available at

Transcription

1 Journal of Crohn's and Colitis (2008) 2, available at Development of pouchitis following ileal pouch-anal anastomosis (IPAA) for ulcerative colitis: A role for serological markers and microbial pattern recognition receptor genes Marc Ferrante a, Sarah Declerck a, Tamara Coopmans a, Gert De Hertogh b, Gert Van Assche a, Freddy Penninckx c, Paul Rutgeerts a, Karel Geboes b, Andre D'Hoore c, Séverine Vermeire a, a Department of Gastroenterology, University Hospital Gasthuisberg, Herestraat 49, B3000 Leuven, Belgium b Department of Pathology, University Hospital Gasthuisberg, Leuven, Belgium c Department of Abdominal Surgery, University Hospital Gasthuisberg, Leuven, Belgium Received 11 September 2007; accepted 17 October 2007 KEYWORDS Ulcerative colitis; Pouch; Pouchitis; Predictor; Genetics; Serology Abstract Background and Aims: Pouchitis, the most common complication after proctocolectomy with ileal pouch-anal anastomosis (IPAA) for ulcerative colitis, has been attributed to altered composition of faecal flora. We investigated the role of antimicrobial and antiglycan antibodies and polymorphisms in microbial pattern recognition receptor genes. Methods: Clinical charts of all 184 patients with ulcerative colitis who underwent IPAA between were reviewed for pre- and post-operative disease course. Results: Follow-up data were available in 172 patients [67 female, median age at proctocolectomy 39.1 years]. During a median follow-up of 6.7 (interquartile range ) years, 80 patients (47%) developed at least one episode of pouchitis. Cox proportional-hazard regression identified extra-intestinal manifestations [HR 1.78 (95%CI ), p =0.020], a Conference presentation: Part of this work was presented during the ECCO Congress 2007 in Innsbruck and DDW 2007 in Washington DC. Abbreviations: ACCA, anti-chitobioside antibodies; ALCA, anti-laminaribioside antibodies; AMCA, anti-mannobioside antibodies; CARD15, caspase recruitment domain 15; CD, Crohn's disease; CD14, cluster of differentiation 14; EIM, extra-intestinal manifestations; EU, Elisa units; gasca, glycominds anti-saccharomyces cerevisiae antibodies; HLA-B27, human leucocyte antigen B27; HR, hazard ratio; IBDU, inflammatory bowel disease type unclassified; IC, indeterminate colitis; IIF, indirect immunofluorescence; IL-1RA, interleukin-1 receptor antagonist; IPAA, ileal pouch-anal anastomosis; IQR, interquartile range; Omp, outer-membrane porin; panca, perinuclear anti-neutrophil cytoplasmatic antibodies; SNP, single nucleotide polymorphism; TLR, toll-like receptor; TNF, tumor necrosis factor; UC, ulcerative colitis; 95%CI, 95% confidence interval. Corresponding author. Tel.: ; fax: address: severine.vermeire@uz.kuleuven.be (S. Vermeire) /$ - see front matter 2007 European Crohn s and Colitis Organisation. Published by Elsevier B.V. All rights reserved. doi: /j.crohns

2 Development of pouchitis following IPAA for UC 143 GT/TT genotype at Toll-like-receptor-1 S87I [HR 1.64 ( ), p= 0.047], anti-chitobioside carbohydrate antibodies [HR 2.03 ( ), p= 0.021] and young age at diagnosis [p =0.003] to be independently associated with pouchitis. Factors associated with chronic pouchitis, diagnosed in 33 patients (19%), were extra-intestinal manifestations [HR 2.45 ( ), p =0.034], backwash ileitis [HR 3.15 ( ), p =0.032], outer-membrane porin antibodies [HR 2.67 ( ), p =0.016] and young age at proctocolectomy [p= 0.008]. Conclusions: The reported association with antibodies and Toll-like-receptor-1 supports the pathophysiological role of the faecal flora in the development of pouchitis European Crohn s and Colitis Organisation. Published by Elsevier B.V. All rights reserved. 1. Introduction Up to 30% of patients suffering from ulcerative colitis (UC) will ultimately need to undergo a total colectomy. 1 The most frequent indications for colectomy include intractable disease and the occurrence of dysplasia or cancer in case of longstanding colitis. A total proctocolectomy with ileal pouch-anal anastomosis (IPAA) has become the surgery of choice for the definitive management of UC, since it avoids a permanent stoma while removing all diseased colonic mucosa. 2 Although the mortality rate after IPAA is acceptable (less than 1%) 2,3, short term morbidity is seen in approximately 30% of patients, including post-operative bleeding, pouch leakage, pelvic abscess and fistula. 