literature review to craft a conceptual model of SUDD HRQOL. This was complemented by3 focus groups of 45 SUDD patients and an expert panel of five gastroenterologists andsurgeons. We developed the DV-QOL items based on our literature search and input fromfocus groups and experts, and obtained feedback from patients about those items in cognitivedebriefing interviews. We administered the items to a cohort of SUDD patients with persistentsymptoms following a confirmed diverticulitis event. We created scales based on factoranalysis and evaluated the scales for reliability and validity. Results: Concept elicitationrevealed a range of illness experiences attributed to SUDD. Coding of 20,490 transcribedwords yielded 52 codes arranged in a network with 4 first-order condition-related concepts:(1) physical symptoms (e.g., pain, bloating); (2) behaviors (e.g., dietary, physical, and socialrestrictions); (3) cognitions and concerns (e.g., lack of control, feeling something wrong);and (4) impact and consequences (e.g., frustration, anxiety). Initially, we developed 46 itemsthat reflected these four concepts. Using data from a cross-sectional validation sample of197 patients, we reduced the DV-QOL to a 24-item instrument. The final instrumentdemonstrated strong internal consistency (Cronbach's Alpha = 0.95) and a reliable four-factor solution (Tucker and Lewis Reliability Coefficient = 0.90). In our validation sample,the DV-QOL significantly discriminated between patients with recent (i.e., within last month)versus distant diverticulitis events (effect size of DV-QOL score difference = 0.66), andcorrelated strongly with both the Short-Form 36 subscales (mean correlation = -0.49) andhospital anxiety and depression (HAD) scores (mean correlation = 0.50). Conclusions:Patients with diverticular disease attribute a wide range of negative psychological, social,and physical symptoms to their condition, even outside of acute attacks. The DV-QOLcaptures these symptoms in a valid and reliable manner.

1052

Daily Mesalamine Fails to Prevent Recurrent Diverticulitis in a Large PlaceboControlled Multicenter TrialWolfgang Kruis, Vassilios Kardalinos, Austin Curtin, Andrey E. Dorofeyev, Salam F.Zakko, Jens Wölkner, Manuel M. Diez Alonso, Harald Peeters, Ioannis E. Koutroubakis,Nicholas J. Talley, Karin Dilger, Ralf Mohrbacher, Roland Greinwald

Background: Open and non-placebo controlled studies suggest that mesalamine might beeffective for the prevention of recurrence of diverticulitis. However, data from placebo-controlled clinical trials has been lacking. Aim: To compare the efficacy and tolerability of twodoses of mesalamine granules vs placebo for the prevention of recurrence of uncomplicateddiverticulitis over 2 years. Methods: We conducted a randomized, double-blind, placebo-controlled phase III study in 330 patients who were successfully treated for uncomplicatedleft-sided diverticulitis, confirmed by computed tomography, within 6 months of enrollment.Patients were randomly assigned to receive either mesalamine granules OD 1.5 g (arm A)or 3 g (arm B) or placebo (arm C) for 96 weeks. The primary endpoint was the proportionof recurrence-free patients at 48 and 96 weeks. Recurrence was defined as any elevation ofC-reactive protein or leucocytosis and diverticulitis-like clinical signs and confirmation bycomputed tomography. The study was prematurely terminated for lack of observed efficacyafter an interim analysis. Results: A total of 324 patients qualified for full analysis set (FAS).Their main baseline characteristics were: 61% female, mean age 55.4 years (SD: 10.6), meantime since most recent attack of diverticulitis 89.4 days (SD: 44.1), 51.2% with only onepreceding episode of diverticulitis. Of the 324 patients, 80 were prematurely terminatedbefore week 48 (group 1) and 160 before week 96 (group 2) due to the general study stopafter the interim analysis. The proportion of the remaining 244 patients who were recurrence-free within 48 weeks of treatment was 46.0% in arm A, 52.0% in arm B, and 57.3% in armC, p=.8416 (arm B vs C). Recurrence was observed in 17.2%, 20.0%, and 20.7%, respectively.Similarly, of the 324 FAS patients 160 prematurely terminated before week 96 and of theremaining 164 patients the proportion who were recurrence-free within 96 weeks of treatmentwas 6.9% in arm A, 9.8% in arm B, and 21.8% in arm C, p=.9728 (arm B vs C). Recurrencewas observed in 27.6%, 33.3%, and 36.4%, respectively. The mean time to recurrence ofdiverticulitis was 116.0 days in arm A, 190.7 days in arm B, and 146.6 days in arm C(FAS). Additional secondary outcomes like mean days with any left lower quadrant painper week showed also no significant differences. Adverse reactions occurred in 17% ofpatients in arm A, 15% in arm B, and 12% in arm C. Conclusions: Mesalamine 1.5 or 3g OD failed to demonstrate superiority over placebo for the prevention of recurrence ofuncomplicated diverticulitis.

