AGA Abstracts Sa1153 Cost-Effectiveness of Nonmelanoma Skin Cancer Screening in Crohn's Disease Patients Philip N. Okafor, Christopher G. Stallwood, Linda T. Nguyen, Debjani Sahni, Deborah L. Cummins, Sharmeel K. Wasan, Francis A. Farraye, Daniel O. Erim Background: Several studies have demonstrated an increased risk of nonmelanoma skin cancer (NMSC) in inflammatory bowel disease (IBD) patients, with the greatest risk in Crohn's disease (CD) patients. There are presently no guidelines for skin cancer screening in IBD patients. We sought to investigate the cost-effectiveness of various skin cancer screening strategies in CD patients. Methods: A Markov state-transition model and Monte Carlo simulation were created to compare lifetime costs, life expectancies and benefits associated with NMSC screening in a hypothetical cohort of 100,000 CD patients followed from age 30 to 100 years. Risk factors for NMSC including age and thiopurine use were accounted for in the model. CD severity was classified into the following states: remission, mild-moderate, moderate-severe and fulminant. Four screening strategies were studied including: annual screening of all patients, screening patients at diagnosis of CD then biannual follow-up with a dermatologist for patients without NMSC, annual screening of patients over 50 years, and as a high-risk group annual screening for all patients over 50 years and all patients previously exposed to thiopurines irrespective of age. Results: Undiscounted mean total lifetime costs (MTLC) for NMSC screening in all CD patients annually were $90,540 (ranging $55,675 to $154,478) while MTLC for not screening were $47,076 (ranging $18,297 to $120,561). This was associated with an increase in quality adjusted life expectancy from 34.2 to 36.2 resulting in $22,872/quality adjusted life year (QALY) gained. Screening patients at diagnosis of CD then biannual follow-up with a dermatologist for patients without NMSC was associated with MTLC of $73,745 resulting in $39,810/QALY gained. Screening only patients over 50 years was associated with MTLC of $82,898 resulting in $22,112/ QALY gained. Finally, combined screening of patients over 50 years and all patients previously exposed to thiopurines irrespective of age was associated with a MTLC of $76,313 with $21,513/QALY gained. Conclusion: At a willingness to pay threshold of $50,000/QALY typically used in decision sciences, all the above strategies prove cost-effective. Annual screening of high-risk patients including all patients over 50 years and patients previously exposed to thiopurines irrespective of age was the most cost-effective strategy. Sa1154 Guidewire Cannulation in Prevention of Post ERCP Pancreatitis. A Meta- Analysis of Randomized Controlled Trials Anupama Inaganti, Naga Swetha Samji, Sarah D. Komanapalli, Rajan Kanth, Mainor R. Antillon, Ramon E. Rivera, Praveen K. Roy Background: Post ERCP acute pancreatitis is the most alarming complication of endoscopic retrograde cholangiopancreatography (ERCP) with frequency of 1-9% although it may increase to 40% in high risk groups. Guidewire is thought to reduce the frequency of post ERCP pancreatitis by reducing papillary trauma but still controversial. We conducted a meta-analysis of several randomized controlled trials (RCTs) that compared guidewire with conventional contrast to evaluate the efficacy in reducing the incidence of PEP. Methods: The Cochrane Library, CINHAL, Pubmed, Embase, and abstracts from major conference proceedings were searched for relevant articles. RCTs comparing guidewire with conventional contrast in preventing post ERCP pancreatitis were included in the analysis. Standard forms were used and data was extracted by 2 independent reviewers. Random effects model was used. Funnel plot was used to assess publication bias. Data regarding patient demographics, indications, risk factors, cannula type, cannulation success and complications was extracted for analysis. Analysis of pooled estimates of PEP, hyperamylasemia and grades of pancreatitis was performed to compare the efficacy of guide with conventional contrast in reducing PEP. Heterogeneity was assessed using I square measure of inconsistency. Results: Ten RCTs were included in the study (7 non crossovers and 3 cross over trials). Of these studies, 3 were abstracts from clinical meetings and 7 were full length publications. 3040 patients were included in the study of which 1146 were females. Mean age of the paients included was 22 to 94 years. The studies were conducted in Italy, Australia, Greece, Korea, Japan and USA. 0.035 inch guidewire was used in 5 studies. Guidewire did not decrease the risk of PEP (RR 0.050 95% CI 0.420-1.000, p =0.050). There was no reduction in risk of severe pancreatitis (RR 0.835 95% CI 0.282-2.473, p = 0.745). No significant risk reduction was noted in mild and moderate pancreatitis. Primary selective biliary cannulation success (reported in 6 studies) was significantly higher in guidewire group compared to conventional contrast group RR (1.125 95% CI 1.030-1.228, p = 0.009). In subgroup analysis, no reduction in PEP was noted with guidewire in noncrossover trials (RR 0.052 95% CI 0.310- 1.005, P=0.052) and crossover trials RR 0.640 (95% CI 0.486 - 1.558, P = 0.640). Number of precuts used was reduced in guidewire group (RR 0.744 95% CI 0.563-0.982, p=0.037). Adverse events were reported in 4 trials and rates of bleeding and perforation were comparable in both the groups. Conclusion: Guidewire does not decrease the risk of PEP when compared to conventional contrast but it does improve primary selective biliary cannulation success rates and rate of precut sphincterotomy was reduced in guidewire group. Sa1155 Personalized Post-Polypectomy Surveillance Based on Gender, Age and a Score for Adenoma Characteristics: A Cost-Effectiveness Analysis Else-Mariette B. van Heijningen, Frank van Hees, Iris Lansdorp-Vogelaar, Ewout W. Steyerberg, Ernst J. Kuipers, Marjolein van Ballegooijen Introduction: Current surveillance guidelines only consider the various aspects of advanced