*S1517 Risk Factors Related to Bleeding in Gastric Tumors after Endoscopic Mucosal Resection Chang Min Cho, Dong Seok Lee, Chang Keun Park, Won Young Tak, Young-Oh Kweon, Sung Kook Kim, Yong Hwan Choi Background/Aim: Endoscopic mucosal resection (EMR) has been widely used for treatment of gastric mucosal tumors because of safety and minimal invasiveness. But the bleeding after EMR has often caused trouble, and it has been regarded as a major complication. Recently, EMR has also been performed on other lesions than conventional indication of EMR, so it is accounted what factors are related to bleeding. We assessed bleeding rates and risk factors related to bleeding after EMR. Methods: We analyzed retrospectively 249 patients with 283 lesions who underwent EMR for flat adenoma (78.8%), hyperplastic polyp (4.9%), and early gastric cancer (16.3%) from January 1999 to August 2003. Male was 161 (64.7%) and female 88 (35.3%), and mean age was 61.4 years. Bleeding during EMR was defined immediate bleeding, bleeding on follow-up endoscopy after EMR was defined delayed bleeding, and bleeding which need to transfusion and angioembolization was defined major bleeding. We followed up endoscopy at one day after EMR. We evaluated risk factors related to bleeding, using univariate and multivariate analysis. Three cases of perforation were excluded. Results: Bleeding after EMR occurred totally in 99 patients (35%). Immediate bleeding occurred in 31.8% and more frequently in case of beginners, upper part, lesions larger than 2cm, or EMR by needle knife. Delayed bleeding occurred in 8.1% and more frequently in case of flat or depressed lesions, carcinoma, piecemeal resection, and occurrence of immediate bleeding. Major bleeding occurred in 10 patients (3.5%). Risk factors related to post-EMR bleeding were experience of operator (beginner vs. expert, p=0.001), anatomical location (body vs. antrum, p=0.009), and method of procedure (needle knife vs. snare or band, p=0.001). Conclusion: Our report showed that the incidence of delayed bleeding was low compared to that of immediate bleeding. We recommend cautious endoscopic management after EMR to prevent bleeding according to experience of operator, anatomical location, and method of procedure. *S1518 Factors Associated with Bleeding After Endoscopic Mucosal Resection of Gastric Tumors Yun Jung Chang, Jong-Jae Park, Ki Ho Park, Gyeng Oh Kim, Hyung Joon Yim, Jin Yong Kim, Jong Eun Yeon, Jae Seon Kim, Kwan Soo Byun, Young Tae Bak, Chang Hong Lee Backgrounds: Endoscopic mucosal resection (EMR) is accepted as a diagnostic and therapeutic procedure for gastric tumors. Bleeding is the most common complications and reported to be 5-15%. The aim of this study is to evaluate risk factors for bleeding associated with EMR of gastric tumors. Methods: Seventy- two patients (39 males, 33 female, mean age 58 years old) who underwent EMR for gastric tumors during the past 1.5 years by single endoscopist were included in this study. Indications for EMR included 15 early gastric cancer, 14 high grade dysplasia, 36 low grade dysplasia, 2 submucosal tumor and 5 histologic diagnosis. The location of the lesion was 14 upper third, 17 middle third, 38 lower third and 3 others. The size of the lesion was classified as #10mm (33, 46%), 11-20mm (31, 43%) and $21mm (8, 11%), and morphology was classified as elevated (23, 32%), flat (29, 40%), depressed (10, 14%), ulcerative (2, 4%) and combined (7, 10%). The method of EMR was C-EMR (EMR with cap) (33, 46%), precut and cut method (23, 32%) and others (16, 22%). Proton pump inhibitor (PPI) of standard dose was prescribed in 28 patients before the procedure and in all the patients after. Bleeding encountered during EMR was defined as ‘‘immediate’’; bleeding after EMR ‘‘delayed’’. The correlations between bleeding (immediate and delayed) and sex, age, indication, location, size, morphology, method, numbers of resection, PPI use before EMR were analyzed using chi-squar and T test. Results: Bleeding occurred in 17 patients (23%) including immediate bleeding in 11 patients (15%) and delayed bleeding in 8 patients (11%). None of the factors were found to be significantly different between the cases with and without bleeding after EMR. The incidence of bleeding tended to be low in patients with PPI use before EMR (P=0.069). Immediate bleeding was observed in 2 patients with delayed bleeding, but was not a significantly correlated factor for delayed bleeding. In 3 patients (4%), perforation occurred at fundus, upper body, and antrum, and was treated by operation, endoclip application and conservative management, respectively. Conclusions: None of the factors were found to be significantly correlated with early and delayed bleeding after EMR. PPI use before EMR seems to be effective for prevention of bleeding and further investigations are needed. *S1519 Chronic Atrial Fibrillation: Considerably Increased Risk for Patients Undergoing the Endoscopy of The Upper Gastrointestinal Tract Lubomir Michalko, Ludovit Gaspar, Stela Hlinstakova The aim of the study: To document the incidence of ECG changes during endoscopy of the upper gastrointestinal tract. We examined 20 patients, mean age 71 years (45 - 86 yrs) with ECG monitoring equipment, mean monitoring time 22 hrs. The endoscopy was performed by endoscope with diameter 7 mm, mean endoscopy procedure duration 3 minutes. Other preexisting diseases: valvular heart disease in 9 patients, hypertension, ischaemic heart disease in 8 patients, diabetes mellitus in 7 patients. Results: Occurence of complex cardiac dysrrythmia during whole monitoring period had been observed in 12 patients, during endoscopy in 6 patients, deterioration of cardiac dysrrhytmia during endoscopy was present in 6 patients. We conclude, that: 1/ endoscopy of the upper gastrointestinal tract in patients having atrial fibrillation induced a complex cardiac dysrrythmia during endoscopy in one third of the patients. 