1053

Early Combined Immunosuppression for the Management of Crohn's Disease:A Community-Based Cluster Randomized TrialReena Khanna, Barrett G. Levesque, Brian Bressler, Guangyong Zou, Larry Stitt, GordonR. Greenberg, Remo Panaccione, Alain Bitton, Pierre Pare, Severine Vermeire, Geert R.D'Haens, Donald G. MacIntosh, William Sandborn, Margaret K. Vandervoort, Joan C.Morris, Brian G. Feagan

INTRODUCTION: Conventional management (CM) of Crohn's disease (CD) consists ofsequential use of corticosteroids, antimetabolites, and tumor necrosis factor (TNF)-antago-nists. Recent evidence indicates that early combined immunosuppression (ECI) with aTNF-antagonist and an antimetabolite may be more effective than CM. We compared theeffectiveness of ECI (Figure 1) to CM in community gastroenterology practices. METHODS:In this cluster randomization trial (Randomized Evaluation of an Algorithm for Crohn'sTreatment or REACT; Clinicaltrials.gov NCT01030809; partial support AbbVie), practicesin Canada (n=34) or Belgium (n=5) were randomly assigned in a 1:1 ratio to ECI or CM.Up to 60 consecutive adult patients (≥18 years of age) with a documented diagnosis of CDin each practice were evaluated for 24 months. The primary outcome was the proportionof patients in remission (Harvey-Bradshaw Score (HBS) ≤4 in the absence of steroids) at 12months, evaluated at the practice level. Secondary measures were the rates of complications,hospitalizations, and surgeries over the entire follow-up period, based on patient-levelanalyses. RESULTS: Twenty-one centers (1084 patients) were assigned to ECI and 18 (898patients) to CM. The mean age of the patients was 44.2 in the ECI group and 44.1 in theCM group. Mean HBS scores were 4.1 in both groups. The proportion of patients in theECI and CM groups who received combination of antimetabolite/TNF-antagonist by 12

S-187 AGA Abstracts

months was 15.1% and 6.5% (P<.001) and 19.7% and 9.6% by 24 months (P<.001). Mean% (SD) remission rates in the ECI and CM groups were 66 (14) and 62 (17) at 12 months(P=.65) and 73 (8) and 65 (17) at 24 months (P=.35). However, highly significant andclinically important differences in the rates of complications, surgeries, and the combinedoutcome of hospitalizations, complications, and surgeries were observed in favor of ECIover 24 months (Figure 2). The 24 month actuarial estimates for the combined outcomewere 27.7% and 35.1% in the ECI and CM groups, respectively (hazard ratio adjusted forCD caseload and country: .74 [.62, .87, P<.001]). CONCLUSION: community-based dataindicate that 1) a symptom based conventional approach to CD management may not beoptimal and 2) ECI may be more effective in preventing CD-related complications.

