Endoscopic hemostasis for bleeding gastric varices treated with combination of variceal ligation and sclerotherapy with N-butyl-2-cyanoacrylate

*4354 ACUTE NON-VARICEAL UPPER GASTROINTESTINAL HAEMORRHAGE - VAI,mATING A RISK STRATIFICATION TOOL TO IDENTIFY PATIENTS AT LOW RISK OF COMPLICATIONS AND CONSIDERED SUITABLE F O R SAFE AND EARLY DISCHARGE, Geeta Srivatsa, Inn Kronborg, Anne-Maree Kelly, Debra Kerr, Shamilah Lachal, Amanda Nicell, Western Hosp, Footscray, VI Australia Objective: The primary objective of this study was to validate a risk stratification tool for the identification of patients with symptoms of upper nonvariceal gastrointestinal haemmorhage at low risk of complications, suitable for safe and early discharge post gastroscopy. Methods: A risk stratification tool was developed combining gastroscepic findings and clinical assessment variables. This was assessed for its predictive value for adverse events and potential cost savings, by a retrospective record analysis of all patients who presented to the Emergency Department in 1999 who had a gastroscope for symptoms of upper gastrointestinal haemorrhage. Data collected included demographic information, clinical assessment variables, cemorbidities, gastroscope findings, clinical diagnosis, length of stay and timing of events. Results: In the retrospective study, eighty-five percent of patients(78 out of 92) achieved low or intermediate risk status. If these patients had been discharged according to the guide, a total of 67 bed-days would have been saved. Average time form presentation to the ED to scope was 24 hours. 18 patients had their scope _>24 hours from presentation, 11 of whom were discharged within a day from procedure. This delay in gastroscepy seemed to increase length of stay. Adoption of the tool would have resulted in no changes in morbidity or mortality, and a 40% reduction in length of stay would have occurred. In the prospective study, eighty-one percent (21 out of 26) of patients achieved low or intermediate risk. Adoption of the tool has not incurred any adverse events as a consequence of early discharge, and average time from presentation to scope declined to 14 hours when compared with previous practice. There has been a reduction in hospital length of stay by an average of 12 hours for patients considered at low and intermediate risk. Conclusion: The proposed risk stratification tool identified retrospectively a population of patients with non-variceal upper gastrointestinal haemorrhage at low risk of complications, suitable for safe and early discharge. The tool is currently being prospectively validated.

*4356

THE EFFECT OF THE BIOPSY NUMBER ON SENSITIVITY AND SPECIFICITY OF ULTRA-RAPID UREASE TEST IN DETECTING HELICOBACTER PYLORI: A PROSPECTIVE STUDY Li-Lin Lira, Bow He, Manuel Salto-Tellez, Khek-Yu He, National Univ Hosp, Singapore Singapore Aim:Prospectively assess sensitivity, specificity and time to positivity of URUT in detecting H.pylori, comparing los 2 biopsies in patients with upper endoscopy. Method:The effect of test duration on the sensitivity and specificity of a self prepared urease test in detecting H.pylori was studied in consecutive patients with upper endoscopy where biopsy indicated. Five antral biopsies were taken: 2 URUT sets for 1 and 2 biopsies,and 2 for H.pylori culture and histology. Times for URUT reading:l, 5, 10, 20, 30, 60,120,130 minutes, and 24 hours after biopsies inserted into the urease test reagent. Reference gold standard was positive histology or culture. Results:H.pylori was positive in 28/53(52.8%) patients. See table for sensitivity data. Specificity was 100% for all time points except at 24 hours: 96% for 1 biopsy and 92% for 2 biopsies. Using a receiver-operating characteristic curve, an optimal combination of sensitivity and specificity was obtained when the urease test was read at 66 minutes. Conclusion:URUT was more sensitive for detecting H.pylori when 2 rather than 1 biopsies were used at all time points up to 60 minutes. The optimal time for reading the URUT for sensitivity and specificity was 60 minutes regardless of the number of biopsies obtained. Sensilivity at various timeS: %

