In the last few years, the number of anticoagulated patients has significantly increased and, as a consequence, so have hemorrhagic complications due to this therapy. We analyzed gastrointestinal (GI) bleeding because it is the most frequent type of major bleeding in these patients, and we hypothesized that they would have lesions responsible for GI bleeding regardless of the intensity of anticoagulation, although excessively anticoagutated patients would have more serious hemorrhages. Objectives: To study the characteristics of anticoagulated patients with GI bleeding and the relationship between the degree of anticoagulation and a finding of causative lesions and bleeding severity. Patients and methods: We prospectively studied 96 patients, all anticoagulated with acenocoumarol and consecutively admitted to hospital between 01/01/2003 and 09/30/2005 because of acute GI bleeding. We excluded patients with severe liver disease, as well as nine patients with incomplete details. Results: The incidence of GI bleeding requiring hospitalization was 19.6 cases/100,000 inhabitants-year. In 90% of patients, we found a causative (85% of upper GI bleeding and 50% of lower GI bleeding) or potentially causative lesion, and 30% of them required endoscopic treatment, without differences depending on the intensity of anticoagulation. No relationship was found between the type of lesions observed and the degree of anticoagulation in these patients. Patients who received more intense anticoagulation therapy had more severe hemorrhages (23% of patients with an INR >= 4 had a life-threatening bleed versus only 4% of patients with INR < 4). Conclusions: We found an incidence of 20 severe GI bleeding episodes in anticoagulated patients per 100,000 inhabitants-year, with no difference in localization or in the frequency of causative lesions depending on the intensity of anticoagulation. Patients receiving more intense anticoagulation had more severe GI bleeding episodes. (C) 2013 Elsevier Espana, S.L. and AEEH y AEG. All rights reserved.