AG

AA

bst

ract

sdisease (CD)] and 102 IBD controls without celiac disease were included in our study. Amajority (92%) of the celiac-IBD patients reported following a gluten-free diet either continu-ously or intermittently. There was no significant difference in the age, gender, or ethnicitybetween celiac-IBD and controls. There was a numerical higher proportion of other autoim-mune diseases amongst the celiac-IBD group (24%) compared to controls (13%, P=0.09).Celiac-IBD patients were significantly less likely to be ever-smokers than controls (27% vs.45%, Odds ratio (OR) 0.44, 95% confidence interval (CI) 0.21 - 0.95). Presence of pancolitiswas more common in celiac-UC patients compared to non-celiac UC controls (63% vs.36%, OR 3.30, 95% CI 1.05 - 10.39). There was also a trend towards increased use ofimmunomodulators among celiac-UC patients than in non-celiac UC controls (OR 2.83, CI95% 0.95 - 8.48). No phenotypic differences were found in celiac-CD patients. There wereno significant differences in IBD-related hospitalizations, or surgeries. Within our healthcaresystem, the prevalence of celiac disease among IBD patients was 1%. In contrast, theprevalence of CD or UC among those with a diagnosis code for celiac disease was significantlygreater at 4% and 3%, respectively (Figure 1). Conclusion: We describe the one of thelargest cohorts to date of IBD patients with celiac disease. Celiac-UC patients were morelikely to have pancolitis and require immunomodulators. Co-existing celiac disease did notinfluence natural history of CD. While patients with IBD may not be at a higher risk forceliac disease, those with underlying celiac disease may be more likely to have CD or UC.

Figure 1: Prevalence of celiac disease among patients with various autoimmune diseases

Mo1250

Increased Risk of Intensive Care Unit (ICU) Admission in InflammatoryBowel Disease (IBD): A Population-Based StudyCharles N. Bernstein, Allan Garland, Christine Peschken, Carol Hitchon, Randy Fransoo,Ruth Anne Marrie

Background: Care in ICUs is costly, and consumes a disproportionate share of medical costs.Little is known about the risk of ICU admission in persons with IBD. We compared theincidence of ICU admission among persons with IBD and a matched cohort from the generalpopulation (GP). Methods: Using provincial administrative data we identified all personswith IBD and an age, sex and geographically-matched GP cohort (matching 5:1). Weconsidered ICU care as that provided in any of the 12 high-intensity (Level 1 or Level 2)adult ICUs in Manitoba. (We have shown that Canadian administrative hospital abstractsaccurately identify the presence and timing of ICU care). For the period 2000-2010, weestimated the annual incidence and the 10-year cumulative incidence of ICU admissionamong prevalent cases of IBD, and the incidence of ICU admission among a cohort withincident IBD dating back to 1989; using Cox proportional hazards models adjusted for ageat diagnosis, sex, comorbidity status and socioeconomic status. Results: At time of first ICUadmission, 47% were female, mean age was 56.7 yrs, mean disease duration was 6.0 yrsand a modified Charlson index was 0 or 1 in 67.6%. Of 4580 incident cases of IBD, 293(6.4%) were admitted to an ICU (vs. 3.8% of GP); the risk for ICU admission was increasedvs. GP (HR 1.79; 95%CI: 1.58-2.02). From 2000-2010, the mean annual incidence was0.74% (95%CI: 0.67-0.81), range: 0.55% to 1.12%). In 2009/10 the IRR for IBD admissionto ICU was 1.81 (95%CI: 1.22-2.46). The standardized 10-year cumulative incidence ofICU admission was 6.5% in IBD and 4.3% in the GP (IRR=1.52; 95% CI 1.36-1.68). Whencompared to GP, the incidence of ICU admission was increased in all age groups, but therelative risk was higher among persons aged 18-39 years (IRR 3.48) than among personsaged 40-59 years (IRR 1.69) and than among persons aged 60 years and older (IRR 1.42).The in-ICU stay, 30 day, and 1 year mortality rates of IBD patients admitted to ICU were8.5%, 12.9%, and 18.6% respectively. We conducted a similar analysis for Manitobans withmultiple sclerosis (MS) and rheumatoid arthritis (RA) and found that compared to the GPadmission to ICU was similarly increased in incident IBD as for RA (HR 1.65; 95%CI: 1.49-1.82) and MS (HR 1.45; 95%CI: 1.19-1.75). Conclusions: The risk of ICU admission isrelatively high among persons with IBD and is an even bigger problem in younger patients.Future studies will explore what the modifiable risk factors for ICU admission are.

