BACKGROUND: A randomized comparative study was conducted of injection therapy with epinephrine-polidocanol (1%) versus hemoclip application, versus injection combined with hemoclip for bleeding peptic ulcers. METHODS: One hundred five patients were randomized and 101 could be evaluated (46 had active spurting or oozing of blood; 55 a visible vessel). Patients were randomized to 1 of the 3 treatment modalities during endoscopy performed within 12 hours of admission. Endoscopy was repeated after 1 day or at recurrence of bleeding and before discharge. In case of recurrent bleeding, patients were retreated with the same modality. RESULTS: Initial failure or the rate of early recurrence of bleeding was highest (but not statistically significant) in the hemoclip group (13/35; 37%), versus the injection (5/34; 15%) and combination (8/32; 25%) groups. Overall failure was significantly (p = 0.01) different among the 3 groups with the highest rate in the hemoclip group (12/35; 34%), versus the injection (2/34; 6%) and combination therapy (8/32; 25%) groups. The use of hemoclips alone appeared to fail because of difficulty with hemoclip placement and incomplete vessel compression. Complications included 1 perforation in the injection group and possibly 1 case of septic arthritis in the combination therapy group. CONCLUSION: In this study, endoscopic treatment of bleeding peptic ulcers with the hemoclip was inferior overall to injection therapy.