Tu1268

Clinical and Genetic Factors Predict Severe Disease: A Novel CompositeSeverity IndexAnne M. Phillips, Ian D. Arnott, Tim Heron, Shirley Cleary, Craig Mowat, Hilary Clark,Nicholas Lewin-Koh, Jack Satsangi

Introduction: Crohn's disease is an incurable progressive disease with marked heterogeneityin outcome. Identifying patients at diagnosis at high risk of poor outcomes would allowtreatment stratification. The most widely used definition of severe disease lacks discriminationwith 85.2% of patients meeting the criteria for ‘disabling’ disease (Beaugerie et al. Gastroenter-ology 130: 650-656). A more discriminant composite severity score would benefit clinicalresearch and aid clinical decision making. Methods: Based on a consensus of 5 clinicians,we designed a composite severity score, assessed over the 5 years after diagnosis, with atotal possible score 1-16. The score was applied retrospectively in 367 CD patients whowere also genotyped for 32 CD susceptibility loci. The severity score that would encompassthe 50% of patients with the highest score was defined, and factors present at diagnosis -both clinical and genetic - were assessed by the chi squared test for their ability to predictbeing in the more severe disease category. CD was classified according to the Montrealclassification. Results: 367 CD patients (median age at Dx 30.8 (IQR 22.9-46.1)) wereassessed. 249 patients had full data available for the first 5 years after diagnosis. The meanseverity score was 6.3 (95% CI 5.9-6.6). Patients with a score >6 were defined as havingmore severe disease. Patients with more severe disease were younger at diagnosis (p=0.0004for trend, odds ratio (OR) A1 v A2=4.42 (95% CI 1.4-13.9)) and had an ileal (p=0.0025,OR=2.2 (95% CI 1.32-3.68)) and upper GI location disease location at diagnosis (p=0.0008,OR=5.3 95% CI 1.9-14.7). Perianal disease at diagnosis approached but did not achievestatistical significance (p=0.052). Only rs9286879 (p=0.0085) and rs17582416 (p=0.0448)were observed more frequently in patients with severe disease. The need for steroids atdiagnosis, having a first degree relative with IBD, needing surgery at diagnosis or smokingstatus at diagnosis did not differ between the groups. Conclusion: This index discriminatesmore effectively than existing definitions between severity groups. Disease location, age atdiagnosis and genetic markers are associated with more severe disease but validation ofthese factors is needed in a separate cohort.Severity score (first 5 years after diagnosis)

Tu1269

New Insights Into the Genetic Basis for Variation in Crohn's Disease (CD)Phenotype in a Population-Based StudyJohn D. Ryan, Mark S. Silverberg, Wei Xu, John R. Walker, Lesley A. Graff, Jason Ediger,Norine Miller, Linda Rogala, Michael Sargent, Rachel Carr, Joanne M. Stempak, PatriciaE. Rawsthorne, Charles N. Bernstein

Aims: In this study we evaluate for associations between clinical variables and IBD-relatedSNPs and specific disease phenotypes in a well-characterized, population-based CD cohort.Methods: Clinical data was obtained on a prospectively collected patient population fromthe Manitoba IBD Cohort Study (n=182, 112 females;70 males. These subjects were enrolledwithin 7 years of diagnosis (mean 4.3 yrs) and have been followed prsopectively every 6months with surveys and annually with in-person interviews. Phenotyping and categorizationof patients was performed using the Montreal Classification. DNA collected on these subjectswas genotyped and the allele frequencies of known IBD-associated SNPs were determined.Analysis was performed to identify the SNPs associated with disease location and behaviour,surgical history, smoking status and psychological variables. Results presented are uncorrec-ted. P value <0.01 was considered significant. Results: The distribution of patients withinMontreal Classification categories was: A1=10%, A2=64%, A3=26%; L1=43%, L2=22%, L3=34%, L4=7%; B1=45%, B2=33%, B3=22%, P=23%. Of the clinical variables, disease behavi-our was strongly associated with surgery p <0.0001, OR=7.29) and smoking status (p=0.01,OR=2.09). Perianal disease was also associated with younger age at diagnosis (p=0.004).Subjects undergoing surgery tended to have higher childhood adversity impact p = 0.06,OR=1.81). Lifetime history of depression p = 0.07, OR=1.75) tended to be more commonin patients with colonic involvement. For the genetic analysis, a SNP in DLG5 was associatedwith colonic and ileocolonic disease, while SNPs in RAC2 and NOX3 were associated withdisease confined to the terminal ileum. SNPs in NR3C1, SCUBE3 and CDH19 were associatedwith disease behaviour. SNPs in TLR9, PARD3 and DLG5 were associated with perianaldisease. 2 SNPs at the 5q31 locus and a SNP at SERPINA1 were associated with early surgery(< 3 years from diagnosis as opposed to no surgery) whereas one SNP in NOX4 was associatedwith no surgical intervention. Conclusions: Psychosocial variables were associated withcolonic disease and early surgery but require further exploration in a larger sample size.Early surgery occurs in those harboring mutations at the 5q31 locus and at a SNP atSERPINA1 Disease location and behaviour was associated with a number of SNPs, somepreviously not reported.

