3.6±1.2 in the elderly and younger group, respectively (p=NS). The cumulative recurrencerate of varices was 21% and 53% at 3 and 5 years of the observation period, respectively,in the elderly group, while it was 55% and 66%, respectively, in the younger group (p<0.001).The cumulative rebleeding rate was also better in the elderly group: 2% at each time pointin the elderly vs. 14% at each time point in the younger group. There was, however, nosignificant difference in the cumulative survival rate between the two groups. Regardingcomplications, the younger group had a higher incidence of postoperative epigastralgia,whereas the elderly group tended to have a higher incidence of intraoperative systemichypertension, postoperative esophageal stricture, and pleural effusion Conclusions: Elderlycases showed a significantly lower incidence of recurrent esophageal varices as comparedto younger cases. Since liver transplantation is rarely indicated in the aged group, pharmacol-ogic and/or endoscopic therapy occupies an important place. Although endoscopic treatmentas applied to elderly patients tended to cause intra-operative blood pressure elevation, therecurrence rate of varices was lower, and severe complications rarely accompanied thistreatment. Hence this procedure seemed to be useful from the view point of quality of life,and as a strategy for the control of hemorrhage.

S1954

Autonomic Nervous System Dysfunction in Cirrhosis - Heart Rate Variability.Does It Act As a Predictor of Variceal Bleed?Joye Varghese, Arun Kumar Muthusamy, Murali Krishnan K, Jayanthi Venkatraman

Background: The tests for autonomic nervous system (ANS) dysfunction are complex andvaried. ANS dysfunction has been reported in cirrhosis of the liver Short term i.e. 5 minutesheart rate variability (HRV) analysis as a simple test of autonomic nervous system functionhas not been studied in cirrhosis. Aim: The present study aimed at applying this test forunderstanding the autonomic nervous system integrity in cirrhotic patients and at determiningthe impact of this test in predicting variceal bleed in cirrhosis. Materials and Methods: HRVwas performed in 14 cirrhotic men, non-bleeders, age range 20 - 50 years, non diabetic,non-alcoholic & non-hypertensive and compared with age matched male healthy controls.HRV was done using RMS polyrite 2.5.2 software in supine rest and standing position.Mean RR interval, SDNN (SD of all the Normal to Normal intervals), and the Lomb frequencystatistics i.e. (I) Low Frequency (LF) & High frequency (HF) power (in percentage (%) andnormalized units (n.u)), (II) Total power and (III) LF/HF ratio were analyzed. Results:Cirrhotic patients had significantly lower (p < 0.05) SDNN (implying reduced HRV) & anincrease in LF power (p value 0.003) i.e. increase in sympathetic tone in supine posture.In upright posture, the HF power was significantly low (p value 0.005) i.e. a loweredparasympathetic tone. The LF/HF ratio was also significantly low (p value 0.01) i.e. nochange in sympathetic tone in upright position. There was no significant difference in thevariables among bleeders and non-bleeders in supine and standing position and duringchange of posture. Conclusion: Patients with cirrhosis liver not only have a significantlyreduced heart rate variability compared to controls but also have an altered sympathovagalbalance with an absent increase in sympathetic tone on standing. But heart rate variabilityhas no influence on the bleed pattern in cirrhotic patients as the above test did not differsignificantly amongst bleeder and non-bleeder cirrhotic patients

S1955

Safety & Eradication Rate of Primary Prophylactic Esophageal Variceal BandLigationWasim Jafri, Mohammad Salih, Amna Subhan, Nida Jafri, Saeed S. Hamid, Hasnain Shah,Shahab Abid, Rozina N. Wasaya

