The pouch behaving badly: management of morbidity after ileal pouch–anal anastomosis

Ileal pouch–anal anastomosis (IPAA), or a ‘pouch’, allows restoration of intestinal continuity after proctocolectomy for ulcerative colitis or familial adenomatous polyposis. Most patients have a good long‐term outcome after IPAA, but in a significant proportion the functional outcome and quality of life are unsatisfactory. We term this outcome ‘the pouch behaving badly’. Managing this, especially one is when unfamiliar with the possible underlying pathologies, is a challenge for both patient and clinician. We aim to outline the clinical approach to the pouch behaving badly, highlighting key aspects of investigation and management.

is performed and the IPAA constructed with a diverting loop ileostomy. The second stage is where the loop ileostomy is closed 6-8 weeks later, restoring intestinal continuity. A modified twostage procedure is where a subtotal colectomy is performed first.
A completion proctectomy and IPAA is performed without a loop ileostomy at a second stage. In a three-stage procedure, a subtotal colectomy is performed first. The second stage is completion proctectomy and formation of the IPAA with a diverting loop ileostomy.
The third stage is closure of the loop ileostomy. A minority of patients will have their proctocolectomy and IPAA formed without an ileostomy as a one-stage procedure.
Most patients will have a good outcome after IPAA, but a significant proportion of patients have a poor quality of life [6]. After creation of a pouch it can take time for adaptation to achieve optimum function [7]. Approximately one in 10 pouches will 'fail' at 10 years, requiring excision or permanent diversion [8,9]. There is a further group of patients who continue to experience a poor quality of life due to a suboptimally functioning pouch -what we term 'the pouch behaving badly'. This can impinge on a patient's psychological, emotional and physical well-being and lead to pouch failure.
Patients present with a variety of symptoms including urgency, high defaecatory frequency, bleeding, abdominal pain, weight loss, watery stool, difficult evacuation, seepage, continence issues and fatigue [10,11]. The possible underlying pathologies have overlapping symptomatology and it can be challenging for both gastroenterologists and surgeons to investigate and manage these patients.
These patients often require long-term medication to control symptoms of a badly behaving pouch, such as analgesia, rotating antibiotics and constipating drugs. Alleviating patients' symptom burden and improving their quality of life is the therapeutic goal, but this can be a difficult place to navigate to. This review will cover the risk factors for the pouch behaving badly, an approach to management and then an overview of the possible underlying pathology.

RIS K FAC TOR S FOR THE P OUCH B EHAVING BADLY
The risk factors for the pouch behaving badly are the same as those which lead to eventual failure of the pouch [10]. Pouch failure is defined as the need for construction of a permanent stoma, with or without excision of the pouch or the need for an abdominoperineal pouch revision [12]. The commonest cause of pouch failure is pelvic sepsis [13][14][15]. Other causes include poor function (of unidentified cause) and chronic pouchitis, uncontrolled by medical therapies [15]. Fistulas, mechanical causes [14] and Crohn's disease (CD) of the pouch may also be significant contributors to cases of pouch failure [16]. Nondiversion at initial surgery, low hospital volumes and female gender have been associated with pouch failure in a Danish national cohort [9].
Primary sclerosing cholangitis (PSC) is associated with lower quality of life after IPAA, with poorer function and increased rates of pouchitis [17]. PSC also increases mortality and the risk of postoperative sepsis after IPAA [18].