S-783 AGA Abstracts

Tu1270

The Risk of Crohn's Disease Progression Towards Intestinal Complications inan Asian CohortWei Lin Tay, Kelvin T. Thia, John C. Allen, San Choon Kong, Khoon-Lin Ling, Choon-JinOoi

BACKGROUND: An emergence of Crohn's disease has been reported in Asia over the pastdecade. While the natural history of Crohn's disease had been well described in Westernseries, data for Asian patients is scant. AIMS: We sought to assess the rate of intestinalcomplications and their associated risk factors in an Asian cohort. METHODS: This is aretrospective study of Crohn's disease patients currently on active follow-up at our centre.Records from the period 12/1975 to 11/2010 were evaluated . Medical records were reviewedand the clinical phenotypes at diagnosis of patients were categorized according to theMontreal classification. Baseline putative risk factors at and within 90 days of diagnosis suchas age, sex, race, disease location and perianal disease were documented. The cumulativeprobabilities of developing intestinal stricturing and/or penetrating events were estimatedusing the Kaplan-Meier method and Cox proportional hazards regression was used toassess associations between baseline risk factors and progression to intestinal complication.RESULTS: The baseline clinical characteristics of 138 patients are as shown in Table 1. Atbaseline, 77.5% had nonstricturing nonpenetrating disease, 13.8% had stricturing disease,and 8.7% had penetrating disease. The cumulative risks of developing either complicationswere 22.5% (95% CI: 15.4-29.6) at or within 90 days, 29.4% (95% CI: 21.4-37.3) at 5years, and 37.0% (95% CI: 27.7-46.3) at 10 years and 43.0% (95% CI: 31.2-54.7) at 20years after diagnosis. Among 107 patients with nonstricturing, nonpenetrating disease atbaseline, 17 were observed to experience a change in disease behavior after the first 90days from diagnosis. The cumulative risks of developing either a stricturing or penetratingcomplication among those with nonstricturing nonpenetrating phenotype at diagnosis were8.9% (95% CI: 2.9-14.9) at 5 years, 18.7% (95% CI: 9.4-28.1) at 10 years, and 26.5%(95% CI: 12.9-40.0) at 20 years. Among the 48 patients in the entire cohort with stricturingand penetrating phenotype, 32 (66.7%) underwent bowel resection surgery for managementof Crohn's disease-related complications. None of the baseline risk factors were found tobe significantly associated with the risk of intestinal complications. CONCLUSIONS: In thisAsian tertiary study, 22.5% of patients with Crohn's disease experienced stricturing orpenetrating complications within 90 days after diagnosis and 43% experienced intestinalcomplications 20 years after diagnosis. The complication rates observed resemble Westerncohort studies. The surgical morbidities were substantial with 2 in 3 patients requiring bowelresection after a change in behavior. Predictive risk factors for complications will need tobe confirmed in a larger Asian cohort.

Tu1271

Determinants Contributing to Fatigue in IBD PatientsLauran Vogelaar, Adriaan van 't Spijker, Ernst J. Kuipers, Christien J. van der Woude

Background: Inflammatory bowel disease (IBD) patients (pts) suffer from a disabling diseasewhich negatively influences their quality of life. Fatigue is a common symptom in IBD ptsand contributes to this decreased quality of life. Aim: To assess underlying factors contributingto fatigue in IBD pts. Patients and methods: Group A included an adult IBD patient groupthat anonymously completed an online survey on fatigue. A second group included patientsfrom the IBD-outpatient clinic, these patients filled out a paper version of this survey (groupB). The survey included the Checklist Individual Strength (CIS) (fatigue: CIS score ≥ 35;non-fatigue CIS score < 35), the Hospital Anxiety and Depression Scale (HADS), the Harvey-Bradshaw Index and the Ulcerative Colitis Index for disease activity and information oncurrent medication use. For statistical analysis descriptive statistics, univariate and multivari-ate logistic regression were used. Results: In total 977 pts were included. In group A 432pts completed the survey (75% female, 57.9% Crohn's disease (CD), 42.1% ulcerative colitis(UC), mean age 40 (SD 12.6). The mean CIS score was 46 (range 13-56). Overall, 90.7%of the patients experienced fatigue. In group A 54 pts (12.5%) had a clinical depressionand 99 pts (22.9%) with clinical anxiety. In contrast to non-fatigue pts, CD pts sufferingfrom fatigue were significantly more depressive (OR 1.57; 95% CI 1.22-2.01), and haddisease activity (OR 1.30; 95% CI 1.07-1.59). The results were similar in UC, however

AG

AA

bst

ract

s