bloating compared to women treated with placebo. There was a similar trend for relief of daily bloating in men, but this difference did not achieve statistical significance, which may be due to the smaller number of men studied. 1042 “Experts” Assess the Accuracy of Celiac Disease Diagnosis Performed in the Community Setting Sonia I. Niveloni, Ana M. Cabanne, Horacio Vázquez, Emilia Sugai, Miriam Galich, Edgardo Smecuol, María Laura Moreno, Maria Ines Pinto Sanchez, Roberto M. Mazure, Hui Jer Hwang, Eduardo Mauriño, Julio C. Bai Background: The diagnosis of celiac disease (CD) relies on well-established histological and serological bases. In the past few years, the awareness of and interest in CD has grown considerably in the medical community. However, this increased focus in CD has not been accompanied by a parallel rise in the expertise. Furthermore, a retrospective analysis has suggested diagnostic deficiencies in the community setting. Aims: We seek to evaluate: 1- the level of agreement among the histological and serological diagnoses made in the commun- ity setting compared to the diagnosis made in an academic specialty center for subjects who were looking for a “second opinion”; 2- the degree of concordance in the histological analysis between two expert pathologists; and 3- the potential impact of misdiagnosis on patients. Methods: Original biopsy slides and pathological reports, which used for CD diagnosis in the community setting, and CD serology test results (IgA tissue transglutaminase antibodies) were analyzed in 65 consecutive patients attending our academic institution for a “second opinion” (39 of these patients were originally diagnosed with CD while in 26 cases diagnosis was ruled out). One expert pathologist (CA) reviewed original histological slides unaware of the results of the external diagnosis, and other findings. A second expert pathologist (GM) reviewed slides also without knowing these assessments. Serological tests were repeated at the institution under similar clinical conditions. Finally, an expert team also blinded of the external diagnosis made the final diagnosis on conventional criteria. Results: In 6 cases (9.2%), the quality of the original slides was considered not evaluable by the expert patholo- gist. Furthermore, we detected a divergent histological diagnosis between the community and academic pathologists in 12 of the remaining 59 cases (20.3%) (Cohen's kappa -κ-: 0.59). According to the expert pathologist assessment, 75% of misdiagnosed cases led to the overdiagnosed of CD. Agreement between expert pathologists was excellent (κ: 0.85). Fifty-four of 63 cases (85.7%) had congruent serology results in both settings (κ: 0.71). Globally, our “expert” team determined that 20% of the cases consulting had a divergent diagnosis than the one identified in the community setting. Overdiagnosis of CD (34.5%) was more common than cases of underdiagnosed (13.0%). Conclusion: One in every five cases that sought a “second opinion” for CD diagnosis in the community setting received a divergent diagnosis by experts. Furthermore, 27.7% of the histopathological diagnoses in the community practice were considered inadequate or misdiagnosed by an expert. In contrast, serology had greater concordance between the two settings. The degree of agreement between expert pathologists suggests that they are relevant for reducing the frequency of misdiagnosis. 1044 Effect of Eradication of Helicobacter pylori on Recurrence After Endoscopic Resection of Gastric Adenoma and Early Gastric Cancer Suh Eun Bae, Hwoon-Yong Jung, Ji-Hoon Jung, Do Hoon Kim, Ji Yong Ahn, Kwi-Sook Choi, Mi Young Kim, Jeong Hoon Lee, Kee Don Choi, Ho June Song, Jin-Ho Kim, Gin Hyug Lee, Young Soo Park OBJECTIVES: Helicobacter pylori infection induces chronic inflammation in the gastric muco- sae and promotes the development of gastric cancer. The effect of H. pylori eradication on recurrence after endoscopic resection of gastric cancer remains uncertain. METHODS: We retrospectively assessed outcomes in 2089 adult patients, aged 28-88 years, who underwent endoscopic resection of gastric adenoma and early gastric cancer from November 2004 to December 2008 at the Asan Medical Center. We excluded 602 patients, including those followed-up <1 year, those with recurrence at ≤3 months (thus excluding patients with primary gastric cancers not detectable at the time of endoscopic treatment), and patients not tested for active H. pylori infection, including by urea breath tests, rapid urease tests or histology. We investigated the demographic and tumor characteristics, clinical status of the stomach, and incidence of recurrence among the three groups, those without active H. pylori infection (Hp- group, n=530, 35.6%), those with successful H. pylori eradication (eradication group, n=669, 44.9%), and those with failed or no H. pylori eradication (non-eradication group, n=288, 19.3 %). RESULTS: Of the 1487 enrolled patients, 1149 (77.7%) were male. Median patient age was 62 years (range, 28-88 years). Overall, 7.9% patients had metachronous recurrences during a median 49 months (range, 29-147 months). Median time to recurrence was 16 months (range, 5-66 months). When we analyzed recurrence- free survival using the Kaplan Meier method, we found that the cumulative incidence of recurrence-free survival differed significantly among the three group. There were no signific- ant differences in recurrence rate (6.0% vs 5.1%, p=0.508) and recurrence-free survival