The objective of this study was to investigate the incidence of oesophageal adenocarcinoma and its contribution to mortality in patients with Barrett's oesophagus, and to identify a subgroup of patients who may benefit from endoscopic surveillance. This was a retrospective study of a cohort of 70 patients diagnosed in the endoscopy unit of a Scottish teaching hospital as having Barrett's oesophagus between 1976 and 1986. Information was obtained from case notes, endoscopy records, histology reports and death certificates. Patients were included if they had: (a) columnar-lined oesophagus of at least 3 cm; (b) histological confirmation; and (c) absence of cancer when endoscopically diagnosed as having Barrett's oesophagus. The main outcome for the patients was development of adenocarcinoma or death. Information was available for 59 of 70 patients (84%). Forty-four patients were confirmed to have Barrett's oesophagus and were followed up for 418 patient-years. Only two patients developed oesophageal adenocarcinoma, resulting in an incidence of one case in 209 patient-years, a 55-fold risk compared with age- and sex-matched population in Scotland. Both these patients had intestinal metaplasia and Barrett's ulcer. The majority (90%) of patients died as a result of causes unrelated to adenocarcinoma of oesophagus. In patients under 70 years with intestinal metaplasia, 189 annual endoscopies would have been required to detect one cancer. Adenocarcinoma is an uncommon cause of mortality in patients with Barrett's oesophagus. Endoscopic surveillance is unlikely to alter overall mortality in patients with Barrett's oesophagus. However, by limiting endoscopic surveillance to patients under 70 years with intestinal metaplasia, the estimated cost per cancer detected can be reduced to ł23000.