2 5 On the long term, patients with a pouch can also be troubled by small bowel obstruction, faecal incontinence, sexual dysfunction and infertility. 2 8 Pouch failure is seen in approximately 5% of patients. 2,3,9 The most frequent long-term complication is the occurrence of pouchitis 10, with cumulative incidence rates varying significantly between studies (7 to 59%). Besides differences in length and type of follow-up, an important reason for this variation is a lack of universally accepted diagnostic criteria for pouchitis. 11 From a clinical perspective, different types of pouchitis exist. 12 Patients can present with one single episode of acute pouchitis, while others have acute relapsing pouchitis or a chronic unremitting course. In our own experience, 25% of patients developed pouchitis (proven by endoscopy and histology) within the first year after closure of ileostomy. 9 During a median follow-up of 6.5 years, 46% of patient developed pouchitis, and more than half of them suffered from acute relapsing on chronic pouchitis. 9 The aetiology of pouchitis is not entirely understood. Bacterial overgrowth, altered balance of luminal bacteria, mucosal ischemia, nutritional deficiencies, lack of shortchain fatty acids and faecal bile acids toxicity have all been suggested as possible etiological factors. 11 Factors that support the role of the microbial flora, are the fact that pouchitis is only seen after restoration of the faecal stream and can be treated successfully by use of antibiotics. 11,12 Bearing in mind the high cumulative incidence rate of pouchitis, several investigators have tried to define predictive factors which could help in selecting these patients in need for more intensive post-operative follow-up or postoperative prophylaxis. Established risk factors are nonsmoking 13,14 and the presence of extra-intestinal manifestations, especially primary sclerosing cholangitis. 9,14 17 Possible risk factors might be the extent of colitis and presence of backwash ileitis prior to surgery 17,18, a young age at diagnosis or surgery 9,14, dysplasia as indication for surgery 13, pre-operative use of steroids 19 and the regular use of NSAIDs in the postoperative phase. 13,17 More recently, an association was found between pouchitis and variations in the interleukin 1 receptor antagonist (IL-1RA) and tumor necrosis factor (TNF) genes. 20,21 Finally, pre-operative panca levels seem to be higher in patients who develop chronic pouchitis compared to patients who do not. 22 The primary aim of our study was to look for factors associated with pouchitis in patients who underwent a restorative proctocolectomy with IPAA for ulcerative colitis or IBD type unclassified (IBDU). Given presumed role of bacteria as main trigger in the onset of pouchitis, we were especially interested in the role of antimicrobial and antiglycan antibodies and polymorphisms in microbial pattern recognition receptor genes. 2. Material and methods 2.1. Patients All patients who underwent a proctocolectomy with IPAA for ulcerative colitis (UC) or IBD type unclassified (IBDU) at the University Hospital Gasthuisberg in Leuven (tertiary referral centre) between January 1990 and December 2004 were identified through the surgical database. A total of 184 consecutive patients [113 male/71 female; median (interquartile range, IQR) age at diagnosis 29.5 ( ) years; median (IQR) age at proctocolectomy 39.0 ( ) years] were identified. However, we were unable to trace 11 out of 184 patients (5%) who had moved house without leaving updated contact details. The ileostomy could be closed in all but one patient. This patient was excluded for further analysis. Clinical charts of all remaining 172 patients were reviewed to trace clinical, endoscopical and histological characteristics, including gender, age at diagnosis, age at proctocolectomy, age at closure of ileostomy, duration of disease prior to proctocolectomy, duration of follow-up after closure of ileostomy, diagnosis prior to proctocolectomy, diagnosis after proctocolectomy, familial history, extraintestinal manifestations, smoking behaviour, therapy at proctocolectomy, extent of disease prior to proctocolectomy, presence of back-wash ileitis. Surgical characteristics included type of surgery (laparoscopic vs. laparotomy), type of anastomosis (stapled without mucosectomy vs. handsewn with mucosectomy), type of pouch (J-pouch or S-pouch), number of stages (1-, 2- or 3-stage surgery), construction of the pouch at time of proctocolectomy. The main clinical and surgical characteristics are enlisted in Table 1. The clinical charts were also reviewed to trace the occurrence of first pouchitis and the occurrence of chronic