2/ Because these changes can evoke serious haemodynamic consequences, the previous cardiological evaluation is necessary. 3/ In one third of patients the endoscopy induced significant myocardial ischaemia. 4/ Age, preexisting cardiac and pulmonary diseases with hypoxemia, ischaemia or dysrrhytmia are the risk factors for appearence of cardiovascular complications during endoscopy of the upper GI tract. *S1520 Safety of Band Ligator Use in the Small Bowel and Colon Kevin B. Barker In the August 2003 Gastrointestinal Endoscopy(58(2); 274-279), Junquera et al. noted in a case series that endoscopic band ligation for bleeding small bowel vascular lesions was safe and efficacious. There have been several other published case reports of band ligators used for bleeding lesions, usually Dieulafoy’s, in the stomach, proximal small bowel and colon. In addition, this method has been used for post-polypectomy bleeding stalks. There has never been a critical look at the anatomic consequences of banding in the thinner sections of bowel. The purpose of this study is to define the anatomic and histologic consequences of applying band ligator devices to the small and large bowel. Fresh surgical specimens, both large and small bowel, that were excised because of neoplastic lesions were transported to our endoscopy unit where one end of the intact bowel was sutured shut. A standard upper endoscope was passed via the open end and the bowel closed tightly with rubber band ties. The bowel was then insufflated and band ligators were applied to unaffected mucosa using standard technique. Photo documentation from the inside and outside the bowel was obtained. Some of the band polyps were cut above the band and some below the band in the fresh state. Some were fixed in formalin and examined microscopically. Histologic sectioning occurred at the level of the bands. The results were striking in that there were large holes (1 cm) in the fresh ileum specimen. There was gross serosal entrapment manifested by visible puckers on the outer surfaces of the specimens, especially in the small bowel and right colon. The left colon, anatomically thicker, was less affected. The histologic evaluation revealed inclusion by the band ligator of the muscularis propria and serosa on the small bowel, the muscularis propria in the right colon, and the submucosa in the left colon. Based on these findings, we conclude that band ligator devices are not safe in the small bowel and right colon but probably are safe in the thicker left colon. *S1521 Interventional Endoscopy at Ambulatory Endoscopy Centers (AEC): Is It Time? Daus Mahnke, Yang K. Chen, Mainor R. Antillon, William Brown, Roger Mattison, Raj J. Shah BACKGROUND: AECs for routine outpatient endoscopy has become more common because of efficiency, financial, and autonomy considerations; whereas ERCP/EUS continue to be performed in hospital-based units due to concerns of higher complication rates, longer sedation requirements, and the need for post- procedure observation or admission. Whether ERCP/EUS can be safely performed in AECs without compromising patient care has not been evaluated. METHODS: Prospective study of consecutive patients (pts) undergoing ERCP/ EUS at a university multi-specialty AEC with 2 suites equipped for ERCP and EUS. Pts received either general anesthesia (GA) or RN-administered conscious sedation with recovery in a post-anesthesia care unit. Admissions required ambulance transfer to the hospital 6 miles away. Complications were defined by published consensus (CONS) criteria, and by comprehensive (COMP) criteria including >23 hr observation, unplanned MD/ER visit or admission for symptoms. RESULTS: From 3/03-10/03, 334 pts (182 F, median age 58 yr) had 448 procedures (236 ERCP/212 EUS). Median ASA class was 2.0 (ASA 1-88, 2- 277, 3-83). 180 (40%) cases were performed with GA. Conscious sedation (median): midazolam 6mg, fentanyl 150ug, meperidine 100mg. 54% received benadryl. ERCP interventions: 26 diagnostic, 31 SOM (16 biliary, 7 pancreatic, 8 biductal), 91 sphincterotomies (21 pancreatic, 70 biliary), 44 stone extractions (38 biliary, 6 pancreatic), 114 stents (79 biliary, 35 pancreatic). EUS: 108 FNA, 89 diagnostic, 10 cyst aspirations, 4 pseudocyst drainages. 30 of 32 admitted pts were pre-scheduled for 23hr observation. Two unplanned admissions had pancreatitis. Follow-up was achieved in 94%. By CONS: 13 complications (12 ERCP - 5.1%, 1 EUS - <1%). By COMP: 37 complications (31 ERCP - 13.1%, 6 EUS - 2.8%) and one 30d mortality unrelated to the procedure. By ASA(COMP): 1 (12.5%), 2 (7.6%), 3 (6.1%). Types of complications: pancreatitis (2 mild, 6 moderate, 3 severe), bleeding (1 mild), perforation (1 mild), infection (3), >23hr observation (10, all post-SOM), MD/ER visit (9), reversal agents (2). CONCLUSIONS: 1) This is the first report on safety of doing ERCP/EUS in an AEC. 2) Unanticipated hospital admissions were rare (2/448); proper selection of high-risk pts for prescheduled admissions makes this approach feasible. 3) ERCP and EUS can be performed safely in an AEC provided mechanisms for admission and anesthesia support are in place. P116 GASTROINTESTINAL ENDOSCOPY VOLUME 59, NO. 5, 2004