Mo1251

Colonic Inflammation Triggers Cytomegalovirus Infection in Steroid NaïveActive Ulcerative Colitis -Significant Role of Mucosal PCR-Takumi Fukuchi, Hiroshi Nakase, Minoru Matsuura, Takuya Yoshino, Takahiko Toyonaga,Satoshi Ubukata, Takaaki Eguchi, Hiroshi Yamashita, Kiyoshi Ashida

Background: Mechanism of cytomegalovirus (CMV) reactivation in patients with ulcerativecolitis (UC) remains unclear. On the contrary, early detection of CMV reactivation in colonictissues by several modalities is required to decide the optimal therapies for patients withrefractory UC. It was reported that real-time polymerase chain reaction assay using colonicmucosa (mucosal-PCR) was promising for detection colonic CMV reactivation in the pointof its high sensitivity (Yoshino, et al. Inflamm Bowel Dis 2007;13:1516-21). In clinical

S-618AGA Abstracts

practice, to detect colonic reactivation in active UC refractory to corticosteroid or immunosup-pressive therapy has been focused because ongoing immunosuppressive therapies under theintestinal inflammation are considered to be involved in CMV reactivation. However, thereare few reports on the frequency of CMV reactivation in colonic tissues and the characteristicsof endoscopic finding related to CMV reactivation in corticosteroid (CS) Naïve patients withactive UC. Aim: To investigate the frequency and endoscopic findings of colonic CMVreactivation in CS Naïve patients with active UC. Methods: Fifty-three patients with activemoderate-severe (clinical activity index (CAI>/=7) UC without CS (29 male, 24 female; meanage 38.0±2.0 years) were examined. Colonoscopic examination was performed in all patientsand biopsy specimens from inflamed colonic mucosa were taken. (i). Colonic CMV reactiva-tion was examined by mucosal-PCR and immunochemistry (IHC). Cases in which CMVwas detected by one or two methods were diagnosed as positive for CMV reactivation. (ii).We compared clinical parameters (age, sex, age, sex, disease duration, and disease location)and endoscopic findings using by ulcerative colitis endoscopic index of sensitivity (UCEIS)(Travis SP et al. Gut 2012;61:535-42) between CMV positive group and negative group.Results: (i). Sixteen (30.2%) of fifty-three patients were diagnosed as positive for CMVreactivation by mucosal-PCR. Of this 16 patients, IHC was positive in only three patients(18.8%) (ii). Disease duration was significantly shorter in CMV positive group than in CMVnegative group (21.5±7.2 vs 66.0±14.1 months, p = 0.049). There were no significantlydifferences in other clinical parameters between two groups. There was no significant differ-ence in the mean of UCEIS score between two groups (8.3±1.1 vs 8.5±1.5, p = 0.689).Conclusion: Our data demonstrated that 30.2% of CS Naïve patients with active UC wasdiagnosed as positive for CMV reactivation. However, there is no significant difference ofUCEIS in active UC between positive and negative for colonic CMV reactivation. These datasuggested that inflammation triggered colonic CMV reactivation in subpopulation of CSNaïve UC. Therefore, prior to starting immunomodulative therapies, mucosal PCR wouldbe recommended for all active UC.