3 144 M. Ferrante et al. Table 1 Clinical and surgical characteristics Subgroup a N=144 Female/male (%) 60/84 (42/58) Median (IQR) age at 31.2 diagnosis (years) ( ) Median (IQR) disease 5.4 duration prior to ( ) proctocolectomy (years) Median (IQR) age at proctocolectomy (years) 39.1 ( ) Original cohort N=172 67/105 (39/61) 31.3 ( ) 5.1 ( ) 39.1 ( ) P-value Diagnosis prior to proctocolectomy b,c Ulcerative colitis (%) 135/144 (94) 161/172 (94) IBD unclassified (%) 9/144 (6) 11/172 (6) Crohn's disease (%) 0/144 (0) 0/172 (0) Diagnosis after proctocolectomy b,c Ulcerative colitis (%) 143/144 (99) 169/172 (99) Indeterminate 1/144 (1) 2/172 (1) colitis (%) Crohn's disease (%) 0/144 (0) 0/172 (0) Indication for proctocolectomy Intractable 138/144 (96) 165/172 (96) disease (%) Dysplasia/malignancy 6/144 (4) 7/172 (4) (%) Familial history of 39/139 (28) 39/140 (28) IBD (%) Extent of colitis prior to proctocolectomy c Proctitis (%) 0/144 (0) 0/170 (0) Left-sided colitis (%) 40/144 (28) 47/170 (28) Extensive colitis (%) 104/144 (72) 123/170 (72) Backwash ileitis prior to 8/133 (6) 8/133 (6) proctocolectomy (%) Extra-intestinal 72/144 (50) 77/170 (45) manifestation prior to proctocolectomy (%) Smoking at proctocolectomy Never (%) 61/130 (47) 61/130 (47) Former (%) 50/130 (38) 50/130 (38) Active (%) 19/130 (15) 19/130 (15) Medication at proctocolectomy Aminosalicylates (%) 72/144 (50) 84/168 (50) Corticosteroids (%) 109/144 (76) 132/168 (79) Azathioprine, 54/144 (38) 64/168 (38) mercaptopurine (%) Methotrexate (%) 1/144 (1) 3/168 (2) Cyclosporine (%) 41/144 (28) 48/168 (29) Infliximab (%) 3/144 (2) 3/168 (2) Antibiotics (%) 23/144 (16) 27/168 (16) Table 1 (continued) pouchitis. In case of absence of follow-up at our tertiary referral centre in the last six months, we contacted the clinician (general practitioner or local gastroenterologist) following the patient. Over the study period of 15 years, ileal pouch-anal surgery was performed by 4 experienced abdominal surgeons. Restorative proctocolectomy with IPAA was defined as a one-, two- or three stage procedure. Importantly, the number of stages did not follow strict guidelines, but the surgical judgement was based on severity of colitis at the time of surgery and clinical condition of the patient. In general, pouch construction was not performed in the presence of risk factors that were identified to be related with leakage: a preoperative white blood cell count N10,000/μl, emergency surgery, or the combination of leucocytosis and preoperative corticosteroid dose equivalent to or more than 200 mg hydrocortisone over 24 h. 23 Follow-up data were obtained by standardized and regular outpatient visits: one early postoperative visit six weeks after closure of ileostomy, followed by 3-monthly visits up to one year, and yearly visits thereafter, unless patients developed symptoms necessitating earlier visits. Follow-up visits included detailed history and clinical examination. If the clinician suspected pouchitis, an endoscopy was performed and biopsies were taken for histological examination. The diagnosis of pouchitis was only made after confirmation on endoscopy and histological examination Genetic analyses Subgroup a N=144 Original cohort N=172 P-value Laparoscopic 69/144 (48) 77/172 (45) assisted surgery Type of anastomosis Stapled without 141/144 (98) 165/172 (95) mucosectomy (%) Handsewn with 3/144 (2) 7/172 (4) mucosectomy (%) Type of pouch J-pouch (%) 143/144 (99) 169/172 (98) S-pouch (%) 2/144 (1) 4/172 (2) Number of stages One-stage surgery (%) 25/144 (17) 26/172 (15) Two-stage surgery (%) 96/144 (67) 115/172 (67) Three-stage surgery (%) 23/144 (16) 31/172 (18) Pouch construction 110/144 (76) 127/172 (74) at the time of proctocolectomy (%) a The subgroup consisted of 144 patients from whom blood was available for genetic and serological analyses. b Based on clinical, radiological, endoscopical and histological findings. 29 c According to the Montreal classification. 30 We performed a robust search of both the literature and the NCBI public database ( to reveal non-

