Objective: We described incidence rates of esophageal adenocarcinoma in Denmark in a 20-yr period and determined the proportion of patients diagnosed with esophageal adenocarcinoma who had a previous diagnosis of Barrett's esophagus, making them potential candidates for endoscopic surveillance. Methods: Rates of esophageal and gastric cancers were collected from the Danish Cancer registry for the period 1970-1991. The registry was used to identify all cases of esophageal adenocarcinoma in the period 1987-1992. Medical records were retrieved and details concerning previous diagnosis of reflux disease and Barrett's esophagus were recorded. Results: The age- and gender-adjusted incidence of esophageal adenocarcinoma increased eight-fold, from 0.3/105/yr in 1970 to 2.3/105/yr in 1990. This increase could not be explained by changes in classification or diagnostic routines. Medical data were retrieved for 524 of the 578 cases of esophageal adenocarcinoma reported during the period 1987-1992. A history of reflux symptoms or a diagnosis compatible with reflux was reported in 113 of 524 patients. A total of 119 patients (23%) had previously been investigated for dyspepsia or reflux symptoms, most often by endoscopy. A previous diagnosis of Barrett's esophagus was found in only 1.3% of the cancer patients. Conclusions: The rate of esophageal adenocarcinoma in Denmark has increased eight-fold over a 20-yr period, and this increase is not explained by changes in classification or diagnostic routines. More than 98% of esophageal adenocarcinomas were found in patients who could not have entered endoscopic surveillance, as Barrett's esophagus had not been diagnosed before the cancer diagnosis. Endoscopic surveillance to detect dysplasia may be an option for the individual patient with Barrett's esophagus, but these screening programs are not likely to reduce the death rate from esophageal adenocarcinomas in the general population.