Objective To study the prevalence of Helicobacter pylori (H, pylori) infection and gastric metaplasia (GM) in the duodenum a large group of patients with duodenal ulcer was evaluated to determine whether these factors are related to the number of ulcer recurrences. Methods Three hundred and seven patients diagnosed by endoscopy as having active duodenal ulcers were studied. At endoscopy, all patients had gastric biopsies taken for histology, the rapid urease test and culture. Three duodenal biopsies were also taken and processed for histology (haematoxylin & eosin, Giemsa, Warthin-Starry, and PAS stain). Results GM and H. pylori in the duodenum was identified in 73% (68-78%) and 66% (60-71%) of the cases, respectively, All patients with H. pylori in the duodenum also had GM at this location, while areas with GM but without H. pylori were described, The kappa statistic for concordance between GM and H pylori at the duodenum was 0.82. The prevalence of GM and H. pylori, depending on the number of ulcer recurrences, was: Ist episode, 34% and 27%, respectively; 2nd episode, 84% and 80%; and greater than or equal to 3rd episode, 90% and 79% (P < 0.001 when comparing 1st vs 2nd or <greater than or equal to>3rd episode). In the multivariate analysis, age and number of ulcer recurrences correlated both with GM and with H pylori in the duodenum. Chronic duodenitis was demonstrated in all duodenal biopsies, 87% being active chronic duodenitis. H. pylori in the duodenum was more frequent in patients with active duodenitis (73%) than in those with inactive duodenitis (13%) (P < 0.001). Conclusions Patients with recurrent ulcer disease have a higher prevalence of both GM and H pylori infection in the duodenum, suggesting that these two factors are related with the chronicity and recurrence of duodenal ulcer disease. H pylori infection in the duodenum always appears in areas of GM, although GM is not necessarily colonized by the organism, H. pylori infection cannot be excluded based only on the results of duodenal biopsies, as false negative results at this area are frequent, Eur J Gastroenterol Hepatol 12:1295-1298 (C) 2000 Lippincott Williams & Wilkins.