reduce healthcare costs by ensuring all IBD patients have timely access to gastroenterolo-gist care.

S1027

Which Drugs Are Safe in Male and Female Patients Considering Conception,and in Pregnant Women and Nursing Mothers, with Crohn's Disease? Resultsof a Multidisciplinary International Expert Panel (EPACT II)Christian Mottet, John-Paul Vader, Christian P. Felley, Florian Froehlich, Jean-JacquesGonvers, Pascal Juillerat, Reinhold Stockbrügger, Pierre F. Michetti, Valérie Pittet

Introduction: High-grade evidence (randomized controlled trials) is lacking regarding theuse of drugs in pregnant and nursing Crohn's Disease (CD) patients. Explicit panel-basedsafety criteria made freely available on the internet (www.epact.ch) may therefore provide thebest available basis for clinical decision-making. Methods: An international multidisciplinaryexpert panel from 9 European countries convened in December 2007 in Geneva, Switzerland,to judge the appropriateness of Crohn's Disease therapy, including safety of therapy formale and female CD patients considering conception, as well as pregnant or nursing CDpatients, using a combined evidence- and panel-based method (RAND AppropriatenessMethod). Treatment options for CD in these patients were submitted to the panelists for anindividual rating on a 9-point scale: 1 = very unsafe, 9 = extremely safe. In the analysis, amedian of 7-9 was considered as safe (S); 4-6 equivocal (E); 1-3 unsafe (U). Results:Panel results are shown in the table below. Conclusion: Mesalazine, prednisone, probiotics,azathioprine and ciproxine are considered safe (the two latter are equivocal during lactation).Infliximab is the only anti-TNF-alpha considered safe for patients wishing to conceive orduring pregnancy, but only in the first trimester or while breastfeeding. These safety criteriamay be helpful in clinical decision-making.

*CC = considering conception. MMF = mycophenolate mofetil. NB: Methotrexate (FDAcategory X) was considered contraindicated and was therefore not subjected to vote.

S1028

Appropriate Maintenance Treatment for Crohn's Disease: Results of aMultidisciplinary International Expert Panel - EPACT IIPascal Juillerat, John-Paul Vader, Christian P. Felley, Valérie Pittet, Jean-Jacques Gonvers,Christian Mottet, Willem Bemelman, Marc Lémann, Tom Öresland, Pierre F. Michetti,Florian Froehlich

ABSTRACT BODY: Introduction: Biological therapy has dramatically changed the manage-ment of Crohn's disease (CD). Appropriateness criteria for maintenance treatment aftermedically-induced remission (MIR) or surgically-induced remission (SIR) of CD have thusbeen updated. Formal evidence is weak formany therapeutic decisions regardingmaintenancetreatment for CD and thus clinically-explicit and situation-tailored criteria made freelyavailable on the internet will be helpful to physicians in daily decision-making in clinicalpractice. Methods: Multidisciplinary international experts (EPACT II, Geneva, Switzerland)discussed and anonymously rated virtual clinical indications on the basis of evidence in theliterature. Median ratings (on a 9-point scale) were stratified according to three assessmentcategories: appropriate (7-9), uncertain (4-6 and/or disagreement) and inappropriate (1-3).Experts were also asked to rank appropriate medication according to their own clinicalpractice, without any consideration of cost. Results: 392 specific indications for maintenancetreatment of CD were rated (200 for MIR and 192 for SIR). Azathioprine, methotrexate and/or anti-TNF antibodies were considered appropriate in 42 indications, corresponding to68% of all appropriate interventions (97% ofMIR and 39%of SIR). The remaining appropriateinterventions consisted of mesalazine and a “wait-and-see” strategy. Factors that influencedthe panel's voting were patient characteristics and outcome of previous treatments. Resultsfavor the use of anti-TNF agents after one failure of any immunosuppressive therapy, whilean earlier use is still controversial. Conclusion: Detailed explicit appropriateness criteriahave been updated for maintenance treatment of CD. New expert recommendations forthe use of the classic immunomodulators as well as anti-TNF agents are now available(www.epact.ch). The validity of these criteria should now be tested by prospective evaluation.

