AGA Abstracts 8 vs 8 years; p=0.02). The maximum eosinophil counts did not vary between the groups (84 vs 80 vs 102; p=0.12). After multivariate analysis, only age and dysphagia were independent predictors of phenotype status. The OR for a 10 year increase in age for a mixed phenotype compared to inflammatory was 1.6 (95% CI 1.4-2.0) and for fibrostenotic was 2.1 (1.7- 2.7) (Figure). The ORs for dysphagia were 3.1 (1.7-5.5) and 7.0 (2.6-18.6), respectively. Conclusions: In this large cohort of subjects with EoE, the likelihood of fibrostenosing disease increased with age. For every ten year increase in age, the odds of having a fibrostenotic EoE phenotype more than doubled. The association of fibrostenosis with age suggests that the natural history of EoE is a progression from an inflammatory to a fibrostenotic disease. Su1834 Topical Steroid Therapy Efficiently Reduces the Risk for Long Lasting Food Impactions Requiring Endoscopic Removal in Eosinophilic Esophagitis Alain Schoepfer, Tanja Kuchen, Ekaterina Safroneeva, Susanne Portmann, Christian Bussmann, Alex Straumann Background: Long lasting food impactions requiring endoscopic bolus removal occur fre- quently in Eosinophilic Esophagitis (EoE) and are risky and frightening for patients. So far it is unknown whether the risk of impaction can be reduced by therapy with swallowed topical steroids. Aim: To assess the frequency of impactions requiring endoscopic bolus removal in EoE patients under therapy with topical corticosteroids compared to a group of untreated EoE patients. Furthermore, we aimed to identify additional risk factors for bolus removal. Methods: Retrospective analysis of the Swiss EoE Database (SEED), extended by a review of patients charts, endoscopy and pathology records. Only EoE patients were included who had a complete visit set consisting of a patient interview, upper endoscopy, histology, and lab values. Esophageal strictures were defined as narrowing of the esophageal lumen of 12mm and less. Results of logistic regressions are presented as Odds Ratio (OR) and 95% confidence interval (95%-CI). Results: Hundred fifty two patients with a total of 511 visits were analyzed (115 males, mean age at EoE diagnosis 39 ± 14 years). The median follow-up time was 5 years (IQR 2-6), patients had on average 3.3 follow-up visits. In the follow-up period, 32 (6.3% of all endoscopies) episodes with endoscopic bolus removal were observed in 25 patients (16.4% of the cohort). Therapy with topical swallowed steroids reduced the risk for bolus removals (OR 0.53, 95%-CI 0.27-1.05, p=0.069). The following risk factors for bolus removal were further evaluated by univariate logistic regression model- ing: clinical activity (OR 1.07, 95%-CI 0.43-2.65, p=0.880), esophageal stricture (OR 2.66, 95%-CI 1.41-5.03, p=0.003), peak eosinophil count . 20 eosinophils/HPF (OR 0.62, 95%-CI 0.29-1.28, p=0.193), blood eosinophilia (OR 0.62, 95%-CI 0.29-1.28, p=0.193), adherence to steroid therapy 3 months before bolus removal (OR 1.14, 95%-CI 0.40-3.25, p=0.805), and esophageal dilation (OR 2.36, 95%-CI 1.16-4.79, p=0.018). Factors with p,0.1 were entered into the multivariate logistic regression model. In the multivariate model therapy with topical steroids was signicantly negatively associated with the risk for endoscopic bolus removal (OR 0.34, 95%-CI 0.16-0.73, p=0.006), whereas the presence of esophageal strictures was positively associated with endoscopic bolus removal (OR 2.56, 95%-CI 1.12- 5.84, p=0.025). Dilation was no longer associated with endoscopic bolus removal (OR 1.79, 95%-CI 0.75-4.28, p=0.190). Conclusions: Treatment with swallowed topical steroids efficiently reduces long lasting bolus impactions requiring endoscopic intervention. A reduced esophageal diameter is associated with an increased risk for bolus impactions. Of note, clinical activity is clearly insufficient to predict the risk of bolus impactions. Su1835 Proteomic Mining of Esophageal Lavage From Children With Eosinophilic Esophagitis Antoine Menoret, Marina L. Fernandez, Francisco A. Sylvester, Jeffrey S. Hyams, Anthony T. Vella, Wael N. Sayej Background: Eosinophilic esophagitis (EoE) is an emerging disease associated with eosinophil infiltration into the epithelial layer of the esophagus. EoE might be caused by exposure to food antigens and/or environmental triggers, prompting a chronic, localized immune response in the esophagus. Aim: To identify proteins from gastric lavage associated with EoE. Methods: We studied children undergoing endoscopy for digestive symptoms