Frequency of Barrett's neoplasia after initial negative endoscopy with biopsy: a long-term histopathological follow-up study

Background: Barrett's adenocarcinoma is being diagnosed increasingly. We examine possible differences between long segment and short-segment Barrett esophagus (LSBE and SSBE) in long-term follow-up on the basis of our histopathology registry. Methods and patients: All Barrett's esophagus patients diagnosed histologically between 1990 and 1995 (n = 1071) were selected. Long-term follow-up data from endoscopy with biopsy were sought on all patients without neoplasia on initial endoscopic biopsy (n = 1003). A total of 255 individuals (25.4%) were regarded as drop-outs (201 lost and 54 without further endoscopy). Of the remaining 748 patients with follow up for more than 5 years, 315 had documented LSBE, 246 had SSBE, and 187 had no length of Barrett esophagus recorded (NLBE). Results: In the study cases (male: female ratio 2.1 : 1, mean age +/- SD 60.9 +/- 14.2 years), the biopsy procedure was fully compliant with guidelines in only 32.5%. Only 5 cases (0.6%) had visible lesions reported on endoscopy, but all were negative for neoplasia. Over a mean follow-up of 78.2 +/- 35.6 months (range 0-240), 7 new cases of low grade intraepithelial neoplasia (LGIN) and 15 cancer cases developed, accounting for a yearly incidence of 0.2% (LGIN) or 0.4% (cancer) after an initial negative endoscopy. When the cases with initial diagnosis of neoplasia were included, this yearly incidence rose to 0.5% (LGIN), 0.3% (high grade intraepithelial neoplasia [HGIN]) or 1.7% (cancer). Differences between SSBE and LSBE were only encountered for cancer incidence. Conclusion: The yearly incidence of Barrett esophagus cancer varies between 0.4% and 1.7%. Despite the limitations of this retrospective and pathology-based study, the observed risk of developing cancer in Barrett esophagus without neoplasia is comparable to that found in other studies, mainly from the US and the UK, and varies between 0.7% and 1.0% of yearly incidence.