The Incidence and Definition of Crohn’s Disease of the Pouch: A Systematic Review and Meta-analysis

Abstract

Background

A subset of patients who undergo total proctocolectomy with ileal pouch–anal anastomosis (IPAA) creation for ulcerative colitis (UC) will later develop Crohn’s disease (CD) of the pouch, which has been associated with significant morbidity. We aimed to analyze the incidence of CD of the pouch and to review the existing diagnostic criteria utilized.

Methods

A systematic search performed through March 1, 2018, identified 12 studies that reported the incidence of CD of the pouch after IPAA for UC or indeterminate colitis (IC). We compiled all diagnostic criteria utilized in these studies and then performed a meta-analysis using random effects modeling to estimate the overall incidence of CD of the pouch in this population.

Results

Among 4843 patients with an IPAA for UC or IC, 10.3% of patients were ultimately diagnosed with CD of the pouch (95% confidence interval [CI], 6.1%–15.4%). The most commonly reported diagnostic criteria were (1) presence of fistula/fistulae, (2) stricture involving the pouch or prepouch ileum, and (3) presence of prepouch ileitis. In a secondary analysis, excluding those studies that included patients with a preoperative diagnosis of IC, the incidence of CD of the pouch was 12.4% (95% CI, 9.0%–16.1%).

Conclusions

The estimated incidence of 10.3% will assist gastroenterologists and surgeons in preoperative counseling regarding the potential to develop CD of the pouch. There is an unmet need for common diagnostic criteria for a more standardized approach to the diagnosis of CD of the pouch.

Crohn’s disease of the pouch occurs in approximately 10% of patients after restorative proctocolectomy with ileal pouch-anal anastomosis for ulcerative colitis or indeterminate colitis.

INTRODUCTION

For patients with medically refractory ulcerative colitis (UC) or for those patients who develop UC-related dysplasia, total proctocolectomy with ileal pouch–anal anastomosis (IPAA) has become the standard approach to surgical therapy.^{1, 2} Despite advances in the medical treatment of UC, approximately 20%–35% of patients with UC will ultimately require colectomy.^3–5 Although the IPAA maintains intestinal continuity and is associated with improvements in quality of life,^{6, 7} concerns about the potential for long-term complications remain.⁸

Although a preoperative diagnosis of Crohn’s disease (CD) has been a relative contraindication to IPAA,⁹ the development of CD of the pouch after an IPAA for UC or indeterminate colitis (IC) has been problematic. Up to 15% of patients will experience pouch failure,^10–12 with CD of the pouch representing one of the most common reasons for pouch failure.¹² Significant heterogeneity exists in the definition of CD of the pouch and the terminology used to describe an inflammatory Crohn’s-like presentation in a patient with an IPAA.⁸ Additionally, the reported incidence of CD of the pouch varies widely,¹³ further underscoring the lack of a standard definition for this condition.

Understanding the true incidence of CD of the pouch and standardizing the diagnostic criteria for this condition remain critical goals in improving the care of patients undergoing IPAA for the treatment of UC or IC. Patients should be counseled regarding the potential risks of developing CD of the pouch during the pre- and perioperative periods, given that many view proctocolectomy with IPAA as a curative surgery for UC.¹⁴ Additionally, those patients who are ultimately diagnosed with CD of the pouch should have an accurate and standardized diagnosis. A more standard approach to diagnosis might aid efforts to improve and expand the limited therapeutic options for this population.^{15, 16}

Through a systematic review and meta-analysis of all published literature, we assessed the reported incidence of CD of the pouch after IPAA for UC or IC. Additionally, we aimed to better understand the published diagnostic criteria for CD of the pouch and to present a summary collection of diagnostic criteria for CD of the pouch.

METHODS

Literature Search

This study was conducted according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines.¹⁷ An electronic literature search was conducted on PubMed, EBSCO, and Web of Science using the following combination of key words: “Crohn’s disease,” “pouch,” “surgery,” “ileal pouch anal anastomosis outcomes,” and “ileal pouch anal anastomosis complications” (see the Supplementary Data for search strings by search engine). Given our interest in any time trends in the incidence of CD of the pouch, no start date limit was used, and the search was performed until March 1, 2018. Searches of bibliographies of original articles and relevant reviews were subsequently performed to identify further studies for inclusion. The search had no language restrictions. After removal of duplicate references, initial screening of article titles and abstracts was undertaken by 2 independent researchers (H.J. and B.K.). Any discrepancies were settled by a third reviewer (E.L.B.). Potentially relevant articles were obtained in full text and reviewed independently. Criteria to determine eligibility for inclusion were determined a priori.

