AGA Abstracts Su1141 ICD-9 Codes Do Not Reliably Identify Cases of Eosinophilic Esophagitis: A Cautionary Tale David A. Rybnicek, William J. Bulsiewicz, Nicholas J. Shaheen, Evan S. Dellon Background: Attempts to conduct population-based studies of eosinophilic esophagitis (EoE) in the United States are hampered by the lack of a validated case definition of EoE for use in administrative databases. Aims: To assess the ability of ICD-9 codes to provide a case definition of EoE in the U.S. using patient-level claims and medical records data. Methods: This was a retrospective study performed at University of North Carolina (UNC). For case definition development, confirmed cases of EoE as per consensus guidelines were identified in the UNC EoE clinicopathologic database from 2005-2006. Using the Carolina Data Warehouse (CDW), a clearing house for claims and administrative data for patients seen in the UNC system, all ICD-9 diagnostic codes were obtained for these cases in order to select inclusion/exclusion ICD-9 codes and assess the viability of a case definition of EoE using administrative data. Then, with the EoE cases as the reference standard, we queried the CDW from 2005-2006 and calculated the sensitivity and specificity of the case definition using the selected ICD-9 codes. After this, we used an iterative process to refine the case definition, re-query the CDW, and optimize sensitivity and specificity. For validation, we planned to use confirmed EoE cases from 2007-2008 as the reference standard and re-query the CDW from 2007-2008 to calculate final sensitivity and specificity. Results: We identified 57 and 121 cases of EoE at UNC to serve as reference standards for the development and validation sets, respectively. Characteristics of EoE cases in the development set (mean age 26 yrs; 72% male; 83% Caucasian; 126 ± 106 eos/hpf) were similar to those in the validation set (mean age 32 yrs; 74% male; 80% Caucasian; 84 ± 80 eos/hpf). The optimized case definition algorithm required: 1) CPT code for EGD with biopsy; 2) exclusion of patients with 16 ICD-9 codes representing non-EoE eosinophilia or symptoms; and 3) inclusion of patients with 18 ICD-9 codes representing symptoms of esophageal dysfunction and EoE. This yielded a sensitivity of 84% (48/57) and specificity of 41% (1627/3936). Given this poor specificity, the case definition was not applied to the validation group. Conclusions: ICD-9 coding did not provide a sensitive and specific administrative case definition of EoE. Because there are no pathology codes for esophageal tissue eosinophilia in routine use in the U.S., the existing relevant CPT and ICD-9 codes (even after considering 530.13 which was approved for EoE in 2008) are too non-specific to reliably identify cases of EoE. Investigators should not use ICD-9 coding to perform database research on EoE at this time. Su1142 Can an Internet Search Engine Provide Clues Into the Evolving Epidemiology of Eosinophilic Esophagitis? David A. Rybnicek, Spencer D. Dorn, Nicholas J. Shaheen, Evan S. Dellon Background: In both acute and chronic conditions, internet search patterns have recently been shown to reflect epidemiologic trends. It is unknown whether this represents a potential tool for studying the evolving epidemiology of eosinophilic esophagitis (EoE). Aim: To evaluate the utility of analyzing internet search trends for describing the changing epidemi- ology of EoE. Methods: Google Insights for Search allows users to determine patterns for particular search terms entered into the Google search engine relative to total internet search volume (termed Search Volume Index, or SVI). We queried this tool to determine the frequency, seasonality, and geographic location of searches conducted using EoE related terms (eg “eosinophilic esophagitis,” “trouble swallowing,” “allergy”) from 2004 to 2010. We also applied a filter that allowed comparison of search term volume relative to total health-related searches. In addition, we compared changes in the SVI to the number and origin of EoE-related scientific publications identified using PubMed. Results: From 2004 to 2010, the frequency of EoE related searches increased dramatically, outpacing other health-related internet searches by 200%. This