Primary Motility  Disorders of the  Esophagus
 The Esophageal
 Esophagogastric  Junction

  Browse by Author
  Browse by Movies
Volume: Barrett's Esophagus
Chapter: Treatments

Can any correlation betweeen gastric polyps development and dysplasia be anticipated in Barrett's patients?

A.C. Svoboda, S. Tillisch-Svoboda, P.L. Morris (Santa Barbara)

The development of large (> 1 cm) non-neoplastic gastric polyps in patients undergoing surveillance endoscopy for Barrett's epithelium treated with continuous omeprazole for over one year, and polyp disappearance or size reduction by omeprazole discontinuation or dose reduction, prompted a review of all Barrett's patients followed by this investigator (Svoboda). The aim was to determine the incidence of gastric polyps in patients with Barrett's epithelium, the importance of length of esophageal columnar epithelium, and the influence of treatment on gastric polyp development. A retrospective review of Barrett's patient's records with attention to length of Barrett's segment, presence of dysplasia, treatment, gastric polyps, colon polyps and cancer, serum gastrin levels, and presence of Helicobacter pylori was carried out.

Case study

A 57-year-old white married moderately overweight male, with daily heartburn despite dietary restrictions and antacid usage, in December 1991 had ulcerative esophagitis (distal 3 cm) at endoscopy. Biopsies demonstrated intestinal metaplasia and Barrett's esophagus and H2 receptor antagonists and prokinetic agents were initiated. The patient returned for continued symptoms and Barrett's esophagus, and erosive esophagitis were confirmed. No gastric polyps were present. Therapy with omeprazole, 20 mg every morning was begun. The patient did so well on omeprazole that he refused to stop the medication despite the fact that it was then only indicated for short-term therapy. Because of concerns about continued omeprazole usage follow-up endoscopy was advised in 6 months. At that time (May 1993) endoscopy showed 7-8 sessile fundal polyps 4-5 mm in diameter diagnosed on biopsy as hyperplastic polyps. Omeprazole was discontinued and nizatidine 150 mg BID was begun. Because of both recurrent symptoms and the polyp history, endoscopy was repeated in 6 months (November 1993) with reintroduction of omeprazole when no gastric polyps were seen. In June 1994 endoscopy demonstrated both more (8-10) and larger (10-15 mm) gastric polyps with features of both fundal and hyperplastic polyps on multiple biopsies. Dose reduction of omeprazole to 20 mg every 2-3 days marginally controlled symptoms and endoscopy in December 1994 showed more (12-15) but smaller (6-9 mm) polyps. During the follow-up period of this patient two additional patients on prolonged omeprazole therapy developed gastric polyps on follow-up. In each patient polyps disappeared or were smaller with omeprazole discontinuation or dose reduction. A review of patients under surveillance for Barrett's esophagus appeared warranted.


A retrospective review of Barrett's patients' records with attention to length of Barrett's segment, presence of dysplasia, treatment, presence of gastric polyps, colon polyps and/or cancer, serum gastrin levels, and presence of Helicobacter pylori. Barrett's esophagus was diagnosed by biopsy proven intestinal metaplasia (IM) with goblet cell hyperplasia at the gastroesophageal junction (GEJ) or in the tubular esophagus.


There were a total of 222 patients (137 males and 85 females). In 80 patients with long-segment Barrett's (over 3 cm of specialized columnar mucosa in the distal esophagus), the sex ratio was 59 males to 21 females, while in 142 patients with short segment Barrett's there were 78 males and 64 females. All moderate and severe dysplasia occurred in long segment Barrett's. In long segment Barrett's the mean age was 66 years with a range of 3383 years. Average duration of known disease was 55 months with a range of 4-219 months. In short segment Barrett's the mean age was similar (68 years) with a range of 17-87 years. The average duration of known disease was 28 months with a range of 2-179 months.

As part of a concurrent study to determine the frequency of IM at the GEJ, 26 of 120 patients (21.6%) referred for endoscopy with no obvious Barrett's had IM on biopsy at the C-E junction.

Dysplasia in metaplastic tissue

Fifty-five Barrett's patients were reported to show dysplasia on at least one examination (26.8%). Low-grade dysplasia was predominant, occurring in 46 patients of whom 26 (31%) had long segment disease and 20 (16.5%) had short segment Barrett's. There was one (1) person with moderate (female) and eight (8) with severe dysplasia (1 female, 7 males). All moderate and severe dysplasia was found in long segment Barrett's. The longer the period of surveillance, the greater the likelihood of dysplasia developing. Dysplasia presence or severity, varied from one exam to the next, presumably because of sampling error although a protocol of four quadrant biopsies every one cm was employed in dysplasia follow-up.

