https://orcid.org/0000-0001-6952-0985
Abstract
In symptomatic patients with ileoanal pouches, pouchoscopy is needed for accurate diagnosis but is invasive. We aimed to assess the utility of non-invasive gastrointestinal ultrasound and faecal calprotectin in ileoanal pouch patients.
Patients with an ileoanal pouch were consecutively enrolled in this cross-sectional study from clinics in Victoria, Australia. The pouchitis disease activity index was used as a reference standard. Video-recorded pouchoscopies were reviewed by three gastroenterologists. Pouch, pre-pouch, and cuff biopsies were reviewed by a single pathologist. Ultrasound was performed by a single gastroenterologist transabdominally and transperineally. Faecal calprotectin was measured from morning stool samples. All examiners were blinded to patients’ clinical history.
A total of 44 participants had a pouchoscopy, of whom 43 had a faecal calprotectin test and 42 had an ultrasound; 17 had pouchitis, 15 had pre-pouch ileitis, and 16 had cuffitis. Pouch wall thickness of <3 mm was 88% sensitive in excluding pouchitis, and pouch wall thickness of ≥4 mm was 87% specific in diagnosing pouchitis. Transabdominal ultrasound had good utility [area under the curve: 0.78] in diagnosing moderate-severe pre-pouch ileitis. Transperineal ultrasound had good utility for the diagnosis of pouchitis [area under the curve: 0.79]. Faecal calprotectin differentiated inflammatory from non-inflammatory pouch disorders, such as irritable pouch syndrome, with an area under the curve of 0.90. Faecal calprotectin <100 µg/g ruled out inflammatory pouch disorders with a sensitivity of 94%.
Faecal calprotectin and ultrasound are accurate and complementary tests to diagnose and localise inflammation of the ileoanal pouch. Prospective studies are needed to validate proposed sonographic indices and calprotectin levels.
1. Introduction
Restorative proctocolectomy with ileal pouch–anal anastomosis [IPAA] is the preferred surgical treatment for most patients with ulcerative colitis [UC] and rectal-involving familial adenomatous polyposis [FAP]. Quality of life after colectomy is generally good.1 However, a significant proportion of patients have pouch-related symptoms characterised by increased frequency of pouch emptying, urgency, and abdominal cramps. These can be secondary to inflammatory disorders of the pouch such as idiopathic pouchitis, which affects up to 50–60% of UC patients and 15–22% of FAP patients,2–4 as well as cuffitis, pre-pouch ileitis, or Crohn’s-like disease of the pouch. In addition, pouch-related symptoms can be secondary to non-inflammatory disorders of the pouch such as irritable pouch syndrome, which accounts for these symptoms in 30–40% of symptomatic pouch patients.5 Since pouch-related symptoms are not specific, an objective tool is needed to correctly diagnose the underlying cause of pouch-related symptoms.
The most widely used objective tool is pouchoscopy, which allows endoscopic and histological inflammation of the pouch to be assessed. Such evaluation, when combined with clinical symptoms, enables the diagnosis of pouchitis, as well as its extent and severity. The most commonly used disease activity index is the Pouchitis Disease Activity Index [PDAI],6 an 18-point composite index comprising symptom [0–6 points], endoscopy [0–6 points], and histology [0–6 points] subscores. A total PDAI score of ≥7, or ≥5 for the modified PDAI that omits histology, are used for the diagnosis of pouchitis.6,7 Pouchoscopy also allows assessment of the pre-pouch ileum and the cuff. However, pouchoscopy has two main limitations. First, the endoscopic items that make up the endoscopic subscore were arbitrarily chosen and, in addition, four of the six scored items—oedema, loss of vascular pattern, granularity, and mucus exudate—perform sub-optimally with only moderate to fair inter-rater reliability, in contrast to friability [spontaneous or contact bleeding], and ulceration, which are reliable.8 Second, because symptoms are a poor guide to the underlying diagnosis [inflammation or not, site of involvement], performing a pouchoscopy on every episode of increased stool frequency is an invasive and unpopular burden on patients with a pouch. An alternative strategy of empirical antibiotic treatment for every symptomatic episode in a patient with a history of pouchitis is associated with potential side effects of unnecessary antibiotic use and is less cost-effective than performing a pouchoscopy.9 There is, therefore, a need for non-invasive alternative tools to objectively assess symptomatic patients.
Gastrointestinal ultrasound [GIUS] is a non-invasive, easy to use, cost-effective, accurate, and reliable test for assessing intestinal inflammation in patients with inflammatory bowel disease.10 Sonographic features such as bowel wall thickness and degree of hyperaemia assessed by colour Doppler are accurate and reliable in assessing the activity and severity of inflammation of the small and large bowel in patients with Crohn’s disease and UC.10 Furthermore, its application by gastroenterologists within a clinic provides immediate results and its value as a point-of-care test is growing. However, its utility in assessing inflammation of the ileoanal pouch, the pre-pouch ileum, and the cuff has not been reported.