A-173 AGA Abstracts

S1029

Clinical Use and Outcomes from IBD Serology Testing in a Tertiary ReferralCenterAlan C. Moss, Adam Cheifetz, Nabeel Chaudhary, Jahvari Junior

Background Testing for a combination of antibodies against ASCA, ANCA, and OmpC hasbecome common-place in IBD practice. Patient selection for testing, and the impact of testresults on patient management, remain variable. Aims To describe patient selection for IBDantibody testing in a tertiary referral center. To measure changes in management after testresults became available. To compare test results between patient sub-groups. MethodsReview of electronic medical records of all patients who completed IBD serology testing(Prometheus) at a single institution. Results Data were available on 213 patients who com-pleted testing. The distribution of pre-testing diagnoses were as follows; no diagnosis 41(19%), Crohn's disease 49 (23%), ulcerative colitis 60 (28%), IBD unclassified 60 (28%).In those with established IBD, 122 (57%) had colonic disease only, 36 (17%) had smallbowel disease only, 13 (6%) had ileo-colonic disease, and 10 (5%) had pouchitis. Theprevalence of elevated antibodies to ASCA, ANCA and OmpC in different patient populationscan be seen in Table 1. The proportion of patients with a change in diagnosis after antibodytesting in each patient sub-group was as follows; no diagnosis 27%, Crohn's disease 48%,ulcerative colitis 20%, IBD unclassified (IBDU) 63%. The odds ratio of a change in diagnosisafter testing positive for ASCA, ANCA or OmpC was not statistically significant in any patientsub-group. A cohort of patients also underwent antibody testing for celiac disease antibodies(n=21), but all were negative. Conclusions IBD serology testing was applied to a wide rangeof patient populations in this center. ASCA antibodies were present in similar proportionsacross all patient populations. Patients with IBDU were most likely to have a change indiagnosis after antibody testing.

S1030

Increased Risk of Psychiatric Disorders in Young Patients with Crohn'sDiseaseEdward V. Loftus, Annie Guérin, Andrew P. Yu, Eric Wu, Jingdong Chao, Parvez Mulani

Introduction: Onset of Crohn's disease (CD) is frequently during young adulthood. Youngpatients (pts) may have a greater risk of developing anxiety, depression, or other psychologicaldisorders because of their chronic illness.1 This study estimated incidence of psychiatricdisorders for young pts with CD vs. young pts without CD. Methods: CD pts aged <18years (yrs) with continuous enrollment 6 months before and 6 months after their first CDdiagnosis dates (index dates) were selected from records of a commercial health insuranceclaims database (MarketScan) dated between January 2000 and December 2006. Pt recordswere studied from CD index date to first diagnosis of a psychiatric disorder (alcohol anddrug abuse, major depression, anxiety disorder [through ICD-9 code], bipolar disorder,suicidal ideation, and eating disorders), end of data availability, or end of continuouseligibility, whichever occurred first. For every selected pt with CD, database records for 5control pts without CD were matched based on yr of birth, sex, and eligibility period. Ptswith psychiatric diagnoses or psychotropic medication use before the index date wereexcluded. Cox proportional hazard models controlling for age, sex, and comorbidities wereused to estimate the effect of CD on the risks of developing psychiatric disorders and useof psychotropic medications. The study also investigated the risk of developing persistentpsychiatric disorders, defined as receiving psychiatric-related care for >1 yr. Results: Thestudy included 2,144 CD pts and 10,720 age-, sex-matched controls. Mean (±SD) age ofpts was 12±5 yrs. Mean duration of follow-up for cases was 2.1 yrs (range, 0.5-6.5) andfor controls was 2.3 yrs (range, 0.5-6.5). Of the CD pts, 8.7% were diagnosed with 1 ofthe studied psychiatric disorders, vs. 4.6% of the CD-free control group. The most prevalentdisorders were depression (5.5% vs. 2.7%) and anxiety (3.8% vs. 1.3%) for CD pts vs.controls. The hazard for CD pts to develop any psychiatric disorder was 60% greater thanthat for control pts (hazard ratio [HR]=1.6, 95% CI: 1.3-2.0). HRs were estimated at 1.7(1.4-2.3) for depression and 2.3 (1.7-3.2) for anxiety. Pts with CD also demonstrated agreater risk of developing any persistent psychiatric disorder (HR=2.8, 95% CI: 1.9-4.2),persistent depression (HR=2.8, 95% CI: 1.7-4.4), and persistent anxiety (HR=4.4, 95%: CI2.2-8.5). Conclusion: In this retrospective analysis, young pts with CD demonstrated anincreased incidence of anxiety and depression vs. pts without CD and a significantly increasedrisk of developing persistent psychiatric disorders. References 1. Inflamm Bowel Dis.2008;14:1569-74.

