Available online at ScienceDirect

Transcription

1 Journal of Crohn's and Colitis (2014) 8, Available online at ScienceDirect Comparison of long-term outcomes of S and J pouches and continent ileostomies in ulcerative colitis patients with restorative proctocolectomy-experience in subspecialty pouch center Saurabh Mukewar a, Xianrui Wu b, Rocio Lopez c, Bo Shen a, a Departments of Gastroenterology/Hepatology, The Cleveland Clinic Foundation, Cleveland, OH, USA b Department of Colorectal Surgery, The Cleveland Clinic Foundation, Cleveland, OH, USA c Department of Quantitative Health Sciences, The Cleveland Clinic Foundation, Cleveland, OH, USA Received 15 October 2013; received in revised form 20 February 2014; accepted 20 February 2014 KEYWORDS Inflammatory bowel disease; Ileal pouch anal anastomosis; Restorative proctocolectomy; S-pouch; Ulcerative colitis Abstract Background: There are no published studies comparing pouch configurations and long-term adverse outcomes. Aim: To evaluate outcomes of ulcerative colitis (UC) patients undergoing restorative proctocolectomy with S-, J- pouches or continent ileostomy (CI). Methods: We conducted a historical cohort study from the prospectively maintained Pouch Registry. Demographic and clinical variables were evaluated with univariate and multivariable analyses. Results: Fourty-five patients with S pouches and 36 with CI (33 K pouches and 3 Barnett continent ileal reservoirs) were compared with 215 J pouches serving as controls (ratio 1:2.5). In multivariable analysis, patients with S pouches were 93% less likely to develop chronic antibiotic-refractory pouchitis (CARP) than subjects with J pouches (odds ratio [OR] = 0.07; 95% confidence interval: b0.001, 0.54; pb0.001). However, no significant difference in the frequency of CARP was found between the CI and J pouch groups (OR = 0.68; 95% confidence interval: 0.17, 2.00, p = 0.40). Patients with S pouches were 8 times more likely (95% confidence interval: 3.7, 17.5; p b 0.001) and patients with CI 5.6 had times more likely (95% confidence Abbreviations: AID, autoimmune disorders; BCIR, Barnett continent intestinal reservoir; CARP, chronic antibiotic-refractory pouchitis; CD, Crohn's disease; CI, continent ileostomy; EIM, extra-intestinal manifestations; IQR, interquartile range; IC, indeterminate colitis; IPAA, ileal pouch anal anastomosis; IPS, irritable pouch syndrome; mpdai, the modified Pouch Disease Activity Index; NSAID, non-steroidal anti-inflammatory drugs; OR, odds ratio; PSC, primary sclerosing cholangitis; SD, standard deviation; TNF, tumor-necrosis factor; UC, ulcerative colitis. Grant support: This study was partially supported by a research grant from the Crohn's and Colitis Foundation of America (to B.S.). Corresponding author at: Department of Gastroenterology/Hepatology-A31, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195, USA. Tel.: ; fax: address: shenb@ccf.org (B. Shen) / 2014 European Crohn's and Colitis Organisation. Published by Elsevier B.V.

2 1228 S. Mukewar et al. interval: 2.4, 13.3; p b 0.001) to have pouch surgery-related complications than those with J pouches. There was no difference in the rate of CD of the pouch, pouch-associated hospitalization and pouch failure between the S- pouch, CI and J- pouch groups. Conclusions: Patients with J pouches appeared to have a greater risk for chronic pouchitis than those with S-pouches and but had a lower risk for developing pouch surgery-related complications than those with S pouches or CI European Crohn's and Colitis Organisation. Published by Elsevier B.V. 1. Introduction Restorative proctocolectomy with ileal pouch-anal anastomosis (IPAA) has become the surgical procedure of choice for patients with ulcerative colitis (UC) who require surgery. Various designs for the pouch configuration have been developed over the past four decades. In 1969, Kock et al. 1 created the continent ileostomy (CI), named Kock pouch or K pouch. In 1978, Parks et al. 2 for the first time, constructed a 3-limb S pouch. This was followed by a 2-limb J pouch design by Utsunomiya et al. 3 in 1980 and a 4-limb W pouch design by Nicholls et al. in Currently, the majority of UC patients who require restorative proctocolectomy elect to have a J pouch, whenever technically feasible. A staged J-pouch surgery with stapled anastomosis without mucosectomy has become a standard procedure. On the other hand, the S pouch is mainly performed in those with technical difficulties for stapled anastomoses, such as patients with mucosectomy (for colitis-associated rectal neoplasia) or a short mesentery. The K pouch is often constructed for patients who have contraindications for J or S pouches, such as those with poor anal sphincter function, as a primary indication, or for selected patients who have failed J or S pouches, as a rescue or secondary operation. Each pouch configuration offers certain advantages over the others. For example, pouch inlet and outlet angulation (namely afferent limb and efferent limb syndromes, respectively) are more common with S pouches than J pouches. 