Faecal calprotectin is sensitive and specific for the assessment of intestinal inflammation in inflammatory bowel disease [IBD].11 The findings reported from three studies have suggested utility of faecal calprotectin in assessing pouchitis,12–14 but the accuracy of the PDAI in those reports is questionable since only a single endoscopist was used, and the cuff and pre-pouch were not evaluated. Hence, whether calprotectin can differentiate inflammatory from non-inflammatory disorders and whether it could differentiate the location of inflammatory causes of pouch disorders was not demonstrated.
We, therefore, aimed to evaluate the utility of GIUS and faecal calprotectin in assessing inflammatory disorders of the pouch, pre-pouch ileum, and cuff in a cross-sectional study using a rigorous methodology to ensure the reference standard used is accurate and reliable.
2. Materials and Methods
2.1. Participant selection
In this cross-sectional study, patients with an ileoanal pouch were consecutively enrolled to take part in this study between August 2018 and April 2020. They were recruited from various outpatient inflammatory bowel disease and surgical clinics in Victoria, Australia. Patients were eligible if they had an ileoanal pouch with bowel continuity for at least 12 months performed because of IBD, and were aged 18 years or older. Exclusion criteria included the inability to provide informed consent.
2.2. Study protocol
The study comprised two visits. At the first visit, the patient’s history was reviewed, current pouch function assessed, blood tests performed, faecal calprotectin sample collected, and both transabdominal and transperineal GIUS performed. At the second visit, patients underwent a pouchoscopy. This study was conducted according to the guidelines of the National Statement on Ethical Conduct in Human Research and all procedures involving patients were approved by the Alfred Health Ethics Committee [HREC 376/18]. Written informed consent was obtained from all participants before study enrolment.
2.3. Categorisation of patients according to pouch history
Patients were categorised according to their pouchitis history into normal pouch [no episodes of pouchitis for ≥2 years], acute antibiotic-responsive pouchitis [less than four episodes a year], chronic antibiotic-dependent pouchitis [four or more antibiotic-responsive episodes or need for ongoing antibiotics], and chronic antibiotic-refractory pouchitis. Patients with a history of pouchitis were also sub-categorised into: microbially mediated pouchitis if they had antibiotic-responsive pouchitis or antibiotic-dependent pouchitis not maintained on aminosalicylates, immunomodulators, or biologics;, or immune-mediated pouchitis if they had antibiotic-refractory pouchitis or antibiotic-dependent pouchitis maintained on aminosalicylates and/or immunosuppressive medications.
2.4. Categorisation of pouch phenotype according to pouchoscopy and histopathology
Pouchoscopies were performed using a gastroscope [Olympus GIF-HQ190; Olympus Corporation, Tokyo, Japan] with the patient under deep sedation 30 min after receiving a phosphate enema. Biopsies were taken from the pouch body, pre-pouch ileum and, when possible, the cuff, fixed in 10% neutral buffered formalin, and sent to the pathology laboratory for histopathological examination. All pouchoscopies were video-recorded. Before the start of the study, three gastroenterologists with >10 years’ experience attended a 1-h study meeting where they received training on the PDAI endoscopic features and standards of reviewing a pouchoscopy, and completed the European Crohn’s and Colitis Organisation online e-course on endoscopic assessment of pouchitis, to ensure a consistent reporting standard. All three were blinded to the clinical history. An internal reviewer [MPS] and external reviewer [SC] scored the endoscopies and a UK-trained reviewer, who had joined our centre the year the study commenced [DG], acted as the tie-breaker when there were disagreements in the scores. Disagreements occurred in three patients with pouchitis, four with pre-pouch ileitis, and thirteen with cuffitis. The final scores were evenly distributed between the scores of the principal endoscopists. The histological subscore was completed by a single experienced pathologist [AP] who was also blinded to the clinical history.
Patients were categorised according pouchoscopy and histopathology, using the following definitions. Pouchitis was diagnosed when the total PDAI was ≥7. Endoscopic inflammation of the pouch was defined as an endoscopy subscore ≥2. Histological inflammation of the pouch was defined as a histological subscore ≥2. Endoscopic and histological evidence of pouchitis was defined as a combined endoscopic and histological score ≥4. Symptomatic patients were defined as those with a clinical PDAI of ≥3. Irritable pouch syndrome was defined as the presence of a clinical PDAI ≥3 with a total PDAI <7. Pre-pouch ileitis was defined as the endoscopic presence of ulcers, erosions, and/or stricture, and/or histological features of ileitis more than 2 cm beyond the pouch inlet if reported by both endoscopists or one endoscopist and the pathologist. Moderate-severe pre-pouch ileitis was defined as endoscopic features that extended >10 cm from the pouch inlet and/or moderate to severe inflammation on histopathological examination based on degree of architectural disruption, degree of mononuclear and polymorph neutrophil infiltration, and presence of erosions or ulcers. Cuffitis was defined according to the Ulcerative Colitis Endoscopic Index of Severity [UCEIS] when vascular obliteration, bleeding, and/or erosions and ulcerations were present.