adenomas separately, while several studies suggest that adenoma characteristics are indepen- dent multiplicative predictors of adenoma recurrence. Also age and gender are not considered. Objective: The aim of this study was to determine optimal surveillance intervals for newly diagnosed adenoma patients based on a more detailed stratification: including age, gender, S-216 AGA Abstracts and combinations of adenoma characteristics. Methods: We used a state-of-the-art microsim- ulation model (MISCAN-colon) to compare various surveillance strategies for their cost- effectiveness. This was done separately for different ages (per 5 years, 40 - 80), gender, and adenoma risk score (0-5). Adenoma risk score was based on observed log of odds ratios of baseline adenoma characteristics for advanced adenoma recurrence in newly diagnosed adenoma patients. The adenoma risk score is constituted as follows: 1 point for each characteristic: adenoma size ≥10 mm, villous histology, proximal location and having 2 to 4 adenomas; and 2 points for ≥5 adenomas. For each combination of score, age and gender, we considered 11 alternative strategies: A varying surveillance interval (1 - 10 years between colonoscopies), and no surveillance (i.e. going back to FIT (fecal immunochemical test) screening in 10 years if appropriate). We considered 3 ages (75, 80, and 85) to stop surveillance. Analyses addressed high quality colonoscopies and first surveillance intervals. The optimal surveillance interval for each group was chosen such that the associated incre- mental costs per life-year gained cut-off were equal among the groups. Results: For patients with an adenoma risk score of 0 (i.e. 1 distal non-villous adenoma), surveillance was not cost-effective irrespective of gender and age; these patients could return to FIT screening 10 years after colonoscopy (Table 1). For the other adenoma risk scores, the recommended surveillance intervals were shorter for men, subjects with a higher adenoma risk score, and those with increasing age. The intervals ranged from 2 years for 80-year old men with an adenoma risk score of 5, to 7 years for 40-year old women with a score of 1. The age to stop surveillance was equal for men and women and depended on the adenoma risk score: 75 for patients with a score of 1, 80 for patients with a score of 2, and 85 for patients with a score of 3-5. Conclusions: The cost-effectiveness of surveillance of adenoma patients is highly dependent on adenoma risk score, gender and age. Adenoma number, large adenoma size, villous histology and proximal location should be combined with age and gender to determine more targeted surveillance. Patients with 1 distal non-villous adenoma can safely return to regular colorectal cancer screening without the need for surveillance. Table 1. Recommended next surveillance-interval (in years) according to adenoma risk score, gender and age, based on the equal incremental cost-effectiveness No surv = no surveillance, return to colorectal cancer screening after 10 years, if age allows (, 75 years) † This age is age at adenoma detection. The ending age holds for subjects with an average life expectancy, decisions should also depend on co-morbidity and otherwise vitality of the subjects Sa1156 Systematic Review and Meta-Analysis of the Prognostic Significance of Circumferential Resection Margin Involvement, As Defined by CAP and Rcp Criteria, on Survival in Patients With Operable Esophageal Cancer David S. Chan, Alex Karran, Tom D. Reid, Paul A. Blake, Wyn G. Lewis Background: The prognostic role and definition of circumferential resection margin (CRM) involvement in operable esophageal cancer (EC) remain controversial. The College of Ameri- can Pathologists (CAP) and Royal College of Pathologists (RCP) define CRM involvement as tumour found at and within 1mm of the cut resection margin respectively. This systematic review and meta-analysis was performed to determine the influence of CRM involvement on survival in operable EC. Method: PubMed, MEDLINE and the Cochrane Library (January 1990 to December 2012) were searched for studies correlating CRM involvement with 5- year mortality. Statistical analysis of dichotomous variables was performed using odds ratio (OR) as the summary statistic. Results: Fifteen studies involving 2691 patients with EC who had undergone potentially curative esophagectomy were analysed. Rates of CRM involvement were 15.3 (173 / 1133) and 36.7 per cent (987 / 2691) according to the CAP and RCP criteria respectively. Overall 5-year mortality rates were significantly higher in patients with CRM involvement compared with CRM negative patients according to both CAP [OR 4.02, 95 per cent confidence interval (c.i) 2.25-7.20, p ,0.001] and RCP (OR 2.75, 95 per cent c.i. 2.17-3.48, p,0.001) criteria. CRM involvement between 0.1 to 1mm was associated with significantly higher 5-year mortality than CRM negative status ( .1mm), [OR 2.05, 95 per cent c.i 1.41-2.99, p ,0.001]. Conclusion: CRM involvement is an important predictor of poor prognosis. CAP criteria differentiate a higher risk group than RCP criteria but overlook a patient group with similar poor outcomes. Consensus regarding the most accurate and prognostically important definition of CRM involvement would be welcome, and in the interim the exact distance of the esophageal tumour from the CRM should form part of routine pathology reporting in EC. Sa1157 The Effect of Cholecystectomy on Resource Utilization in Biliary Dyskinesia Nitin Aggarwal, Klaus Bielefeldt We have recently demonstrated an increase in elective cholecystectomies for Biliary Dyskinesia (BD). Given the functional nature of the disorder, we hypothesized that the apparently lower threshold for cholecystectomy would not translate to a decreased rate of resource utilization. METHODS: We analyzed electronic medical records of patients seen in multidisciplinary clinics for digestive disorders between January 2008 and January 2012, collecting demo- graphic information, data on nature and duration of symptoms, diagnostic evaluation and