5,6 However, there are no published studies addressing the longterm outcome of different pouch designs, in particular, a comparison of the occurrence of chronic antibiotic-refractory pouchitis (CARP), Crohn's disease (CD) of the pouch, pouch surgery-related complications, and pouch failure based on the pouch configuration. At our Pouch Center, we noticed that patients with an S pouch often had mechanical issues, such as afferent and efferent limb syndromes. On the other hand, we also found that pouch inflammation was not common in patients with S pouches or with CI. Therefore, we hypothesized that the long-term outcomes of ileal pouches vary with the configuration of pouch. The aim of the study was to evaluate long-term outcomes of UC patients undergoing restorative proctocolectomy with an S pouch, J pouch or CI. 2. Patients and methods 2.1. Patients The patients for this historical cohort study were identified from our Institutional Review Board (IRB)-approved Pouchitis Registry from 2002 to Study and control groups The study group consisted of patients with S or CI. CI consisted of patients with K pouches or Barnett continent intestinal reservoir (BCIR). Patients with J pouches, who were randomly selected from the 1550-case Pouch Registry, served as the control group, with a study to control ratio of approximately 1: Inclusion and exclusion criteria Patients with UC, indeterminate colitis (IC), or Crohn's colitis who underwent restorative proctocolectomy and intestinal pouches with S, K (or BCIR) or J configurations were included. BCIR was considered as a form of CI and was also included. Patients with familial adenomatous polyposis and other pouch configurations were excluded Demographic and clinical variables Demographics, history of smoking, and clinical characteristics of IBD, including duration of IBD, disease extent, history of the use of anti-tumor-necrosis factor (TNF) biologics (before and/ or after colectomy), the postoperative long-term use of nonsteroid anti-inflammatory drugs (NSAID), family history of IBD, extra-intestinal manifestations (EIM), concurrent autoimmune disorders (AID), and indication for colectomy, were retrieved. The clinical variables were defined as follows: smoking, smoking more than seven cigarettes a week since the surgery; preoperative diagnosis, including UC, IC, and Crohn's colitis; extent of colitis, defined on the basis of anatomical distribution of IBD at the time of colectomy and determined by a combined assessment of endoscopic, radiographic, macroscopic, and histopathological data; duration of IBD, defined at the time from the diagnosis of IBD to the completion of IPAA with ileostomy closure; family history of IBD, CD or UC in the first-degree relatives; preoperative biological therapy, any use of biological agent for IBD before colectomy; indication for proctocolectomy and IPAA, the primary reason for surgery based on clinical presentation and preoperative diagnostic studies; EIM, including musculoskeletal manifestations peripheral arthritis, ankylosing spondylitis and sacroilitis, hepatobiliary diseases primary sclerosing cholangitis, skin manifestations pyoderma gangrenosum, erythema nodosum and aphthous stomatitis, ocular manifestations uveitis, scleritis and episcleritis, vascular disorders thromboembolic events; autoimmune diseases defined by the presence of AID including adult-onset asthma autoimmune thyroid disease, type I diabetes, rheumatoid arthritis, autoimmune hemolytic anemia, celiac disease, Sjögren or

3 Pouch type and long-term outcomes 1229 Raynaud's disease with a history or coexisting presence of autoantibodies. The diagnosis of AID was confirmed by clinical evaluation and documentation of ICD-9-CM codes and current and/or history of disease directed medical therapy. Both electronic and paper charts from referring clinicians were reviewed by one of the senior investigators (B.S.) to diagnose AID. We also evaluated significant comorbidities including congestive heart failure, coronary artery bypass graft, chronic obstructive pulmonary disease, renal insufficiency, liver failure, non-pouch-related cancer and stroke. Data pertaining to the pouch surgery (staging of pouch surgery and indications for colectomy), duration of pouch, post-operative use of biologics and/or immunomodulators, and number of visits to the Pouch Center were compared between study and the control groups. Pouch related variables were defined as follows: duration of the pouch, the interval between the ileostomy closure and the time of pouch failure or the most recent follow-up at the Pouch Center; post-operative use of immunomodulators, the use of immunomodulators for pouch-related disorders or AID after pouch construction; post-operative use of biological agents, the use of anti-tnf biologics for pouch-related disorders or AID after pouch construction Diagnostic criteria for pouch disorders Patients were considered to have a normal pouch if they were asymptomatic, with no endoscopic and no or minimum histologic inflammation and fewer than three episodes of antibiotic-responsive pouchitis (symptoms responding to a 2-week, single-agent antibiotic) per year with the last episode occurring at least 6 months prior. 7 Pouchitis was defined as a condition with the modified Pouchitis Disease Activity Index (mpdai) scores N 5. 8 Acute pouchitis was defined as pouchitis that responded to a short course of a single antibiotic, which included antibioticresponsive (b 4 episodes per year) and antibiotic-dependent pouchitis (N 4 episodes per year, and requires long-term, continuous antibiotic or probiotic therapy to maintain remission). CARP was defined as a condition where a patient failed to respond to a 4-week course of a single antibiotic (metronidazole or ciprofloxacin), requiring prolonged therapy of N4 weeks consisting of N 2 antibiotics, oral or topical 5-aminosalicylate, corticosteroid therapy, immunomodulator therapy or therapy with anti-tnf biologics. CD of the pouch was diagnosed if there were nonsurgery-related perianal fistulae, granulomas on histology or inflammation and ulcerations in the afferent limb or in the small bowel on endoscopy in the absence of NSAID use. All patients with CD of the pouch underwent clinical, endoscopic, radiographic, and histologic evaluation. 