Background: The role of Helicobacter pylori infection in the development of Barrett's oesophagus and its complications is uncertain. The aim of the present study was to determine the importance of H. pylori infection in this disease by comparing the frequency of oesophageal and gastric H. pylori infection in a group of patients with Barrett's oesophagus or adenocarcinoma, with the frequency of infection in a control group without Barrett's disease.
Methods: The study group included 160 patients (123 male, 37 female; mean age: 61.2 years) who were classified (according to the highest grade pathological lesion in the oesophagus) as having Barrett's intestinal metaplasia (IM; 88 patients), Barrett's oesophagus with low-grade dysplasia (LGD; 28 patients), high-grade dysplasia (HGD; five patients), Barrett's indefinite for dysplasia (n = 4), and Barrett's adenocarcinoma (33 patients). A total of 91 of these patients had gastric antral specimens available for study. The control group consisted of 214 consecutive, prospectively enrolled symptomatic patients (122 male, 92 female; mean age: 57.2 years) who underwent upper gastrointestinal endoscopy and in whom Barrett's oesophagus or Barrett's adenocarcinoma was not found. A modified Warthin-Starry method was used to detect H. pylori infection.
Results: Oesophageal H. pylori infection was found in eight of 160 (5%) patients with Barrett's oesophagus or Barrett's adenocarcinoma. Holicobacter pylori organisms in the oesophagus were found only on non-intestinalized cardiac or oxyntocardiac mucosa. All patients with oesophageal H. pylori infection and an antral biopsy available for study had antral H. pylori infection. Gastric antral H. pylori infection was significantly less prevalent in patients in the Barrett's study group (15/91, 16.5%) than in the non-Barrett's control group (67/214, 31.3%; Fisher's exact test, P = 0.01). Patients from the control group with an endoscopic diagnosis of duodenal ulcer, gastric ulcer, gastritis, or duodenitis had a significantly higher prevalence of infection compared with the Barrett's group, but there was no difference in the infection prevalence in patients in the Barrett's group and patients with reflux oesophagitis, hiatal hernia, no endoscopic abnormality, or any other diagnosis.
Conclusions: Oesophageal H. pylori infection is uncommon in patients with Barrett's IM, dysplasia, or adenocarcinoma, and may be restricted to non-intestinalized columnar epithelium. Gastric H. pylori infection may have a protective effect for the development of Barrett's oesophagus.