1 Biopsy

2 Biopsies

1m 5m 10 m 20 m 30 m 60 m 120 m 180 m 24 h

3.6 32.1 46.4 64.3 71.4 82.1 82.1 85.7 96.4

10.7 50 60.7 75 82.1 85.7 85.7 85.7 96.4

m=Minules, h=hours

*4355

EVALUATION OF THE OLYMPUS SIF-Q140 ENTEROSCOPE FOR THE INVESTIGATION OF SUSPECTED SMALL BOWEL PATHOLOGY. Deepak S. Desai, Jamie S. Barkin, Rafael Amaro, Univ of Miami/Mt Sinai Medical Ctr, Div of Gastroenterology, Miami, FL; Jack Lubin, Univ of Miami, Sch of Med/Mt Sinai Medical Ctr, Dept of Pathology, Miami, FL Enteroscopy is the procedure of choice for evaluation of patients with obscure GI bleeding and suspected small bowel (SMB) disease. A newly developed SIF-Q140 (Olympus Corporation) small bowel enterascope 0 70% larger image and higher reso(SBE) with greater field of view (140), lution is 250cm in length and is commercially available. Purpose: Prospective evaluation of the clinical utility of the SIF-Q140 SBE with overtube(OT). Population: 33 consecutive patients(PTS)-21 with occult GI bleeding and 12 with suspected SMB pathology. Method: Data included distance advanced into the SMB beyond ligament ofTreitz (LT) via x-ray confirmation; distance of OT advancement measured from x-rays; resistance (RT) to insertion of OT, rated no, mild, moderate, or severe; ability to perform biopsy with non-spiked forceps including RT (none, mild, moderate, severe) and adequacy of specimen evaluated as follows: muscularis mucosa (MM) present (yes/no), number of contiguous villous crypt gland complexes perpendicular to MM, squashed/fragmented (none, minimal, severe), adequacy for interpretation (yes/no); adequacy of mucosal visualization, rated excellent, good, fair or poor; tip defiection(TD) at maximum insertion; PT tolerance, graded good, fair or poor and complications. Results: Insertion beyond the LT mean 134.6cm (range 80-195). OT advancement into 2nd duodenum in 30/31 (2 PTS OT not passed due to varices & severe kyphosceliosis), one PT OT into stomach. RT to OT insertion moderate in 1, mild in 2 and none in 28 PTS. RT to the passage of biopsy forceps (biopsy obtained in 21 PTS) was severe (could not pass forcep at full insertion) in 2, moderate in 1, mild 1 and none in 17 PTS. Adequacy of specimens: MM present 18/21 PTS; number of contiguous villous crypt gland complexes perpendicular to MM mean 2.55/specimen; squashed/fragmented none 1, minimal 18, severe 2; adequate for interpretation 19/21. SMB visualization excellent in 32 and good in one. Tip deflection at maximum insertion was excellent in 29 and good in four. Patient tolerance was good in all cases. Complications included 2 PTS with trauma (gastric & 2 nd duodenum) due to OT passage without any significant clinical consequences. Conclusion: The SIF-Q140 SBE allows excellent SMB visualization with full tip deflection to a maximum depth of 195cm beyond the LT with good PT tolerance. SMB biopsies adequate for interpretation were obtained in 90% of PTS. Summary: The SIF-Q140 is an improved enteroscope.

AB210

GASTROINTESTINAL ENDOSCOPY

*4357

ENDOSCOPIC HEMOSTASIS FOR BLEEDING GASTRIC VARICES TREATED WITH COMBINATION OF VARICEAL LIGATION AND SCLEROTHERAPY WITH N-BUTYL-2-CYANOACRYLATE. Kenichiro Watanabe, Kazuyo Okamoto, Tomofumi Ogusu, Rye Shimoda, Keiji Matsunaga, Hiroyuki Sakata, Ryuichi Iwakiri, Kazuma Fujimote, Saga Medical Sch, Saga Japan Bleeding of gastric varices is a lethal complication for liver cirrhosis and several modalities including surgical and endoscopic approach were applied for hemostasis. The aim of the present study was if the endoscopic scleretherapy with injection of N-butyl-2-cyanoacrylate combined with variceal ligation was useful for hemostasis of bleeding gastric varices. Twenty six patients with bleeding gastric varices underwent endoscopic treatment with variceai ligation and sclerotherapy with N-butyl-2-cyanoacrylate. Patients underwent endoscopic variceal ligation only for the bleeding spot just before sclerotherapy with N-butyl-2-cyanoacrylate. Injection was continued until varices were engorged, and mean volume of N-butyl-2-cyanoacrylate in the initial therapy was 5.2 ml. Among these patients, 11 had active bleeding and 15 had recent bleeding within 24 h with white or red plaques on gastric varices. All varices presented nodular or tumorous forms. After these therapy, patients were followed endescopically with re-treatment as necessary. The hemostasis rate at one week after treatment with N-butyl-2-cyanoacrylate was 88.6% (23/26). Among the patients of achievement of initial hemostasis for one week, 30.4% (7/23) experienced recurrent bleeding during 5 and 53 months after the initial treatment. Four patients of re-bleeding were treated with N-butyl-2-cyanoacrylate. Finally, the rate of ultimate hemostasis was 79.1%. Three patients, who were not achieved initial hemostasis by combination therapy, were dead. There were few long-term complications except for N-butyl2-cyanoacrylate cast extrusion related mucosal defects. To date, 10 patients have died, mostly as a result of uncontrollable esophago-gastric varices and/or liver failure, 16 are still alive. The survival rate for 2 years was 73.1%, which was equivalent to that of the patients of bleeding esophageal varices treated with sclerotherapy in our hospital (69.0%). These data indicated that endoscopic variceal ligation for bleeding gastric varices achieved transient hemostasis, which was helpful for following sclerotherapy with N-butyl-2-cyanoacrylate. Combination of these two methods is one of recommended choices for the effective treatment of bleeding gastric varices.