Selection Criteria

Original research studies of longitudinal follow-up of adult patients with a preoperative diagnosis of UC or IC who underwent IPAA creation and listed a definition for CD of the pouch were included. Studies that were reviews, that included only pediatric patients or IC patients, or that did not list a definition for CD of the pouch were excluded. Two independent researchers read through the full texts to determine eligibility for inclusion (B.K. and E.L.B.), and a third reviewer settled any discrepancies (H.H.H.). Two study investigators (B.K. and E.L.B.) independently assessed the quality of included studies using the Cochrane Risk of Bias Tool.¹⁸

Data Extraction

Eligible citations were divided into 3 groups for extraction by reviewers (B.K., E.L.B., and H.H.H.). Extracted data included the title, first author, last author, year of publication, journal of publication, center of study, or data source. Additional data extracted from available studies included the total number of UC patients, total number of UC and IC patients, total number of patients with an eventual diagnosis of CD of the pouch, the mean or median time to diagnosis of CD of the pouch, and the definition of CD of the pouch utilized in the study. Definitions for CD of the pouch were compiled from each article. The salient aspects of the definitions were compiled and summarized in the systematic review. When multiple studies utilized the same study population or data source, the study with the most comprehensive data was included in the systematic review and meta-analysis.

Statistical Analysis

In performing the meta-analysis to determine the incidence of CD of the pouch after total proctocolectomy with IPAA, the raw number of patients developing CD of the pouch was identified as the metameter in available studies. We conducted a random effects meta-analysis of proportions using the Freeman-Tukey transformation to obtain a pooled incidence of CD of the pouch accounting for heterogeneity and variance. We assessed for heterogeneity using a Q statistic and I² analysis. All analyses were conducted using Stata 14.0 (College Station, TX, USA).

We made the a priori decision to analyze by year of publication to assess for any differences in reported incidence of CD of the pouch by time period. Additionally, we performed a metaregression to assess for any significant differences in the incidence of CD of the pouch over time. We also performed a subanalysis of studies that only included those patients with UC. Due to the small number of eligible studies assessing incidence of CD of the pouch, a reliable assessment of publication bias could not be performed.

RESULTS

We identified 181 titles. After a predefined selection process (Fig. 1), 27 full texts were selected for potential inclusion (Supplementary Table 1). In the review of these 27 full texts, the realization was made that multiple manuscripts were presented from the same medical center or central database, likely leading to the inclusion of the same pouch patient population or subpopulations of the overall pouch cohort of the center in multiple studies. Selecting 1 paper with the most inclusive patient sample from each center or data source resulted in a final list of 12 manuscripts included in the systemic review and meta-analysis.

FIGURE 1.

Flow diagram summarizing identification and selection of eligible studies.

Standardized Incidence of Crohn’s Disease of the Pouch

Of the 12 studies included in the systematic review, 11 studies included an eligible population of patients with UC/IC undergoing an IPAA and longitudinal analysis of patients who ultimately developed CD of the pouch.^{14, 19–28} One study was excluded given that the total population of patients with UC/IC undergoing an IPAA was not reported.¹³ In this cohort of 4843 patients, 330 patients were reported to have developed CD of the pouch. The standardized incidence for development of CD of the pouch among the entire population was 10.3% (95% confidence interval [CI], 6.1%–15.4%) (Fig. 2). Among the 11 included studies, the incidence of CD of the pouch ranged from 3.1%²² to 21.3%.²⁸

FIGURE 2.

The incidence of Crohn’s disease of the pouch.

As hypothesized, there was significant heterogeneity present among the studies, with an I² value of 94.8 (Q = 190.9; df = 10; P < 0.001). On analysis with the 1-study-removed technique, the results were similar. In a subgroup analysis separating studies by year of publication, a higher standardized incidence of CD of the pouch was reported in those studies published between 2016 and 2018 (13.1%; 95% CI, 5.9%–22.6%) than in those studies published between 1997 and 2012 (8.7%; 95% CI, 4.4%–14.2%) (Supplementary Fig. 1A and B). Using metaregression, there was no significant trend noted when analyzing the incidence of CD of the pouch by year (P = 0.209).

In a separate secondary analysis, we excluded the 3 studies that included patients with IC. In the remaining 9 studies that included only patients with a preoperative diagnosis of UC, the incidence was 12.4% (95% CI, 9.0%–16.1%) (Supplementary Fig. 2).