S1031

Increased Risk of Lymphoma Among Inflammatory Bowel Disease PatientsTreated with Azathioprine in a Dutch Nationwide StudyChristine Vos, Noor Bakkal, Robert C. Minnee, Dirk J. De Jong, Gerard Dijkstra, PieterStokkers, Ad A. van Bodegraven, Marie J. Pierik, Christien J. van der Woude, BasOldenburg, Daniel Hommes

Introduction: A meta-analysis by Kandiel et al. (GUT 2005) suggest a fourfold increasedrisk of lymphoma in IBD patients treated with azathioprine/6-MP. In addition, the GETAIDgroup recently confirmed this association when they reported the results of the CESAMEstudy showing a standardized incidence ratio for lymphoma of 1.8 (DDW 2008). Hence, weundertook a nationwide database search to investigate similar associations in the Netherlands.Methods: To identify the patients with both IBD and lymphoma, the Dutch National Databaseof Pathology (PALGA) was used. From this registry eligible patients were selected between

AG

AA

bst

ract

s

AG

AA

bst

ract

s1997 and 2004. The data base search criteria used, included “ulcerative colitis”, “Crohn'sdisease”, “inflammatory bowel disease”, “inflammation”, “malignant lymphoma”, “oeso-phagus”, “stomach”, “small intestine” or “colon”. The final selection of cases was made uponverifying individual medical records from all cooperative hospitals. In addition, data on IBDcharacteristics, -outcome, medication and lymphoma were collected in CRFs. If possible,pathology specimens were tested for EBV. The Dutch, age-adjusted incidence of malignantlymphoma was retrieved from the Central Bureau of Statistics of the Netherlands (2004).The total number of IBD cases was determined per participating hospital. Results: Thedatabase search identified 259 cases with both IBD and malignant lymphoma in a totalestimated cohort of 16945 IBD cases. Only 186 case histories were available for full analysis,a total of 35 lymphomas were identified, 151 cases were excluded because of lacking theobjective diagnosis of IBD or lymphoma. The expected and age-adjusted lymphoma incidencewas 6.28, with an observed incidence of 35, the relative risk for lymphoma was 5.57. Ofthe 35 cases, 17 patients were exposed to AZA/6MP (min 3 months - max 12 yrs) and 18cases never used AZA/6MP. Remarkably, 10/13 (77%) lymphoma tissue specimens of IBDpatients using AZA/6MP tested positive for EBV versus only 1/17 (6%) EBV+ in patientswith no AZA/6MP use. Also, all EBV positive lymphoma's were diffuse large B cell lymphoma'swhereas EBV negative tumors were lymphoma's of different origin. Infliximab was not usedin any of the studied cases. Conclusion: This retrospective nationwide cohort study suggestsan increased risk for malignant lymphoma in IBD patients. In IBD patients using AZA/6MPthere seems to be a strong correlation with the development of EBV-positive diffuse largeB cell lymphoma's.

S1032

5-Aminosalicylic Acid Dependency in Crohn's DiseaseDana Duricova, Natalia Pedersen, Margarita Elkjaer, Pia S. Munkholm

Background: In the Danish Crohn Colitis Database during the treatment era of 5-Aminosal-icylic acid (5-ASA), steroids and surgery, it has been revealed that 8 years from diagnosis44% of Crohn s disease (CD) patients were characterized with a mild disease course, 20%with an aggressive (relapse every year) and 36% with a moderate disease course (relapseevery other year). Aim: The outcome of the first treatment course with 5-ASA monotherapy(1.5-4.8 g/day) was retrospectively studied in a consecutive cohort of 345 CD patientsdiagnosed 1953-2007. The immediate and long-term outcome of 5-ASA treament wasdescribed. Methods: A phenotyped model was used to assess treatment response (Table 1).Results: Out of 345 patients, 165 (48%) had monotherapy with 5-ASA. In 50% of them 5-ASA monotherapy was initiated within one year after diagnosis with a range 0-49 years.Complete or partial response was obtained in 75% and no response in 25% of patients.Among initial responders (complete/partial), prolonged response was obtained in 47% (59)of patients, 5-ASA dependency in 31% (38) and 18% (22) of patients lost initial responseto 5-ASA and had to shift to surgery (73%) or immunomodulator (27%). Five patients (4%)were not assessed in long-term outcome due to short treatment course. Female gender wasassociated with higher probability to develop prolonged response or 5-ASA dependency(OR 2.68, 95%CI: 1.06-6.77, p=0.03). The median duration (range) of 5-ASA course was34 months (1-304) in prolonged responders, 63 (6-336) in 5-ASA dependent patients and5 (2-10) in non-responders. Conclusion: It is indicated that the “right” phenotype of CDpatientsmay profit from 5-ASA treatment. Seventy-eight percent of initial responders obtainedlong-term benefit with 31% becoming 5-ASA dependent, resulting in 5 up to 28 years ofremission on 5-ASA in 50% of them. Prospective studies are warranted to assess the roleof 5-ASA in CD.Table1. A phenotyped model of treatment response

