OBJECTIVE To evaluate the accuracy of (13)C-urea breath test (UBT) to detect Helicobacter pylori infection in patients hospitalized with peptic ulcer bleeding and treated with proton pump inhibitors (PPIs). METHODS Patients hospitalized with peptic ulcer bleeding, and treated with omeprazole, had a first UBT performed the day after resuming oral feeding. Patients with a negative UBT during hospitalization underwent a repeated UBT 15 days after stopping PPIs. RESULTS The first UBT during hospitalization was positive in 86% of 131 patients. Time between admission and performance of the test was longer in patients with negative versus positive UBT (5.2 +/- 0.7 versus 4.3 +/- 0.5 days; p < .001). The repeated UBT became positive in 15 of 18 (83%) patients with a negative first UBT. In the multivariate analysis, the only variable associated with a negative first UBT was the time elapsed between admission and performance of the test (odds ratio = 6.6; 95%CI = 2.9-15.1). CONCLUSIONS Most H. pylori-positive patients with ulcer bleeding have a positive UBT (performed just after resuming oral feeding) despite previous treatment with high-dose PPIs. Nevertheless, to preclude false-negative results due to PPI therapy, the UBT should be performed as early as possible. If the infection cannot be demonstrated with this first UBT, H. pylori still needs to be definitively excluded with a second UBT performed after stopping PPIs.