Epidemiology of Barrett’s esophagus and adenocarcinoma*

The definition of Barrett’s esophagus (BE) is made when visible red columnar epithelium is seen in the esophagus, with intestinal metaplasia seen on biopsy. Most epidemiologic data relates to long-segment BE, 3 cm or longer. Short segments or tongues < 3 cm are now increasingly recognized as having a risk for development of adenocarcinoma. Biopsies taken just below a normally located squamocolumnar junction in consecutive patients having endoscopy will show intestinal metaplasia in about 15% of cases.1 The cancer risk in such patients with intestinal metaplasia of the cardia is probably low, but has not been documented adequately. Questionnaires sent to random members of the population have shown that about 18% of adults have heartburn at least once a week, and that 7% have heartburn every day.2,3 Prospective studies in which people with frequent (at least weekly) heartburn had endoscopy showed that about 5% have BE (> 3 cm, with intestinal metaplasia).4–6 If about 1 in 5 persons have frequent reflux symptoms, and about 1 in 20 with reflux symptoms have BE, then the prevalence of BE in the general adult population can be estimated at about 1 in 100 persons. Not all patients with BE have reflux symptoms, however. In patients with a short history of cancer-related symptoms (dysphagia, bleeding, weight loss), in whom endoscopy shows adenocarcinoma and a previously undiagnosed BE, only about 60% have a long previous history of reflux symptoms.7,8 In a large series of patients having upper gastrointestinal endoscopy for various clinical indications, the prevalence of BE was found to be rare in childhood and to rise progressively with age, consistent with an acquired disorder.9 After the age of 60 years, BE was found in about 1% of patients having routine endoscopy. The prevalence in males was twice that in females in this series. In a Mayo Clinic autopsy series, seven cases of BE were found in 733 consecutive autopsies, so that about 1 in 100 older people had BE.10 An estimate of the clinically diagnosed prevalence was also obtained in the same population of Olmsted County, Minnesota. Patients with BE previously diagnosed by endoscopy and biopsy still living in the county in 1998 were included. The age- and sex-adjusted prevalence estimate based on the autopsy series was 376 per 100 000.11 This was about five times greater than the clinically diagnosed prevalence of 80 per 100 000. The conclusion was that most BE cases in the population have not been diagnosed, and so are not available for surveillance for early detection of dysplasia or adenocarcinoma. A population based in Olmsted County, Minnesota11 showed that the incidence of new clinical diagnosis of BE rose from less than 1 case per 100 000 per year in 1965 to 11 cases per 100 000 per year in 1997. This may be due to increased detection of BE rather than any real increase in incidence; over the same time interval, the number of upper endoscopic examinations performed in County residents rose to a similar extent from 50 per 100 000 in 1965 to 1400 per 100 000 per year in 1997. At least 10 families have been reported in which two or more relatives had BE.12 In two studies, the prevalence of reflux symptoms in first degree relatives of BE patients was shown to be 2.2 and 4.8 times greater than in controls.13,14 We performed endoscopy in 100 symptomatic BE relatives and in 100 unrelated persons with similar reflux symptoms, and found BE in 8% of relatives and in 5% of symptomatic controls.6 It appears that the BE risk for any individual relative is not much increased but, because more relatives have reflux symptoms, our preliminary estimate is that the BE risk for all BE relatives as a group may be five times greater than for controls. Inherited reflux appears to be more significant than direct inheritance of BE. Data from multiple population-based cancer registries in Europe, North America and in Australia have shown that adenocarcinoma was rare until about 1970, when a marked and continuing increase in incidence began.15,16 In Olmsted County, Minnesota, the incidence of esophageal adenocarcinoma rose10-fold from about 0.2–2.1 cases per 100 000 population in the past 25 years.16 The incidence of adenocarcinoma of the cardia (esophagogastric junction) has increased to a somewhat lesser extent over the same time interval. In the US in 1988–90, the incidence of esophageal adenocarcinoma was four times higher in white males than in black males.15 The incidence was eight times higher (2.5 vs. 0.3/100 000/year) in white males compared with white females. The reason for the white male predominance is not clear, although BE is more prevalent in whites than blacks in the US.17 The incidence of adenocarcinoma also varies between different European countries, being highest in England (males, 6.9/100 000/year) and lowest in Finland (males, 0.9/100 000/year) and the Czech Republic (males, 0.5/100 000/year). Again, the cause of the differing incidence is not known. In a US National Cancer Institute study, smoking was associated with a 2.2 times increased risk for esophageal adenocarcinoma and for a 2.6 times increased risk for adenocarcinoma of the cardia.18 The increased risk persisted up to 30 years after stopping smoking. Another factor is obesity. In a Swedish study,19 the heaviest quarter of the population had a 7.6 times increased risk for esophageal adenocarcinoma and a 2.3 times increased risk for adenocarcinoma of the cardia, when compared with the lightest quarter of the population. Helicobacter pylori infection may be a protective factor. In a study of over 100 000 persons in Norway20 in which blood had been stored for serology for a median of 12 years, the presence of H. pylori was associated with a five times increased risk for adenocarcinoma of the distal stomach, but with a five times decreased risk for adenocarcinoma of the cardia. At least in the US, the prevalence of obesity is increasing and the prevalence of H. pylori infection is decreasing, and these two factors may contribute to the increasing incidence of adenocarcinoma of the esophagus and cardia.