Clinical assessment
The patient with a badly behaving pouch may present with multiple symptoms [10] and therefore a clear picture of the symptomatology should be constructed. Key questions to ask in the history are outlined in Table 1 A history of early pouch-related septic complications should be noted as well as pouch function following closure of the diverting loop ileostomy and restoration of bowel continuity. Preoperative corticosteroid use should be noted, as this is associated with pouch-related septic complications [19]. Other aspects of the history, including medication, psychological morbidity, smoking status and travel, should be covered.
Assessing the patient's quality of life in an objective and consistent way is an important part of the clinical assessment. It informs the patient and clinician of the overall impact of the disease and also allows treatment efficacy to be determined along the patient's journey. Tools such as the Pouch Dysfunction Score [20] and the Oresland score [21] are specific to quality of life after IPAA.
Examination should focus on the abdomen and perianal area and include digital examination of the anus and pouch-anal anastomosis.
Often the perianal skin is sore in patients with poor pouch function due to seepage, frequency of stools and cleaning practices. Examination under anaesthesia may allow more detailed assessment, particularly if the patient is in pain. A note should be made of the sphincter integrity and tone. The pouch-anal anastomosis should be assessed with regard to patency, integrity and height (and hence the length of the rectal cuff). The presence of sepsis, fistula or induration should be noted.

Investigations
Investigations should be guided by the clinical assessment of the patient. An overview of relevant investigations is given in Table 2. We find it useful to perform some initial blood tests to rule out pathology that may have been overlooked in managing a difficult pouch.
Patients routinely have a full blood count, haematinics, biochemistry, liver function tests, coeliac serology, thyroid function tests and examination of inflammatory markers. Faecal calprotectin levels are a clinically useful marker of inflammatory activity [22]. A stool culture including Clostridium difficile toxin is recommended. Pouchoscopy is a useful investigation which can identify a number of underlying pouch pathologies. It will identify pouchitis, prepouch ileitis, cuffitis and strictures, and allow biopsy to assess for cytomegalovirus (CMV) and to identify granulomas pathognomic of CD.
Pelvic MRI is the investigation of choice to assess the tissues around the pouch. It will identify chronic pelvic sepsis which can masquerade as primary idiopathic pouchitis or as CD of the pouch [23,24].
MR or CT enterography allows assessment of the remaining small bowel for CD if this is suspected and identifies areas of stricturing.
Endoanal ultrasound allows assessment of the anal sphincter muscles and manometry provides objective testing of function. This is helpful in those patients with frequent small-volume incontinence after a pouch. It should be considered where damage to the anal sphincter may have occurred, for example after vaginal delivery or during the IPAA procedure.
A defaecating pouchogram shows the size and anatomy of the pouch, especially small-volume pouches (that cause frequency), as well as documenting pouch emptying. This may reveal functional outflow obstruction such as anismus (which causes difficult evacuation and tenesmus) or prolapse. However, what constitutes a 'normal' defaecating pouchogram after a pouch is not known. A nondefaecating pouchogram will also show the size and shape of the pouch as well as leaks.

Diagnosis
Many patients have more than one cause of poor pouch function; conversely up to 20% of patients, despite assessment and investigation, will have no cause found [10].
Diagnostic clinical algorithms guide management. There is a proportion of patients who after repeated investigations and assessments do not have a clear diagnosis but live with significant morbidity [10]. Both the clinician and patient need to recognize when this point is reached, and the focus must change from further futile investigations to one of symptomatic relief and support for the patient. Offering the patient with a 'badly behaving pouch' a carefully constructed end ileostomy may lead to an improvement in quality of life. Giving patients this option early in the overall time course may optimize quality of life and reduce morbidity.

C AUS E S OF THE P OUCH B EHAVING BADLY
Here we give an overview of the pathology that can cause a pouch to behave badly. We divide the pathology into surgical, inflammatory, mechanical, functional and dysplastic causes (see Table 3).

Anastomotic leak and pelvic sepsis
Pelvic sepsis occurs when the anastomosis or a suture or staple line separates in the postoperative period. An identifiable site of leakage or contained leakage may not be found in a minority of cases [10]. Pelvic sepsis is responsible for up 50% of cases of pouch failure [15].

