OBJECTIVE To determine the efficacy and safety of transcatheter embolization of nonvariceal gastrointestinal hemorrhage. METHODS Descriptive analytical study. METHODS Radiology Department, The Aga Khan University Hospital, Karachi, Pakistan, from October 1999 to August 2004. METHODS The study included 27 patients who underwent angiography for suspected gastrointestinal hemorrhage. The presumptive diagnosis was based on the findings of endoscopy in 08 patients, scintigraphy in 05 patients and on clinical findings only in 14 patients. Nine patients had gastroduodenal hemorrhage, 04 patients had hemobilia and 14 patients had lower gastrointestinal bleeding. Five patients had comorbidities. One patient had uremia, 2 had hepatitis C, 1 had cirrhosis and 1 had coagulopathy. Technical success, clinical success, complications and patient survival were assessed. RESULTS Nine patients had upper gastrointestinal hemorrhage, 4 patients had hemobilia and 14 patients had lower gastrointestinal bleeding. Etiology of bleeding was known in 15 patients and unknown in 12 patients. Clinical success was achieved in 23 (85.18%) of 27 patients. Two patients had early rebleeding. One of these was managed surgically and recovered. The other had severe coagulopathy and was unfit for surgery. Two patients had bowel infarction, one of these was managed surgically and recovered and the other died of complication of procedure. One patient had small splenic infarcts and was managed conservatively. Overall 85.18% success was achieved for the embolization procedure. Two of 27 patients died. One patient died because of recurrence of bleeding and other died because of infarction of right hemi-colon which occurred as complication of embolization. CONCLUSIONS Transcatheter embolization is effective and safe in patients with nonvariceal gastrointestinal hemorrhage.