Among these patients, 223 patients were satisfied with our UGIB criteria. We assessed these 223

patients by GBS, RS and AIMS65. We defined unfavorable outcome as requiring a clinical intervention or death in-hospital within 30 days. And then the predictability of each scoring system for clinical outcome compared between each other. Results: In-hospital mortality was 2.2% (5/223). 43.9% (98/223) of all patients was unfavorable outcome. The re-bleeding rate was 9.4% (21/223). Of scoring systems, GBS was significantly increased in unfavorable outcome patients (p = 0.0119) and was most superior than other scoring systems in predicting the need for packed red blood cell (PRBC) transfusion (p = 0.0134). Overall among the three scoring systems there was no significant difference in predicting re-bleeding, the need of therapeutic intervention and death. Conclusion: Assessment Selumetinib research buy for acute upper gastrointestinal bleeding (AUGIB) has been performed through several scoring systems. Clinicians can predict the need for intervention and unfavorable clinical outcome by these scoring

systems. Our study showed there was no significant difference in predicting clinical outcome among these systems. But GBS was superior than other scoring systems to predict unfavorable outcome patients and need for transfusion. Key Word(s): 1. BMS-907351 in vitro gastrointestinal bleeding; 2. Rockall score; 3. Glasgow-Blatchford score; 4. AIMS65 Presenting Author: WEI-CHEN TAI Additional Authors: SENG KEE CHUAH, KENG LIANG WU Corresponding Author: WEI-CHEN TAI Affiliations: Kaohsiung Chang Gung Memorial Hospital, Kaohsiung Chang Gung Memorial Hospital Objective: Infections in cirrhotic patients with upper gastrointestinal bleeding are a common complication causing severe complication and mortality. Antibiotic prophylaxis has been recommended for cirrhotic patients with variceal hemorrhage but little is known about the effect for peptic ulcer bleeding. This study aimed to evaluate the antibiotic prophylaxis on prognosis in cirrhotic patients with peptic ulcer bleeding after endoscopic hemostasis and to identify risk

factors predictive of re-bleeding, bacterial infection and in-hospital mortality. Methods: The medical records of 426 patients with acute peptic ulcer bleeding who had received endoscopic hemostasis between January 2008 and January 2014 were reviewed. Two hundred and thirty-five patients were enrolled MCE after strict exclusion criteria. Patients who received prophylactic intravenous ceftriaxone were classified as group A (n = 88) while those who did not receive antibiotics were classified as group B (n = 147). The outcomes were length of hospital days, bacterial infection, rebleeding and in-hospital mortality. Multivariable analysis was performed to determine predictors of death, ulcer rebleeding and infection development. Kaplan-Meier survival analysis was used to compare the mortality between two groups and in subgroups between patients with compensated and decompensated cirrhosis.