trend closely paralleled the rapid increase in number of EoE scientific publications (Figure 1). Globally, the U.S., India, Canada, Australia, and the U.K. had the highest SVIs. Within the U.S., there was not a clear overall geographic trend in search volume. The states with the highest SVIs in 2010 were Ohio, Pennsylvania, Massachusetts, North Carolina, and Illinois, which roughly correlated with the location of the clinical and research centers with the most EoE-related publications that year (Cincinnati, Chicago, Chapel Hill, Philadelphia, San Diego). Finally, in contrast to the search term “allergy,” which demonstrated regular seasonal peaks in search volume in the spring and fall, there were not seasonal spikes in “eosinophilic esophagitis” searches (Figure 2). Conclusions: Internet search volume for terms related to EoE have rapidly risen and outpaced other health-related internet searches. This trend mirrors the increasing number of EoE scientific publications. Unlike allergy, search frequency for EoE does not demonstrate seasonal variation. Though a finer level of granularity for searches is needed, these results suggest that internet search trend analysis may be a potential tool to study EoE epidemiology and related health care utilization. Su1143 Prevalence of Eosinophilic Esophagitis in Japanese Adults Yasuhiro Fujiwara, Motoko Muraki, Yukie Kohata, Hirokazu Yamagami, Tetsuya Tanigawa, Kenji Watanabe, Toshio Watanabe, Kazunari Tominaga, Tetsuo Arakawa Background: Eosinophilic esophagitis (EoE), a chronic, immune/antigen-mediated eso- phageal disease, is diagnosed by the presence of typical symptoms such as dysphagia or food impaction and intraepithelial eosinophil infiltration of more than 15 per high-power field (HPF). A recent consensus recommendation suggests that EoE should be distinguished from proton-pump inhibitor (PPI)-responsive esophageal eosinophilia in patients who are responsive to PPI treatment. Although many studies about EoE have been reported in Western countries, a few studies have been reported in Asia. The aim of this study was to examine the prevalence of EoE in Japanese adults in a multicenter study. Methods: Thirteen hospitals and clinics with specialized Gastroenterology units participated in this survey between September 2010 and March 2011. Esophageal biopsy was performed when the patients had typical EoE symptoms or when endoscopic findings revealed typical EoE S-436 AGA Abstracts appearance such as linear furrow, ring, white exudates, and strictures. Esophageal eosino- philia was defined as intraepithelial infiltration of eosinophilia of more than 15 per HPF. Patients with esophageal eosinophilia were administered rabeprazole 10 mg per day for 8 weeks. Results: A total of 10,853 subjects who underwent further examination or routine check-up underwent esophagogastroduodenoscopy (EGD). Among them, 64 (0.6%) patients were thought to have EoE, and their biopsy specimens were obtained. Six (9.3%) of these 64 patients showed intraepithelial infiltration of eosinophilia of more than 15 per HPF. Among these 6 patients, 3 showed no symptoms and the other 3 were treated with PPI for 8 weeks; subsequently, their symptoms and histology were re-evaluated. Finally, 2 patients were diagnosed with PPI-responsive esophageal eosinophilia and 1 patient with EoE. Conclu- sion: The prevalence of EoE was 0.01% in the patients at the gastroenterology unit, suggesting that EoE is rare in Japanese adults. Su1144 Acute Esophageal Food Bolus Impaction. a Prospective Study Hedwig van der Sluis, Egbert-Jan van der Wouden, Freek Moll, Michael G. Havenith, Maarten Meijssen, Juda Vecht Introduction and aim of the study: Acute esophageal food bolus impaction (AEFBI) has an estimated incidence of 13 episodes per 100.000. This makes AEFBI one of the most common endoscopic emergencies in gastroenterology. To date, AEFBI has been studied mainly retro- spectively. In addition, there has been a growing interest in food bolus obstruction caused by eosinophilic esophagitis. The aim of this study was to determine the etiology and outcome of AEFBI in a prospective manner. Patients and methods: New consecutive patients (≥18yrs) who needed endoscopic intervention because of AEFBI were evaluated. Patients