Figure 1: Therapeutic algorithm utilized for patients in the ECI group

Figure 2. Hospitalizations, complications (abscesses, new fistulas, extra-intestinal manifesta-tions of CD, serious adverse events), and surgeries for patients in the ECI and CM groupsover 24 months

1054

Elevated Fecal Calprotectin Predicts Relapse in Inflammatory Bowel DiseasePatients After Stopping TNFα-Blocking AgentsPauliina Molander, Martti A. Färkkilä, Ari Ristimäki, Kimmo Salminen, HelenaKemppainen, Timo Blomster, Ritva Koskela, Airi Jussila, Henna Rautiainen, MarkkuNissinen, Johanna Haapamäki, Perttu E. Arkkila, Urpo Nieminen, Juha Kuisma, JariPunkkinen, Kaija-Leena Kolho, Harri Mustonen, Taina Sipponen

Background & Aims: Few studies have shown that the assessment of subclinical inflammationby using surrogate markers such as fecal calprotectin (FC), may indicate the risk of relapsebefore the actual clinical relapse occurs. In this prospective multicenter study, our aim wasto determinate whether a consecutive elevated FC concentration after stopping TNFα-blocking therapy predicts relapse before the actual clinical or endoscopic relapse occurs.Methods: Fifty-two patients with IBD (17 Crohn's disease [CD], 30 ulcerative colitis [UC]and 5 inflammatory bowel disease type unclassified [IBDU]) in clinical, endoscopic and FC-based (<100μg/g) remission after at minimum one year of TNFα- blocking therapy wererecruited. They were followed up to approximately one year after the discontinuation ofTNFα-blocking therapy and FC was collected monthly for the first six months and everyother month thereafter up to one year. Colonoscopy was performed at 4 and 12 monthsafter stopping therapy and at the time of clinical relapse. In CD, clinical relapse was definedby Harvey Bradshaw Index (HBI) and endoscopic relapse by the Simple Endoscopic Score

AG

AA

bst

ract

s

AG

AA

bst

ract

sfor Crohn's Disease (SES-CD). In UC or IBDU, clinical and endoscopic relapse was definedby Mayo score. Results: Forty-nine patients (16 CD, 33 UC/IBDU) provided the requestedstool samples and comprised the study group. Of the 49 patients 15 (31%, 4 CD [25%],11 UC/IBDU [33%]) relapsed during the one year follow-up and 34 (69%, 12 CD [75%],22 UC/IBDU [66%]) remained in remission. Significant increase in median FC levels wasseen two (120 μg/g, 0-1867, n = 15, p = 0.0014), four (108 μg/g, 7-650, n = 8, p = 0.0056),and six (120 μg/g, 0-431, n = 6, p = 0.0029) months before the actual relapse. In contrastto those with sustained endoscopic remission, patients relapsing showed constantly elevatedFC levels. A cutoff of > 140 μg/g (ROC analysis) predicted relapse with a 79% specificityand a 53% sensitivity. At the moment of clinical and endoscopic relapse 4 out of 15 patients(27%, 1 CD, 3 UC) had FC < 100 μg/g. Interestingly, no difference was found in the relapserates in patients with FC concentration under 50 or 50-100 at baseline, (p = 0.325) nor inpatients with or without acute or chronic inflammatory infiltrate in biopsy specimens atbaseline (p = 0.221). No significant difference was seen in median FC levels in patients havingendoscopic remission and those with mild endoscopic activity at follow-up colonoscopies.Conclusion: FC seems to rise and remain elevated before the actual clinical or endoscopicrelapse, suggesting that FC can be used as a surrogate marker for predicting relapse inpatients with IBD, and to identifying patients requiring a close follow-up in clinical practice.

1055

Recommendations for Creating a High Quality Endoscopy Report forInflammatory Bowel Disease Patients: Results From a RAND-AppropriatenessPanelShane M. Devlin, Gil Y. Melmed, Peter M. Irving, David T. Rubin, Asher Kornbluth,Patricia L. Kozuch, Laura H. Raffals, Fernando S. Velayos, Miles Sparrow, LeonardBaidoo, Brian Bressler, Adam S. Cheifetz, Jennifer Jones, Gilaad Kaplan, Jennifer Greene,Corey A. Siegel

