AGA Abstracts review and meta-analysis to compare durability and efficacy of immunomodulator and anti- TNF agents in clinical trials of Crohn's disease. Method: A systematic review of PubMed and the Cochrane databases (1950-2012) was conducted to identify induction, maintenance, and combined induction-and-maintenance placebo-controlled randomized controlled trials (RCT) of immunomodulators or biologics currently used for Crohn's disease or with positive phase III trials. RCTs primarily evaluating fistula closure were excluded. Durability was defined as dropout due to adverse event (AE), which includes both side-effects from treatment and disease exacerbation. Efficacy was defined as clinical remission. Meta-analysis using the DerSimonian and Laird random-effects model was performed to establish relative risk (RR) of efficacy and harm with 95% confidence intervals (CI). Number needed to treat (NNT) and number needed to harm (NNH) with 95% CIs were calculated for each drug class. NNH/NNT was calculated to determine the number of dropouts due to AE for each treatment benefit. Heterogeneity and publication bias were assessed. Results: Twenty-nine manuscripts were included in the analysis, representing fourteen trials evaluating anti-TNF agents (inflixi- mab, adalimumab, certolizumab), five of anti-trafficking agents (natalizumab, vedolizumab), seven of AZA/6MP, and four analyzing MTX. Maintenance 6-MP/AZA had significantly excess dropouts due to AE with a NNH of 12.1 (95% CI 1.3-833.3) and RR=9.4 (95% CI 1.1- 78.0), as did induction MTX with a NNH of 6.7 (95% CI 0.8-500.0) and RR=8.0 (95% CI 1.1-58.5). Combined induction and maintenance 6-MP/AZA and MTX trials revealed a similar but non-significant trend with RR=2.0 (95% CI 0.3-12.2) and RR=4.2 (95% CI 0.5- 38.3), respectively. Induction anti-TNF therapy did not have excess AE dropouts (RR=0.9, 95% CI 0.6-1.3). Maintenance anti-TNF therapy was not associated with excess AE dropouts (RR=1.1, 95% CI 0.5-2.4), though there was evidence of heterogeneity (p ,0.001). Anti- trafficking therapy did not have excess AE dropouts in either pooled induction (RR=0.9, 95% CI 0.6-1.2) or maintenance trials (RR=0.5, 95% CI 0.3-0.7). The ratio NNH/NNT favored biologics over immunomodulators for both induction and maintenance strategies. Conclusions: Thiopurines and methotrexate were significantly associated with excess drop- outs due to adverse events; this was not seen in trials of anti-TNF or anti-trafficking agents. This suggests that patients with Crohn's disease are more likely to remain on treatment with biologics than immunomodulators. Su1143 Inflammatory Bowel Disease Patients Have a High Prevalence of Anal Human Papillomavirus and Anal Dysplasia Ross Cranston, Jonathan Baker, Carly Mowrey, Laura Janocko, Shaun Burneisen, Joann Fultz, Jana G. Hashash, Ian McGowan, Miguel Regueiro Background: Defective cell mediated immunity (CMI) increases the risk of anal dysplasia and cancer associated with high-risk human papillomavirus (HR-HPV) infection. Treatment of inflammatory bowel disease (IBD) often relies on immunosuppressants that impair CMI and are associated with opportunistic infections. There is concern that immunosuppressants in female IBD pts are associated with higher rates of cervical HPV. Whether anal HPV and its sequelae are increased in male and female IBD pts is not known. As part of a larger study evaluating the impact of antiTNF treatment on anal HPV and dysplasia, we report our baseline data on anal HPV and dysplasia prevalence. Aim: To determine the rate of anal HPV infection and dysplasia in IBD patients Naïve to antiTNF treatment. Methods: Sexually active adult IBD patients (pts) were recruited at the time of routine colonoscopy. Immediately prior to colonoscopy pts had anal and vaginal swabs collected for HPV typing (PCR amplifica- tion using biotinylated GP5+/GP6+ primers and hybridization for 37 HPV genotypes), anal cytology, and high-resolution anoscopy (HRA) with biopsy of all lesions with visual criteria suspicious for high-grade anal dysplasia. Results: Twenty-nine pts were enrolled and 3 were excluded (1 pregnant, 2 females unable to provide urine for a pregnancy test). Of the 26 pts, 15 (58%) were male and 11 (42%) female, average age was 37.1 years (range 21-65 years), and 24 (92%) were White. Sixteen (62%) had Crohn's disease and 10 (38%) had ulcerative colitis. All pts were Naïve to antiTNF therapy and 19 were taking one or more IBD treatments; 11 were taking azathioprine (AZA), 6-mercaptopurine (6MP), or corticoste- roids alone or in combination with other medication. The average length of IBD diagnosis was 4.8 years (range 0 to .10 years). Anal cytology was normal in 13 (50%), abnormal in 11 (42%), and unsatisfactory in 2 (8%). 21/26 (81%) pts had anal HPV detected (5 single, 16 multiple). Twenty (78%) had one or more HR-HPV types. Eight (73%) female pts had vaginal HPV detected (2 single, 6 multiple) all with one or more HR-HPV type. All 11 pts on AZA, 6MP or prednisolone had one or more anal high-risk HPV detected. No patient with undetectable anal HPV was taking an immunosuppressant. Seventeen (65%) pts had biopsies: 2 (8%) no dysplasia, 10 (38%) anal intraepithelial neoplasia (AIN) grade 1, and 4 (15%) had AIN 2. One biopsy was lost in processing. 