Etiology of intestinal metaplasia at the gastroesophageal junction

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

Background: Intestinal metaplasia occurs in the esophagus as a consequence of gastroesophageal reflux disease and in the stomach secondary to H. pylori infection. The etiology of intestinal metaplasia limited to the gastroesophageal junction or cardia (CIM) is disputed. We hypothesized that CIM has dual etiologies: gastroesophageal reflux in some, H. pylori infection in others, and that cytokeratin immunostaining can help to differentiate between these two etiologies. Methods: We defined CIM as the presence of intestinal metaplasia within cardiac mucosa on biopsy from an endoscopically normal-appearing gastroesophageal junction. Thirty patients with CIM who had multiple biopsy specimens taken from the esophagus, gastroesophageal junction, and stomach were identified. Tissue blocks from biopsy specimens taken at the gastroesophageal junction were sectioned and immunostained for cytokeratins 7 and 20. The cytokeratin 7/20 staining of the CIM in each patient was determined to be either a Barrett's or non-Barrett's pattern. H. pylori infection was assessed by Giemsa staining of antral biopsy specimens. Results: H. pylori infection was present in 16 patients. A Barrett's cytokeratin 7/20 staining pattern in the CIM was present in only 46% of the H. pylori–positive patients, as compared to 86% in the 14 patients with CIM and no H. pylori (p = 0.025). Objective evidence of reflux disease was present in 71% of patients with CIM and no H. pylori, as compared to 31% of patients with H. pylori. Conclusions: The two different patterns of cytokeratin 7/20 staining found in patients with CIM support the concept of dual etiologies for CIM. A Barrett's staining pattern was associated with objective evidence of gastroesophageal reflux and the absence of H. pylori, suggesting that cytokeratin 7/20 immunostaining is useful to determine the likely etiology of CIM.