2.5. GIUS examination
GIUS was performed using a Canon Aplio i800 [Canon Medical Systems, Tochigi, Japan] by a single gastroenterologist [AF] with more than 8 years and 5000 scans of sonographic experience, who was blinded to the patients’ clinical history and endoscopic findings. Neither fasting nor bowel preparation was required for the sonographic examination, but patients were asked to avoid emptying the bladder in the 3 h before the assessment. The pouch was assessed in a similar fashion to the way the rectal views are obtained, with transabdominal and a transperineal approaches.10,15 Briefly, for the transabdominal approach, convex [1–8 MHz] and linear [3–11 MHz] probes were used with the patient supine, whereas convex [1–8 MHz] and microconvex [4–11 MHz] probes were used for the transperineal approach with the patient in the left lateral position and the probe placed over the anus (Figure 1). The pre-pouch ileum was defined as the first segment of small bowel loop at least 2–5 cms proximal to the pouch body. Indices measured included: [a] bladder fullness categorised qualitatively as ‘full’, ‘half full’, or ‘empty’ transabdominally; [b] quality of view defined as ‘good’ when the mucosa-lumen interface and the muscularis-serosa interface were clearly identified, ‘partial’ when one of the interfaces was not clearly identified, and ‘poor’ when both interfaces were not clearly identified; and [c] sonographic features of the bowel/pouch inflammation. The features assessed included the following.
![A] Transabdominal ultrasound of the pouch with a transverse view. The pouch lies deep or posterior to the bladder and the uterus or prostate. The pouch lumen is identified by the white gas line. The mucosa is identified as the hypoechoic area interfacing 0the gas line. The submucosa is the echogenic area superficial to the mucosa. The muscularis is the hypoechoic area superficial to the submucosa and deep to the hyperechoic serosa. B] Schema of transabdominal ultrasound in a transverse plane. C] Transperineal ultrasound of the pouch with a longitudinal view of the cuff and pouch. The pouch lumen is identified by the white gas line. The mucosa is identified as the hypoechoic area interfacing the gas line. The submucosa is the echogenic area superficial to the mucosa. The muscularis is the hypoechoic area superficial to the submucosa and deep to the hyperechoic serosa. D] Schema of transperineal ultrasound demonstrating sagittal or longitudinal plane [transducer placed over anus in anteroposterior direction, with anterior side on the right]. IS, internal sphincter; ES, external sphincter.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/16/1/10.1093_ecco-jcc_jjab125/1/m_jjab125f0001.jpeg?Expires=1749498075&Signature=0Tcf4i9JWI2QZnAqKaiS~IfXjWnTQ3nJiz6ZO5Cnb9Sr8J8wzZ2L~8za3Rlh~GNNdoR5KVionLU~Hr4ZMO~TftEgqfY327~ojiNUTe-OkjBZJve7bqIbrgKALUz0UqMRxs3mUdQI6XxJp0BaZ5E9n5TKDIBxTCwwKF7Rcw7GFNmCNtIpDOQfBVn1GxfOw2ACsef6yvIWGUHLrmwTs1bdFCRVgS7kp44kR1L~jcHnpW8rW~PbePJqvqHi987iCPWBz5wi7TqAgw5tvLGfYton1IEnXYzllB40jZdi1DxlI-lV9zeFAfQPYN5uoMMzhNn~OMoS6chmWr-ajY569HAAnA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
A] Transabdominal ultrasound of the pouch with a transverse view. The pouch lies deep or posterior to the bladder and the uterus or prostate. The pouch lumen is identified by the white gas line. The mucosa is identified as the hypoechoic area interfacing 0the gas line. The submucosa is the echogenic area superficial to the mucosa. The muscularis is the hypoechoic area superficial to the submucosa and deep to the hyperechoic serosa. B] Schema of transabdominal ultrasound in a transverse plane. C] Transperineal ultrasound of the pouch with a longitudinal view of the cuff and pouch. The pouch lumen is identified by the white gas line. The mucosa is identified as the hypoechoic area interfacing the gas line. The submucosa is the echogenic area superficial to the mucosa. The muscularis is the hypoechoic area superficial to the submucosa and deep to the hyperechoic serosa. D] Schema of transperineal ultrasound demonstrating sagittal or longitudinal plane [transducer placed over anus in anteroposterior direction, with anterior side on the right]. IS, internal sphincter; ES, external sphincter.