5,9 Irritable pouch syndrome (IPS) was defined as the presence of abdominal pain, pelvic discomfort and diarrhea with no inflammation of the afferent limb, pouch or cuff on endoscopy. These symptoms had persisted for at least 12 weeks in 12 months prior to the data entry. Pouch-related mechanical and structural complications included inlet or outlet obstruction, pouch prolapse, nipple valve dysfunction or strictures (from the K pouch or BCIR), trapped ovarian syndrome, pouch bezoar or foreign body, pouch leak and anastomotic sinus. Hospitalization was defined as admission and hospital stay for more than 24 h due to a pouch order as a primary diagnosis, which did not include admission for elective pouch excision or diversion surgery. Pouch failure was defined as the need for permanent diversion or revision of the ileal pouch, with or without pouch excision Clinical practice pattern The Cleveland Clinic Pouch Center has been a national and international referral center for various ileal pouch disorders. At the Cleveland Clinic, the J pouch has been the preferred configuration for majority of the patients over the study period. The S pouch was generally created in patients with difficulty with reach or patients with rectal dysplasia undergoing mucosectomy. The K pouch was constructed under following circumstances: 1) the primary K pouch: The pouch was constructed for patients with poor anal sphincter function, in whom the construction of J or S pouches was contraindicated. 2) The secondary K pouch converted from failed J or S pouches: A K pouch was constructed using the old J or S pouches 10 ; or 3) secondary neo-k pouch: In patients with J- or S-pouch failure, the old pouch was excised and a new K pouch was constructed. BCIR, not a standard practice at our institution, was constructed at outside institutions, typically as primary pouches. All patients had routine follow-up at the Pouch Center staffed by an IBD specialist (B.S.). Typically, we followed patients with CARP, CD of the pouch and refractory cuffitis every 3 6 months, acute pouchitis or cuffitis every 6 12 months, and yearly for those with normal pouches Outcome measurement Primary outcome was the development of CARP, CD of the pouch, pouch-related mechanical/structural complications, pouch-associated hospitalization and pouch failure at the time of the most recent visit to the Pouch Center Statistical analysis Data are presented as mean ± standard deviation (SD), median (interquartile range [IQR]) or N (%). A univariable analysis was performed to assess differences between J, S and K pouches. Analysis of variance (ANOVA) or the nonparametric Kruskal Wallis tests were used to compare continuous factors and Pearson's chi-square tests were used for categorical variables. Ad-hoc pair-wise comparisons were done using a significance criterion of (0.05/3) in order to account for multiple comparisons. In addition, a multivariable logistic regression analysis was used to assess the impact of pouch type on post-operative development of CARP, pouch-related mechanical/structural complications and CD of the pouch. The final models were chosen using an automated stepwise variable selection method performed on 1000 bootstrap samples. Pouch type and smoking were included in the models and all other factors were considered for inclusion; variables with inclusion rates of at least 30% were kept in the models. None of the patients with S pouches developed CARP; hence a penalized maximum

4 1230 S. Mukewar et al. likelihood and the likelihood ratio tests were used for the regression model. In addition, a time-to-pouch failure analysis was performed. Follow-up time was defined as duration of pouch from creation to either failure or last follow-up visit. Kaplan Meier plots were constructed and log-rank tests were used to compare the groups; multivariable Cox Regression analysis was used to adjust for smoking. P b 0.05 was considered statistically significant. All analyses were performed using SAS (version 9.2, The SAS Institute, Cary, NC) and R (version , The R Foundation for Statistical Computing, Vienna, Austria). 3. Results 3.1. Demographics and IBD-related features Forty-five patients had S pouches and 36 had CI (including 21 primary K pouches, 3 BCIR, 4 secondary K pouches, and 8 secondary neo-k pouches). The control group consisted of 215 patients with J pouches. The demographic and clinical features of the S, CI and J pouch groups were compared (Tables 1 and 2). Patients with J pouches were more likely to Table 1 Demographic and clinical characteristics. Factor S Continent ileostomy J p-value (N = 45) (N = 36) (N = 215) Mean age at diagnosis of ulcerative colitis, years a 22.2 ± ± ± Male gender 16(35.6) b 17(47.2) 122(56.7) c Race 0.42 Caucasian 45(100.0) 36(100.0) 205(95.3) African American 0 0 4(1.9) Others 0 0 6(2.8) Smoking 0.99 Never 39(86.7) 30(83.3) 180(83.7) Ex-smoker 4(8.9) 4(11.1) 24(11.2) Current smoker 2(4.4) 2(5.6) 11(5.1) Chronic use of NSAIDs 4(8.9) 2(5.6) 21(9.8) 0.72 Family history of IBD a 0.31 None 32(72.7) 29(80.6) 178(82.8) Crohn's disease 2(4.5) 3(8.3) 13(6.0) Ulcerative colitis 10(22.7) 4(11.1) 24(11.2) Pre-operative diagnosis Crohn's colitis 1(2.2) 4(11.1) 3(1.4) Indeterminate colitis 3(6.7) 4(11.1) 14(6.5) Ulcerative colitis 41(91.1) 28(77.8) b 198(92.1) d Extra-intestinal manifestations a 18(40.0) 13(37.1) 81(37.7) 0.95 Primary sclerosing cholangitis 2(4.4) 0 8(3.7) 0.47 Autoimmune diseases 8(17.8) 1(2.8) 22(10.2) Co-morbidities 3(6.7) 2(5.6) 16(7.4) 0.91 Extent of colitis a 0.21 Pancolitis 43(97.7) 35(100.0) 202(94.0) Left-sided colitis or proctitis 1(2.3) 0 13(6.0) Fulminant colitis a 4(9.1) 2(5.7) 22(10.2) 0.7 Pre-operative biological use a 8(18.2) 4(11.1) 32(15.0) 0.68 Duration of IBD from diagnosis to pouch surgery (years) a 5.0[3.0,10.5] d 8.0[5.0,15.0] c 6.0[3.0,12.0] Indication for colectomy a Refractory to medical management 37(84.1) d 36(100.0) b,c 184(85.6) d Dysplasia 7(15.9) 0 31(14.4) A significance level of was used for pairwise ad-hoc comparisons. a Data not available for all subjects. Missing values: age at diagnosis of UC = 4, family history of IBD = 1, indication for colectomy = 1, extent of UC = 2, fulminant colitis = 2, pre-operative biological therapy = 2, duration of IBD from diagnosis to pouch surgery (years) = 2. Values presented as Mean ± SD with ANOVA; Median [P25, P75] with Kruskal Wallis test, or N (%) with Pearson's chi-square test unless otherwise stated. b Significantly different from J. c Significantly different from S. d Significantly different from continent ileostomy.