4 Development of pouchitis following IPAA for UC 145 Table 2 Results of genetic analyses (n = 144) Marker Freq (%) Marker Freq (%) Marker Freq (%) Marker Freq (%) TLR1 R80T TLR4 D299G TLR10 W525R CARD15 SNP8 GG 120 (83) AA 129 (90) CC 127 (88) (86) GC 22 (15) AG 15 (10) CT 17 (12) (13) CC 2 (2) GG 0 (0) TT 0 (0) 22 1 (1) Allele G 262 (91) Allele A 273 (95) Allele C 271 (94) Allele (93) Allele C 26 (9) Allele G 15 (5) Allele T 17 (6) Allele 2 21 (7) HWE HWE HWE HWE TLR1 S87I TLR5 I487L TLR10 I298V CARD15 SNP12 GG 69 (48) CC 138 (96) GG 138 (96) (97) GT 64 (44) CA 6 (4) GA 6 (4) 12 5 (3) TT 11 (8) AA 0 (0) AA 0 (0) 22 0 (0) Allele G 202 (70) Allele C 282 (98) Allele G 282 (98) Allele (98) Allele T 86 (30) Allele A 6 (2) Allele A 6 (2) Allele 2 5 (2) HWE HWE HWE HWE TLR6/1 T469A TLR6 S249P TLR10 S163A CARD15 SNP13 AA 144 (100) CC 44 (31) GG 138 (96) (95) AG 0 (0) CT 67 (46) GT 6 (4) 12 7 (5) GG 0 (0) TT 33 (23) TT 0 (0) 22 0 (0) Allele A 288 (100) Allele C 155 (54) Allele G 282 (98) Allele (98) Allele G 0 (0) Allele T 133 (46) Allele T 6 (2) Allele 2 7 (2) HWE HWE HWE HWE TLR2 P631H TLR7 L11Q CD14 promotor IL-1RA CC 134 (93) AA 107 (75) TT 35 (24) (86) CA 10 (7) AT 15 (10) TC 79 (55) 13 2 (1) AA 0 (0) TT 22 (15) CC 30 (21) 14 1 (1) Allele C 278 (97) Allele A 229 (80) Allele T 149 (52) (9) Allele A 10 (3) Allele T 59 (20) Allele C 139 (48) 33 4 (3) HWE HWE b0.001 HWE All other 0 (0) TLR2 R753Q TLR9 T882A TNF promotor GG 133 (92) GG 144 (100) GG 109 (76) GA 11 (8) GA 0 (0) GA 34 (23) AA 0 AA 0 (0) AA 1 (1) Allele G 277 (96) Allele G 288 (100) Allele G 252 (88) Allele A 11 (4) Allele A 0 (0) Allele A 36 (12) HWE HWE HWE HWE: Hardy Weinberg Equilibrium. synonymous single nucleotide polymorphisms (SNPs) in genes encoding for Toll-like receptors (TLR) 1 to 10. Besides 50 SNPs in TRL1-10, we also investigated previous reported SNPs in caspase recruitment domain CARD15, cluster of differentiation 14 (CD14) and TNF, as well as a variable number tandem repeat in IL-1RA. 20,21,24 26 All SNPs were analyzed using PCR- RFLP. In a cohort of 192 healthy controls, 37 SNPs did not exceed a prevalence of 1%. The remaining 19 polymorphisms were analyzed in 144 patients. The remaining 28 patients refused the extra blood withdrawal. Blood for genetic analysis was taken at time of informed consent and enrolment. Primers and restriction enzymes are available upon request (Supplementary File 1) Serological analyses Sera of 145 patients, obtained at time of informed consent and enrolment, were analyzed for expression of IgG anti- Saccharomyces cerevisiae (gasca), IgG anti-laminaribioside (ALCA), IgA anti-chitobioside (ACCA), IgG anti-mannobioside (AMCA) and anti-outer membrane porin (Omp) antibodies in a blinded fashion, by one experienced lab technician. All assays were performed in our own laboratory by using ELISA as described by the manufacturers, Glycominds Ltd (Lod, Israel) and INOVA Diagnostics Inc (San Diego, CA, USA). Cutoff values for positivity were defined at 50 Elisa Units (EU) for Table 3 Results of serological analyses (n = 145) Marker Median (IQR) Positive (%) (Cut-off value) gasca 7.68 ( ) 12/145 (8) (50 EU) ALCA ( ) 8/145 (5) (70 EU) ACCA ( ) 21/145 (14) (90 EU) AMCA ( ) 12/145 (8) (90 EU) Omp ( ) 57/145 (39) (25 EU) panca 65/145 (45) (1/40) IQR: interquartile range; EU: Elisa Units.

5 146 M. Ferrante et al. gasca, 70 EU for ALCA, 90 EU for both ACCA and AMCA, and 25 EU for Omp. 27 Perinuclear anti-neutrophil cytoplasmatic antibodies (panca) were determined by indirect immunofluorescence (IIF) using ethanol-fixed neutrophil slides (INOVA Diagnostics) 27 with a cut-off value for positivity set at 1/40. Figure 1 Pouchitis-free survival after closure of ileostomy (A), stratified by extra-intestinal manifestations (B), ACCA antibodies (C), TLR1 S87I genotype (D) and TLR6 S249P genotype (E).

6 Development of pouchitis following IPAA for UC Definitions The diagnosis of pouchitis was based on typical clinical symptoms, in combination with a suggestive endoscopic image of inflammation of the pouch, confirmed by histology. 11,28 Symptoms include watery, sometimes bloody diarrhoea, abdominal cramps, faecal urgency and tenesmus, general malaise and fever. Typical endoscopy findings are characterized by hyperaemic and/or hemorrhagic friable and granular mucosa with excessive mucopurulent areas and Figure 2 Chronic pouchitis-free survival after closure of ileostomy (A), stratified by extra-intestinal manifestations (B), backwash ileitis (C), ACCA antibodies (D), Omp antibodies (E) and TLR2 R753Q genotype (F).

