No blood transfusion (n [ 840) (95% CIs) Received Blood Transfusion (n[1078) (95% CIs) Esophagitis/Gastritis/duodenitis 20% 11% Mallory-Weiss Tear 4% 2% Other 5% 8% Lower GI Tract Lesions Diverticulosis 5% 5% Colitis 6% 2% Anorectal 4% 3% Neoplasia 2% 2% Other 9% 7% No source 11% 9% 157 Early Angiographic Embolization After Endoscopic Hemostasis to High Risk Bleeding Peptic Ulcers Improves Outcomes James Y. Lau* 1 , Ka-Tak Wong 1 , Philip W. Chiu 1 , Rapat Pittayanon 2 , Rungsun Rerknimitr 2 , Ingrid L. Holster 3 , Ernst J. Kuipers 3 , Kaichun Wu 4 , Kim W. L. Au 1 , Francis K. L. Chan 1 , Joseph J. Y. Sung 1 1 Institute of Digestive Diseases, Hong Kong, China; 2 King Chulalongkorn University Hospital and Chulalongkorn University, Bangkok, Thailand; 3 Erasmus University Hospital, Rotterdam, Netherlands; 4 Xijing Hospital, Fourth Military University, Xian, China Background: Further bleeding after endoscopic hemostasis to bleeding peptic ul- cers is associated with an at least 3-fold increase in mortality. We sought to determine if pre-emptive angiographic embolization (AE) to high risk ulcers can reduce further bleeding and improves patients’ outcome. Method: Patients with bleeding gastro-duodenal ulcers that required endoscopic hemostasis were enrolled if they fulfilled one of the following criteria; Forrest I a bleeding, ulcer size hypotensive (systolic blood pressure ! 90 mmHg) or hemoglobin % 9g/dl. They were randomized to undergo angiographic embolization within 12 hours of endoscopy plus an intravenous infusion of high dose proton pump inhibitor (IV- PPI) or IV-PPI alone. Our primary endpoint was re-bleeding within 30 days of randomization. Results: Between January 2010 to October 2013, 222 patients were randomized (AE 109, no AE 113). One and 3 patients were excluded after randomization. Baseline characteristics were similar between groups; mean age 65.3 vs. 67.7), in patient bleeding (19.2 vs. 18.6%), mean transfusion before randomization (2.53 vs. 2.6u), mean systolic BP (109 vs. 105mmHg). In the AE group, mean time to angiography was 4.2 (SD 3.3) hour. Of 109 randomized to AE, 16(14.7%) did not receive AE; 6 of 16 re-bled and 3 died (2 after surgery and 1 from major bleeding), 10 without further bleeding and one died. On an ITT analysis, further bleeding occurred in 8/109 (7.3%) and 12/113 (10.6%) of AE and no AE group respectively (PZ0.39, OR, 95%CIZ0.67; 0.26-1.70). On a PP analysis, further bleeding occurred in 4/91 (4.4%) patients in the AE group compared to 12/110 (10.9%) patients on iv-PPI alone (PZ0.089; OR, 95%CIZ2.66, 0.8-8.5). There was no 30 day mortality in the AE group (0/91 vs. 5/110, PZ0.065). Hospitalization and blood transfusion were less in the AE group (6.9 vs. 8.7 days, PZ0.2 and 1.2 vs. 1.9U, PZ0.09 respectively). Reduction in re-bleeding after AE was seen mostly in patients with ulcersR1.5cm in size (2/40 vs. 10/43, PZ0.027) [Figure 1]. Conclusions: Early angiographic embolization after endoscopic hemostasis to pa- tients with gastro-duodenal ulcers R1.5cm in size reduces re-bleeding. 158 Transarterial Angiographic Embolization vs. Surgery in Patients With Bleeding Peptic Ulcers Uncontrolled At Endoscopy; a Multicenter Randomized Trial James Y. Lau* 1 , Ka-Tak Wong 1 , Philip W. Chiu 1 , Ingrid L. Holster 2 , Ernst J. Kuipers 2 , Rapat Pittayanon 3 , Rungsun Rerknimitr 3 , Lars R. Lundell 4 , Kim W. L. Au 1 , Enders K. NG 1 , Joseph J. Y. Sung 1 1 IDD, Prince of Wales Hospital, Hong Kong, China; 2 Erasmus University Hospital, Rotterdam, Netherlands; 3 King Chulalongkorn University Hospital and Chulalongkorn University, Bangkok, Thailand; 4 Helsinki University Hospital, Helsinki, Finland Methods: Patients with massive bleeding uncontrolled during endoscopy were randomized to receive immediate TAE or surgery. Results: Between May 2007 and October 2013, a total of 31 patients were randomized (TAE nZ17, surgery nZ14). Their baseline characteristics were as follows; mean age (SD) Z 72.2 (12.1) vs. 73.6 (8.3), Men Z 10/17 and 9/14, admission Hb Z 7.6 (1.5) and 6.9 g/dl (3.1) and prior transfusion Z 10.9 (8.5) and 9.4 u (7.2). 3 in each group were in-bleeders. Duodenal/ gastric ulcer ratios were 13/4 vs. 11/3. Of 17 randomized to TAE, there were 4 initial treatment failures; failure to cannulate celiac artery 1, failure to access gastroduodenal artery 1, no empirical embolization in the absence of contrast extravasation and pseudoaneurysm 1 (patient rebled on day 3) and 1 patient rushed to surgery before TAE became available. Thirteen received embolization. Of them, 5 rebled; 4 underwent surgery (Polya gastrectomy 2, ulcer plication 2) and one received further TAE. Of 14 randomized to surgery, one patient with end-stage renal failure was considered unfit and did not undergo surgery. Types of surgery per- formed were ulcer plication or excision 8, gastrectomy 