[i] Bowel/pouch wall thickness in millimetres. Two measurements of wall thickness were done in the transabdominal and transperineal approaches, and the mean taken. In the transabdominal approach, cross-sectional or horizontal views were used for assessing the pouch, and both cross-sectional and longitudinal views were used for assessing the pre-pouch ileum. In the transperineal approach, both cross-sectional and longitudinal views were used for assessing the pouch and cuff.
[ii] Presence or absence of hyperaemia, using colour Doppler signal and the Limberg score which incorporates bowel wall thickness and degree of hyperaemia as first described by Limberg in 1999.16 Mild hyperaemia was defined as a few short vessels within the wall; moderate hyperaemia as long-stretched vessels within the wall; and severe hyperaemia as multiple long-stretched vessels within the wall and into the surrounding mesentery.
[iii] Preservation of wall stratification. This was categorised as preserved when all five sonographic layers of the bowel were seen, reduced when there was a reduction in the number of the layers seen, and absent when no layers where seen between the lumen-mucosal interface and the muscularis-serosa interface.17
[iv] Degree of peristalsis following the methodology described by Menys et al.18
[v] Presence or absence of mesenteric hyperechogenicity.
[vi] Presence or absence of lymphadenopathy, defined as the presence of lymph nodes greater than 4 mm in the short axis diameter.19
2.6. Laboratory analyses
Peripheral venous blood samples obtained at the first visit were assessed in the routine pathology laboratory for full blood count, urea and electrolytes, liver function tests, C-reactive protein [CRP], erythrocyte sedimentation rate [ESR], and iron studies.
For faecal calprotectin analysis, patients collected all the stool from the first bowel motion on the morning of their first visit, ensuring no urinary contamination occurred, and kept it at 4 oC until delivered to the laboratory later that day, where a sample was aliquoted into prefilled extraction devices for calprotectin analysis. The remainder of the stool was stored at -20 oC until required for analysis. Faecal calprotectin was measured by ELISA [CAL0100, CALPRO AS, Lysaker, Norway] as per manufacturer’s instructions. The results were express a micrograms per gram of faeces.
2.7. Statistical methods
Data were analysed using SPSS v27 statistical analysis software [IBM, Armonk, NY, USA]. Continuous variables were compared using the independent two-sample t test and Mann‐Whitney U test according to the normality of their distribution. Nominal variables were compared using the chi square test and Fisher’s exact test. Multiple logistic regression was used to assess the relationship of bladder fullness and body mass index [BMI] on the quality of views obtained by ultrasound. Associations between two continuous, and between continuous and ordinal, variables were assessed using Spearman’s rank order correlation. Area under the curve [AUC] for receiver operating characteristic [ROC] analysis was used to analyse diagnostic performance of faecal calprotectin and the sonographic features measured. All tests were two-sided and considered statistically significant at p ≤0.05.
3. Results
3.1. Participants
Of 50 consecutive patients invited to participate in this study, 44 had a pouchoscopy, of whom 43 had a faecal calprotectin and 42 had a GIUS within a median of 38 (interquartile range [IQR] 14–60) days. Patients’ characteristics, pouchitis phenotype, and pouchitis status are provided in Table 1. All patients had a J pouch. Ulcerative colitis was the main indication for proctocolectomy [89%], followed by indeterminate colitis [11%]. The majority of patients [63%] were male, with a mean age of 50 years and pouch age of 13.5 years. Body mass index [BMI] was <25 kg/m2 in 55%, 25.1–29.9 kg/m2 in 30%, and >30 kg/m2 in 13%.
Patient characteristics and pouch phenotype.
