s of the 4th Congress of ECCO the European Crohn’s and Colitis Organisation S97 of the disease. We therefore wanted to investigate whether granuloma formation should be seen as a stable trait in CD patients and also studied if this was influenced by a genetic predisposition. Materials and Methods: From a cohort of 466 CD patients who underwent a bowel resection in the period between 1 1 1991 and 31 12 2007 we identified 88 patients with two or more bowel resections in that time period (55.7% male, median [interquartile range, IQR] age at diagnosis 20.92 [17.05 28.51] years, median [IQR] age at first surgery 32.81 [22.93 41.64] years, median [IQR] time between surgeries 5.76 [3.8 8.96] years. Revision of pathology slides (average of 10 slides per resection specimen) was performed to classify patients as granuloma positive or negative at each surgery. Epithelioid granulomas were defined as a well circumscribed collection of at least five epithelioid macrophages (activated histiocytes with a homogenous eosinophilic cytoplasm) with or without multinucleated giant cells. All patients were furthermore genotyped for 56 selected SNPs in human homologues of yeast autophagy (Atg) genes, using the Sequenom MassARRAY® platform. Results: In 71.6% of patients (n = 63) no change in granuloma status was observed with subsequent surgeries. Among these patients, 58.7% were classified as granuloma positive and 41.3% as granuloma negative. We could not find significant differences in age at first surgery, disease duration to first surgery or time between surgeries between patients whose granuloma status changed and the patients in whom it remained unchanged. Genotype and allele frequencies for the studied polymorphisms did not differ between patients whose granuloma status did or did not change. Conclusions: Granuloma formation appears to be a stable trait in the majority of CD patients. This raises the hypothesis that an underlying intrinsic defect may make a patient prone to develop granulomas. Genetic variants in autophagy genes do not seem to play a role in this resepect. P219 Overall and organ-specific extra-intestinal cancer in inflammatory bowel disease: a meta-analysis of population-based studies N. Pedersen Jr.1 *, D. Duricova2, M. Elkjaer3. 1Herlev University Hospital, Taastrup, Denmark, 2Charles University, Prague, Czech Republic, 3Herlev University Hospital, Herlev, Denmark Aim: To describe the risk of overall and organ-specific extraintestinal cancer (EIC) in inflammatory bowel disease (IBD), Crohn’s disease (CD) and ulcerative colitis (UC) based on population-based cohort studies. Methods: The MEDLINE (January 1980 February 2008) and abstracts from international conferences (UEGW, DDW 2004 2007) were searched for related articles using defined search criteria. Additionally, reference lists of the articles were reviewed. Papers fulfilling the inclusion criteria were studied on population size, time of follow-up, type of cancer and observed to expected cancers. The overall pooled risk estimates (standardized incidence ratio (SIR), observed/expected) were calculated using STATA meta-analysis software. Results: Eight papers fulfilled the inclusion criteria and reported on extra-intestinal cancer in IBD (n = 5), CD (n = 6) and UC (n = 5). The pooled SIR for EIC was not increased in IBD overall (SIR, 1.1; 95%CI 0.97 1.27) or in CD (SIR 1.13, 95%CI 0.89 1.40) and UC (SIR 1.04, 95%CI 0.92 1.16) separately. However, patients with IBD had significantly increased risk of cancer of the liver-biliary system (SIR 1.94, 95%CI 1.07 3.54) and skin (squamous cell, SIR 1.79, 95%CI 1.01 3.16; melanoma, SIR 1.17, 95%CI 1.66 2.08). In separate analyses, CD patients had a significantly increased risk of cancer of the upper gastrointestinal tract (SIR 2.87, 95%CI 1.66 4.96), stomach (SIR 2.05, 95%CI 1.06 3.97), lung (SIR 1.82, 95%CI 1.18 2.81), urinary bladder (SIR 2.03, 95%CI 1.14 3.63) and skin (SIR 2.35, 95%CI 1.43 3.86), but a decreased risk of corpus uteri cancer (SIR 0.84, 95%CI 0.36 0.96). Patients with UC had a significantly increased risk of cancer of the liver-biliary system (SIR 2.58, 95%CI 1.58 4.22) and a decreased risk of lung cancer (SIR 0.39, 95%CI 0.20 0.74). Conclusions: IBD patients are not at increased overall risk of developing extra-intestinal cancer but the organ-specific pattern differs from that of the general population. CD patients are at increased risk of developing upper gastrointestinal cancer, lung, urinary bladder, and squamous cell carcinoma, and at decreased risk of corpus uteri cancer. UC patients have an increased risk of liver-biliary cancer counterweighted by a decreased risk of lung cancer. P220 The NOD2 variants rs2066843 and rs2076756 are novel independent Crohn’s disease susceptibility gene variants associated with severe penetrating disease phenotype resulting in frequent need for surgery J. Seiderer1 *, J. Glas1, J. Diegelmann1, G. Pasciuto1, C. Tillack1, S. Pfennig1, M. Jurgens1, A. Konrad1, H. Torok1, U. Schiemann2, T. Griga3, W. Klein4, J.T. Epplen4, T. Mussack5, P. Lohse1, B. Goke1, T. Ochsenkuhn1, M. Folwaczny5, B. Muller-Myhsok6, S. Brand1. 1University of Munich Klinikum Groshadern, Munich, Germany, 2Inselspital Bern, Bern, Switzerland, 3Knappschaftskrankenhaus, Dortmund, Germany, 4Ruhruniversitat, Bochum, Germany, 5Klinikum der Universitat Munchen, Munich, Germany, 6Max-Planck-Institut fur Psychiatrie, Munich, Germany Background and Aims: A number of studies indicates a strong association of the three common NOD2/CARD15 variants p.R702W, p.G908R, and p.Leu1007fsX1008 with Crohn’s disease (CD) while the role of other NOD2 variants has not been investigated in detail. We therefore aimed to test the association and phenotypic consequences of two recently identified NOD2 variants (SNPs rs2066843 and rs2076756) in a large German cohort of inflammatory bowel disease (IBD) patients. Methods: Genomic DNA from 2750 individuals (CD: n = 832; ulcerative colitis (UC): n = 451; controls: n = 1478) was analyzed for the NOD2 variants rs2066843, rs2076756, rs2066844, rs2066845 and rs2066847. Results: The SNPs rs2066843 and rs2076756 were significantly associated with CD (p 40 years [A3; p = 0.032 for rs2066843, p = 0.007 for rs2076756 vs. wild-type (WT)]. WT patients were more likely to have colonic disease (L2; p = 0.041 for rs2066843, p = 0.032 for rs2076756 vs. homozygous) with a trend of more ileocolonic involvement (L3) in homozygous patients (rs2066843: p = 0.066; rs2076756: p = 0.058 vs. WT). Moreover, both SNPs were significantly associated with the need for disease-related surgery in CD patients homozygous for these variants (rs2066843: p = 0.015; rs2076756: p = 0.003 vs. WT). Patients homozygous for rs2076756 had a more severe disease behaviour with less non-stricturing, non-penetrating disease (B1, p = 0.023 vs. WT) and more stricturing, penetrating disease (B3, p = 0.015 vs. WT). In addition, in these patients higher rates of fistulas were found (p = 0.015 vs. WT). Conclusion: Our study identified SNPs rs2066843 and rs2076756 as novel independent CD susceptibility variants associated with later disease onset and a severe penetrating disease phenotype resulting in frequent need for surgery. by gest on A ril 5, 2016 http://eccoxfordjournals.org/ D ow nladed from