A prospective, randomized comparative study was performed to compare the hemostatic effect of endoscopic injection with normal saline, 3% NaCl solution, 50% glucose/water solution, and pure alcohol during a period of 2 years. Only patients with peptic ulcers and active bleeding or nonbleeding visible vessels were considered. Each group comprised 50 patients. The four groups were matched at random for age, sex, location of bleeders, stigmata of recent hemorrhage, shock, hemoglobin, and concomitant illness. No statistically significant difference was observed among patients injected with normal saline, 3% NaCl solution, 50% glucose/water solution, or pure alcohol in achieving initial hemostasis (82%, 90%, 86%, and 92%, respectively); rebleeding rates (7.3%, 24.4%, 14%, and 10.9%, respectively); ultimate hemostasis (78%, 68%, 78%, and 84%, respectively); number of emergency operations (5, 7, 4, and 3, respectively); and average number of days in the hospital (6.7, 6.1, 6.1, and 5.8, respectively). A tendency toward a lower blood transfusion requirement was observed in the pure alcohol group, but this tendency failed to achieve statistical significance. One patient had a perforated ulcer develop 5 days after injection of 3 ml 50% glucose/water. Otherwise, no major complication was observed. We suggest that endoscopic injection with any of the above solutions can be used as the first-line modality for the arrest of peptic ulcer hemorrhage.