Barrett's oesophagus: an audit of surveillance over a 17-year period

Objective To audit whether our patients with Barrett's oesophagus (BO) enter into our endoscopic surveillance programme and whether they continue with it after entry. We have determined the incidence of oesophageal adenocarcinoma among our surveyed patients. Design We retrospectively audited prospectively collected data from our BO surveillance programme over the years 1987-2003. Setting An inner city teaching hospital. Results During these years, 466 patients with BO were diagnosed (392 long segment, >= 3cm), 29 had oesophageal adenocarcinoma at diagnosis, 232 [195 with intestinal metaplasia (IM) on biopsy] had at least one follow-up endoscopy, and 205 have not been re-endoscoped. In 27 out of 205 no I M was present. Of the remaining 178 out of 205 with IM, 30 were within 2 years of diagnosis and 148 have not been re-endoscoped for the following reasons: age (51), non-attendance (35), not referred back by general practitioner (30), non-oesophageal cancer (14), severe concurrent illness (12), death (three), refused follow-up (two), left the area (one). The 195 patients with IM who entered endoscopic surveillance consisted of 108 men and 87 women (aged 62.9 years, range 31-96), were followed for a total of 1068 patient-years (average 5.5 years), and had 556 endoscopies (average 2.9 per patient). Ninety-seven out of 195 patients remain under active endoscopic surveillance but 98 discontinued for the following reasons: age (31), non attendance (21), death (21 including one from oesophageal adenocarcinoma), refused follow up (seven), concurrent illness (six), left the area (four), no IM on repeat biopsies (three). Of the 195 patients with IM, four developed low-grade dysplasia, two high-grade dysplasia and four adenocarcinoma (incidence 0.37%); 178 out of 195 have been maintained on proton pump inhibitor (PPI) therapy. Conclusions The majority of patients with BO either do not enter or do not continue in an endoscopic surveillance programme. This needs to be acknowledged when the workload and cost of BO surveillance programmes are considered. The incidence of adenocarcinorna was low compared with many published series, and we speculate whether this is the result of maintenance PPI therapy. Eur J Gastroenterol Hepatol 18:271-276 (c) 2006 Lippincott Williams & Wilkins.