Background: Esophageal Variceal Band ligation (EVBL) is recommended for Primary prophy-laxis of variceal bleed but its long term safety & benefits are unknown. Aims: To determinethe safety and eradication rate of EVBL as primary prophylaxis for esophageal varices.Methods: Patients with cirrhosis and large Esophageal Varices (> 5 mm in diameter) onscreening upper Gastrointestinal Endoscopy (EGD) were enrolled for primary prophylacticesophageal variceal band ligation. Primary end points were safety and eradication rate ofEVBL. Secondary end points were development and/or progression of portal gastropathyand gastric varices. Results: A total of 156 patients with mean age of 50.28 ±10 years wereenrolled. There were 103(67.3%) males. Most common etiology of cirrhosis was HCV in106 (69.3%) followed by HBV in 19(12.4%) cases. Majority of the patients had Child's B/C (45.8%/43.1%). Concomitant gastric varices and portal gastropathy was seen in 33(22.3%)and 105(68.6%) cases respectively. Post eradication 1-year follow up was completed by 129patients and they were analyzed for safety and eradication rate of EVBL; eradication achievedin 106(82.1%) with 2-4 sessions in 91 % cases. During EVBL mild bleeding occurred in6(3.9%) cases which did not require admission or vasopressors. Post EVBL retrosternal pain,transient dysphagia and fever reported in 96.7%, 94.8% & 3% cases respectively. Progressionof Portal gastropathy was seen in 8.5% and formation or worsening of gastric varices wasseen in 6.5% cases. Variceal bleed occurred in 7(4.6%) cases during eradication and 4(2.6%)patients died due to causes other than variceal bleed. While 12(7.8%) patients lost followup. One (0.9%) patient experienced variceal bleed during 1 year follow up. Conclusion:Esophageal Variceal Band ligation as primary prophylaxis of variceal bleed is safe and effectivewith acceptable adverse effect profile. Key words: Esophageal varices, Primary prophylaxis,Esophageal Variceal Band Ligation

S1956

The Frequency of Endoscopic Diagnosis Other Than Esophageal Varices inPatients with Portal Hypertension and Esophageal VaricesLuis F. Lara, Shou Jiang Tang, Jayaprakash Sreenarasimhaiah, Samir Gupta, Marlyn J.Mayo, Peter F. Malet, William M. Lee, Don C. Rockey

Introduction: Variceal bleeding is a frequent cause of GI bleeding in patients with cirrhosis,particularly in patients with high risk endoscopic stigmata, and a high C-P-T or MELDscore. It is thought that bleeding from peptic ulcer disease is the most common cause of

T : 11501$$CH304-02-08 17:16:45 Page 791Layout: 11501B : o

A-791 AASLD Abstracts

UGI bleeding in patients with cirrhosis, despite a general lack of data to support thispostulate. We hypothesized that UGIH hemorrhage in cirrhotics is mostly due to varices.Methods: Patients with GI bleeding that had endoscopy between 1/06 and 8/07 wereprospectively captured in the UTSW GI Bleed database, including demographic and clinicaldata. Clinical characteristics were compared between patients without esophageal varices(EV), and those with EV. The presence of other endoscopic diagnosis were evaluated in theEV group. Results: Out of 473 patients with GI bleeding, eighty (17%) were diagnosed withEV based on clinical data, stigmata of cirrhosis, previous variceal bleed or histology. Sixty-one (76%) of the 80 patients had EV only, whereas 19 (24%) patients had another endoscopicdiagnosis beyond EV. This included one or more of the following: portal hypertensivegastropathy (PHG) (n=14), gastric ulcer (GU) (n=5), esophagitis (n=4), and one each ofMallory-Weiss tear, esophageal erosions, gastric erosions, and intestinal ischemia. In the EVgroup, 25% smoked and 28% consumed alcohol, not different than the non-EV cohort.The INR was surprisingly similar (1.6 vs 1.3; p=NS), as were hemoglobin before and afterendoscopy, and transfusions, but platelets were significantly different 107 vs 338 x 103/uL,p< 0.001. 24% of patients with EV had active bleeding (hematemesis, melena or > 6g HCTdecrease), and 44 of the 80 patients (55%) with EV were banded due to active bleeding orthe presence of high risk stigmata. The Rockall score was significantly higher, 6.7 in patientswith a gastric ulcer and EV compared to 4.1 with EV alone, (p=0.02). There were 5 deaths(6%), all in the EV only group. The five patients with a GU had low grade stigmata, anddid not require endoscopic therapy. They were placed on acid suppression and beta-blockers.Conclusion: In a population of patients with GI bleeding and esophageal varices, 24% hadanother endoscopic diagnosis besides the EV that could explain their clinical presentation.PHG was the most common diagnosis, followed by gastric ulcers. In patients with EV, ahigh Rockall score should raise the suspicion of a gastric ulcer.