but who had normal appearance of the mucosa and normal histology (controls) and children with EoE defined by characteristic endoscopic appearance and ≥ 15 eosinophils/hpf. Proteomic approaches were used to compare EoE patients versus controls. During endoscopy gastric fluid was collected and then the esophagus was washed with 30 mL of saline and aspirated from the stomach to generate an esophageal saline lavage (EoL). The EoL was fractionated by anion exchange chromatography, and the immune stimulatory potential of each fraction was tested in vitro in a monocyte-based bioassay for secretion of pro-inflammatory cytokines. More than twenty fractions induced monocyte secretion of TNF- α and IL-1β. Results: We studied S-486 AGA Abstracts 24 children (10-17 years): 2 new EoE, 11 treated EoE (8 in remission and 3 with active disease), 1 patient with acid reflux, 1 with Barrett's esophagus and 9 normal controls. EoL fractions that induced cytokine secretion yielded different patterns of elution between EoE and control. Secondly, we used the ProteomeLab PF 2D (Beckman Coulter) to make proteomic maps which yielded 7 unique chromatographic peaks in EoL from an EoE patient while 29 peaks were restricted to the normal control EoL (Figure 1). Conclusion: EoL can be used to obtain proteomic information from patients suffering from EoE and that some of this material is immune stimulatory. We hope to uncover specific biomarkers and triggers that can be targeted clinically in the future. Su1836 The Esophageal Biopsy "Pull" Sign: A Novel and Highly Specific Endoscopic Finding in Eosinophilic Esophagitis Jessica H. Gebhart, Kelly E. Hathorn, Nicholas J. Shaheen, Evan S. Dellon Background: We have recently observed that esophageal biopsies in patients with eosino- philic esophagitis (EoE) can feel firm, with resistance appreciated when pulling the forceps to obtain the tissue sample. To our knowledge, this endoscopic sign has not been described in the literature. Aim: To describe the esophageal biopsy "pull" sign in EoE and assess the diagnostic utility of this finding. Methods: This is a sub-analysis of a prospective study at UNC from 2011-2012 enrolling consecutive adult patients undergoing outpatient EGD. Research protocol esophageal biopsies were taken to determine the maximum eosinophil count (eos/hpf; hpf=0.24mm 2 ). Incident cases of EoE were diagnosed as per consensus guidelines (symptoms of esophageal dysfunction, ≥15 eos/hpf, no response to a PPI trial, and other causes excluded). Controls were subjects who did not have EoE. During endoscopy, prior to knowledge of case/control status of the patient, the presence of resistance on pulling the biopsy forceps was recorded as a positive "pull" sign. Patients with EoE were compared to non-EoE controls, and the sensitivity, specificity, and predictive values of the pull sign were calculated. For patients with follow-up after steroid or dietary therapy, the pull sign was reassessed. Results: Of the 94 patients enrolled in the study, 50 (53%) had EoE and 44 were non-EoE controls. Compared to controls, EoE subjects were likely to be younger (37 vs 48 years; p ,0.001), male (72% vs 48%; p=0.02), and white (96% vs 82%; p=0.03). The maximum eosinophil count was 76 eos/hpf in the EoE group and 2 eos/hpf in controls (p,0.001). 43 EoE patients (86%) had a positive pull sign compared with none of the controls (p,0.001). EoE cases who had the pull sign were more likely to be male (79% vs 29%; p=0.006), less likely to have an endoscopically normal-appearing esophagus (2% vs 29% p=0.007), and tended to have a higher eosinophil count (60 vs 33 eos/hpf; p=0.07) compared with cases who were pull sign negative. Sensitivity of the pull sign for diagnosis of EoE was 86%, specificity was 100%, positive predictive value was 100%, and negative predictive value was 86%. Post-treatment data were available for 25 EoE cases with a baseline positive pull sign. Of the 16 with a histologic and symptomatic response, 11 (69%) had resolution of the pull sign. Of the 9 non-responders, 7 (78%) had a persistent pull sign (p= 0.03). Eosinophil count was also higher with a positive pull sign (61 vs 7 eos/hpf; p=0.004). Conclusions: The esophageal "pull" sign is a novel and highly specific endoscopic finding in EoE; a patient without this sign is very unlikely to have EoE. In addition, the "pull" sign frequently becomes negative after successful treatment. Further research is underway to assess this finding for association with subepithelial esophageal fibrosis known to occur in EoE.