6/14 (43%) patients with anal dysplasia were taking immunosuppressants. Conclusions: Although less than 50% of patients were taking immunosuppressants, high rates of both anal HPV and anal dysplasia, including precancerous lesions, were detected. Screening for anal dysplasia with HRA should be considered in this population. The impact of antiTNF on anal HPV is currently being evaluated and will be presented in the future. Su1144 A Predictive Model of the Risk of Developing Multiple Cancers With Tumor Necrosis Factor Alpha Inhibitor Therapy Among Patients With Inflammatory Bowel Disease Derrick J. Stobaugh, Parakkal Deepak, Eli D. Ehrenpreis Background: The risk of developing multiple synchronous/metachronous cancers after any exposure to Tumor Necrosis Factor alpha (TNF- α) inhibitors in inflammatory bowel disease (IBD) is unclear. We sought to examine this risk, by creating a predictive model using the Food and Drug Administration Adverse Event Reporting System (FAERS). Methods: More than 3.1 million files between January 2003 and June 2012 were downloaded from the FAERS and analyzed using SPSS 20 (IBM Co. Armonk, NY, USA). They were queried for adverse events with TNF- α inhibitors (Adalimumab, Certolizumab, Infliximab) and control drugs (5-aminosalicylic acid and sulphasalazine), using both trade and generic names, for IBD. These were further queried for outcomes of cancer (solid cancers, leukemia, lymphomas and/or skin cancers) as well as control reactions (syncope, hernia, deafness and vertigo, S-410 AGA Abstracts which were predefined to have no association with study drugs or control drugs) utilizing reactions terms from the Medical Dictionary for Regulatory Activities. Any reports with comorbidities of Human Immunodeficiency Virus infection, organ transplantation or rheuma- tologic diseases associated with a risk of cancer as well as concomitant exposure to a non- TNF-α inhibitor biologic medication were excluded from analysis. We then constructed a model (Figure 1) to simulate the odds ratio (OR) of patients developing multiple cancers using STELLA (isee, Hanover, NH). Each patient was simulated to have continued exposure to the study or control drugs over a period of 9.5 years (data time period in FAERS). Rates of developing these cancers were determined from data collected from the FAERS. Data was expressed as a changing OR over time. Results: The predicted odds of developing the first cancer after any exposure to TNF- α inhibitors are elevated starting at 1 year and plateaus afterwards over the simulated time (OR 2.43 at 9.5). The odds of a second cancer after having the first cancer is however greater than the odds for the initial cancer (OR 10.97 at 9.5 years) and the odds for developing a third cancer (OR 18.81 at 9.5 years) were higher than a second cancer (Figure 2). The odds for developing the second and third synchronous/ metachronous cancer also demonstrated an increase over the time period of simulation, unlike the risk of developing a first cancer. Conclusion: Any exposure to TNF- α inhibitors is associated with increased odds of developing cancer compared to controls. Development of a first cancer increases the odds of further developing a second or a third cancer, with increasing odds with continued exposure to TNF- α inhibitors. Based on our data, those who have developed cancer while on TNF- α inhibitors should be closely monitored for a second cancer. Figure 1: Image of the model constructed in STELLA Figure 2: Reported odds of developing cancers with Tumor Necrosis Factor alpha inhibitors among inflammatory bowel disease patients Su1145 Clinical Characteristics and Colectomy Rate in the Patients With Ulcerative Colitis Who Are Intolerant to 5-Aminosalycylic Acid Agents: A Single-Center Cohort Study Hiroto Kinoshita, Reiko Kunisaki, Tomohiko Sasaki, Hideaki Kimura, Katsuaki Tanaka, Shin Maeda Background and aims; 5-aminosalycylic acid (5-ASA) compounds are used as the primary treatment for ulcerative colitis (UC). Although 5-ASA is generally well tolerated, some patients are intolerant to this drug. Many case reports on serious adverse effects have been reported. However, there have been few studies on clinical characteristics and prognosis of 5-ASA intolerant patients. The aim of this study is to elucidate the clinical characteristics and the colectomy rate of the 5-ASA intolerant patients at our tertiary referral center. Methods; Data were obtained using a retrospective chart review of 554 consecutive UC patients who were first referred to our center from January 2008 to August 2012 with a mean follow up of 24 +/- 21 S.D. months. Patients were considered to be intolerant to 5-ASA if they discontinued the drug due to any kind of adverse effect. Results; A total of 58 patients (10.4%) were considered to be intolerant to 5-ASA. Adverse effects that lead to discontinuation of the drug are summarized in the table 1. Common clinical manifestation of 5-ASA intolerance was a combination of one or more of the following symptoms: exacerbation of colitis, fever, and severe fatigue. Initial dose of 5-ASA was significantly higher in the patients with 5-ASA intolerance compared with the patients who tolerated 5-ASA (P = 0.008, t-test). Although 22 out of 58 cases of 5-ASA intolerance were improved after cessation of 5-ASA, the other 36 cases needed corticosteroid to treat the adverse effects, most of that were exacerbation of colitis and/or fever. The cumulative probability of surgery in patients with 5-ASA intoler- ance was 13.4% at 1 year and 19.5% at 2 years, which was significantly higher than that in patients who tolerated 5-ASA, 3.6% at 1 year and 6.2% at 2 years (P = 0.0082, Wilcoxon test). Conclusion; In this retrospective cohort analysis at a tertiary referral IBD center, rate of intolerance to 5-ASA in patients with UC was 10.4%, which was relatively higher than the previous reports. Higher initial dose of 5-ASA can be a risk factor for incidence of adverse effect and discontinuation, implying that it is safer to start with standard dose rather than with high dose. The colectomy rate of 5-ASA intolerant patients was higher than that of those who are on 5-ASA maintenance therapy, suggesting the importance of early diagnosis of 5-ASA intolerance and of timely intensification of treatment in 5-ASA intolerant patients. This is the first report on the prognosis of 5-ASA intolerant patients. Table 1. Adverse effects that lead to discontinuation of 5-ASA agents