S1033

Diagnostic Delay in Patients with Inflammatory Bowel DiseaseStephan R. Vavricka, Sabrina Spigaglia, Alain Schoepfer, Valérie Pittet, Pierluigi Ballabeni,Pierre F. Michetti, Michael Fried, Peter Bauerfeind, Gerhard Rogler

Background and Aim: The diagnosis of inflammatory bowel disease (IBD) comprising Crohn'sdisease (CD) and ulcerative colitis (UC) continues to present difficulties. The unspecificsymptoms and the limited value of tests are the most common causes of delay in diagnosis.Studies have reported diagnostic delays in different gastrointestinal diseases such as celiacdisease and hemochromatosis and small studies have investigated diagnostic delay in pediatricIBD patients. However, diagnostic delay in adult IBD patients has not sufficiently beenstudied in recent years. The aim of the present study was to determine the time intervalsbetween onset of symptoms and diagnosis of IBD and find prognostic factors of delay.Methods: IBD patients from an adult, population-based cohort in Switzerland (Swiss IBDcohort study) were included in this study and were asked about their time of first symptoms,the time of diagnosis, visits to general practitioners and gastroenterologists and family historyof IBD. The Swiss IBD cohort collects prospectively data on a large sample of IBD patientsacross Switzerland through physicians and patient questionnaires. Results: A total of 909IBD patients (47.4% male, 52.6% female) aged 43.7 ± 14.6 (range 17-89 y) were included.523 patients presented with CD, 352 patients with UC and 24 patients with indetermined

A-174AGA Abstracts

colitis. Time from first symptoms due to IBD to diagnosis was 113 ± 254 weeks. For CDpatients time from first symptoms to diagnosis was significantly longer than for UC patients(154 ± 110 weeks vs. 55 ± 110 weeks, p<0.0001). There was no gender difference in bothdiseases. The time from the first symptoms to the visit at a general practitioner was 12.8 ±30.8 weeks (CD 14 ± 36 weeks vs. UC 11 ± 22 weeks, p=ns). The time between diagnosisand first visit to a gastroenterologist was 10 ± 32 weeks (CD 12 ± 39 weeks vs. 7 ± 14weeks, p=ns). In the 126 patients with a positive family history the time from first symptomsto diagnosis was significantly lower than in patients without family history of IBD (p<0.001).Conclusion: In our Swiss cohort, the diagnostic intervals of recognizing inflammatory boweldisease are still unacceptably long. More public awareness work has to be done. Diagnosisis delayed mainly on the level of general practitioners indicating that specific educationalmeasures and programs need to be established.

S1034

Symptoms of Functional Gastrointestinal Disorders (FGIDs) in IBD Patientsand the Inter-Relationships of Psychological Co-Morbidity, Quality of Life andDisease ActivityDaniel R. Van Langenberg, Robert V. Bryant, Gerald Holtmann, Jane M. Andrews

