To assess the efficacy of injection therapy with alcohol on prevent rebleeding and emergency surgery in patients with gastroduodenal ulcers and nonbleeding visible vessels, we have performed a prospective controlled trial involving 39 patients who were classified into two groups according to the time of the day on which emergency endoscopy was performed: group 1 (25 patients) in which endoscopic hemostasis with absolute alcohol was performed, and group 2 (14 patients) in which conventional therapy was applied (blood transfusions, antacids, and ranitidine). The two groups were comparable with regard to age, sex, and type of bleeding. The rebleeding rate/emergency surgery rate of 8%/4%, respectively, for group 1 was lower than the 57%/50% for group 2 (p less than 0.001). Our results suggest that endoscopic hemostasis with alcohol should be considered as the initial treatment of choice in patients who present with major upper gastrointestinal hemorrhage and are found to have an ulcer with a nonbleeding visible vessel.