Objective: To determine the clinical outcomes of transcatheter arterial embolization(TAE)in endoscopically unmanageable non-variceal upper gastrointestinal bleeding.To determine factors influencing the risk of rebleeding in TAE.Method: Data of 52 patients(30 men,22 women; mean age,64.4 years)with acute non-variceal upper GI bleed who underwent TAE following failed endoscopic management were retrospectively reviewed.Patients demographics,clinical parameters on presentation,biochemistry results,comorbidities and embolization parameters were all analyzed,and clinical outcomes were subsequently calculated.Primary outcomes measured were complications and re-bleeding rates,length of hospital stay and 30-days mortality rates.Secondary outcomes were safety,efficacy and predictors of successful TAE.Result: TAE successfully achieved haemostasis in all 52 patients.Lesions treated were gastric ulcers(28.6%),duodenal ulcers(69%)and Dieulafoy lesion(2.4%).Rate of complication and rebleeding were 4.9%and 24.8%respectively.The median length of hospital stay was 14 days,ranging from a minimum stay of 1 day to 165 days.Thirty-day mortality rate was 14.3%with the mean number of days of death occurring within 30 days being 18 days.Age,systolic blood pressure at presentation,liver failure,coagulopathy and having multiple comorbidities were found to be significantly predictors of rebleeding(p<0.05).Conclusion: TAE is an effective and safe salvage therapy for patients with upper GI bleed that endoscopy fails to achieve haemostasis.Patients who have multiple co-morbidities were found to have the highest risk of rebleeding.Further research needs to address whether surgery could be a better alternative for those patients who are deemed to be high risk at under going TAE based on age,systolic blood pressure at presentation,coagulopathy and having multiple comorbidities.