How should endoscopic surveillance of Barrett's esophagus in adults be conducted?
C.G. Bremner (Los Angeles)
The annual risk for the development of esophageal cancer in Barrett's esophagus (BE) is approximately 800/100,000. This is greater than the annual risk of developing lung cancer which is 500/100,000 in men over the age of 65 years [1, 2]. The rate of developing adenocarcinoma in BE is one cancer for 125 to 175 patient years . It is not surprising that cost-effective limitations have entered into the philosophy of endoscopic surveillance for BE. Parsonnet and Adams  state that it is difficult to be explicit about cost effectiveness, because it is impossible to put a price on life. Achkar and Carey  estimated that a yearly endoscopic surveillance program would cost $62,000 and 78 lost working days to discover one cancer at follow-up. The benefit of early diagnosis and appropriate surgical therapy for adenocarcinoma is evident from reports of remarkable 5-year-survival results (Table I). Patients who develop carcinoma while on a surveillance program are diagnosed at an earlier symptom-stage than patients who are diagnosed by symptom-driven investigation, explaining the better survival. In one series , 58% of patients under surveillance had stage 0 and 1 disease compared to only 17% of the patients not under surveillance. Van de Veen et al.  do not agree with surveillance after a review of their experience with BE diagnosed over a 12 year period. Few patients developed cancer, which gave an incidence of 1 in 170 patient years. The survival of patients with BE in their series was not different from a matched control population, and they conclude that systematic endoscopic surveillance in patients with BE is not indicated. Most centers would disagree with their conclusions because of reports of cure in surveillance directed early diagnosis and treatment.
At the present time there is a dilemma regarding the endoscopic negative, biopsy goblet cell positive patient. Should these patients enter into an annual surveillance or not? Until the results of longitudinal studies are available, it is important to continue a surveillance program for these patients so as to give directive to future protocols.
Which patients should undergo diagnostic endoscopy
in order to diagnose Barrett's esophagus?
The duration of symptoms gives some guidance because patients who have a long history of heartburn are more likely to develop BE . The mean duration of symptoms in patients with BE was 19 years, and 9 years in patients without BE.
Which patients should enter a surveillance program?
Intestinal metaplasia with goblet cells in the esophagus is the hall-mark on which BE is diagnosed. This is a premalignant condition, and therefore all patients who have intestinal metaplasia should theoretically be enlisted into a surveillance program. Routine biopsy of the squamo-columnar junction in symptomatic patients who do not have macroscopic BE has resulted in an unexpected high yield of patients with intestinal metaplasia.
How frequently should endoscopic surveillance be carried out?
At the present time no conclusive recommendations can be made because there is insufficient data to give a suitable directive. Reid et al.  recommend a two-year endoscopic evaluation, if a rigorous jumbo biopsy regimen and brush cytology are negative for dysplasia. Reid  also suggests that if specimens taken for flow cytometry are normal, without any aneuploidy or increased G2/tetraploid fractions, a surveillance interval of 3 years is probably safe. Their practice is also to repeat surveillance in 6 to 24 months for patients whose biopsy specimens show indefinite or low-grade dysplasia. The practice at the University of Washington differs from some other centers also in the management of high grade dysplasia. Because there have been reports of regression of high-grade dysplasia (HGD)  in a follow-up of 22 patients (7 had no further HGD on follow-up; 2 developed adenocarcinoma), their protocol includes repeat endoscopy with multiple biopsies after 3 months of intensive medical therapy. If HGD persists, biopsies are repeated every 3-6 months. Reid et al.  and Levine et al.  also claim that their multiple jumbo biopsy technique can detect high-grade dysplasia from early adenocarcinoma.
There have been several criticisms of this approach. Firstly, it is often difficult to distinguish HGD from a well differentiated adenocarcinoma, and the biopsy specimen may not be deep enough to demonstrate invasion below the muscularis mucosa. Secondly, the numbers of biopsy specimens taken result in a prohibitive cost. Finally, cure is possible if an early adenocarcinoma is resected. Delay in diagnosis may preclude a cure. Once there is invasion beyond the muscularis mucosa, metastatic spread to lymph nodes is likely, and the chance of cure is reduced.
On the other hand, there is a considerable mortality risk ranging from about 2-30% (average 10%) for esophagectomy, which should always be done in a dedicated center which has shown a consistently low mortality rate. Most centers repeat endoscopic surveillance at yearly intervals and follow the proposals of the World Congress of Gastroenterology in 1990  (Table II). Richter  has proposed a blind non-endoscopically directed cytologic sampling as the principal method for surveying patients once the diagnosis of BE has been made. This method would cost a fraction of the endoscopic-biopsy method which he suggests could still be reserved for patients who have cytologic abnormalities. This approach would be reasonable, provided an experienced cytologist was available, and provided a multi-centered trial continues to demonstrate the safety and cost effectiveness of such a program.
9. Peters JH, Clark GWB, Ireland AP, Chandrosoma P, Smyrk TC, DeMeester TR. Outcome of adenocarcinoma arising in Barrett's esophagus is in endoscopically surveyed and non-surveyed patients. J Thorac Cardiovasc Surg 1994;108:813-822.
11. Tuohy CD, Allen V, Sampliner RE. Can symptoms alone differentiate patients with Barrett's esophagus from patients with gastroesophageal reflux disease lacking Barrett's? Gastroenterology 1990;98:A141.
12. Reid BJ, Haggitt RC, Rubin CE. Barrett's esophagus and esophageal adenocarcinoma. In: Hill LD, Kozarek R, McCallium R: eds. The esophagus: medical and surgical management. Philadelphia: W.B. Saunders, 1998:157.
15. Reid BJ, Weinstein WM, Lewin RJ, et al. Endoscopic biopsy can detect high-grade dysplasia or early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic lesions. Gastroenterology 1988;94:81-90.
16. Levine DS, Haggitt RC, Blount PL, Rabinovitch PS, Rusch VW, Reid BJ. An endoscopic biopsy protocol can differentiate high grade dysplasia from early adenocarcinoma in Barrett's esophagus. Gastroenterology 1993;105:40-50.