Background/Aims: Antrum has been proposed for the best biopsy site to diagnose Helicobacter pylori (H. pylori) infection. The aim of this study was to determine the best biopsy site for maximum diagnostic yield. Methods: H. pylori infection was determined by rapid urease test and histology from the biopsy specimens obtained from 992 patients (652 gastritis, 130 gastric ulcer, 148 duodenal ulcer, and 62 gastric adenocarcinoma). Three biopsy specimens were obtained from each of three biopsy sites (the antral greater curvature, mid-body greater curvature, and fundus). Results: H. pylori positivity was significantly higher in mid-body than in antrum or fundus (52.8% vs. 40.0%, 40.3%). The diagnostic yield from the mid-body alone was not significantly different from those of any combinations of other two biopsy sites. The results of rapid urease test were very similar to those of histologic diagnosis. H. pylori positivity in the patients with gastric ulcer and gastric cancer was significantly higher in mid-body than in antrum (55.4%, 45.2% vs. 26.9%, 27.4%). In duodenal ulcer patients, the positivity was significantly higher in mid-body than in antrum or in fundus (68.9% vs. 54.7%, 56.2%). Conclusions: The greater curvature of mid-body, not the antrum, is the best site to diagnose H. pylori infection irrespective of the kinds of gastroduodenal diseases. (Korean J Gastroenterol 2001;37:406-411)