Mo1252

Inflammatory Bowel Disease Is Associated With an Increased Risk ofMelanoma: A Systematic Review and Meta-AnalysisSiddharth Singh, Sajan Jiv Singh Nagpal, Mohammad H. Murad, Siddhant Yadav,Sunanda V. Kane, Darrell S. Pardi, Jayant Talwalkar, Edward V. Loftus

Background and Aims: Several observational studies have shown that inflammatory boweldisease (IBD) is associated with an increased risk of non-melanoma skin cancer, particularlywith the use of thiopurines. It is unclear whether IBD modifies the risk of melanoma.Methods: We conducted a systematic search of Medline, Embase and Web of Science upto October 2012, and manually reviewed abstracts from major gastroenterology conferences(2004-2012) as well as reference lists of all articles. Keywords used in the search included"cancer" combined with "inflammatory bowel disease", "Crohn's disease" or "ulcerativecolitis". Cohort studies which reported incident cases of melanoma after diagnosis of IBD,as well as an estimate of incidence rate ratio (IRR) or standardized incidence rate (SIR),were included in the meta-analysis. Pooled relative risk (RR) estimates with 95% confidenceintervals (CIs) were calculated using the random-effects model. Statistical heterogeneity wasassessed with the Cochran's Q statistic and the I2 statistic. Results: Eleven studies (151,547patients with IBD - 87,430 with Crohn's disease [CD] and 64,117 with ulcerative colitis[UC]) reporting 118 cases of melanoma from 1940-2009 were included in the analysis. Thepooled crude incidence rate of melanoma in patients with IBD was 33.7 (95% CI, 23.9-46.3) per 100,000 person-years (based on data from 6 population-based cohorts). Overall,IBD was associated with 43% increased risk of melanoma (n=11 studies; RR, 1.43; 95% CI,1.08-1.88). The risk was increased in both CD (n=6 studies; RR, 1.80; 95% CI, 1.17-2.75)and UC (n=7 studies; RR, 1.23; 95% CI, 1.01-1.50), with no significant heterogeneity acrossstudies (CD, Q statistic p=0.24, I2=26.0%; UC, Q statistic p=0.54, I2=0%). There was nosignificant difference in the risk of melanoma in population-based or hospital-based cohortstudies. The risk of melanoma was increased in studies performed in the pre-biologic era(before 1998) (n=8 studies; RR, 1.52; 95%CI, 1.02-2.25), although not significantly increasedin studies in the biologic era (after 1998) (n=2 studies; RR 1.08; 95% CI, 0.59-1.96); 1study fromOlmsted county included patients from 1940-2004. Subgroup analyses are shownin Table 1. There was no evidence of significant publication bias. Conclusion: Based on thismeta-analysis, both CD and UC are associated with an increased risk of melanoma, indepen-dent of the use of biologic therapy. The effect of immunomodulator therapy could not beestimated. Patients should be appropriately counseled on the risk of melanoma with thediagnosis of IBD.Subgroup Analysis

Mo1253

National Trends and Inpatient Outcomes of Inflammatory Bowel Disease WithConcomitant Chronic Liver DiseaseDouglas L. Nguyen, M. Mazen Jamal

Background: There is little information on the frequency of chronic liver disease amonghospitalized patients with Inflammatory Bowel Disease (IBD). In this study, we seek to definecommon etiologies contributing chronic liver disease among IBD patients and identifyingpotential risk factors predictive of increased mortality in this population. Methods: Weanalyzed the Nationwide Inpatient Sample from 1988-2006 to determine the frequency ofchronic liver disease among patients with IBD and to determine their in-hospital outcomes.A multivariate analysis was performed to identify factors predictive of increased inpatientmortality. Results: From 1988 to 2006, the age-adjusted rate of chronic liver disease amonghospitalized patients with IBD has nearly tripled from 0.75 per 100,000 persons in 1988-2001 to 2.15 per 100, 000 persons in 2004-2006. The most common etiologies contributingto chronic liver disease among Ulcerative Colitis patients were: primary sclerosing cholangitis(50.62%), cryptogenic cirrhosis (20.17%), non-alcoholic fatty liver disease (12.73%), chronichepatitis C (11.00%), primary biliary cirrhosis (4.15%), and chronic hepatitis B (1.33%).In contrast, the most common etiologies contributing to chronic liver disease in Crohn'spatients were: primary sclerosing cholangitis (26.54%), chronic hepatitis C (23.99%), cryp-togenic cirrhosis (23.99%), non-alcoholic fatty liver disease 20.26%), chronic hepatitis B(3.05%), and primary biliary cirrhosis (2.17%). Compared to IBD patients without concomi-tant liver disease, there was a 2.5-fold higher rate of inpatient morality among IBD patientswith concomitant liver disease (2.76% vs. 1.29%, p ,0.01). The multivariate analysis showedthat factors predictive of inpatient mortality among IBD patients with liver disease include—age .50 (OR 1.83 95% CI 1.44, 2.34), spontaneous bacterial peritonitis (OR 2.93 95% CI2.24, 3.83), hepatic encephalopathy (OR 3.77, 2.75, 5.16), presence of cirrhosis (1.93 OR95% CI 1.51, 2.48), and Clostridium difficile colitis (OR 2.36 95% 1.29, 4.32). Geographiclocation, hospital size, and variceal hemorrhage were not predictive of increased mortality.Conclusions: The age-adjusted rate of chronic liver disease among hospitalized IBD patientshas nearly tripled since 1988. There is a higher rate of inpatient mortality among patientswith concomitant IBD and chronic liver disease. Factors predictive of increased moralityinclude older age, Clostridium difficile infection, hepatic encephalopathy, spontaneous bacte-rial peritonitis, and the presence of cirrhosis. Therefore, early recognition and managementof complications related to portal hypertension among patients with IBD and chronic liverdisease is particularly important in order to reduce inpatient mortality and morbidity.