5 Pouch type and long-term outcomes 1231 Table 2 Long-term pouch outcomes. Factor S Continent ileostomy J p-value be male than those with S pouches (56.7% vs. 35.6%; p = 0.028). A greater number of patients with J pouches had a pre-operative diagnosis of classic UC than CI patients (92.1% vs. 77.8%; p = 0.014). Patients with CI (median = 8.0 years; IQR: 5.0, 15.0) had a significantly longer duration of IBD than those with S pouches (median = 5.0 years; IQR: 3.0, 10.5; p = 0.035). None of the patients who had CI had colectomy (N = 45) (N = 36) (N = 215) Mean age at pouch surgery a 30.2 ± ± ± Number of stages for pouch surgery a b b 1(2.9) c,d 8(3.7) b 2 32(72.7) 16(45.7) 154(71.6) 3 5(11.4) 2(5.7) 43(20.0) 4 or redo 7(15.9) 16(45.7) 10(4.7) Median duration from pouch to last visit (years) 15.0[11.0,21.0] b,c 20.5[13.5,28.5] c,d 9.0[5.0,14.0] b,d b0.001 Median duration from first Pouch Center visit to date a 5.0[1.00,7.0] 4.0[1.5,5.0] 5.0[2.0,6.0] 0.33 Post-operative use of immunomodulators 3(6.7) 4(11.1) 16(7.4) 0.72 Post-operative use of biologics 1(2.2) 2(5.6) 22(10.2) 0.17 Final pouch diagnosis Acute pouchitis 7(15.6) c 5(13.9) c 78(36.3) b,d Chronic antibiotic-refractory pouchitis 0 c 3(8.3) 28(13.0) d Crohn's disease of the pouch 11(24.4) 9(25.0) 40(18.6) 0.51 Pouch-related mechanical and structural complications 20(44.4) c 12(33.3) c 20(9.3) b,d b0.001 Number of visits to pouch center 1.00[1.00,2.0] 1.00[1.00,3.0] 1.00[1.00,3.0] 0.75 Post-op hospitalizations 4(8.9) 5(13.9) 25(11.6) 0.78 Pouch failure 3(6.7) 3(8.3) 17(7.9) 0.95 Cause of pouch failure 0.86 No failure 42(93.3) 33(91.7) 198(92.1) Pouch-surgery related complications 2(4.4) 2(5.6) 4(1.9) Chronic antibiotic-refractory pouchitis 0 0 1(0.47) Crohn's disease of the pouch 1(2.2) 1(2.8) 10(4.7) Cancer 0 0 2(0.93) Median follow up (years) a 15.0[10.0,21.0] b,c 21.0[14.0,29.0] c,d 9.0[4.0,14.0] b,d b0.001 a Data not available for all subjects. Missing values: age at pouch creation = 1, stage of pouch surgery = 2, first pouch center visit to date = 31, follow-up of pouch failure (years) = 1. Values presented as Mean ± SD with ANOVA; Median [P25, P75] with Kruskal Wallis test, or N (%) with Pearson's chi-square test unless otherwise stated. b Significantly different from continent ileostomy. c Significantly different from J. d Significantly different from S. for dysplasia, as compared to 14.4% and 15.9% of those with J or S pouches, respectively (p = 0.047). There were no differences in the age at the diagnosis of IBD, ethnic background, family history of IBD, history of smoking, NSAID use, EIM, PSC, extent of colitis, history of fulminant colitis, pre-operative anti-tnf biologic therapy, age at pouch surgery and current age of patients between the study and control groups Pouch configuration and outcomes Figure 1 Long-term adverse outcomes in patients with different pouch configurations. Stage of surgery was significantly different between the three groups, largely due to the nature of indication for colectomy and for the CI (p b 0.001). Final diagnosis of the pouch was also significantly different (Fig. 1). Patients with S or CI were less likely to have CARP than those with J pouches (15.6% in the S vs. 13.9% in the CI vs. 36.3% in the J groups; p = 0.002). CARP was more common in patients with J pouches than those with S pouches (13.0% vs. 0%; p = 0.013). Patients with S pouches or CI were more likely to have pouch surgery-related complications than those with J pouches (44.4% in S vs. 33.3% in CI vs. 9.3% in J; p b 0.001). There was no difference in pouch surgery-related complications between the patients with S pouches and those with CI (p = 0.31). The

6 1232 S. Mukewar et al. Table 3 Categorization of pouch-related mechanical and surgical complications. Factor S Continent ileostomy J p-value (N = 20) (N = 12) (N = 20) Inlet or distal pouch prolapse 4(20.0) 0 1(5.0) 0.12 Afferent limb syndrome 1(5.0) 0 1(5.0) 0.73 Inlet obstruction from intrinsic stricture 2(10.0) 2(16.7) 4(20.0) 0.67 Outlet obstruction from bowel angulation or long limb 7(35.0) 0 2(10.0) Outlet obstruction from anastomotic stricture 7(35.0) NA 3(15.0) Nipple valve dysfunction NA 5(41.7) NA NA Anastomotic sinus 0 NA 0 The tip of the J leak NA NA 1(5.0) NA Pouch sinus 2(10.0) a 0 a 12(60.0) b,c b0.001 Pouch bezoars or foreign bodies 0 1(8.3) Nipple valve stricture NA 1(8.3) NA NA Trapped ovarian syndrome 1(5.0) a Significantly different from J. b Significantly different from S. c Significantly different from continent ileostomy. median duration of pouch was significantly different between the three groups (15.0; IQR: 11.0, 21.0 years in the S vs. 20.5; IQR: 13.5, 28.5 years in the K vs. vs. 9.0; IQR: 5.0, 14.0 years in the J; p b 0.001). There were no differences in the use of anti-tnf biologics or immunomodulators after pouch surgery, number of visits to the Pouch Center and hospitalization and pouch failure for pouch-related disorders between the three groups Chronic antibiotic-refractory pouchitis Table 4 presents the multivariable analysis for development of CARP. After adjusting for smoking, patients with S pouches were 93% less likely to develop CARP than subjects with J pouches [S vs. J: likelihood