S1957

The Changes in Clinical Outcomes of Variceal Bleeding in Cirrhotic Patients ;a 10-Year Experience in Gangwon Province, South KoreaYoung Don Kim, Ki Tae Suk, Chang Soon Choi, Koon Hee Han, Woo Jin Jeong, HyeYoung Choi, Moon Young Kim, Hye Jeong Kim, Jae Woo Kim, Sung Jung Kim, GwangHo Baik, Jin Bong Kim, Gab Jin Cheon, Soon Koo Baik, Dong Joon Kim

Background/Aims : Variceal bleeding is one of the important causes of upper GI bleeding.And the early mortality rate of variceal bleeding has been reported up to about 30% incirrhotic patients. The aim of our study is to evaluate whether there were any changes inclinical outcomes and clinical characteristics of acute variceal bleeding patients during thelast 10 years, and to clarify prognostic factors for predicting mortality. Methods : We reviewedthe medical records of upper GI bleeding patients who visited emergency centers of 3 tertiaryhospitals in Gangwon province, South Korea, during 3 periods (August 1996 ~ July 1997,August 2001 ~ July 2002, August 2006 ~ July 2007). Clinical outcomes and clinical character-istics of 3 groups by period were compared. Subgroup analysis for variceal bleeding wasalso performed. Results : During the study period, a total of 1,753 upper GI bleeding caseswere reported. The etiology were peptic ulcer (n=842, 48.0%), variceal bleeding (n=639,36.5%), Mallory-Weiss tear (n=139, 7.9%), hemorrhagic gastritis (n=33, 1.9%), and others(n=100, 5.7%). The variceal bleeding rate by period did not decrease (26.0% vs. 44.3% vs.35.7%, p<0.05) and the mortality rate was significantly higher than the non-variceal bleedinggroup (11.1% vs. 2.0%, OR 6.205, 95% CI 3.805-10.116). There were 751 cirrhotic patients,and the etiology were alcohol (n=501, 66.7%), HBV (n=198, 26.4%), HCV (n=22, 2.9%),and others (n=30, 4.0%). In the variceal subgroup analysis (mean age 52.3±11.0, 570 male(89.2%)), there were 574 esophageal (89.8%) and 65 gastric variceal bleedings (10.2%).The mortality rate of variceal bleeding had a decreasing tendency by period but had nostatistical significance (15.0% vs. 11.7% vs. 8.9%, p=0.229). The prognostic factors in thevariceal group were Child-Pugh class, initial systolic blood pressure, and number of transfusedpacked RBC unit (p<0.05). Conclusions : Variceal bleeding is still the second most commoncause of upper GI bleeding, and their prevalance and mortality rate has not significantlydecreased during the past 10 years in Gangwon province, South Korea.

S1958

Validity Scores of Different Reagent Strips and Automated Cell Count for theDiagnosis of SBPDanai Limmathurotsakul, Narudee Bhokaisawan, Panarat Thaimai, Rungsun Rerknimitr,Sombat Treeprasertsuk, Varocha Mahachai, Pinit Kullavanijaya

Aims: The standard criteria for diagnosis of spontaneous bacterial peritonitis (SBP) are anascitic fluid polymorphoneuclear (PMN) cell count of ≥ 250/mm3 and/or a positive asciticfluid bacterial culture. However, the results from these criteria are not promptly availablein the emergency situation. Automated cell count and reagent strip test have been used fora more rapid diagnosis of other infections including UTI and meningitis. We, therefore,evaluated their usefulness by testing the validity scores of automated cell count and reagentsstrip tests for SBP diagnosis in ascites from patients with cirrhosis. Methods: 150 consecutiveparacentesis in cirrhotic patients were performed. All ascitic fluid samples were sent foranalysis with automated cell count and three reagent strips:1) Aution sticks,A.MenariniDiagnostic, Firenze, Italy 2) Combur10 Test M, Roche, Mannheim, Germany 3)Multis-tix10SG, Bayer Corporation, Elkhart, USA. Manual cell count result for PMN of ≥ 250/mm3 was referred as a gold standard. The sensitivity, specificity, positive predictive values(PPV), negative predictive values (NPV), and accuracy for diagnosis of SBP by differenttechniques were compared. Results: SBP was diagnosed by manual cell counts in 16 speci-mens. The sensitivity, specificity, PPV, NPV and accuracy of 1+ cut off scale by Aution stickstest, Multistix test, Combur test and automated cell count result are shown as table. Summary:With similar percentage of sensitivity when compare to Aution stick and Combur test,automated cell count provides better results in other validity scores. Among the three strips,all validity scores are comparable except multistix that contains the lowest sensitivity.Conclusion: Automated cell count is as sensitive as many reagent strips for a rapid diagnosisof SBP. However, it provides better specificity, PPV, NPV and accuracy.Validity scores for SBP diagnosis by automated cell count and different reagent strips

AA

SL

DA

bst

ract

s