were excluded when analysis of food bolus impaction already occurred in the past or when evaluation was not possible due to co-morbidity. Included patients underwent a second esophagogastroduodenoscopy in order to inspect the esophagus. During endoscopy a min- imum of five biopsies were randomly taken from the esophageal squamous epithelium and examined by two independent pathologists. If the number of eosinophils was equal to or higher than 20 per high-power field, then AEFBI was considered to be caused by an eosinophilic esophagitis. When AEFBI was insufficiently explained by endoscopy, additional analysis by esophageal manometry and a barium esophagogram was performed. Follow-up time was 6 months. Results: From October 2009 until June 2011 59 patients needed endoscopic intervention because of AEFBI. All patients were successfully treated by flexible endoscopy. 11 patients were excluded from evaluation (4 patients because they were previ- ously analyzed for esophageal food impaction, 5 patients because of co-morbidity and 2 patients did not show up for the evaluation). 18 of the 48 remaining patients were female, 30 patients were male, mean age was 66 years (range 18-93). A final diagnosis was made in 45 patients (94%). In 30 patients (63%) AEFBI was caused by gastroesophageal reflux disease with or without a peptic stenosis. A motility disorder was found in 5 patients (10%), candida esophagitis in 5 patients (10%), radiation-induced stenosis in 3 patients (6%) and denture-related problems in 1 patient (2%). In two patients (4%) AEFBI was caused by an eosinophilic esophagitis. None of the patients had an esophageal malignancy. Four patients (8%) had recurrent food bolus impaction during follow-up. Conclusions: Acute esophageal food bolus impaction indicates underlying esophageal pathology in most patients which can be found by additional workup. In this prospective study most esophageal food bolus obstructions were caused by gastroesophageal reflux disease with or without a peptic stenosis. Eosinophilic esophagitis remains a rare cause of AEFBI. Su1145 Is it Advised to Take Biopsies in an Endoscopic Normally Appearing Oesophagus to Exclude an Eosinophilic Oesophagitis in Children? Bruno Hauser, Tierry Devreker, Gigi Veereman, Elisabeth De Greef, Yvan Vandenplas OBJECTIVES AND STUDY: Eosinophilic oesophagitis (EE) generally presents with reflux like symptoms, a specific endoscopic appearance and a histological extensive infiltration of eosinophils in the oesophagus. Some individuals can have a normal endoscopic appearance of the oesophagus making histology the only reliable diagnostic tool. We performed a prospective study to look for the presence of histological EE (HEE) in normally appearing oesophagus of children with gastro-oesophageal reflux symptoms. METHODS: One hundred and forteen children (56 boys) with a mean age of 8 years (0.5-17 years) presenting with gastro-oesophageal reflux symptoms underwent a fibroduodenoscopy with forceps biopsies taken at the distal (DOB) and proximal (POB) oesophagus between January and December 2010. Endoscopic appearance of the oesophagus was suggestive for EE when longitudinal furrows or vertical lines and/or horizontal rings and/or exudate were present. Refluxoeso- phagitis (RE) was defined according to the Los Angeles classification. A cut-off of 20 eosinophils per high power field was used to differentiate histological RE (HRE) from HEE (>20/HPF). RESULTS: Seven children had an endoscopic appearance of EE: DOB and POB confirmed HEE in 3/7, showed a HRE in 1/7 and a mycosis in 1/7, and were normal in 2/ 7 patients. Thirty-five children had an endoscopic appearance of RE: DOB and POB were normal in 29/35, showed a HRE in 6/35 and a HEE in 0/35 patients. Six patients had an endoscopic appearance of combined EE and RE: DOB and POB confirmed HEE in 1/6, showed a HRE in 2/6 and were normal in 3/6 patients. Finally, the endoscopic appearance of the oesophagus was normal in 66 patients: DOB and POB were normal in 49/68, showed a HRE in 17/68 and a HEE in 0/68 patients. CONCLUSIONS: The diagnosis of eosinophilic oesophagitis was never made by oesophageal biopsies in an endoscopic normally appearing oesophagus in children with gastro-oesophageal reflux symptoms.