HOME
Return to Table of Contents
15 SYSTEMATIC REVIEW OF ADENOCARCINOMA RISK IN BARRETT'S ESOPHAGUS J Santana1, M Khan1, N Vakil2, P Moayyedi1 INTRODUCTION: Surveillance is recommended for Barrett's esophagus as this condition is associated with an increased risk of esophageal adenocarcinoma. The magnitude of the risk is uncertain as a previous systematic review suggested there was publication bias in the literature. Furthermore, definitions of Barrett's have changed to include short segment Barrett's esophagus (SSBE) and to exclude those in whom intestinal metaplasia (IM) is not present regardless of length of the Barrett's segment. We have therefore conducted an up to date systematic literature review to identify current esophageal adenocarcinoma risk and assess the risk in those with SSBE and those without IM.
1Gastroenterology Division, McMaster University, Hamilton, Ontario; 2College of Health Sciences, Marquette University, Milwaukee, Wisconsin, USA
METHODS: An electronic database search (Medline and Embase) was conducted to July 2005. Eligibility and data extraction was performed by two independent reviewers. Rates of conversion to esophageal adenocarcinoma were synthesised using a random effects model (DerSimonian and Laird). Funnel plots were produced for each comparison, and Egger's test for funnel plot asymmetry was used to investigate whether the results were affected by publication bias.
RESULTS: 46 papers were eligible evaluating 11,056 cases of Barrett's esophagus in 42,880 patient-years follow-up, with 255 cases of esophageal cancer. The overall incidence ratio was 0.75% per year (95% CI 0.56% to 1.0% per year). There was no funnel plot asymmetry (P=0.19) in these data. Nine papers compared the esophageal cancer risk in long segment Barrett's (LSBE) versus SSBE. There were 66 cancers in 5295 patient years follow-up in the LSBE group compared with 9 cancers in 1311 patient years follow-up in the SSBE group (0.67% per year increased risk in LSBE group: 95% CI 0.39 to 0.94%). Three papers compared the esophageal cancer risk in Barrett's with and without IM at baseline. There were 41 cancers in 7834 patient years follow-up in the IM group compared with three cancers in 1966 patient years follow-up in the no IM group (0.59%/year increased risk in IM group: 95% CI 0 to 1.2%).
CONCLUSION: This up to date systematic review suggests the annual esophageal adenocarcinoma risk in Barrett's patients is 0.75% (95% CI 0.56% to 1.0%) and this is not subject to publication bias. Current practice of including SSBE but excluding patients without IM regardless of length of Barrett's is illogical given the relative cancer risks.