ratio 0.07; 95% confidence interval: b0.001, 0.54; p b 0.001]. On the other hand there was no evidence to suggest a significant difference between the CI and J pouch groups (CI vs. J: odds ratio [OR) 0.68; 95% confidence interval: 0.17, 2.00, p = 0.40] Pouch-related mechanical and structural complications Table 3 presents a summary and Table 5 presents the multivariable analysis for pouch-related mechanical and structural complications. After adjusting for smoking, EIM and age at time of pouch surgery, patients with S pouches were 8 times more likely to have pouch-related mechanical and structural complications than subjects with J pouches (95% confidence interval: 3.7, 17.5; p b 0.001). Also, patients with CI were 5.6 times more likely to have pouch-related mechanical and structural complications than patients with J pouches (95% confidence interval: 2.4, 13.3; p b 0.001) Crohn's disease of the pouch Table 6 presents the multivariable analysis for the risk for the development of CD of the pouch. After adjusting for smoking, family history of CD, age at the time of pouch construction and pre-op diagnosis, there was no evidence to suggest that pouch type was associated with the occurrence of CD of the pouch Pouch failure There was no evidence of a significant difference in pouch failure between the three groups (Fig. 2). On multivariable analysis, pouch type was not associated with risk for pouch failure (Table 7) Cleveland Clinic pouches vs. non-cleveland Clinic pouches Table 4 Risk for chronic-antibiotic refractory pouchitis: multivariable logistic regression analysis. Factor Odds ratio p-value (95% confidence interval) Continent ileostomy 0.68 (0.17, 2.00) 0.40 vs. J pouch S vs. J pouch 0.07 (b0.001, 0.54) Smoking 2.4 (1.01, 5.60) There were no significant differences in type of pouches, pouch failure rate, CARP and CD of the pouch. However, the rate of pouch procedure-related complications was lower in the Cleveland Clinic pouches (Table 8). 4. Discussion This is the first study comparing the various long-term adverse outcomes of the different pouch designs in patients with IBD. The majority of studies comparing pouch designs 4,11 17 have either focused on functional outcomes or short-term

7 Pouch type and long-term outcomes 1233 Table 5 Risk for pouch surgery-related complications: multivariable logistic regression analysis. Factor Odds ratio (95% confidence interval) p-value Continent ileostomy 5.6 (2.4, 13.3) b0.001 vs. J pouch S vs. J pouch 8.0 (3.7, 17.5) b0.001 Extra-intestinal 0.38 (0.18, 0.81) manifestations Age at time of 0.86 (0.75, 1.00) pouch construction (5 year increase) Smoking 1.4 (0.53, 3.50) 0.51 complications. Our study showed that in the long run, patients with S pouch were 8 times or CI were 5.6 times more likely to have pouch-related mechanical/structural complications than those with J pouches. On the other hand, patients with S pouches were 93% less likely to develop pouchitis than patients with J pouches. The trade-off of mechanical/structural complications and pouchitis provides a new insight for patients, colorectal surgeons, and IBD specialists to select the appropriate pouch configuration after colectomy for UC. Pouch-related mechanical and structural complications such as anastomotic leaks, fistulae, sinuses, prolapse, strictures, afferent limb syndrome and efferent limb syndrome are seen in patients undergoing pouch construction. While few of these complications can be managed conservatively, most of these require surgical or endoscopic intervention. A recent meta-analysis 18 comparing J, S and W pouch showed no difference in pouch-related mechanical/structural complications (anastomotic leak, anastomotic stricture, pelvic sepsis and small bowel obstruction). However, the included studies had fewer patients (less than 300 in total), shorter follow-up and the lack of well-defined criteria for long-term pouchrelated disorders. For example, patients with pouch stricture secondary to CD of the pouch could have been mislabeled as a pouch-related mechanical/structural complication. In addition, the studies primarily discussed immediate post-operative Table 6 Crohn's disease of the pouch: multivariable logistic regression analysis. Factor Odds ratio (95% confidence interval) p-value Continent ileostomy 1.3 (0.54, 3.20) 0.56 vs. J pouch S vs. J pouch 1.4 (0.62, 3.00) 0.45 Family history of 5.0 (1.8, 13.8) Crohn's disease Age at time of 0.87 (0.76, 0.98) pouch construction (5 year increase) Smoking 0.88 (0.36, 2.20) 0.78 Pre-op diagnosis of ulcerative colitis 0.51 (0.21, 1.20) 0.14 surgical complications. Results from our current study showed a greater long-term risk of pouch-related mechanical/structural complications in patients with S pouches or CI than those with J pouches. We suspect that patients with S pouches or CI are prone to these complications owing to the anatomy of the pouch and the way in which the pouch is constructed. The construction of S or CI can be technically demanding. Complications such as afferent and efferent limb syndromes were more commonly seen in the patients with S pouches as a consequence of triplicated design and potential placement of a long efferent limb. On the other hand, the main pouch-related mechanical/structural complication in the CI was malfunction of the nipple valve. In contrast, J pouches were more prone to the development of anastomotic sinuses, probably due to the risk of anastomotic tension during its construction. Colorectal surgeons may find it difficult for the pouch body to reach the rectal stump due to the length of mesentery or a body habitus. This can cause anastomotic tension or compromise of perfusion of the distal pouch leading to anastomotic leak (acute) or sinus (chronic) and ischemic pouchitis. 