Mo1254

Reduction in Colectomy- and Health Care-Related Costs in Ulcerative ColitisPatients Treated With Adalimumab Compared With Standard TherapyMichael V. Chiorean, Mei Yang, Joanne Rizzo, Parvez Mulani, Jingdong Chao

Background: Anti-tumor necrosis factor agents are effective for treating patients with ulcera-tive colitis (UC). Aim: Our aim in this study was to determine the effect of treatment withadalimumab (ADA) vs. immunosuppressants or steroids (IMS) on health care utilization andcost of care in patients with UC. Methods: Adult patients (.18 years) were selected fromthe Truven Health MarketScan® Commercial Claims and Encounters databases if they had≥1 inpatient or ≥2 outpatient claims corresponding to a diagnosis of UC ( InternationalStatistical Classification of Diseases, 9th Revision [ICD-9], code/modifier 556.x) during a 5-year interval (2005-2009). Patients with concurrent diagnosis of Crohn's disease (ICD-9code, 555.x) and those receiving other biologics were excluded. Patients receiving ADA orIMS were grouped into 2 mutually exclusive cohorts: an ADA arm (with or without IMS)and an IMS-only arm. Patients had to be continuously enrolled during the 12-month periodbefore (baseline) and after (follow-up) the biologic or IMS initiation date. Changes frombaseline in UC-related hospitalizations (rate and length of stay [LOS]), outpatient services,corresponding health care costs, and colectomy (rate and LOS) were compared between theADA and IMS arms. Results: Of the 9,230 patients with UC eligible for the study, 143received ADA, and 9,087 received IMS only. There were no differences in age or sex betweenthe 2 groups, although ADA-treated patients had more comorbidities and greater steroidusage at baseline compared with the IMS-treated patients (69.2% vs. 32.1%, P ,.05).From baseline to follow-up, ADA-treated patients had significant reductions in UC-relatedhospitalizations and LOS, whereas IMS-treated patients had increases in these outcomes(table). UC-related outpatient service utilization and costs, as well as total inpatient andoutpatient service costs, decreased significantly for ADA-treated patients compared with asmall increase for IMS-treated patients. ADA-treated patients also had a significantly greaterreduction in hospital LOS for colectomy. Conclusion: In this retrospective claims databaseanalysis, when compared with IMS therapy, ADA therapy was associated with greater reduc-tions in UC-related hospitalization rates, length of hospitalization for colectomy, and totalinpatient and outpatient medical service costs.Difference in Outcomes: 12-Month Post- vs. Pre-treatment Initiation

S-619 AGA Abstracts

Mo1255

Complications and Length of Stay After Cholecystectomy Are IncreasedAmong Men With IBDHala Al-Jiboury, Adam V. Weizman, Dror Berel, Marc Berns, Marla Dubinsky, Stephan R.Targan, Eric A. Vasiliauskas, David Q. Shih, Manreet Kaur, Andrew Ippoliti, Dermot P.McGovern, Gil Y. Melmed