| Characteristics | No pouchitis [n = 27] | Pouchitis [n = 17] | p-value | |
|---|---|---|---|---|
| Age, years, mean [SD] | 51 [18] | 49 [15] | 0.61 | |
| Male n [%] | 20 [77] | 10 [49] | 0.31 | |
| Pouch age, years, mean [SD] | 13 [6.5] | 14 [78] | 0.72 | |
| Indication for proctocolectomy | Ulcerative colitis | 24 [89] | 15[88] | 0.85 |
| Indeterminate colitis | 3 [11] | 2 [12] | 0.99 | |
| Pouch phenotype n [%] | Normal | 9 [33] | 0 [0] | 0.07 |
| Antibiotic-responsive | 5 [19] | 4 [24] | ||
| Antibiotic-dependent | 9 [33] | 10 [59] | ||
| Antibiotic-refractory | 1 [4] | 2 [12] | ||
| Crohn’s-like disease | 3 [11] | 1 [6] | ||
| Pouchitis medications | Antibiotics | 3 [11] | 4 [24] | 0.24 |
| Aminosalicylates | 0 [0] | 0 [0] | - | |
| Steroids | 3 [11] | 2 [12] | >0.99 | |
| Immunomodulators | 6 [22] | 3 [17] | >0.99 | |
| Biologics | 5 [27] | 2 [12] | >0.99 | |
| Clinical PDAI, median [IQR] | 2 [1–2] | 3 [2–4] | 0.002 | |
| Endoscopic PDAI, median [IQR] | 2 [1–3] | 4 [2.5–5] | <0.001 | |
| Histological PDAI, median [IQR] | 1 [0–1] | 2 [2–3.5] | <0.001 | |
| Total PDAI, median [IQR] | 4 [3–5 | 9 [8–10] | <0.001 | |
| Faecal calprotectin µg/g, median [IQR] | 109 [51–260] | 553 [353–1156] | <0.001 | |
| CRP, median [IQR] | 4[1.75–8] | 5[1–13] | 0.58 |
| Characteristics | No pouchitis [n = 27] | Pouchitis [n = 17] | p-value | |
|---|---|---|---|---|
| Age, years, mean [SD] | 51 [18] | 49 [15] | 0.61 | |
| Male n [%] | 20 [77] | 10 [49] | 0.31 | |
| Pouch age, years, mean [SD] | 13 [6.5] | 14 [78] | 0.72 | |
| Indication for proctocolectomy | Ulcerative colitis | 24 [89] | 15[88] | 0.85 |
| Indeterminate colitis | 3 [11] | 2 [12] | 0.99 | |
| Pouch phenotype n [%] | Normal | 9 [33] | 0 [0] | 0.07 |
| Antibiotic-responsive | 5 [19] | 4 [24] | ||
| Antibiotic-dependent | 9 [33] | 10 [59] | ||
| Antibiotic-refractory | 1 [4] | 2 [12] | ||
| Crohn’s-like disease | 3 [11] | 1 [6] | ||
| Pouchitis medications | Antibiotics | 3 [11] | 4 [24] | 0.24 |
| Aminosalicylates | 0 [0] | 0 [0] | - | |
| Steroids | 3 [11] | 2 [12] | >0.99 | |
| Immunomodulators | 6 [22] | 3 [17] | >0.99 | |
| Biologics | 5 [27] | 2 [12] | >0.99 | |
| Clinical PDAI, median [IQR] | 2 [1–2] | 3 [2–4] | 0.002 | |
| Endoscopic PDAI, median [IQR] | 2 [1–3] | 4 [2.5–5] | <0.001 | |
| Histological PDAI, median [IQR] | 1 [0–1] | 2 [2–3.5] | <0.001 | |
| Total PDAI, median [IQR] | 4 [3–5 | 9 [8–10] | <0.001 | |
| Faecal calprotectin µg/g, median [IQR] | 109 [51–260] | 553 [353–1156] | <0.001 | |
| CRP, median [IQR] | 4[1.75–8] | 5[1–13] | 0.58 |
SD, standard deviation; PDAI, Pouchitis Disease Activity Index; IQR, interquartile range; CRP, C-reactive protein.
Patient characteristics and pouch phenotype.
| Characteristics | No pouchitis [n = 27] | Pouchitis [n = 17] | p-value | |
|---|---|---|---|---|
| Age, years, mean [SD] | 51 [18] | 49 [15] | 0.61 | |
| Male n [%] | 20 [77] | 10 [49] | 0.31 | |
| Pouch age, years, mean [SD] | 13 [6.5] | 14 [78] | 0.72 | |
| Indication for proctocolectomy | Ulcerative colitis | 24 [89] | 15[88] | 0.85 |
| Indeterminate colitis | 3 [11] | 2 [12] | 0.99 | |
| Pouch phenotype n [%] | Normal | 9 [33] | 0 [0] | 0.07 |
| Antibiotic-responsive | 5 [19] | 4 [24] | ||
| Antibiotic-dependent | 9 [33] | 10 [59] | ||
| Antibiotic-refractory | 1 [4] | 2 [12] | ||
| Crohn’s-like disease | 3 [11] | 1 [6] | ||
| Pouchitis medications | Antibiotics | 3 [11] | 4 [24] | 0.24 |
| Aminosalicylates | 0 [0] | 0 [0] | - | |
| Steroids | 3 [11] | 2 [12] | >0.99 | |
| Immunomodulators | 6 [22] | 3 [17] | >0.99 | |
| Biologics | 5 [27] | 2 [12] | >0.99 | |
| Clinical PDAI, median [IQR] | 2 [1–2] | 3 [2–4] | 0.002 | |
