Background: Acute upper gastrointestinal bleeding (UGIB) is common and carries significant morbidity worldwide. Effective
risk assessment for UGIB is required in order to deliver the optimal therapeutic plans.
Objectives: To describe clinical characteristics and treatment outcomes of acute UGIB in Thailand and to evaluate predictors
for rebleeding and complications.
Material and Method: Consecutive patients with acute UGIB who underwent esophagogastroduodenoscopy at Rajavithi
Hospital, Bangkok, between 2012 and 2015 were retrospectively analyzed. Important clinical data, endoscopic findings and hospital course were reviewed. Multivariate analysis was performed to identify the predictors of rebleeding and complications within 4 weeks.
Results: 286 patients were included of whom 180 were non-variceal UGIB (NVUGIB) and 106 were variceal UGIB (VUGIB). Males accounted for 71.7% of participants and had amean age of 53.6 years. Of patients with NVUGIB, 43.4% were taking NSAIDs/ASA, and the most common causes of bleeding were peptic ulcers (62.8%) and gastritis (32.2%). All patients with VUGIB had cirrhosis, and 54.7% were Child-Pugh B/C. When compared to NVUGB, patients with VUGIB were more likely to have active bleeding on presentation, longer prothrombin time, and lower serum albumin and platelet counts. Endoscopic treatments were more commonly performed in VUGIB patients than in NVUGI Bones (62.3% vs. 20.6%, p<0.001). The overall rebleed in grate was 7.3% and mortality was 1%; with no significant difference between NVUGIB and VUGIB. Hospital complications (39.6% vs. 11.7%, p<0.001) and units of blood transfusion (1.85 vs. 1.46 units, p<0.001) were significantly higher in patients with VUGIB than in those with NVUGIB. In the NVUGIB cohort, lower serum sodium and bleeding from duodenal ulcers were independent predictors of rebleeding, where as female gender, hemodynamic instability, and rebleeding were independent predictors of complications. In the VUGIB cohort, lower platelet count was an independent predictor of rebleeding, and lower serum sodium was an independent predictor of complications. Based on the AIMS65 system, the overall rebleeding rates were 5.3% (8/151), 7.0% (6/86), 18.2% (6/33), 7.1% (1/14), 0% (0/1) and 0% (0/1), and complication rates were 9.3% (14/151), 23.2% (20/86), 48.5% (16/33), 78.6% (11/14), 100% (1/1) and 100% (1/
1), corresponding to the AIMS65 score of 0, 1, 2, 3, 4 and 5 respectively.
Conclusion: The overall outcomes of UGIB were good, with better outcomes in NVUGIB than in VUGIB. AIMS65 score and
serum sodium may be useful in predicting rebleeding and complications in UGIB.

Keywords: Upper gastrointestinal bleeding, UGIB, Gastrointestinal hemorrhage, Peptic ulcers, Esophageal varices,
Predictors, AIMS65 score, Rebleeding, Complications