7 148 M. Ferrante et al. superficial erosions. Histological examination shows acute inflammation including neutrophilic infiltration and mucosal ulceration in addition to chronic inflammation, including villous atrophy, crypt hyperplasia, and a chronic inflammatory cell infiltration. 11,12,28 Pouchitis was further categorized as acute, acute relapsing or chronic, as described previously. 12 Pouchitis was considered acute if symptoms responded rapidly to medication and if the duration was less than 4 weeks. Starting from 3 or more acute episodes, pouchitis was defined acute relapsing. Pouchitis was considered chronic if symptoms lasted for more than four weeks, despite standard therapy Statistical analysis patients were classified as inflammatory bowel disease type unclassified (IBDU). 30 In 9 of these patients the diagnosis of IBDU could be changed to ulcerative colitis after thorough examination of the resection specimen and previous biopsies. Postoperatively, two patients were diagnosed with indeterminate colitis (IC). None of the patients who underwent a proctocolectomy with IPAA had a pre- or perioperative diagnosis of Crohn's disease IPAA surgery and follow-up Surgical characteristics are summarized in Table 1. Twentysix patients (15%) underwent a one-stage procedure, while The cumulative incidence of pouchitis and chronic pouchitis was estimated by means of Kaplan Meier analyses. We used both Breslow and LogRank tests to compare hazard rates in populations defined by one variate at the time. Since the length of follow-up varied significantly between patients, the most appropriate test to look for predictors is the Breslow test, also known as Gehan Wilcoxon test, where time points are weighted by the number of cases at risk at each time point. However, we also included the most commonly used LogRank test, were all time point are weighted equally. This test has optimal power to detect alternatives where the hazard rates in the populations are proportional to each other. A Cox proportional hazards survival regression including all variables with Breslow p b 0.100, was performed to identify independent predictors. We checked departures from the proportional hazards model by introducing timedependent explanatory variables, using log(t) and adjusted by the average of the logs (0.5). There was no indication that the proportionality of hazard assumption was violated. Since blood samples were only available in 144 out of 172 patients, Cox proportional hazards survival regression was only performed in this subgroup of patients. However, as shown in Table 1, the main characteristics of this subgroup did not differ from the overall cohort. Other performed analyses included Chi Square statistics, Fisher's Exact and Mann Whitney U tests. All statistical analyses were performed using the SPSS 15.0 (SPSS Inc, Chicago, IL, USA) and SAS 9.1 statistical software packages (SAS Institute Inc., Cary, NC, USA). The threshold for statistical significance was predefined as p b Results 3.1. Demographics Demographic and clinical characteristics of the 172 patients [67 female/105 male; median (IQR) age at diagnosis 31.3 ( ) years; median (IQR) age at proctocolectomy 39.1 ( ) years] are summarized in Table 1. These characteristics did not differ significantly in the subgroup of 144 patients who consented for both genetic and serological blood analyses (Table 1). Prior to proctocolectomy, ulcerative colitis was diagnosed in 94% of the patients, based on clinical, radiological, endoscopical and histological findings. 29 The remaining 11 Figure 3 Pouchitis (A) and chronic pouchitis-free survival (B) after closure of ileostomy stratified by number of independent nominal risk factors. The independent nominal risk factors identified and included for pouchitis were extra-intestinal manifestations, ACCA antibodies and a GT/TT genotype at TLR1 S87I. The independent nominal risk factors identified and included for chronic pouchitis were extra-intestinal manifestations, backwash ileitis and Omp antibodies.

8 Development of pouchitis following IPAA for UC 149 thirty-one patients (18%) underwent a three-stage procedure (subtotal colectomy first, followed by completion proctocolectomy and IPAA, and finally ileostomy closure). The remaining 115 patients (67%) underwent a two-stage procedure. In 101 of these, the pouch construction was performed at the time of proctocolectomy (first stage), while in the remaining 14 patients pouch construction was performed after prior subtotal colectomy. The median (IQR) duration between proctocolectomy and closure of ileostomy was 11.7 ( ) weeks Genetic and serological analyses Genetic and serological analyses were performed in 144 and 145 patients, respectively. All but one SNP were in Hardy Weinberg equilibrium. The main results are highlighted in Tables 2 and Pouchitis During a median (IQR) follow-up of 6.7 ( ) years, 80 patients (47%) developed at least one episode of pouchitis (Fig. 1A). The prevalence of pouchitis steadily increased over time, with cumulative incidence rates of 25%, 32%, 36%, 40% and 45% at 1, 2, 3, 4 and 5 years, respectively. Of all patients developing pouchitis, 32 (40%) developed less than three episodes of pouchitis during follow-up, 15 (19%) had acute relapsing pouchitis and 33 (41%) developed chronic pouchitis (Fig. 2A). A Kaplan Meier analysis revealed that patient with EIM prior to proctocolectomy (Breslow p =0.006, LogRank p=0.011), and positive ACCA antibodies (Breslow p=0.041, LogRank p = 0.040) developed their first pouchitis episode more rapidly (Fig. 1B C). Furthermore, patients with a GT/TT genotype at TLR1 S87I (Breslow p=0.065, LogRank p=0.051), and a CC/CT genotype at TLR6 S249P (Breslow p = 0.070, LogRank p= 0.113) showed a tendency to develop their first pouchitis more rapidly (Fig. 1D E). Patients who developed pouchitis were significantly younger, both at diagnosis [median (IQR) age ( ) vs ( ) years, p = 0.005] and at proctocolectomy [median (IQR) age ( ) vs ( ) years, p = 0.003]. Duration of disease prior to proctocolectomy was similar in both cohorts. The Cox proportional hazard survival regression, identified presence of EIM [Hazard ratio 1.78 ( ), p = 0.020], presence of a GT/TT genotype at TLR1 S87I [Hazard ratio 1.64 ( ), p = 0.047], ACCA antibodies [Hazard ratio 2.03 ( ), p=0.021] and a young age at diagnosis [p = 0.003] as independent factors associated with pouchitis. As Fig. 3A shows, the pouchitis risk significantly increased if more risk factors (EIM, GT/TT genotype at TLR1 S87I and ACCA antibodies) were present (Breslow p b 0.001, LogRank pb0.001). In univariate analysis, chronic pouchitis was more rapidly diagnosed in patients with extra-intestinal manifestations (Breslow p= 0.004, LogRank p = 0.010), backwash ileitis (Breslow pb 0.001, LogRank p b 0.001), positive ACCA antibodies (Breslow 0.048, LogRank 0.091), positive Omp antibodies (Breslow p = 0.020, LogRank 0.024), and a GA/AA genotype at TLR2 R753Q (Breslow p= 0.009,LogRank p = 0.040) (Fig. 2B-F). Patients with pre-operative use of aminosalicylates tended to develop chronic pouchitis more rapidly (Breslow p =0.082, LogRank p= 0.102). Patients who developed chronic pouchitis were significantly younger at proctocolectomy [median (IQR) age ( ) vs ( ) years, p=0.040]. Age at diagnosis and duration of disease prior to proctocolectomy was similar in patients who developed chronic pouchitis and patients who did not. The Cox proportional hazard survival regression, identified presence of EIM [Hazard ratio 2.45 ( ), p = 0.034], presence of backwash ileitis [Hazard ratio 3.15 ( ), p = 0.032], Omp antibodies [Hazard ratio 2.67 ( ), p = 0.016] and a young age at proctocolectomy [p = 0.008] as independent factors associated with chronic pouchitis. As Fig. 3B shows, the chronic pouchitis risk significantly increased if more risk factors (EIM, backwash ileitis and Omp antibodies) were present (Breslow p = 0.001, LogRank p=0.002). 4. Discussion Pouchitis after proctocolectomy with ileal pouch-anal anastomosis for ulcerative colitis is still a major challenge. There is increasing evidence for a pathophysiological role of the microbial flora in the development of pouchitis. We looked for factors associated with pouchitis in 172 patients who underwent a restorative proctocolectomy with IPAA for UC or IBDU, and were especially interested in the role of antimicrobial and antiglycan antibodies and microbial pattern recognition receptor genes. During a median (IQR) follow-up of 6.7 ( ) years, 80 patients (47%) developed at least one episode of pouchitis. A Cox proportional hazard survival regression identified presence of EIM, a GT/TT genotype at TLR1 S87I, ACCA antibodies and a young age at diagnosis as independent factors associated with pouchitis. During the same follow-up period, 33 patients (19%) developed chronic pouchitis. Independent factors associated with chronic pouchitis, were presence of EIM, backwash ileitis, Omp antibodies and a young age at proctocolectomy. The main benefits of this trial are the extended follow-up (median 6.7 years), the availability of longitudinal data and the separate analysis of patients with chronic pouchitis. In contrast to other investigators 17,19 we did not analyze the subgroup of patients with acute and acute relapsing pouchitis separately, since these patients might still develop chronic pouchitis later on. After Cox proportional hazard survival regression, we identified a GT/TT genotype at TLR1 S87I to be associated with the development of pouchitis. To our knowledge, this is the first time that a pattern recognition receptor gene has been implicated in the development of pouchitis. Interestingly, TLR1 is a co-factor of the TLR2 lipopolysaccharide receptor complex 31, supporting the hypothesis that bacterial recognition is important in the onset of pouchitis. We were not able to confirm the previously reported associations with a polymorphism in the promoter region of TNF 21 neither with the IL-1RA gene. 20,21 Like Lammers et al we did not find an association with the three common CARD15 mutations. 32 Since we were limited to SNPs which could be analyzed through PCR-RFLP, it is still possible that other pattern recognition receptor genes are associated with pouchitis.