Time to diagnosis of CD of the pouch was reported in 4 studies: mean time to diagnosis was 17 months and 20 months in 2 studies,^{20, 23} and median time to diagnosis was 22 months and 36 months in 2 studies.^{22, 26} A fifth study reported a range of time to diagnosis between 14 months and 8 years after IPAA creation.¹⁹

Defining Crohn’s Disease of the Pouch

In our systematic review of the definition of CD of the pouch, we included 12 original research manuscripts that were published between 1997 and 2018.^{13, 14, 19–28} Details of the included studies are shown in Table 1. These studies included a total of 4843 patients with UC or IC who underwent a proctocolectomy with IPAA.

TABLE 1.

Full-Text Manuscripts Included in Systematic Review and Meta-analysis of Crohn’s Disease of the Pouch Incidence and Definition

Study	Center	Study Period	UC/IC Patients	CD Pouch Cases	CD Pouch Incidence
Goldstein et al.19 1997	William Beaumont Hospital	1981–1995	74	8	10.8%
Peyregne et al.20 2000	Centre Hospital Lyon–Sud	1985–1997	43	4	9.3%
Rossi et al.21 2002	Rush–Presbyterian–St. Luke’s Medical Center	1989–2000	68	4	5.9%
Melton et al.22 2010a	Cleveland Clinic	1983–2007	2814	87	3.1%
Haveran et al.23 2011	Penn State–Hershey Medical Center	1990–2009	382	32	8.3%
Coukos et al.24 2012a	Boston Medical Center	2006–2010	142	21	14.8%
Tyler et al.25 2012	Mount Sinai Hospital, Toronto	N/A–2007	399	50	12.5%
Zaghiyan et al.26 2016	Cedars-Sinai	1997–2007	237	40	16.9%
Diederen et al.27 2017a	Emma Children’s Hospital, the Netherlands	2000–2015	303	14	4.6%
Lightner et al.13 2017b	Mayo Clinic	1982–2016	Not reported	35	N/A
Yanai et al.28 2017	Tel Aviv Sourasky Medical Center	1981–2013	253	54	21.3%
Shamah et al.14 2018	Lenox Hill	1960–2015	128	16	12.5%

^aStudy included patients with both ulcerative colitis and indeterminate colitis.

^bStudy included in analysis the definition of Crohn’s disease of the pouch but excluded incidence analyses given that the total population of patients with UC/IC undergoing an IPAA was not reported.

All studies defined CD of the pouch based on features that developed at least 3 months after ileostomy take-down to avoid confusion with postsurgical complications (Table 2). All 12 manuscripts used the presence of fistulae to define Crohn’s disease of the pouch. Four of the 12 manuscripts did not detail locations of fistulae.^{21, 22, 24, 28} When the location of fistulae was delineated, perianal fistulae were the most commonly reported (7 studies),^{13, 14, 23, 25–27} followed by pouch^{19, 26, 28} and vaginal fistulae^{13, 20, 23} (3 studies each). Other locations of fistulae included anastomotic (31 months after surgery),²⁰ anovulvar,²⁰ and small bowel fistulae¹⁴ (1 study each).

TABLE 2.

The Definitions of Crohn’s Disease of the Pouch Utilized by Each Study Included in the Systematic Review and Meta-analysis

First Author	Year	Definition
Goldstein19	1997	“Crohn’s-like complications include…anal fissures, perianal abscesses, marked mucosal-based enteritis with transmural extension, strictures, fistulas of the small bowel close to the proximal end of the pouch, pouch fistulas that involve small bowel or other internal organs, or blind-ended pouch fistulas that extend for long distances within soft tissues or into skeletal muscle, usually into the pelvic soft tissues posterior to the pouch”
Peyregne20	2000	“Discovery of pouch and afferent limb ulcerations during endoscopy performed for clinical symptoms doubtful of pouchitis…typical lesions of CD, one each of pouch-vaginal fistula, anastomotic fistula, anoperineal disease, ileal stricture located 10 cm above the pouch (treated with stricturoplasty) and anovulvar fistula”
Rossi21	2002	“Endoscopic, radiographic or histologic evidence: CD in 2 patients with fistulas and 2 patients with presacral abscess…. Nine had severe stenosis and 4 additional patients had abscesses”
Melton22	2010
Haveran23	2011	“Fistulas to the pouch and perineum or small bowel to the vagina; stricturing: proximal small bowel or pouch inlet stricture with associated inflammation; severe pouchitis: ≥4 episodes a year for 2 consecutive years or the need for continuous antibiotics”
Coukos24	2012
Tyler25	2012	“(1) Development of a perianal fistula >1 year after ileostomy closure; (2) a stricture proximal to the pouch that was not related to a surgical complication; (3) evidence of endoscopic inflammation (ulceration, erythema, friability) in the afferent limb/prepouch ileum or more proximal small intestine”
Zaghiyan26	2016
Diederen27	2017	“Non-surgery related perianal fistulas, granulomas on histology, inflammation and/or ulceration in the afferent limb or small intestine on endoscopy, in the absence of NSAID use”
Lightner13	2017	“Debilitating inflammation not responsive to antibiotics, and/or inflammation or stricturing in the afferent limb or other parts of the small bowel, and/or fistulizing disease to the perineum or the vagina”
Yanai28	2017
Shamah14	2018	“Debilitating inflammation not responsive to antibiotics, and/or inflammation or stricturing in the afferent limb or other parts of the small bowel, and/or fistulizing disease to the perineum, the small bowel, or the pouch itself”

