2 65 23.1 1.5 3 68 36.8 2.9 4 28 53.6 7.1 5 18 61.1 0 6 13 69.2 20 >6 24 83.3 21 When patients with and without IM were compared, there was no statistically significant difference in frequency of strictures, esophagitis, age, sex. CONCLUSIONS: The frequency of intestinal metaplasia and dysplasia correlates with increasing length of endoscopic columnar-lined esophagus. Endoscopic features which may cause symptoms are not associated with either IM or extent of CLE. Any columnar epithelium in the tubular esophagus should be biopsied, regardless of other endoscopic findings. In CLE of > 5 cm, dysplasia is present in over 20% and biopsies should be taken for diagnosis of dysplasia.