time between the eradication and Hp- groups. However, recurrence rates (17.0% vs 51%, p<0.01; hazard ratio 3.6 in a multivariate Cox proportional hazard model) and recurrence-free survival differed significantly between the non-eradication and Hp- groups. In addition, recurrence rates (17.0% vs 6.0%, p<0.01; hazard ratio 2.8) and recurrence-free survival differed significantly between the non-eradication and eradication groups. CONCLUSIONS: The results of this large-scale retrospective study suggest that H. pylori eradication reduces metachronous gastric cancer in high risk patients with a history of gastric adenoma or EGC and severe atrophy and metaplasia. Together with the results of previous studies, our findings indicate a causal relationship between H. pylori infection and gastric cancer. Therefore, H. pylori eradication may prevent the development of gastric cancer, especially in high- risk patients. Multivariate Cox proportional hazards model for recurrence free survival S-183 AGA Abstracts There was significant difference in recurrence free survival rate between non-eradicated group and Hp- group (Hazard Ratio 3.6). Also there was significant difference in recurrence free survival between non-eradicated group and eradicated group (Hazard ratio 2.8). Signific- ant other risk factor of recurrence free survival was only sex. 1045 A Neutral Intragastric pH is Required for Eradication of Helicobacter pylori by Amoxicillin Elizabeth A. Marcus, George Sachs, Nobuhiro Inatomi, David R. Scott Background: The current success of triple therapy for H. pylori eradication is <75%, largely due to clarithromycin resistance. Ampicillin is bactericidal by inhibiting cell wall synthesis; therefore H. pylori must be dividing for ampicillin to be effective. Ampicillin resistance has rarely been reported. Understanding the bacterial responses to acidity at the gastric surface where the pH is often close to 3.0, may allow development of novel, more effective treatment regimens. Aim: To mimic the gastric environment using a novel In Vitro culture system that maintains constant pH and urea levels for up to 16 hrs to demonstrate the efficacy of ampicillin therapy at different environmental pHs and to explain the need for acid inhibition for eradication In Vivo. Methods: H. pylori were incubated in dialysis chambers suspended in 1.5L of BHI media at pH 3.0, 4.5, or 7.4 with 5mM urea, with or without ampicillin (2mg/ml>>MIC), for 8 or 16 hrs. Changes in gene expression were determined by tran- scriptomal analysis and confirmed by qPCR. Viability and survival were measured by colony counting and quantitative Live/Dead® analysis. The bacterial load of H. pylori infecting the gerbil was determined in the presence and absence of the highly efficacious K-competitive H,K-ATPase inhibitor, TAK-438 (Ki = 10nM). Results: The medium pH remained constant over the course of the experiments. There was an increase in CFUs at 8 and 16 hr at pH 4.5 and 7.4 but no change at pH 3.0. In the presence of ampicillin, CFUs declined at pH 4.5 and 7.4 but not at pH 3.0. With ampicillin, at pH 4.5 and most clearly at pH 7.4, the ratio of dead to live bacteria increased over time whereas there was no change at pH 3.0. There was decreased expression of many genes involved in cell envelope synthesis at pH 3.0, especially penicillin-binding proteins, confirmed by qPCR. High dose TAK-438 treatment of H. pylori infected gerbils increased the bacterial load >10 fold, predominantly in the antrum and raised the gastric pH to ~7.0. Conclusions: Ampicillin has no bactericidal effect at pH 3.0 but has an effect at medium pH 4.5 and 7.4. This is likely due to decreased expression of cell envelope genes at pH 3.0, including penicillin binding protein genes. Therefore, at pH 3.0, likely close to the pH frequently encountered by the bacteria at the gastric surface, the bacteria are non-dividing and ampicillin insensitive. The long doubling time of H. pylori (4-6 hrs) also likely contributes to a decrease in efficacy of ampicillin. An effective PPI even with modern formulations raises the gastric pH >4.0 for only 16 hrs/day, leaving a period of time where amoxicillin/ampicillin is likely ineffective. A more effective inhibitor of acid secretion that maintains gastric pH near neutrality for 24 hrs/day from day 1, such as TAK-438, should enhance the efficacy of H. pylori eradication therapy. 1046 Predictors of Success of Helicobacter pylori Eradication Treatment in a Multicentric Randomized Clinical Trial (SWOG S0701) in Latin America Rolando Herrero, Eduardo Salazar-Martinez, Rachael Sexton, Douglas R. Morgan, Ricardo Dominguez, Javier Torres, Catterina Ferreccio, Luis Eduardo Bravo, Maria E. Martinez, Edgar M. Peña, Pelayo Correa, Manuel Valdivieso, Mercedes Meza-Montenegro, Garnet L. Anderson, William D. Chey, Edwin R. Greenberg, John Crowley, Laurence H. Baker Background: Success of treatments to eradicate Helicobacter pylori (H.p.) infection varies by geographic areas and patient characteristics. However, the variables influencing response to treatment have not been completely elucidated. In a randomized clinical trial conducted in 7 centers in Latin America, we assessed determinants of H.p. treatment success. Methods: We recruited 1852 adults (ages 21-65 yrs) from 7 centers in Chile, Colombia, Costa Rica, Honduras, México, and Nicaragua and screened them for H.p. infection with a urea breath test (UBT). A total of 1471 (79.4%) individuals were UBT positive, of whom1463 were AGA Abstracts