Investigation Rationale
Thyroid function, coeliac serology, haematinics, full blood count, routine biochemistry, liver function tests and inflammatory markers

TA B L E 2 Investigations for the badly behaving pouch and their rationale
The diagnosis of a leak may be made early in the postoperative period and present in the classic manner: abdominopelvic pain, fever, tachycardia and generalized peritonitis. However, patients may have mild or no symptoms (particularly as many have a diverting loop ileostomy) and this may delay diagnosis [25]. Some leaks are diagnosed at routine pouchogram scheduled before loop ileostomy closure, having displayed no symptoms [26].
A proportion of patients will manage to live with chronic pelvic sepsis. They will present later in the postoperative period, even

Fistulation
The development of a pouch fistula is a morbid complication associated with a high rate of pouch failure [29,30]. Fistulas may occur between the pouch and perianal region or the pouch and vagina, or may be complex and connect other spaces (such as the bladder or abdominal wall).
There is a wide range of time to presentation of fistulas in the literature, with the average time from IPAA to fistula presentation reported as less than 1 year [31], 2 years [32] and 6.9 years [33].
The incidence of pouch fistula was reported as 4.7% in the 2017 Ileoanal Pouch Registry Report [34]. It is important to identify the aetiology of the fistulizing process as this will affect the principles of management [35].
There is controversy in the aetiological classification of pouch fistula. We have proposed a rational classification based on aetiology [35] (see Figure 2). Previously, these fistulas were commonly thought to arise from an anastomotic leak or CD. A number of features have been suggested to distinguish fistulas arising from CD and anastomotic leak. These features tend to perpetuate the belief that fistulas arising soon after pouch creation or following reversal of ileostomy are secondary to an anastomotic leak and those that present later represent CD (Table 4). This has led to an overdiagnosis of pouch fistula due to CD, with the implication that many patients are treated with biologicals without substantial evidence. Lightner found that 90% had evidence of postoperative septic complications [24]. Aetiological classification of these fistulas should take into account that fistulas may arise from pelvic sepsis years after pouch creation. Most of the pouch fistulas arising from an anastomotic leak (Type 1) originate from the pouch-anal anastomosis, the tip of the J-pouch or the longitudinal staple line. There will be evidence of sepsis on MRI fistula such as peri-pouch sepsis, pelvic or supralevator collection [35].  [24,36]. There may be other sequelae of CD, including mouth ulcers. In the absence of features definitive for CD, a fistula may occur in the presence of IBD related to primary idiopathic pouchitis or cuffitis [37].
Type 3 cryptoglandular fistulas may present at any time after pouch creation and will arise with an internal opening at the dentate line [36,38]. Type 4 fistulas arising from a healthy pouch many years following pouch creation should raise suspicions of a fistulizing cancer of the pouch body or anal transition zone, especially in the presence of a pouch created for FAP or in the presence of a UC pouch with historical evidence of dysplasia. Expedient investigation and diagnosis may prevent the requirement for morbid exenterative surgery in these patients.
Essentially, pouch fistulas mainly arise from an anastomotic leak, cryptoglandular disease, malignancy and rarely IBD.
The management of fistulas should be tailored to the aeti-