Introduction: Cholecystectomy is the most common abdominal surgery in the United States.Recent studies have suggested that patients with inflammatory bowel disease (IBD) areat higher risk of post-cholecystectomy complications (Navaneethan 2012). We aimed tocharacterize and identify risk factors for complications after cholecystectomy at a tertiaryIBD center. Methods: Using discharge codes, 118 patients who underwent cholecystectomyat our institution (1/1/2000 to 10/31/2012) with concomitant IBD were identified for medicalrecord review. Data collection included demographics, bodymass index (BMI), comorbidities,prior surgeries, smoking history, IBD phenotype, and concurrent therapies. Primary end-points included a) acuity of cholecystectomy (elective vs emergent) b) postoperative complica-tions (infections, bile leak, ileus, small bowel obstruction, intensive care (ICU)) c) lengthof stay (LOS). Statistical analysis included univariate and multivariable methods. Results:We identified 58 subjects with confirmed IBD (Crohn's disease n=41, ulcerative colitis n=17) and cholecystectomy (55% female, mean age 50 yrs (range 24 to 84), mean BMI 24.9(range 16 to 41)). At the time of surgery, 50% were not on any immunosuppressantmedication, 31% were on immunomodulators, 31% on steroids, and 21% on anti-TNFs.Those on any immunosuppressant had a longer LOS after surgery compared to those noton immunosuppressant therapy (4.7 days vs 3.6 days, p=0.04). Emergent cholecystectomy(compared to elective) was associated with a longer duration of IBD (483 mo. vs 176 mo,p=0.03), older age (p=0.003), male gender (p=0.03), and ever-smoking (p=0.02). Increasingage was also associated with transfer to the ICU (p=0.02) and postoperative ileus (p=0.04).Patients with Crohn's disease were more likely to have concurrent pancreatitis than thosewith UC (44% vs 18%, p=0.05). Only 7% women vs 28% men had histologic acutecholecystitis (p=0.06); the remainder had chronic changes only. Men were more likely toexperience postoperative ileus (p=0.03) and ICU admission (p=0.006), and had a longerLOS (mean 6.1 days vs 2.6 days, p=0.02) than women. After adjusting for multiple covariates,male gender remained associatedwith longer LOS after cholecystectomy. Immunosuppressantuse, and specifically anti-TNF therapy, was not associated with increased postoperativecomplications. Conclusion: Men with IBD are more likely than women to experience postop-erative complications after cholecystectomy, leading to a longer LOS after surgery. Otherfactors associated with postoperative outcomes include older age, smoking, and longerduration of IBD. Further characterization of risk factors is warranted to clarify preventablecauses of increased post-cholecystectomy complications in patients with IBD.

Graph: Post-cholecystectomy complications are increased among men with IBD relativeto women

Mo1256

Role of Fecal Calprotection in Predicting Ileocolonic Endoscopic Recurrence inPostoperative Crohn's DiseaseChristian Primas, Gertrud Frühwald, Sieglinde Angelberger, David Allerstorfer, PavolPapay, Alexander Eser, Cornelia Gratzer, Clemens Dejaco, Walter Reinisch, GottfriedNovacek, Harald Vogelsang

Introduction: Fecal calprotectin nicely correlates with intestinal disease activity in Crohn'sdisease. It was the aim of this study to evaluate its role in predicting endoscopic recurrence inpostoperative Crohn's disease. Material and Methods: 62 patients who underwent ileocolonicresection due to Crohn's diease at the General Hospital of Vienna were prospectively followedup. Ileocolonoscopy was done 6 to 18 months postoperatively and scored after Rutgeerts.Endoscopic recurrence was defined as i2b (at least i2 in terminal ileum) or higher. Endoscopicpictures were reviewed by 2 independent reviewers unaware of the calprotectin results. 5patients had to be excluded because the terminal ileum was not reached. 57 patients wereevaluated. The median age was 41 (range 22 to 68), 32 male (56,1%), 25 female. MontrealClassification: A2 (35%), A3 (65%); L1 (10%), L2 (15%), L3 (67,5%), L4 (7,5); B1 (2,5%),B2 (40%), B3 (57,5%); none had perianal disease. Stool specimens were collected immediately

AG

AA

bst

ract

s