Patients with Barrettʼs Esophagus Have Higher Prevalence of Colon Polyps and Adenomas

Abstract

Purpose: Barrett's esophagus (BE) and colorectal neoplasms have similar risk factors. Previous studies have shown variable prevalence of colon polyps in patients with BE. Our aim was to determine the prevalence of colon polyps and adenomas in patients with BE compared to patients without BE. Methods: In this case-control study, the study group included patients with biopsy proven BE aged 50-75 years, who underwent a colonoscopy at Cleveland Clinic during the period, January 2002 to December 2012. The control group was frequency matched 50-75 year old patients, who underwent a colonoscopy for any indication during the same time period and also had an EGD with no evidence of BE. Exclusion criteria for both groups were family history of colon cancer, prior history of colon polyps, prior resection of colon, inflammatory bowel disease, and those with incomplete colonoscopy and inadequate bowel preparation. Patient demographics, endoscopic and colonoscopic details including biopsy results were collected. Polyps were classified into hyperplastic and non-hyperplastic which included adenomas and sessile serrated adenomas. Adenomas were defined as inclusive of tubular adenomas, tubulovillous adenomas, and adenomas with high-grade dysplasia. Proximal colon was defined as inclusive of cecum, ascending colon, transverse colon including splenic flexure and colon distal to this was defined as distal colon. Analysis of variance (ANOVA), non-parametric Kruskal-Wallis test, Fisher's exact test, or Pearson's chi-square were used. In addition, generalized linear models (GLM) with a logit link were used to model compare presence of polyps in any colonoscopy performed while accounting for multiple procedures per patient; an autoregressive (AR1) covariance structure was used to model the intra-subject correlation. Results: A total of 519 patients were included in the study; 173 with BE in study group and 346 without BE in control group. Both groups were frequency matched based on gender and age at index colonoscopy. 75% of patients were male and mean age at index colonoscopy was 61 ± 8 years. On the index colonoscopy, patients with BE were more likely to have polyps than controls (45% vs 32%; p=0.003). Patients underwent between 1 and 5 colonoscopies performed during follow-up. On multivariate analysis after adjusting for age, gender, and presence of diabetes (Table 1), patients with BE were 80% more likely to have any type of polyps, and 50% more likely to have adenomas found during any colonoscopy.Table: Table. Effect of BE on Colonoscopy Findings: Multivariable GLM†Conclusion: Patients with BE had higher prevalence of colon polyps and adenomas on colonoscopy. These findings have important clinical implications for screening and surveillance in these patients.