Primary idiopathic pouchitis
Primary idiopathic pouchitis is a nonspecific inflammatory condition of the ileal pouch with a poorly understood pathophysiology. It is characterized by symptoms of frequency, urgency, incontinence, seepage, abdominal cramping and pelvic pain [39].
Patients may present with extraintestinal manifestations such as arthralgia, and this can aid diagnosis of primary idiopathic pouchitis [40]. There may be an association between PSC and pouchitis [40]. Other points to note in the history include if the patient is a nonsmoker [41], use of nonsteroidal anti-inflammatory drugs (NSAIDs) [41,42] and previous extensive colonic disease [42].
Pouchitis is seen less often after IPAA for FAP than for UC [43,44].
Many patients will have single episode of pouchitis, but between 10% and 15% of patients with acute pouchitis will develop chronic pouchitis. This is a poorly defined diagnosis but the most common definition is where the symptoms of pouchitis persist after 4 weeks of treatment [46]. There are further recognized problems with the diagnosis of pouchitis. Firstly, the patient's symptoms, flexible pouchoscopy findings and the histology findings do not always correlate. Additionally, how pouchoscopy is performed and reported is not consistent [47].
The mainstay of treatment for acute primary idiopathic pouchitis is antibiotics -most commonly ciprofloxacin or metronidazole [48,49]. Chronic primary idiopathic pouchitis is treated with a combination of antibiotics. Patients may require long-term antibiotics that need rotating. The therapeutic benefit of long-term antibiotics must be weighed against the risks of antibiotic resistance and side effects [50]. Oral budesonide and oral beclomethasone have been shown to be effective alternatives to antibiotics in chronic pouchitis [51,52]. They are unsuitable for long-term use due to side effects [53]. Other therapies such as biologicals should be considered. A systematic review of the management of chronic refractory pouchitis found a pooled remission rate of 53% with biological therapy [46]. Importantly, secondary causes of pouchitis should also be considered: infections (such as Clostridium difficile, Campylobacter, Salmonella, Candida and cytomegalovirus [54]), pelvic sepsis, CD (with the above caveats noted), faecal stasis, ischaemia or drugs (particularly NSAIDs) [54].

Prepouch ileitis
Prepouch ileitis affects 6% of patients with IPAA [55] and usually occurs with primary idiopathic pouchitis [55][56][57]. It presents with similar symptoms to pouchitis. Diagnosis is with endoscopy of the prepouch ileum or small bowel imaging, and treatment principles are similar to pouchitis. Prepouch ileitis should be differentiated from CD. There is also evidence that prepouch ileitis does not predict CD of the pouch at long-term follow-up [57].

Cuffitis
In forming a pouch, a small amount of residual rectum (the 'cuff') is retained to which the ileal pouch is anastomosed. In a hand-sewn technique, the rectal mucosa can be excised entirely, although islands of rectal mucosa are left even with this technique (a mucosectomy). Cuffitis is inflammation of the cuff anastomosed to the ileal pouch.
Patients with cuffitis will present with symptoms similar to that of pouchitis, with frequency, urgency and passage of blood in the faeces. Cuffitis can be difficult to diagnose as flexible endoscopy may not visualize the area well; careful digital examination and pouchoscopy can aid diagnosis.
Shen et al [58] treated patients with cuffitis with mesalazine suppositories, with good tolerance and response. In 2013 Wu showed that 33% of all patients with cuffitis responded to 5-ASA or steroid treatment, 18% were dependent on treatment and 48% were refractory to treatment [59]. Importantly these studies were based on small numbers with heterogeneity in the definition of cuffitis and hence should be interpreted with caution.
Crohn's disease of the pouch CD of the pouch is probably overdiagnosed [24]. It is possible that a pouch is performed with a preoperative diagnosis of CD but it has been estimated that pouch excision rates for this are 45%-55% in patients who have a preoperative diagnosis of CD [60,61]. In those who originally underwent restorative proctocolectomy for presumed UC, 2%-8% had their original diagnosis changed to CD. Importantly, the criteria utilized to diagnose CD are varied. Some studies have defined CD of the pouch as including: inflammation of the pouch that is resistant to antibiotic treatment, stricturing of the afferent limb, stricturing of the small bowel or fistulating disease [62][63][64][65].
CD of the pouch should only be considered when there is conclusive histology (i.e. granulomas supporting CD) and/or the presence of characteristic skip lesions in the small bowel. To help understand the likelihood of CD of the pouch, the timing of the CD-like problems of the pouch can often aid diagnosis. As a general rule, fistulas that occur within 2 years of pouch formation are likely to be related to the surgery itself, whereas CD-like changes beyond this may represent CD, but this requires thorough investigation and is less likely.
The semantics are often unhelpful for the patient; fistulas and CDlike complications regardless of cause will need a multidisciplinary approach and careful assessment.