3, vagotomy and pyloroplasty 1 and gastrojejunostomy 1. One patient was diagnosed with cancer to head of pancreas with liver metastases during laparotomy. After surgery, further bleeding occurred in 3 of 13 patients. These were treated by TAE. Analysis was by ITT. There was no death in either group on day 30. Treatment failures were more frequent in the TAE group (9/17 vs. 3/14, PZ0.052, OR; 95%CIZ4.7; 0.9-23.7). Patients in the TAE group needed more transfusion (median, range; 2 u, 0-18 vs. 0 u, 0-9, PZ.058) but shorter ICU stay (0, 0-6 vs. 2, 0-8, PZ.037). Hospitalization in both groups was similar (9, 3-36 vs. 12, 3-27, PZ.91). Conclusions - TAE is associated with higher rate of treatment failures when compared to surgery. Outcomes following either approach are similar. 159 Cost-Effectiveness Analysis Comparing Upper Endoscopy vs. Angiographic Embolization in Recurrent Non-Variceal Upper GI Bleeding Ari Garber*, Mazen Albeldawi, Maged K. Rizk, Paresh P. Mehta, Rocio Lopez, John J. Vargo, Sunguk Jang Cleveland Clinic, Cleveland, OH Background: Non-variceal upper GI bleeding (NVUGIB) remains a significant source of morbidity and mortality with incidence rates ranging from 103-172 events per 100,000 adults, hospital admissions approaching 400,000 annually, recurrence rates estimated at 20% and mortality approaching 10-14%. Total annual health care costs for NVUGIB have been estimated in excess of 2 billion dollars. The purpose of this study was to ascertain the comparative cost-effec- tiveness and health outcomes of upper endoscopy vs. angiographic embolization in recurrent NVUGIB. Methods: This was a retrospective study at a tertiary care referral center between January 2008 and November 2011. All patients O 18 years of age admitted with a diagnosis of acute NVUGIB were included. Rebleeding was defined as NVUGIB following index endoscopic therapy within 30 days of an initial bleed. Demographics, presentation, comorbidities, medications, serum and he- modynamic parameters were collected. Data was analyzed as cost-effectiveness ratios with p-values corresponding to t-tests and Fieller’s confidence intervals for the ratio of two means. Results: A total of 251 patients underwent EGD for NVUGIB (mean age 64 Æ 13 years, 39% female, 74% Caucasian). The most com- mon source of bleeding was peptic ulcer disease (42%). A total of 123 (49%) patients had a recurrent bleed after index endoscopy. A total of 43 angiographic procedures were performed. The average time to an episode of rebleeding was 3.5 Æ 6.1 days. The majority of patients only had 1 episode of rebleeding (58%). Average length of ICU stay was 15 days. The in-hospital mortality rate was 18% for all NVUGIB. The total admission cost was 52% higher in subjects with rebleeding as compared with those without recurrent NVUGIB (Figure 1). In subjects with recurrence, angiographic embolization cost was 7-times higher in comparison with endoscopic intervention. Total treatment costs for patients who received both EGD and angiographic procedures was 5.2-times higher than that for pa- tients who only had endoscopic interventions regardless of the number of en- doscopies performed (95% CI: 4.07 - 6.95; p !0.001). After adjusting for confounding factors, subjects who underwent EGD at any given bleeding episode were 80% more likely to achieve hemostasis in comparison to those who under- went angiographic embolization (ORZ 1.8; 95% CI 1.04 - 3.3). Conclusions: Repeat EGD with therapeutic intervention is superior in terms of both cost- effectiveness and health outcomes when compared to angiographic embolization for recurrent non-variceal upper GI bleeding. 160 Comparative Effectiveness of Novel Techniques for Barrett’s Esophagus Screening in the Community: a Prospective Randomized Trial Sarmed S. Sami* 1 , Kelly T. Dunagan 2 , Michele L. Johnson 2 , Cathy D. Schleck 3 , Alan R. Zinsmeister 3 , Louis M. Wong Kee Song 2 , Kenneth K. Wang 2 , David a. Katzka 2 , Krish Ragunath 1 , Prasad G. Iyer 2 1 Digestive Diseases Biomedical Research Unit, University of Nottingham, Nottingham, United Kingdom; 2 Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN; 3 Biomedical Statistics and Informatics, Mayo Clinic, Rochester, MN Introduction and Aims: Barrett’s Esophagus (BE) is the strongest precursor of esophageal adenocarcinoma. Participation patterns and clinical effectiveness of BE community screening using unsedated transnasal endoscopy (uTNE) versus sedated endoscopy (sEGD) is not known. Feasibility of screening in a mobile van closer to the patient’s home is also unknown. We aimed to assess the comparative effec- tiveness of these techniques in BE screening. Methods: A population cohort of subjects, R 50 years of age, without previous history of endoscopic evaluation, was www.giejournal.org Volume 79, No. 5S : 2014 GASTROINTESTINAL ENDOSCOPY AB113 Abstracts