Upper gastrointestinal endoscopy was carried out with Olympus gastroscopes. Between one and six biopsy specimens (average 2 2 for both study and control groups) were taken from the antrum. The specimens were immediately fixed in neutral formalin and later routinely processed for histological examination.Serial sections, varying in number between two and 32 (average 5-6 for the study group and 4 8 for the controls), were prepared from each biopsy specimen. Sections from all cases were stained with haematoxylin and eosin. A number of other sections were also stained with Warthin-Starry or modified Giemsa5 for the more specific demonstration of H pylori, especially when examination of stained sections for the presence of these bacteria was inconclusive.All sections were examined for evidence of chronic inflammation, judged by an increase of mononuclear inflammatory cells in the lamina propria, and active chronic inflammation, indicated by the presence of neutrophils as well as an excess of mononuclear cells in the mucosa. The presence or absence of lymphoid follicles, intestinal metaplasia, and Hpylori was then noted.The significance of the results was assessed by the fourfold tables of the x2 test, unless otherwise specified. Probabilities were calculated with one degree of freedom.

Sections from 200 cholecystectomy specimens were searched microscopically for areas of gastric metaplasia and Helicobacter (Campylobacter) pylori colonization. The presence of gastric metaplasia was confirmed in 31 cases by the presence of tall columnar cells with abundant apical neutral mucin. Examination of haematoxylin and eosin and, in some cases, Warthin-Starry stained sections failed to demonstrate H. pylori in any of these 31 cases or in 29 other cases with 'bladder neck-type mucous gland metaplasia'. The failure of H. pylori to colonize areas of gastric metaplasia in the gall bladder confirms their intolerance to high concentrations of bile.

No abstract