Gastric polyps

Gastric polyp presence was recorded at each endoscopy with an estimate of number of polyps and size range. Gastric fundal polyps frequently occur between folds and adequate inflation of the stomach and careful retroflexion of the endoscope is necessary to avoid missing sessile lesions. The authors had a prior interest and previous publications relative to non-neoplastic gastric polyps providing a control group drawn from a very similar demographic population. This prior study was concerned with the frequency of parietal cell antibodies in endoscopic patients. Of 300 consecutive patients over the age of 40 having endoscopy, 14 or 4.7% were found to have non-neoplastic polyps generally against a background of partial atrophy or atrophic gastritis agreeing well with the 3-5% incidence in the endoscopic literature.

Gastric polyp occurrence in current study

At the time of diagnosis of Barrett's and prior to initiation of treatment, non-neoplastic (fundic gland or hyperplastic) polyps were present in 7 of 222 patients or 3.15%. While less frequent than controls this was not statistically significant. After proton pump inhibitor (PPI) therapy (omeprazole) the changes were dramatic. Polyps develop in 32 of 215 patients or 14.88% (p = 0.0001!). Of long segment Barrett's patients with polyps 14 (77%) were on PPIs. Of 62 patients without gastric polyps, 28 were on PPIs or 45%. This difference is significant (p = 0.03). Similarly in short segment Barrett's the number of those with polyps on PPIs (13/21 or 61.9%) was greater than those without polyps on PPIs (48/121 or 39.7%, p = 0.10) suggesting a trend. Polyps tended to be multiple, sessile and evenly distributed (on endoscopic biopsy) between fundic gland and hyperplastic polyps. Other investigators have subsequently reported this association of proton pump use and gastric polyp development, especially in Barrett's patients. Presumably just as in pernicious anemia or atrophic gastritis with achlorhydria, high gastrin levels exert a trophic effect on gastric epithelium. However, in this study there were only two instances of serum gastrin elevation, one in short segment and the other in long segment Barrett's. Neither of these had gastric polyps or dysplasia. These polyps do not have a significant potential to become neoplastic and should not deserve endoscopic surveillance. However, additional examination of our data raised serious concerns. Not only did there appear to be an association between use of PPIs and subsequent non-neoplastic gastric polyp development, but those with long segment Barrett's who developed polyps were those most likely to develop dysplasia. Of the 80 patients with long segment Barrett's, 18 developed polyps and dysplasia was present in 15 of these 18 (83.3%). Of those 62 without polyps, dysplasia was present in only 22 (25.5%). This is highly significant (p = 0.001!). This risk in short segment Barrett's has not yet been demonstrated, in part because of short follow-up.


If gastric polyp development is associated with PPI usage and if polyps are a warning sign of dysplasia developing, can PPI usage be associated with dysplasia and progression to neoplasia? PPI usage is associated with gastric polyp development in both short and long segment Barrett's. Patients with Barrett's esophagus on prolonged use of omeprazole are more likely than controls to develop non-neoplastic gastric polyps (p < .0001). This is especially true with long segment (classical) Barrett's (p 0.0001 versus 0.131) Patients with long segment Barrett's are more likely than short segment Barrett's to develop dysplasia. (p < .0001). All moderate and severe dysplasia occurred in long segment Barrett's. A finding of gastric polyps is associated with a greater risk of esophageal dysplasia. The development of gastric polyps during acid suppression therapy (especially with PPIs) in patients with long segment Barrett's is associated with a high and statistically significant risk of esophageal dysplasia. Careful surveillance is warranted and alternative therapy should be considered if polyps and dysplasia co-exist.


1. Svoboda AC. Is polypectomy indicated for gastric polyps? Am J Gastroenterol 1980.

2. Tillisch S, Svoboda A, Garrett R. Gastric polyps; use common sense, not surveillance. Am J Gastroenterol 1986;81:851.

3. Svoboda A, Tillisch-Svoboda S. Barrett's patients developing gastric polyps during acid suppression are at greater risk of esophageal dysplasia. Endoscopy 1996;28:S20.

4. El-Zimaity H, Jackson F, Graham D. Fundic gland polyps developing during omeprazole therapy. Am J Gastroenterol 1997;92:1858.

5. Choudhry V, Boyce HW, Coppola D. Proton pump inhibitor-associated gastric polyps. Am J Clin Path 1998;110:615.

Publication date: August 2003 OESO©2015