Surveillance in Barrett's esophagus: an audit of practice

Dig Dis Sci. 2010 Jun;55(6):1615-21. doi: 10.1007/s10620-009-0917-y. Epub 2009 Aug 11.


Goals: Determine the rates of follow-up, incident rate of cancer during surveillance, and changes in pathologic grade of patients with Barrett's esophagus during surveillance in a gastroenterology practice without a formal surveillance program.

Background: Barrett's esophagus is a pre-malignant condition. Surveillance endoscopy (SE) is recommended in order to detect and treat high-grade dysplasia and esophageal adenocarcinoma early and prevent deaths. SE has not been shown to have mortality benefit and several studies have questioned its cost-effectiveness. Most gastroenterology practices do not have a surveillance program for Barrett's esophagus. The few that exist are only in very specialized and funded programs. Little information exists on outcomes in patients with Barrett's esophagus outside of these well-structured surveillance programs.

Study: A retrospective analysis of a cohort of patients with Barrett's esophagus diagnosed and surveyed between 1995 and 2005 at a Veterans Affairs medical center. Data were collected on age, body mass index, and race. Patients who missed their SE by 6 months or more and those that missed their SE by twice the recommended intervals or more were identified and analyzed for changes in pathologic grades.

Results: A total of 472 patients were diagnosed with Barrett's esophagus or had SE between 1995 and 2005. Three hundred and five patients only had one endoscopy and biopsy. They did not have follow-up surveillance endoscopies and biopsies. Two patients were excluded from the final analysis: one had an esophagectomy after an index diagnosis of high-grade dysplasia, and one had a diagnosis of esophageal adenocarcinoma 2 days after an initial impression of Barrett's esophagus. There were 165 patients with Barrett's metaplasia or dysplasia who had SE more than once and were included in the final analysis. Overall, 53.3% had no change in pathologic grade, 35.2% regressed to a lower pathologic grade, and 11.5% progressed to a higher grade. None (0/165, 0%) progressed to esophageal adenocarcinoma; 3.6% (6/165) progressed to high-grade dysplasia and 11.5% (19/165) regressed to normal mucosa. Forty-four patients missed their SE by 6 months or more. Of these, 50% regressed, 40.9% had no change, and 9.1% progressed. Four patients regressed to normal mucosa, one progressed to high-grade dysplasia and none progressed to esophageal adenocarcinoma. Twenty-three patients missed their SE by twice the recommended intervals or more. Of these, 60.9% regressed, 34.8% did not change, and 4.3% progressed. None progressed to esophageal adenocarcinoma or high-grade dysplasia but three regressed to normal mucosa. After adjusting for age and body mass index, patients with low-grade dysplasia are nearly seven times more likely to miss their endoscopy by twice the recommended intervals or more (OR 6.56, P-value 0.03).

Conclusions: Most veteran patients with Barrett's esophagus do not undergo surveillance endoscopies. Veteran patients with Barrett's esophagus undergoing SE rarely progress to high-grade dysplasia or esophageal adenocarcinoma. Veteran patients with Barrett's esophagus who have longer SE up to twice the recommended intervals because they missed their scheduled SE did not have a worse outcome when compared to the general Barrett's esophagus surveillance group. Veteran patients with low-grade dysplasia are most likely to miss their SE by twice the recommended intervals or more, though the reason for this is unknown.

MeSH terms

  • Aged
  • Barrett Esophagus / diagnosis*
  • Barrett Esophagus / epidemiology
  • Barrett Esophagus / pathology
  • Biopsy
  • Disease Progression
  • Esophageal Neoplasms / diagnosis*
  • Esophageal Neoplasms / epidemiology
  • Esophageal Neoplasms / pathology
  • Esophagoscopy* / statistics & numerical data
  • Female
  • Humans
  • Incidence
  • Logistic Models
  • Male
  • Mass Screening / methods*
  • Mass Screening / statistics & numerical data
  • Medical Audit
  • Middle Aged
  • Patient Acceptance of Health Care
  • Precancerous Conditions / diagnosis*
  • Precancerous Conditions / epidemiology
  • Precancerous Conditions / pathology
  • Predictive Value of Tests
  • Retrospective Studies
  • Time Factors
  • United States / epidemiology
  • United States Department of Veterans Affairs