Stricture was the second most commonly utilized feature to define CD of the pouch, with 10 of 12 studies reporting stricture as part of their definition.^{13, 14, 19–25, 28} Afferent loop and small bowel strictures were the most commonly mentioned (7 studies).^{13, 14, 19, 20, 22–25} The presence of pouch strictures was utilized in the diagnosis of CD in 3 studies,^{22, 23, 28} and 1 study mentioned anal strictures.¹⁹ Prepouch ileitis was mentioned in 9 of 12 studies.^{13, 14, 19, 20, 22, 25–28} The presence of pouchitis was included in the definition of CD of the pouch in 4 studies.^{13, 14, 19, 23} Finally, the presence of granulomas on histology was required in 2 studies as a defining feature of CD of the pouch.^{24, 27} The relative frequencies of these included features within the definitions of CD of the pouch are depicted in Figure 3.

FIGURE 3.

Diagnostic criteria used to identify patients with Crohn’s disease of the pouch.

DISCUSSION

Given the continued need for total proctocolectomy with IPAA as a treatment for patients with medically refractory UC and UC-related dysplasia, understanding outcomes after surgery is critical, particularly the long-term risk for developing CD of the pouch. In performing this systematic review and meta-analysis, we made multiple key observations regarding CD of the pouch in this population. Despite a preoperative diagnosis of UC or IC (in a minority of cases), approximately 10% of patients undergoing proctocolectomy with IPAA will develop CD of the pouch. We also confirmed that significant heterogeneity exists in the definition of CD of the pouch across studies. Establishing the incidence and standardizing a definition of CD of the pouch are needed to study this disease, to discuss the diagnosis with patients, and ultimately to improve our approach to therapy choices.²⁹

Establishing the incidence of CD of the pouch is important for both preoperative counseling and monitoring patients after an IPAA for UC. Although significant variability in the incidence of CD of the pouch has been reported, the overall incidence of 10.3% observed in this meta-analysis is within the estimated cumulative frequencies used when counseling patients and cited in prior review articles.^{8, 9, 30} This estimated incidence represents the first standardized incidence of CD of the pouch and was calculated using data from more than 4800 adult patients with a confirmed diagnosis of UC or IC undergoing IPAA, representing the largest reported patient cohort to date.

Based on existing literature presented in this systematic review, several unifying features were identified that may be helpful in the future creation of a more standardized definition for patients with CD of the pouch. The presence of a fistula or multiple fistulae arising from the pouch or afferent limb was consistently identified as a feature of CD of the pouch across all included studies. The timing of development of a fistula may also be critical, with fistulae developing more than 1 year after IPAA surgery being more likely to represent CD of the pouch and not a postoperative complication. Additionally, the development of a stricture or stenosis in the pouch body, pouch inlet, or afferent limb more than 1 year after surgery appears to be almost uniformly concerning for an underlying diagnosis of CD of the pouch. In almost all studies, a stricture at the ileorectal anastomosis was not considered to be associated with CD of the pouch. Prepouch ileitis has been associated with several inflammatory conditions of the pouch, including extensive pouchitis,³¹ primary sclerosing cholangitis,³² other immune-mediated pouch disorders,³³ and the use of nonsteroidal anti-inflammatory drugs.³⁴ In addition, some have argued that prepouch ileitis may be a distinct entity.³⁴ It remains important to both consider and rule out these other pouch-related inflammatory conditions when utilizing inflammation of the afferent limb to diagnose CD of the pouch.