19 In fact, ischemic pouchitis is likely a cause of CARP in some patients. 19 In contrast, the construction of S pouch maygiveanadditional2 3 cm length of the efferent limb to reach the anal transitional zone. Similarly, the reach is not a main issue during construction of CI. Twenty-three to 46% patients develop pouchitis years after pouch surgery. 20,21 Various risk factors have been associated with pouchitis genetic polymorphisms of interleukin-1- receptor antagonist and NOD2/CARD 25,26 non-carrier status of TNF allele, 22,27 extensive UC, 21,28,29 backwash Table 7 Factor Figure 2 Kaplan Meier curve: pouch survival. Pouch failure: multivariable Cox Regression analysis. Hazard ratio (95% confidence interval) p-value J vs. S pouch 1.4 (0.40, 5.0) 0.59 Smoking 1.4 (0.46, 4.3) 0.55

8 1234 S. Mukewar et al. ileitis, 29 pre-proctocolectomy thrombocytosis, 30,31 preoperative corticosteroid use, 31 the presence of EIM, especially PSC, 20,28,31 34 the presence of perinuclear antineutrophil cytoplasmic antibodies, being a non-smoker 7,31,38 and the regular use of NSAIDs. 7,28 However, none of these studies assessed pouch configuration as a risk factor for CARP. Meta-analysis by Lovegrove et al. 18 showed no difference between patients with J or S pouches for the risk for CARP. However, pouchitis was not stratified into acute pouchitis and CARP in this study. Clinically, acute pouchitis and CARP behave as different clinical entities, with latter being more like IBD. Our observations are contrary to these results. We evaluated various conventional risk factors, which have been known to be associated with chronic pouchitis amongst the two groups extensive UC, backwash ileitis, pre-operative corticosteroid use, the presence of EIM, PSC, history of smoking and NSAID use and could not identify any difference. There were no significant differences in the use of steroids, biologics or fulminant colitis amongst the groups. Therefore, we paid attention to non-conventional factors and one of them was pouch configuration, based on our clinical observation of a large number of patients in our Pouch Center. We found that J pouch configuration was independently associated with pouchitis. The obvious difference in the frequency of pouchitis between J vs. other S pouches or CI suggested that mechanical reasons could play a role in the pathogenesis of CARP. One of them could be a greater stretch of mesenteric vasculature during J pouch construction. CD of the pouch is one of the most dreaded adverse event following the ileal pouch surgery. It is estimated that 2.7% to 13% of pouch patients develop CD of the pouch. Smoking, 7 family history of CD, 9,39 ethnicity, 40,41 longer duration of pouch, 42 pre-operative diagnosis of IC 42 and seropositivity for anti-saccharomyces cerevisiae-iga antibody 39 are some of the risk factors associated with this condition. Although the majority of these studies did not include the pouch configuration as a study variable, 9,39,40,42 one study by our group assessed for the pouch configuration and failed to show it as a risk factor for CD of the pouch. 7 Studies comparing different pouch configurations, including the recent meta-analysis, did not include CD of the pouch as an outcome. This is likely because CD of the pouch is being increasingly recognized, and most of these studies were conducted in 1980s and 90s and had a shorter duration of follow-up of patients and a smaller sample size. Results from our study confirm that the pouch configuration was not a risk factor for CD of pouch. According to the results of a meta-analysis of nearly 9000 patients, pouch failure occurred in 8.5% of the patients. 43 CD of the pouch, chronic pouchitis and pelvic sepsis 44,47 are the main causes for pouch failure. A prospective study of 634 patients 48 noted a greater pouch failure rate in J and S pouches than W or K pouches. However, our observations are contrary to their study, as we did not identify any difference in pouch failure rate amongst the groups. The discrepancy may result from the sample size, duration of follow-up, and practice setting. The results of this study have several clinical implications. Our study shows that the risk for developing CARP and pouch-related mechanical/structural complications varies with pouch design, suggesting the role of surgical techniques and mechanical factors in their etiopathogenesis. The tradeoff for the construction of a S pouch or CI with a lower risk for CARP and a higher risk for pouch-related mechanical/surgical complications vs. J pouches with a higher risk for CARP and a lower risk for the mechanical/surgical complications should