Study Objective: To assess the efficacy and safety of treatment with endoscopic multipolar electrocoagulation in patients who have ulcers with nonbleeding visible vessels. Design: Prospective, randomized, sham-controlled trial; patients were followed until their discharge from the hospital. Setting: Urban, nonreferral county hospital. Patients: Consecutive sample of 75 patients who had a bloody nasogastric aspirate sample, melena, or hematochezia; unstable vital signs, a transfusion of at least two units of blood in 12 hours, or a drop in the hematocrit of at least 0.06 in 12 hours; and endoscopic evidence of an ulcer with a nonbleeding visible vessel. Intervention: Sham or real multipolar electrocoagulation at the time of diagnostic endoscopy. Measurements and Main Results: Compared with the control group, the group receiving multipolar electrocoagulation showed marked improvement in the following variables: rebleeding (18% compared with 41%, P < 0.05; difference, 23%; 95% CI, 3% to 43%); need for emergency surgery (8% compared with 30%, P < 0.05; difference, 22%; CI, 5% to 39%); mean number of hospital days (4.3 ± 0.4 compared with 6.2 ± 0.7, P < 0.05; difference, 1.9; CI, 0.4 to 3.4); and cost of hospitalization ($3790 ± $410 compared with $5730 ± $650, P < 0.05; difference, $1940; CI, $400 to $3480). The mean transfusion requirement in the treatment group was 1.6 ± 0.3 as compared with 3.0 ± 0.6 units in the control group (P = 0.13; difference, 1.4; CI, 0 to 2.8). The overall mortality was extremely low: Only 1 (1%) of 75 patients died. Bleeding was induced in 7 (18%) of the 38 patients treated with electrocoagulation, and 1 patient required urgent surgery. Conclusions: Endoscopic treatment with multipolar electrocoagulation is beneficial in patients who present with major upper gastrointestinal hemorrhage and are found to have an ulcer with a nonbleeding visible vessel.