Introduction We present the long-term outcome of Barrett’s oesophagus (BO) observed over 37 years at a District General Hospital. We comment on the risk of developing oesophageal adenocarcinoma (OAC), its outcome, and on the value of endoscopic surveillance.
Methods The study includes all patients diagnosed with BO from 1.1.1977 to 31.12.2011 and followed-up until 31.12.2013 (37 years). Patients with prevalent OAC were excluded from analysis. Data were prospectively collected and updated at every visit and all deaths recorded. Patients were followed-up by surveillance endoscopy or if contraindicated by age or co-morbidity then by clinic visit or telephone survey, with urgent endoscopy reserved for recurrence of symptoms for all patients. To compare outcome we divided OAC patients according to their method of follow-up at the time OAC was diagnosed. To adjust for the confounding factor of age when creating survival estimates we applied a Cox proportional hazards regression which included gender, age, and age squared (for there is evidence that the relationship with age and survival is not linear).
Results Total numbers: 1977–2011 n = 1381. BO numbers steadily increased but plateaued towards the end of the study. BO diagnosis was confirmed by histology in 88%. Outcome: In a total FU of 10366 patient-years (mean 7.5y range 0–36y), 54/1381 (3.9%) developed OAC (mean interval 9 years, SD 5.5, range 13 months–25.4 years). Thus 1 OAC developed per 192 patient-years of FU i.e. 0.52% per year. Mortality: Deaths from all causes were 417/1381 (30.2%). Surveillance and survival: Mortality was significantly lower in the 37 patients under endoscopic surveillance at the time OAC was diagnosed (51% vs 88% p = 0.0141) due in part to the older age and co-morbidity of the other 17 in whom serial endoscopy was contraindicated (mean age 67y vs 75y p = 0.002, respectively). However, after adjusting for age there was no significant difference in mortality between the two groups (p = 0.08). Importantly though, the estimated hazard rate ratio was lower in the surveillance group (0.64, 95% CI 0.30 to 1.48).
Conclusion Reduction of risk: The risk of dying from OAC was reduced by about one-third in those under endoscopic surveillance at the time OAC was diagnosed. Identifying such a reduction of risk is a very useful indicator of the value of surveillance. Numbers and implications: This substantial long-term study was still not large enough to show a significant reduction in death once adjusted for confounding factors. Our results stress the need for individual centres to contribute to data warehouses such as national registries.
We now have clear evidence that BO surveillance confers benefit. With today’s technology which offers a wider scope of both palliative and curative treatments, and a continued systematic and disciplined approach, we have reason to expect that endoscopic surveillance will give even better results in the future.
Disclosure of Interest None Declared