be carefully balanced. Our study has several limitations. Although the cohort had a range of healthy and diseased pouch conditions, there might have been a referral bias due to the nature of our practice. As pouchitis clinic is a major referral center, we see patients referred from the Department of Colorectal Surgery of the Cleveland Clinic as well as from outside hospitals and practices. A sizeable number of patients were referred to the pouchitis clinic for pouch related problems from elsewhere. Finally, CD of the pouch and pouch failure can occur many years after the construction of pouch. Despite the fact that this was a long-term study, patients with S pouches or CI had a longer follow-up than those with J pouches. Reliability in follow-up of pouch patients may have been better in the recent years, with new guidelines and recommendations. Thus, patients with J pouches might have had a better follow-up than those with S pouches or CI. In conclusion, this study shows that different pouch designs are associated the development of CARP and pouch-related mechanical/structural complications. The findings are of prognostic benefit and can help patients and clinicians choose a proper pouch design, balancing the trade-off of advantages and disadvantages. Table 8 pouches. Characteristic Comparison of pouch configurations and major adverse outcome between Cleveland Clinic and Non-Cleveland Clinic NonCleveland Clinic pouches (N = 129) Cleveland Clinic pouches (N = 167) Type of the pouch 0.49 J pouch 91 (70.5) 124 (74.3) S pouch 19 (14.7) 26 (15.6) Continent ileostomy 19 (14.7) 17 (10.2) Pouch failure 14 (10.9) 9 (5.4) Chronic antibiotic-refractory pouchitis 14 (10.9) 17 (10.2) 0.85 Crohn's disease of the pouch 24 (18.6) 36 (21.6) 0.53 Pouch procedure-related complications 32 (24.8) 20 (12.0) p-value

9 Pouch type and long-term outcomes 1235 Disclosure The authors declared no financial conflicts of interest. References 1. Kock NG. Intra-abdominal reservoir in patients with permanent ileostomy. Preliminary observations on a procedure resulting in fecal continence in five ileostomy patients. Arch Surg 1969; 99(2): Parks AG, Nicholls RJ. Proctocolectomy without ileostomy for ulcerative colitis. Br Med J 1978;2(6130): Utsunomiya J, Iwama T, Imajo M, Matsuo S, Sawai S, Yaegashi K, et al. Total colectomy, mucosal proctectomy, and ileoanal anastomosis. Dis Colon Rectum 1980;23(7): Nicholls RJ, Pezim ME. Restorative proctocolectomy with ileal reservoir for ulcerative colitis and familial adenomatous polyposis: a comparison of three reservoir designs. Br J Surg 1985; 72(6): Shen B, Remzi FH, Lavery IC, Lashner BA, Fazio VW. A proposed classification of ileal pouch disorders and associated complications after restorative proctocolectomy. Clin Gastroenterol Hepatol 2008;6(2): Wu XR, Kiran RP, Mukewar S, Remzi FH, Shen B. Diagnosis and management of pouch outlet obstruction caused by common anatomical problems after restorative proctocolectomy. J Crohns Colitis 2014;8(4): Shen B, Fazio VW, Remzi FH, Brzezinski A, Bennett AE, Lopez R, et al. Risk factors for diseases of ileal pouch-anal anastomosis after restorative proctocolectomy for ulcerative colitis. Clin Gastroenterol Hepatol 2006;4(1): Shen B, Achkar JP, Connor JT, Ormsby AH, Remzi FH, Bevins CL, et al. Modified pouchitis disease activity index: a simplified approach to the diagnosis of pouchitis. Dis Colon Rectum 2003;46(6): Shen B, Remzi FH, Hammel JP, Lashner BA, Bevins CL, Lavery IC, et al. Family history of Crohn's disease is associated with an increased risk for Crohn's disease of the pouch. Inflamm Bowel Dis 2009;15(2): Lian L, Fazio VW, Remzi FH, Shen B, Dietz D, Kiran RP. Outcomes for patients undergoing continent ileostomy after a failed ileal pouch-anal anastomosis. Dis Colon Rectum 2009;52(8): de Silva HJ, de Angelis CP, Soper N, Kettlewell MG, Mortensen NJ, Jewell DP. Clinical and functional outcome after restorative proctocolectomy. Br J Surg 1991;78(9): Neilly P, Neill ME, Hill GL. Restorative proctocolectomy with ileal pouch-anal anastomosis in 203 patients: the Auckland experience. Aust N Z J Surg 1999;69(1): Nasmyth DG, Williams NS, Johnston D. Comparison of the function of triplicated and duplicated pelvic ileal reservoirs after mucosal proctectomy and ileo-anal anastomosis for ulcerative colitis and adenomatous polyposis. Br J Surg 1986;73(5): OreslandT,FasthS,NordgrenS,HallgrenT,HultenL.Aprospective randomized comparison of two different pelvic pouch designs. Scand J Gastroenterol 1990;25(10): Romanos J, Samarasekera DN, Stebbing JF, Jewell DP, Kettlewell MG, Mortensen NJ. Outcome of 200 restorative proctocolectomy operations: the John Radcliffe Hospital experience. Br J Surg 1997;84(6): Schoetz Jr DJ, Coller JA, Veidenheimer MC. Ileoanal reservoir for ulcerative colitis and familial polyposis. Arch Surg 1986; 121(4): Svaninger G, Nordgren S, Oresland T, Hulten L. Incidence and characteristics of pouchitis in the Kock continent ileostomy and the pelvic pouch. Scand J Gastroenterol 1993;28(8): Lovegrove RE, Heriot AG, Constantinides V, Tilney HS, Darzi AW, Fazio VW, et al. Meta-analysis of short-term and longterm outcomes of J, W and S ileal reservoirs for restorative