9 150 M. Ferrante et al. Furthermore, we found that ACCA and Omp antibodies were independent factors associated with pouchitis and chronic pouchitis, respectively. This finding again supports the pathophysiological role of the microbial flora. The previously reported association between high panca levels and development of chronic pouchitis 22,33 was not confirmed in our cohort. However, we did not use a quantitative ELISA method. The remaining anti-glycan antibodies ALCA, AMCA and gasca were also not associated with the development of pouchitis. The lack of association with ASCA has also been demonstrated by other investigators. 21,22,33,34 Recently, Fleshner et al reported an association between CBir1 levels and acute pouchitis 33, but the CBir1 kit is not commercially available yet. Finally, we could not observe a correlation with the cumulative amount of antibody response towards the 6 different serological markers tested in this cohort (data not shown). Concerning clinical factors, the presence of extraintestinal manifestations was associated with both development of pouchitis and chronic pouchitis. Similar results have been found by several other investigators. 14,17,19,35,36 The previously reported strong association with primary sclerosing cholangitis in particular could not be confirmed in our cohort, most likely due to the small number of patients with this extra-intestinal manifestation (n = 4). Interestingly, 3 out of 4 patients with primary sclerosing cholangitis developed pouchitis within 2 years after closure of ileostomy. The role of backwash ileitis has also been suggested in the past. 18,37 However, while Schmidt et al and Abdelrazeq et al reported a higher frequency of pouchitis in patients with terminal ileal inflammation, we only demonstrated an association with chronic pouchitis. Both extra-intestinal manifestations and backwash ileitis might be associated with a bacterial overload. Experiments in animal models showed a failure to develop joint disease in HLA-B27 transgenic rates which were reared in germ-free conditions 38 and an attenuation of the severity of joint disease was seen after treatment with metronidazole. 39 Finally, we found a role for young age at diagnosis and young age at proctocolectomy in pouchitis and chronic pouchitis, respectively. Similarly, Stalhberg et al reported that patients who developed pouchitis were significantly younger at diagnosis of UC compared to patients who did not develop pouchitis. 14 Like other investigators, we did not observe an association with duration of disease. 13 The main limitations of our study are its retrospective character and the fact that this is not a gene-tagged approached. Furthermore, antibody responses towards Pseudomonas fluorescens-associated sequence I2 (anti-i2) 40 and flagellin CBir1 (anti-cbir1) 41 were not evaluated, since these kits are not commercially available yet. Finally, most of the blood samples were not taken prior to proctocolectomy, which might have influenced the results of the serological markers. Although investigators always considered serological response stable over time, our group recently showed that the antibody response towards gasca, ACCA, AMCA and Omp increased over time in a cohort of patients with Crohn's disease (CD). 27 Data on the stability of markers in patients with UC are not available at the moment. In conclusion, we demonstrated an association between serological markers (ACCA and Omp), a non-synonymous SNP in TLR1 and the development of pouchitis. Together with the effective use of antibiotics for the treatment of pouchitis and the observation that pouchitis only develops after restoration of the faecal stream, our results suggest a major pathophysiological role of the microbial flora. Further prospective trials are necessary to confirm our findings. Acknowledgement Marc Ferrante and Sarah Declerck collected all the data. Marc Ferrante analyzed all data and wrote the article with the help of all other co-authors. Tamara Coopmans performed most of the genetic and serological analyses. Gert De Hertogh and Karel Geboes are the main pathologist who examined the resection specimen and biopsies. Freddy Penninckx and Andre D'Hoore are the abdominal surgeons who performed most of the surgeries and the post-operative follow-up. Paul Rutgeerts, Gert Van Assche and Séverine Vermeire are the main gastroenterologists who were in charge of the medical treatment in all patients. All authors read and approved the final manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi: /j.crohns References 1. Langholz E, Munkholm P, Davidsen M, Binder V. Colorectal cancer risk and mortality in patients with ulcerative colitis. Gastroenterology 1992;103(5): Fazio VW, Ziv Y, Church JM, et al. Ileal pouch-anal anastomoses complications and function in 1005 patients. Ann Surg 1995;222(2): Fazio VW, Tekkis PP, Remzi F, et al. Quantification of risk for pouch failure after ileal pouch anal anastomosis surgery. Ann Surg 2003;238(4): Keighley MR, Grobler S, Bain I. An audit of restorative proctocolectomy. Gut 1993;34(5): Meagher AP, Farouk R, Dozois RR, Kelly KA, Pemberton JH. J ileal pouch-anal anastomosis for chronic ulcerative colitis: complications and long-term outcome in 1310 patients. Br J Surg 1998;85(6): Damgaard B, Wettergren A, Kirkegaard P. Social and sexual function following ileal pouch-anal anastomosis. Dis Colon Rectum 1995;38(3): MacLean AR, Cohen Z, MacRae HM, et al. Risk of small bowel obstruction after the ileal pouch-anal anastomosis. Ann Surg 2002;235(2): Waljee A, Waljee J, Morris AM, Higgins PD. Threefold increased risk of infertility: a meta-analysis of infertility after ileal pouch anal anastomosis in ulcerative colitis. Gut 2006;55(11): Ferrante M, Declerck S, De Hertogh G, et al. Outcome after proctocolectomy with ileal pouch-anal anastomosis for ulcerative colitis. Inflamm Bowel Dis 2007 Oct 31 [Published online]. 10. Pemberton JH, Kelly KA, Beart Jr RW, Dozois RR, Wolff BG, Ilstrup DM. Ileal pouch-anal anastomosis for chronic ulcerative colitis. Long-term results. Ann Surg 1987;206(4): Mahadevan U, Sandborn WJ. Diagnosis and management of pouchitis. Gastroenterology 2003;124(6): Sandborn WJ, Pardi DS. Clinical management of pouchitis. Gastroenterology 2004;127(6): Shen B, Fazio VW, Remzi FH, et al. Risk factors for diseases of ileal pouch-anal anastomosis after restorative proctocolectomy