| Endoscopic PDAI, median [IQR] | 2 [1–3] | 4 [2.5–5] | <0.001 | |
| Histological PDAI, median [IQR] | 1 [0–1] | 2 [2–3.5] | <0.001 | |
| Total PDAI, median [IQR] | 4 [3–5 | 9 [8–10] | <0.001 | |
| Faecal calprotectin µg/g, median [IQR] | 109 [51–260] | 553 [353–1156] | <0.001 | |
| CRP, median [IQR] | 4[1.75–8] | 5[1–13] | 0.58 |
| Characteristics | No pouchitis [n = 27] | Pouchitis [n = 17] | p-value | |
|---|---|---|---|---|
| Age, years, mean [SD] | 51 [18] | 49 [15] | 0.61 | |
| Male n [%] | 20 [77] | 10 [49] | 0.31 | |
| Pouch age, years, mean [SD] | 13 [6.5] | 14 [78] | 0.72 | |
| Indication for proctocolectomy | Ulcerative colitis | 24 [89] | 15[88] | 0.85 |
| Indeterminate colitis | 3 [11] | 2 [12] | 0.99 | |
| Pouch phenotype n [%] | Normal | 9 [33] | 0 [0] | 0.07 |
| Antibiotic-responsive | 5 [19] | 4 [24] | ||
| Antibiotic-dependent | 9 [33] | 10 [59] | ||
| Antibiotic-refractory | 1 [4] | 2 [12] | ||
| Crohn’s-like disease | 3 [11] | 1 [6] | ||
| Pouchitis medications | Antibiotics | 3 [11] | 4 [24] | 0.24 |
| Aminosalicylates | 0 [0] | 0 [0] | - | |
| Steroids | 3 [11] | 2 [12] | >0.99 | |
| Immunomodulators | 6 [22] | 3 [17] | >0.99 | |
| Biologics | 5 [27] | 2 [12] | >0.99 | |
| Clinical PDAI, median [IQR] | 2 [1–2] | 3 [2–4] | 0.002 | |
| Endoscopic PDAI, median [IQR] | 2 [1–3] | 4 [2.5–5] | <0.001 | |
| Histological PDAI, median [IQR] | 1 [0–1] | 2 [2–3.5] | <0.001 | |
| Total PDAI, median [IQR] | 4 [3–5 | 9 [8–10] | <0.001 | |
| Faecal calprotectin µg/g, median [IQR] | 109 [51–260] | 553 [353–1156] | <0.001 | |
| CRP, median [IQR] | 4[1.75–8] | 5[1–13] | 0.58 |
SD, standard deviation; PDAI, Pouchitis Disease Activity Index; IQR, interquartile range; CRP, C-reactive protein.
3.2. Phenotype according to pouch history
Clinical categorisation of the patients is also shown in Table 1. Of the 35 patients with a history of pouchitis, 18 were classified as predominantly microbially mediated and 17 as predominantly immune mediated.
3.3. Pouch phenotype
Of the 44 patients, 17 [39%] had pouchitis [total PDAI ≥7], 15 [34%] had pre-pouch ileitis, considered at least moderate-severe in 10 and associated with pouchitis in eight patients [53%], and 16 had cuffitis, of whom nine also had pouchitis and two had pre-pouch ileitis. Hence, four had isolated pouchitis, five [43%] isolated pre-pouch ileitis, and five isolated cuffitis. Five had irritable pouch syndrome.
The clinical PDAI subscores did not correlate significantly with the endoscopic or histological subscores. However, the endoscopic subscores had a moderate positive correlation with the histological subscores [r = 0.39; p <0.0001]. All the patients with a diagnosis of pouchitis had an endoscopic subscore of ≥2, consistent with endoscopic inflammation of the pouch [p <0.001]; 81% of pouchitis patients had a histological subscore of ≥2 [p = 0.001]; and 94% of the patients with pouchitis had a combined endoscopic and histological subscore of ≥4 [p >0.001]. More patients with immune mediated pouchitis had moderate-severe pre-pouch ileitis compared with microbially mediated pouchitis [47% vs 11%; p = 0.027]. More patients with chronic, antibiotic-refractory pouchitis had moderate-severe pre-pouch ileitis compared with those with chronic antibiotic-dependent pouchitis [71% vs 26%; p = 0.069].
3.4. Gastrointestinal ultrasound
Of the 44 patients who underwent a pouchoscopy, 42 [26 no pouchitis and 16 pouchitis] had a GIUS via transabdominal and transperineal approach. In three patients, the transperineal examination used only the convex or microconvex probe, due to availability of probes on the examination days.