INTRODUCTION: Inflammatory bowel disease (IBD) patients with coexistent FGIDs appearto have significantly impaired quality of life. Furthermore, it has been proposed that psycholo-gical comorbidity is highly prevalent in this group. We aim to determine the burden ofFGIDs in a cohort of IBD patients and define the relationships amongst concomitant FGIDs,psychological comorbidity, quality of life and disease activity. METHODS: Over six monthsall IBD patients attending a single tertiary centre were prospectively identified by ICD-10coding. Diagnosis and current disease activity were ascertained on case note review. Eachwas sent a questionnaire on current quality of life (QoL) (short-IBDQ 10), anxiety/depression(HADS), and GI symptoms as recalled from when their IBD was quiescent (BDQ-6, adaptedAustralian version). Statistical analyses included Pearson's correlations, t tests and Fisher'sexact tests as appropriate. RESULTS: 162/256 patients responded to questionnaire (63%).95 had Crohns Disease (58.6%), 63 ulcerative colitis (38.8%), 52.5% female, median age43.2 years. By Rome III criteria, 34% had no FGID, 32% had one FGID & 34% had twoor more FGIDs. 31% met criteria for functional gastroduodenal disorders, 42% functionalbowel disorders and 31% functional anorectal disorders. None met criteria for functionalesophageal disorders. Specifically, 20.4% of patients met criteria for IBS and these had highermean anxiety/depression scores and lower QoL than those without IBS (9.3 vs 6.8, 7.3 vs5.0, 35.8 vs 46.8, all p<0.003). Overall, the number of FGIDs per patient correlated positivelywith both higher depression and anxiety scores (D r = 0.37, p<0.001: A r = 0.18, p=0.02)and negatively with QoL (r = -0.41, p <0.001). Those with clinically significant anxiety anddepression (HADS ≥8) were more likely to have coexistent FGIDs (A 78% vs 22%, D 89%vs 11%, each p<0.001) than those with HADS <8. Higher IBD disease activity negativelycorrelatedwithQoL (r= -0.27, p<0.001) but did not correlate with either anxiety or depressionscores. However, those with active disease in the study period were more likely than thosewith inactive disease to report symptoms of coexistent FGIDs (69% vs 31%, p<0.001).CONCLUSIONS: FGIDs are highly prevalent in IBD patients, even with quiescent disease.Whilst separation of organic vs functional symptoms is problematic, those meeting criteriafor FGIDs have greater psychological comorbidity and poorer QoL; IBS in particular appearsto define an at-risk group. Whilst IBD activity impairs QoL and leads to more reportedsymptoms, it does not appear to directly influence the rate of psychological comorbidity.

S1035

The Risk of Severe Crohn's Disease: Validation of a Clinical Prediction RuleGuillermo Bastida, Olga Merino, Mariam Aguas, Manuel Barreiro, Yamile Zabana, DanielGinard, Daniel Ceballos, Jose María Huguet, Fernando Muñoz, Valle Garcia, IgnacioCatalán, Alfredo J. Lucendo, Cristobal de la Coba, Xavier Aldeguer Mante, Antonio Palau,Carme Loras

Introduction: Early treatment with immunomodulators could modify the natural history ofCrohn's disease (CD). It is important to define a subgroup of patients at risk for severe CDat the time of diagnosis. A published study was performed with this aim on a French cohort.Aim: to validate the predictive markers identified in the French study and to assess a largernumber of factors associated to the development of disabling CD over the entire course ofthe disease. Methods: We included patients with complete follow-up at least the 5-yearperiod after diagnosis. We excluded operated patients or treated with immunosuppressorswithin the first month of the disease. CD was considered as disabling when at least one ofthe following criteria was present: more than two steroid courses required and/ or dependenceon steroids, further hospitalization for flare-up or complications of the disease, need forimmunosuppressive therapy and intestinal resection or surgical operation for perianal disease.After univariate analysis a logistic multivariate regression model was performed. Results: Atotal of 511 patients were included. Mean age 29.6 (SD 11.6) years, median follow up 109months, range 60-438. The rate of disabling CD within 5 years after diagnosis was 70.2%.Age under 40 years (OR: 1.95 (95% CI: 1.1-3.42)), the need for steroids (OR: 1.6 (95%CI: 1.05-2.3)), ileo-colonic location (OR: 1.84 (95% CI: 1.1-2.9)), and perianal lesions atdiagnosis (OR: 1.8 (95% CI: 1.01-3.5)) were confirmed as independent predictive markers.Positive predictive value was 0.82 and 0.89 in patients having 3 or 4 risk factors at diagnosis,respectively. Over the entire course of the disease the rate of disabling CD was 84.5%. Bylogistic regression, 6 factors were independently predictive of a disabling CD course: Strictur-ing or penetrating behaviour (OR: 2.3 (95% CI: 1.2-4.2)), active smoker (OR: 1.7 (95% CI:1.1-2.9)), age below 40 years (OR: 2.3 (95% CI: 1.2-4.2)), the need for steroids (OR: 1.8(95% CI: 1.1-2.9)), ileo-colonic location (OR: 2.3 (95% CI: 1.2-4.3)), and perianal lesionsat diagnosis (OR: 5.8 (95% CI: 1.8-19.2)). Positive predictive value was 0.91 in patientshaving 4 risk factors at diagnosis; all patients with 5 or 6 risk factors developed a disablingCD. Conclusions: We could confirm and validate the association between the developmentof disabling disease and age below 40 years, perianal lesion at diagnosis and the need ofsteroids to treat the first flare of CD. Other factors as stricturing or penetrating behaviouror active smoker should be corroborated. These patients could be the target for early treatmentwith anti-TNF and/or immunosuppressive drugs.