The present prospective randomized trial has been carried out to assess the efficacy of the endoscopic injection of epinephrine plus polidocanol, in the prompt treatment of upper gastrointestinal bleeding due to peptic lesions. Over a period of one year, 72 patients were admitted to the study. Thirty-six of these underwent endoscopic treatment (group A), while the others received medical treatment (group B). Definitive hemostasis was achieved in 29 patients in group A, and in 21 in group B (p less than 0.05). Seven patients in group A and 12 in group B underwent emergency surgery for persistent or recurrent bleeding (p = NS). Three patients from group B received endoscopic treatment owing to relative surgical contraindications. It was successful in all of them. Patients from group A needed a significantly lower number of units of packed red cells (p less than 0.05). Our results show that it is justified to employ endoscopic injection as the first therapeutic step in upper GI bleeding due to peptic gastroduodenal lesions.