3.4.1. Transabdominal gastrointestinal ultrasound assessment of the pouch
Using the linear probe, only 20% of the transabdominal views were of good quality, but with the convex probe, 69% were of good quality, 21% were of partial quality, and 10% were of poor quality. Examples of good quality scans are shown in Figure 2A and B. Good quality views were obtained in 21 [84%] of 25 patients when the bladder was full, which was 6.6 times more likely than when the bladder was half [n = 15] or completely empty [n = 2; p = 0.022]. Likewise, BMI <25 kg/m2 was associated with good quality views in 20 of 23 [86%] compared with nine of 19 [47%] when the BMI was ≥25 kg/m2 [p = 0.02].
![Ultrasound images of ileal pouches and pre-pouch ileum. Transabdominal ultrasound of the pouch using the convex probe in A] a patient with pouchitis based on total PDAI of ≥7; and B] a patient with no pouchitis based on a total PDAI <7. Transperineal ultrasound of the pouch using the convex probe in C] a patient with pouchitis based on total PDAI of ≥7; and D] a patient with no pouchitis based on a total PDAI <7. Transabdominal ultrasound of the pre-pouch pouch showing E] a bowel wall thickness of 4.4 mm; and F] colour Doppler of the pre-pouch ileum, showing mild degree of hyperaemia. G] Schema of transabdominal ultrasound of the pre-pouch ileum in view. PDAI, Pouchitis Disease Activity Index.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/16/1/10.1093_ecco-jcc_jjab125/1/m_jjab125f0002.jpeg?Expires=1749498075&Signature=D5bgKK85dtvDxryo7icpS0~XfV9XN10TapKcMBlqzntf7JYAaIJ9e43PYPTuUYDZ8GlxdIlEYG4sjP3aBmE39~e8pk5-DmQ4EjO8F9JMoPNOSUG1Z5jmrImTOQRTL6RZZPzq0ILXigguS4AYR-AIYLVoBC2vHG3KpE4okvqZO2V~jRkbGC6W7qbKO3Afb3q5UBRYkuf~NJnon4aE1tpr7kjunB0UMdfEA-ChFb0OswIvb4pKqsvk8qssFJILucm20wuqTQKIJPp2JL4kBANG0jpbGXvkdmYRWtTzTXfV15cj5rm5bmDUYC40sfhTNSa37SVffHLBGJbRg-yRhHwI3w__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Ultrasound images of ileal pouches and pre-pouch ileum. Transabdominal ultrasound of the pouch using the convex probe in A] a patient with pouchitis based on total PDAI of ≥7; and B] a patient with no pouchitis based on a total PDAI <7. Transperineal ultrasound of the pouch using the convex probe in C] a patient with pouchitis based on total PDAI of ≥7; and D] a patient with no pouchitis based on a total PDAI <7. Transabdominal ultrasound of the pre-pouch pouch showing E] a bowel wall thickness of 4.4 mm; and F] colour Doppler of the pre-pouch ileum, showing mild degree of hyperaemia. G] Schema of transabdominal ultrasound of the pre-pouch ileum in view. PDAI, Pouchitis Disease Activity Index.
Since only 20% of the linear probe views were of good quality, only the measurements obtained by the convex probe [1–8 MHz] were analysed. There were no statistically significant differences in any of the assessed GIUS indices of pouch body inflammation between the pouchitis and no pouchitis groups, even when only the measurements obtained with the good quality views were analysed [data not shown]. Similarly, none of the sonographic indices of inflammation were significantly different in those with endoscopic and acute histological evidence of pouchitis [combined endoscopic and histological subscore ≥4] compared with those with a subscore <4.
3.4.2. Transabdominal gastrointestinal ultrasound assessment of the pre-pouch ileum
Using the high-frequency linear probe 3–11 MHz transabdominally, good views of the pre-pouch ileum were seen in 97% of the patients. As shown in Table 2, the bowel wall thickness was 3.0 [2.1–4.6] mm in patients with pre-pouch ileitis, which tended to be higher than 2.2 [2–2.4] mm in those with no pre-pouch ileitis [p = 0.072]. The presence and degree of hyperaemia was significantly higher in patients with pre-pouch ileitis [p = 0.046]. Furthermore, the Limberg score was significantly higher in patients with pre-pouch ileitis compared with those without [p = 0.013]. Of the additional features of inflammation assessed, only abdominal lymphadenopathy was significantly more common in patients with pre-pouch ileitis compared with patients without pre-pouch ileitis [p = 0.034]. All four features had moderate discriminative performance for detecting pre-pouch ileitis; on ROC analysis, bowel wall thickness had an AUC of 0.67 [0.48–0.84], hyperaemia 0.62 [0.43–0.81], Limberg score 0.69[0.51–0.87], and lymphadenopathy 0.65[0.47–0.84]. Examples of pre-pouch findings are shown in Figure 2C–E.