Stricture
Stricturing of the pouch tends to develop at two locations -the pouch-anal anastomosis and at the pouch inlet between the prepouch ileum and pouch itself. Other less common locations include the pouch body (midpouch) and the afferent limb at the previous diverting ileostomy site [66,67]. Sometimes this latter site can be erroneously thought to be a CD-related stricture.
The incidence of pouch strictures has been reported to be as high as 38% [68]. Large retrospective cohorts found early stricture to occur in 5% and late stricture in 11% [12,69].
A systematic review of pouch strictures highlights how the cause of the stricture can be gleaned from the anatomical site. Subsequent management is then tailored in a stepwise fashion. Pouch inlet strictures are likely to be inflammatory or fibrotic. Inflammatory strictures can be treated initially with medical therapy (such as thiopurines or biologicals) and fibrotic strictures can be managed with endoscopic dilatation or surgical approaches [67].
Pouch-anal anastomotic strictures will respond to digital and instrumental dilatation; this can be highly successful but requires repeat dilatations [68]. The likely cause is postoperative sepsis or surgical causes such as tension or ischaemia at the anastomosis.
Sometimes there is no clear cause [67]. Medena catheters will also allow patients to evacuate at home but patients must be taught how to use them [70].
Midpouch strictures are poorly understood in terms of their aetiology and there is no clear evidence for definitive management.
They may be treated with a mixture of endoscopic and surgical approaches [67].

Anal sphincter insufficiency
The anal sphincter should be assessed preoperatively before offer- A further problem can be a long cuff and cuffitis (see Figure 3).
However, a noninflamed rectum may also cause problems with adequate emptying of the pouch. This is increasingly common in laparoscopic stapled pouches [76].

Neoplasia
Dysplasia can occur in both UC and FAP despite IPAA. The stapled technique leaves a rectal cuff which may undergo chronic inflammation and dysplastic change. Rectal mucosa may also regrow after mucosectomy and hand-sewn anastomosis. Dysplasia may also occur in the pouch body itself or arise from the anal squamous epithelium.
Pouch-related adenocarcinoma is rare in UC. A systematic review of the evidence found the pooled incidence rate to be 0.35% 20 years after IPAA [77]. Neoplasia at colectomy was the strongest risk factor for development of cancer. Irritable pouch syndrome All patients should get advice on diet, fluid and electrolyte replacement, constipating medication and barrier creams to optimize function even in the absence of any pathology. However, up to a fifth of patients will have no demonstrable pathology accounting for their symptoms despite clinical, endoscopic, radiological and histological investigations [10]. These patients are given a diagnosis of exclusion: irritable pouch syndrome. Patients with irritable pouch syndrome are more likely to be taking anxiolytics, antidepressants and opioid medication than those with inflammatory pouch conditions or normal pouches [80]. Treatment is aimed at alleviating symptoms with constipating drugs to reduce frequency and urgency. Referral to a dietician to improve the diet, exclude certain foods and alter timing of meals may also help. Biofeedback may have a role [81].

CON CLUS ION
The badly behaving pouch is a challenging clinical problem to solve for both gastroenterologists and surgeons. The patient may have been on a difficult journey with poor pouch function for many years and will be desperate for improvement. The importance of a clear and detailed history, careful examination and specialist investigation has been demonstrated. There are a number of potential pathologies with overlapping symptoms that can lead to misdiagnosis and potentially dangerous treatments. The importance of patient-centred care is stressed -the aim here is to improve quality of life and that may be achieved with a diverting loop ileostomy.

ACK N OWLED G EM ENTS
The authors wish to thank Mr Stephen Preston, Multimedia Consultant, for his illustration.

CO N FLI C T O F I NTE R E S T S
None of the authors have relevant disclosures.

E TH I C A L S TATEM ENT
Ethics approval was not sought nor required for this work.

FU N D I N G I N FO R M ATI O N
The authors received no specific funding for this work.

AUTH O R CO NTR I B UTI O N S
MD, JS, GW, SC, OF, SC and AH conceived and designed the review. All authors wrote the paper. All authors critically appraised the paper and edited the final version. AH is the guarantor of the paper.