Acute upper gastrointestinal bleeding remains a difficult emergency problem, and, despite recent pharmacological advances, the choice and results of medical treatment are the subject of debate. To determine the effectiveness of two potent inhibitors of gastric acid secretion, somatostatin and cimetidine, in the control of upper gastrointestinal bleeding of non-variceal origin, we initiated a multicentric, randomized, prospective therapeutic trial in 56 patients presenting with acute hemorrhage due to gastric and duodenal ulcers and erosions defined by uniform and precise criteria. The two drugs were administered i.v. for 48 hours at a dose of 250 mcg/h (somatostatin) and 1,600 mg/24 h (cimetidine). Vital signs, laboratory values, and gastric aspirate were checked frequently in accordance with a strict schedule; the number of blood transfusions was also noted. Endoscopy for the assessment of bleeding before admission to the trial and the end of treatment was performed in every patient. Bleeding stopped in 28 of the 30 (93.3%) patients treated with somatostatin, and in 16 of the 26 (61.5%) of those receiving cimetidine (p less than 0.01). The blood requirement of the patients treated with somatostatin and cimetidine was, on average, 1.10 +/- 1.16 and 2.46 +/- 4.03 units per patient, respectively (p less than 0.05). Somatostatin was also significantly superior - in the patients in whom the treatments were successful - with respect to the time taken to achieve this result. In conclusion, our results, together with those already published, point to a definite therapeutic effectiveness of somatostatin in upper gastrointestinal bleeding as here defined, and would appear to justify a more extensive clinical use under controlled conditions.