The symptoms often reported by patients with CD of the pouch are not specific and may be seen in other inflammatory pouch-related conditions, including pouchitis, cuffitis, and irritable pouch syndrome.³⁰ However, in a patient suspected to have CD of the pouch, endoscopic evaluation is often the firstline method of diagnostic evaluation. In all 12 of the included definitions of CD of the pouch, suggestive features identified on pouchoscopy represented a critical method of diagnosis. Only 2 of the included studies required histologic abnormalities for a diagnosis, further indicating the importance placed on the appearance of the mucosa (including the anatomy involved and the complications identified such as strictures and fistulae). In prior studies of patients with an IPAA performed for CD, only 10%–12% of patients demonstrated granulomas on pathology.^{30, 35} Additionally, in a single-center study of patients undergoing pouch excision for refractory symptoms, only 20% demonstrated pathologic features consistent with a diagnosis of CD on the gross surgical specimen after excision.¹³

Given our interest in identifying the incidence of CD of the pouch and the diagnostic approaches utilized, we did not specifically examine the risk factors for developing CD of the pouch after total proctocolectomy with IPAA in this study. However, in prior studies, longer duration of IPAA³⁶ and smoking^{25, 36} have been identified as potential risk factors for the development of CD of the pouch. The duration of IPAA may be important with regards to our study findings, as a numerically higher incidence of CD of the pouch was demonstrated in the studies published between 2016 and 2018. This may indicate a longer experience with IPAA or a lead-time bias given the inability to account for onset of symptoms or to definitively state when diagnosis of CD of the pouch occurs. Alternatively, this may also represent increased awareness of this pouch-related condition. However, a minimum number of studies reported the duration of IPAA before diagnosis of CD of the pouch, and thus these findings could not be further analyzed. Additionally, due to the heterogeneity of the data, we did not attempt to further classify CD of the pouch by phenotypes, an approach that has been utilized in prior studies.^{35, 37}

Our study has multiple strengths, including the large available sample size created by pooling the experience of 12 medical centers. However, as with many meta-analyses, our study has important limitations to acknowledge. There was significant heterogeneity present among the included studies, and this heterogeneity could not be entirely explained by the long time period of included data. Although CD of the pouch is responsible for considerable morbidity among patients undergoing total proctocolectomy with IPAA as a treatment for UC, this remains a relatively small population. Significant heterogeneity exists in the sample sizes of included studies based on the volume of patients undergoing an IPAA or the follow-up time included in a study. Additionally, the lack of a standard diagnosis for CD of the pouch likely contributes significantly to the overall heterogeneity among studies. We excluded patients with a preoperative diagnosis of CD; however, a minority of patients with a preoperative diagnosis of IC were included. It was necessary to include these patients within the study based on the characteristics of individual included studies. This may introduce some bias, but if a differential effect were created by including these patients, we would expect that the incidence of CD of the pouch would potentially increase. Furthermore, in our planned secondary analysis excluding those studies in which patients with a preoperative diagnosis of IC were excluded, the incidence of CD of the pouch was higher than in the entire study population, which may be due to a smaller sample size because of the exclusion of the study by Melton et al.²² Although the heterogeneity present in our analyses is a limitation, this heterogeneity is reflective of the larger issues involved in diagnosing CD of the pouch and understanding the clinical course of patients undergoing a total proctocolectomy with IPAA for UC or IC.

In conclusion, in a systematic review and meta-analysis of more than 4800 patients with a confirmed diagnosis of UC or IC undergoing an IPAA, the incidence of CD of the pouch was 10.3%. Although significant heterogeneity exists in the definitions used to diagnose patients with CD of the pouch in the published literature, several unifying features of CD of the pouch were suggested that may be informative to the future creation of a standard definition for CD of the pouch. Among these features, the presence of a fistulae, stricture, or inflammation above the level of the pouch (prepouch ileitis) seems to be most consistent with the current definition of CD of the pouch and may form the foundation for a uniform set of diagnostic criteria in future studies.

Conflicts of interest: E.L.B., B.K., H.R.J., and H.H.H. report no relevant disclosures.

Supported by: This work was supported by the Nickolas Bunn Boddie, Sr., and Lucy Mayo Boddie Foundation (to E.L.B.), the Crohn’s and Colitis Foundation (grant number 567497 to E.L.B. and grant number 568735 to B.K.), and the National Institutes of Health (P30DK034987).