PREDICTIVE FACTORS AND OUTCOMES FOR LONG AND SHORT-SEGMENT BARRETT'S ESOPHAGUS

A Ramji, H Chaun, L Halparin, J Amar, S Whittaker, R Enns

Division of Gastroenterology, Department of Medicine, St. Paul's Hospital, UBC, Vancouver, British Columbia

Barrett's esophagus (BE) is a premalignant condition that requires endoscopic surveillance. There is considerable debate regarding significance of length of BE.
OBJECTIVE: To assess factors predictive of long and short-segment BE, and to evaluate the effect of Barrett's length on the risk of dysplasia.
METHODS: All patients with biopsy-proven BE diagnosed between 1995-2002 at a tertiary-care hospital in British Columbia. Patients were divided into short-segment (SS) and long-segment Barrett's esophagus (LSBE). Long-segment was defined as endoscopically visualized BE ³ 3 cm in length. Demographic data including age, gender, indication for endoscopy, presence of hiatus hernia, esophagitis and development of dysplasia was recorded. Univariate analysis was performed utilizing the chi-square, Fischer's test and student t-test methods, SPSS version 11.0.
RESULTS: BE was biopsy-proven in 222 patients (mean age 58 yrs, 73% male). LSBE was diagnosed in 54% of patients. There was no gender difference between the 2 groups. An endoscopic diagnosis of BE was not made in 48% of patients who were found to have biopsy-proven Barrett's epithelium. Endoscopically-assessed esophagitis was more commonly diagnosed in SSBE (32% vs. 19%, p=0.027 [CI 95% 0.26-0.93]). Reflux disease was the most common reason for endoscopy (69% of patients), but was not predictive of length of Barrett's. There was a trend towards increased SSBE in those patients with peptic ulcer disease, although this did not reach statistical significance (p=0.053). Patients undergoing endoscopy for possible GI bleed were more likely to have LSBE (p=0.021). Esophageal stricture or presentation of dysphagia did not predict length of BE. Twenty patients were diagnosed with dysplasia (13 high-grade). There was no significant difference in development of dysplasia between patients with SSBE or LSBE.
CONCLUSION: The most common presenting symptoms of reflux, dysphagia or possible stricture do not predict length of BE, necessitating endoscopic evaluation in all groups. SSBE can appear as esophagitis endoscopically, and therefore requires a low threshold for biopsy. The risk of dysplasia is not different in patients with LSBE or SSBE.