proctocolectomy. Colorectal Dis 2007;9(4): Shen B, Plesec TP, Remer E, Kiran P, Remzi FH, Lopez R, et al. Asymmetric endoscopic inflammation of the ileal pouch: a sign of ischemic pouchitis? Inflamm Bowel Dis 2010;16(5): Penna C, Dozois R, Tremaine W, Sandborn W, LaRusso N, Schleck C, et al. Pouchitis after ileal pouch-anal anastomosis for ulcerative colitis occurs with increased frequency in patients with associated primary sclerosing cholangitis. Gut 1996;38(2): Fazio VW, Ziv Y, Church JM, Oakley JR, Lavery IC, Milsom JW, et al. Ileal pouch-anal anastomoses complications and function in 1005 patients. Ann Surg 1995;222(2): Aisenberg J, Legnani PE, Nilubol N, Cobrin GM, Ellozy SH, Hegazi RA, et al. Are panca, ASCA, or cytokine gene polymorphisms associated with pouchitis? Long-term follow-up in 102 ulcerative colitis patients. Am J Gastroenterol 2004;99(3): Brett PM, Yasuda N, Yiannakou JY, Herbst F, Ellis HJ, Vaughan R, et al. Genetic and immunological markers in pouchitis. Eur J Gastroenterol Hepatol 1996;8(10): Carter MJ, Di Giovine FS, Cox A, Goodfellow P, Jones S, Shorthouse AJ, et al. The interleukin 1 receptor antagonist gene allele 2 as a predictor of pouchitis following colectomy and IPAA in ulcerative colitis. Gastroenterology 2001;12(4): Meier CB, Hegazi RA, Aisenberg J, Legnani PE, Nilubol N, Cobrin GM, et al. Innate immune receptor genetic polymorphisms in pouchitis: is CARD15 a susceptibility factor? Inflamm Bowel Dis 2005;11(11): Paye F, Penna C, Chiche L, Tiret E, Frileux P, Parc R. Pouchrelated fistula following restorative proctocolectomy. Br J Surg 1996;83(11): Sandborn WJ. Pouchitis following ileal pouch-anal anastomosis: definition, pathogenesis, and treatment. Gastroenterology 1994; 107(6): Achkar JP, Al-Haddad M, Lashner B, Remzi FH, Brzezinski A, Shen 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): Okon A, Dubinsky M, Vasiliauskas EA, Papadakis KA, Ippoliti A, Targan SR, et al. Elevated platelet count before ileal pouchanal anastomosis for ulcerative colitis is associated with the development of chronic pouchitis. Am Surg 2005;71(10): Fleshner P, Ippoliti A, Dubinsky M, Ognibene S, Vasiliauskas E, Chelly M, et al. A prospective multivariate analysis of clinical factors associated with pouchitis after ileal pouch-anal anastomosis. Clin Gastroenterol Hepatol 2007;5(8): Hata K, Watanabe T, Shinozaki M, Nagawa H. Patients with extraintestinal manifestations have a higher risk of developing pouchitis in ulcerative colitis: multivariate analysis. Scand J Gastroenterol 2003;38(10): 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): Shepherd NA, Hulten L, Tytgat GN, Nicholls RJ, Nasmyth DG, Hill MJ, et al. Pouchitis. Int J Colorectal Dis 1989;4(4): Fleshner PR, Vasiliauskas EA, Kam LY, Fleshner NE, Gaiennie J, Abreu-Martin MT, 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):671 7.

10 1236 S. Mukewar et al. 36. 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): Hui T, Landers C, Vasiliauskas E, Abreu M, Dubinsky M, Papadakis KA, et al. Serologic responses in indeterminate colitis patients before ileal pouch-anal anastomosis may determine those at risk for continuous pouch inflammation. Dis Colon Rectum 2005;48(6): Merrett MN, Mortensen N, Kettlewell M, Jewell DO. Smoking may prevent pouchitis in patients with restorative proctocolectomy for ulcerative colitis. Gut 1996;38(3): Melmed GY, Fleshner PR, Bardakcioglu O, Ippoliti A, Vasiliauskas EA, Papadakis KA, et al. Family history and serology predict Crohn's disease after ileal pouch-anal anastomosis for ulcerative colitis. Dis Colon Rectum 2008;51(1): Tyler AD, Milgrom R, Xu W, Stempak JM, Steinhart AH, McLeod RS, et al. Antimicrobial antibodies are associated with a Crohn's disease-like phenotype after ileal pouch-anal anastomosis. Clin Gastroenterol Hepatol 2012;10(5): Mukewar S, Wu X, Lopez R, Kiran RP, Remzi FH, Shen B. Ethnicity and the risk of development of Crohn's disease of the ileal pouch. J Crohns Colitis 2013;7(5):e Delaney CP, Remzi FH, Gramlich T, Dadvand B, Fazio VW. Equivalent function, quality of life and pouch survival rates after ileal pouch-anal anastomosis for indeterminate and ulcerative colitis. Ann Surg 2002;236(1): Hueting WE, Buskens E, van der Tweel I, Gooszen HG, van Laarhoven CJ. Results and complications after ileal pouch anal anastomosis: a meta-analysis of 43 observational studies comprising 9,317 patients. Dig Surg 2005;22(1 2): Prudhomme M, Dehni N, Dozois RR, Tiret E, Parc R. Causes and outcomes of pouch excision after restorative proctocolectomy. Br J Surg 2006;93(1): Winther KV, Jess T, Langholz E, Munkholm P, Binder V. Survival and cause-specific mortality in ulcerative colitis: follow-up of a population-based cohort in Copenhagen County. Gastroenterology 2003;125(6): Nicholls RJ. Review article: ulcerative colitis surgical indications and treatment. Aliment Pharmacol Ther 2002;16(Suppl 4): Sagap I, Remzi FH, Hammel JP, Fazio VW. Factors associated with failure in managing pelvic sepsis after ileal pouch-anal anastomosis (IPAA) a multivariate analysis. Surgery 2006;140(4): Tulchinsky H, Hawley PR, Nicholls J. Long-term failure after restorative proctocolectomy for ulcerative colitis. Ann Surg 2003; 238(2):