10 Development of pouchitis following IPAA for UC 151 for ulcerative colitis. Clin Gastroenterol Hepatol 2006;4(1): Stahlberg D, Gullberg K, Liljeqvist L, Hellers G, Lofberg R. Pouchitis following pelvic pouch operation for ulcerative colitis. Incidence, cumulative risk, and risk factors. Dis Colon Rectum 1996;39(9): Penna C, Dozois R, Tremaine W, et al. Pouchitis after ileal pouchanal anastomosis for ulcerative colitis occurs with increased frequency in patients with associated primary sclerosing cholangitis. Gut 1996;38(2): Kuisma J, Jarvinen H, Kahri A, Farkkila M. Factors associated with disease activity of pouchitis after surgery for ulcerative colitis. Scand J Gastroenterol 2004;39(6): Achkar JP, Al Haddad M, Lashner B, et al. Differentiating risk factors for acute and chronic pouchitis. Clin Gastroenterol Hepatol 2005;3(1): Schmidt CM, Lazenby AJ, Hendrickson RJ, Sitzmann JV. Preoperative terminal ileal and colonic resection histopathology predicts risk of pouchitis in patients after ileoanal pull-through procedure. Ann Surg 1998;227(5): Fleshner P, Ippoliti A, Dubinsky M, et al. A prospective multivariate analysis of clinical factors associated with pouchitis after ileal pouch-anal anastomosis. Clin Gastroenterol Hepatol 2007;5(8): Carter MJ, Di Giovine FS, Cox A, et al. The interleukin 1 receptor antagonist gene allele 2 as a predictor of pouchitis following colectomy and IPAA in ulcerative colitis. Gastroenterology 2001;121(4): Aisenberg J, Legnani PE, Nilubol N, et al. Are PANCA, ASCA, or cytokine gene polymorphisms associated with pouchitis? Longterm follow-up in 102 ulcerative colitis patients. Am J Gastroenterol 2004;99(3): Fleshner PR, Vasiliauskas EA, Kam LY, et al. High level perinuclear antineutrophil cytoplasmic antibody (PANCA) in ulcerative colitis patients before colectomy predicts the development of chronic pouchitis after ileal pouch-anal anastomosis. Gut 2001;49(5): Heyvaert G, Penninckx F, Filez L, Aerts R, Kerremans R, Rutgeerts P. Restorative proctocolectomy in elective and emergency cases of ulcerative colitis. Int J Colorectal Dis 1994;9(2): Brett PM, Yasuda N, Yiannakou JY, et al. Genetic and immunological markers in pouchitis. Eur J Gastroenterol Hepatol 1996;8(10): Hugot JP, Chamaillard M, Zouali H, et al. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature ;411(6837): Lammers KM, Vergopoulos A, Babel N, et al. Probiotic therapy in the prevention of pouchitis onset: decreased interleukin-1beta, interleukin-8, and interferon-gamma gene expression. Inflamm Bowel Dis 2005;11(5): Ferrante M, Henckaerts L, Joossens M, et al. New serological markers in inflammatory bowel disease are associated with complicated disease behaviour. Gut 2007;56(10): Moskowitz RL, Shepherd NA, Nicholls RJ. An assessment of inflammation in the reservoir after restorative proctocolectomy with ileoanal ileal reservoir. Int J Colorectal Dis 1986;1(3): Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol Suppl 1989;170: Silverberg MS, Satsangi J, Ahmad T, et al. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can J Gastroenterol 2005;19(Suppl A): Pierik M, Joossens S, Van Steen K, et al. Toll-like receptor-1,-2, and-6 polymorphisms influence disease extension in inflammatory bowel diseases. Inflamm Bowel Dis 2006;12(1): Lammers K, Ouburg S, Crusius B, et al. Synergistic effect of Cd14 and Tlr9 polymorphisms in the susceptibility to develop chronic relapsing pouchitis. Gastroenterology 2005;128(4): A Fleshner P, Vasiliauskas E, Dubinsky M, et al. Both preoperative PANCA and CB1R1 flagellin expression in ulcerative colitis (UC) patients influence pouchitis development after ileal pouch-anal anastomosis (IPAA). Gastroenterology 2006;130(4): A Dendrinos KG, Becker JM, Stucchi AF, Saubermann LJ, Lamorte W, Farraye FA. Anti-saccharomyces cerevisiae antibodies are associated with the development of postoperative fistulas following ileal pouch-anal anastomosis. J Gastrointest Surg 2006;10(7): Lohmuller JL, Pemberton JH, Dozois RR, Ilstrup D, van Heerden J. Pouchitis and extraintestinal manifestations of inflammatory bowel disease after ileal pouch-anal anastomosis. Ann Surg 1990;211(5): Aisenberg J, Wagreich J, Shim J, et al. Perinuclear antineutrophil cytoplasmic antibody and refractory pouchitis. A case-control study. Dig Dis Sci 1995;40(9): Abdelrazeq AS, Lund JN, Leveson SH. Implications of pouchitis on the functional results following stapled restorative proctocolectomy. Dis Colon Rectum 2005;48(9): Rath HC, Herfarth HH, Ikeda JS, et al. Normal luminal bacteria, especially bacteroides species, mediate chronic colitis, gastritis, and arthritis in HLA-B27/human beta(2) microglobulin transgenic rats. J Clin Invest ;98(4): Rath HC, Bender DE, Holt LC, et al. Metronidazole attenuates colitis in HLA-B27/β2 microglobulin transgenic rats: a pathogenic role for anaerobic bacteria. Clin Immunol Immunopathol 1996;76: S Sutton CL, Kim J, Yamane A, et al. Identification of a novel bacterial sequence associated with Crohn's disease. Gastroenterology 2000;119(1): Targan SR, Landers CJ, Yang H, et al. Antibodies to CBir1 flagellin define a unique response that is associated independently with complicated Crohn's disease. Gastroenterology 2005;128(7):