Background & Aims: There has been recent focus on improving the quality of reporting forscreening colonoscopy by standardizing reporting elements. However, there is no consensuson standardization of endoscopy reporting for patients with inflammatory bowel disease(IBD). Although endoscopic disease activity is central to therapeutic decision-making, otherdetails about the procedure can also be critical to creating a complete picture of the patients'current disease status. Our aim was to develop a template for a high quality endoscopyreport that can be used broadly by endoscopists. Methods: We applied the RAND/Universityof California Los Angeles Appropriateness Method to rate the importance and feasibility ofelements for inclusion in endoscopy reports for patients with IBD. First, a literature reviewwas conducted on endoscopy in IBD including quality indicators for colonoscopy, polypoidlesions in IBD, endoscopic approach to dysplasia surveillance, endoscopic disease activityindices in IBD, and endoscopy of the ileo-anal pouch. This review was presented to theBuilding Research in Inflammatory Bowel Disease Globally (BRIDGe) group, a panel of 13gastroenterologists clinically experienced in IBD. In addition, two external IBD/endoscopyexperts were consulted (DR and AK). Next, based on the review a total of 106 candidatereporting elements were proposed initially spanning "disease background", "findings","dysplasia surveillance", "Crohn's disease with an ileo-colonic anastomosis", "Crohn's diseasewith a colo-colonic anastomosis" and "pouchoscopy". Panelists used the modified Delphimethod to anonymously rate the importance and feasibility of candidate elements on a 1-9 scale (1-3 not important/feasible, 4-6 moderately important/feasible, 7-9 very important/feasible). Disagreement was assessed using a validated index(DI). The panelists then met inperson to discuss areas of disagreement, which was followed by a second round of anonymousrating . Elements rated a median of 7 or higher on importance after re-rating were retained.Results: 56 elements were retained on the basis of importance, of which six were duplicatedin two or more scenarios. Therefore, a total of 50 reporting elements were retained acrossthe four themes of "disease background", "findings and interventions", "Crohn's disease withan ileo-colonic anastomosis", and "pouchoscopy". Disagreement was noted on 9 elements,of which 3 were included in the model. A summary of representative elements is presentedin Table 1. Conclusions: A comprehensive list of recommended elements for high qualityIBD endoscopy reporting stratified by clinical scenario has been described, using a rigorousand evidence-based approach. These elements can be incorporated into endoscopy reportingsoftware platforms. Standardized endoscopy reporting may improve the quality of care in IBD.Table 1: Representative Sample of Included Elements

S-188AGA Abstracts

*Table includes a sample selection of only 19 of the 50 included elements. A full list willbe presented. DI: disagreement index.

1056

The Prevalence and Persistence of the Use of Biologic Therapy inInflammatory Bowel Disease in a Population-Based SettingLaura Targownik, Zoann Nugent, Aruni Tennakoon, Harminder Singh, Charles N.Bernstein

Introduction: Biologic therapies, specifically infliximab (IFX) and adalimumab (ADA) playan important role in the management of inflammatory bowel disease (IBD). Most dataregarding their utilization and efficacy have come from randomized controlled trials andIBD-research centres, which may not reflect their use and effectiveness in the generalpopulation. Therefore, we aimed to establish a population based dataset to evaluate the useof biologic therapy in IBD and the outcomes associated with their use Methods: All personsmeeting a validated administrative definition for IBD in the province of Manitoba between1984 and 2012 were included in the analyses, with incident cases being defined as thosewith at least 3 years of registration with no patient contacts for IBD. We used a comprehensiveuniversal drug utilization database to track all dispensations of infliximab and adalimumab.We calculated the point prevalence of IFX or ADA use (defined as use in the previous 90days) according to both calendar time and time since diagnosis. Person time accumulatedprior to the approval date for biologics in Canada for IBD (June 27, 2001) was excluded.We also determined the actuarial likelihood of continuing on a biologic agent once initiated,and assessed trends in the time to initial biologic prescription from diagnosis. Results: Theprevalence of IFX use has been increasing steadily over time; in 2012, 4.1% and 2.1% ofIBD subjects were using IFX or ADA respectively. Persons under age 25 were more likelyto be using biologic therapy than those >25 (11.3% vs. 4.6%), as were persons with CDvs. UC (9.5% vs. 3.1%). The likelihood of being exposed to either IFX or ADA by 1, 2, 5,and 10 years of diagnosis were 2.7%, 4.8%, 8.9%, and 13.5%. The time to first exposure