Gastrointestinal ultrasound findings in the pre-pouch ileum according to the inflammatory activity.
| Parameter | Pre-pouch ileitis | Severity of pre-pouch inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent | Present | p-value | No or mild | Moderate-severe | p-value | ||
| Number of patients | 27 | 15 | 32 | 10 | |||
| Bowel wall thickness, median [IQR], mm | 2.2 [2–2.4] | 3.0 [2.1–4.6] | 0.075a | 2.2 [2–2.4] | 3.7 [2.2–5] | 0.008a | |
| Hyperaemia | Absent | 89% | 67% | 0.046b | 91% | 50% | 0.002b |
| Mild | 11% | 7% | 9% | 10% | |||
| Moderate | 0% | 13% | 0% | 20% | |||
| Severe | 0% | 13% | 0% | 20% | |||
| Limberg score, median [IQR] | 0[0] | 1[0–3] | 0.013a | 0[0] | 1.5[0–3] | 0.0005a | |
| Stratification | Preserved | 82% | 67% | 0.43b | 81% | 60% | 0.37b |
| Reduced | 15% | 20% | 13% | 30% | |||
| Absent | 4% | 13% | 6% | 10% | |||
| Submucosal prominence | Absent | 93% | 73% | 0.164c | 94% | 60% | 0.021c |
| Present | 7% | 27% | 6% | 40% | |||
| Peristalsis | Good | 85% | 60% | 0.210b | 87% | 40% | 0.011b |
| Reduced | 8% | 20% | 7% | 30% | |||
| Absent | 8% | 20% | 7% | 30% | |||
| Mesenteric hyper-echogenicity | Absent | 96% | 80% | 0.122c | 97% | 70% | 0.036c |
| Present | 4% | 20% | 3% | 30% | |||
| Lymphadenopathy | Absent | 85% | 53% | 0.034c | 81% | 50% | 0.094c |
| Present | 15% | 47% | 19% | 50% | |||
| Parameter | Pre-pouch ileitis | Severity of pre-pouch inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent | Present | p-value | No or mild | Moderate-severe | p-value | ||
| Number of patients | 27 | 15 | 32 | 10 | |||
| Bowel wall thickness, median [IQR], mm | 2.2 [2–2.4] | 3.0 [2.1–4.6] | 0.075a | 2.2 [2–2.4] | 3.7 [2.2–5] | 0.008a | |
| Hyperaemia | Absent | 89% | 67% | 0.046b | 91% | 50% | 0.002b |
| Mild | 11% | 7% | 9% | 10% | |||
| Moderate | 0% | 13% | 0% | 20% | |||
| Severe | 0% | 13% | 0% | 20% | |||
| Limberg score, median [IQR] | 0[0] | 1[0–3] | 0.013a | 0[0] | 1.5[0–3] | 0.0005a | |
| Stratification | Preserved | 82% | 67% | 0.43b | 81% | 60% | 0.37b |
| Reduced | 15% | 20% | 13% | 30% | |||
| Absent | 4% | 13% | 6% | 10% | |||
| Submucosal prominence | Absent | 93% | 73% | 0.164c | 94% | 60% | 0.021c |
| Present | 7% | 27% | 6% | 40% | |||
| Peristalsis | Good | 85% | 60% | 0.210b | 87% | 40% | 0.011b |
| Reduced | 8% | 20% | 7% | 30% | |||
| Absent | 8% | 20% | 7% | 30% | |||
| Mesenteric hyper-echogenicity | Absent | 96% | 80% | 0.122c | 97% | 70% | 0.036c |
| Present | 4% | 20% | 3% | 30% | |||
| Lymphadenopathy | Absent | 85% | 53% | 0.034c | 81% | 50% | 0.094c |
| Present | 15% | 47% | 19% | 50% | |||
IQR, interquartile range.
aMann‐Whitney U test.
bChi square test.
cFisher’s exact test.
Gastrointestinal ultrasound findings in the pre-pouch ileum according to the inflammatory activity.
| Parameter | Pre-pouch ileitis | Severity of pre-pouch inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent | Present | p-value | No or mild | Moderate-severe | p-value | ||
| Number of patients | 27 | 15 | 32 | 10 | |||
| Bowel wall thickness, median [IQR], mm | 2.2 [2–2.4] | 3.0 [2.1–4.6] | 0.075a | 2.2 [2–2.4] | 3.7 [2.2–5] | 0.008a | |
| Hyperaemia | Absent | 89% | 67% | 0.046b | 91% | 50% | 0.002b |
| Mild | 11% | 7% | 9% | 10% | |||
| Moderate | 0% | 13% | 0% | 20% | |||
| Severe | 0% | 13% | 0% | 20% | |||
| Limberg score, median [IQR] | 0[0] | 1[0–3] | 0.013a | 0[0] | 1.5[0–3] | 0.0005a | |
| Stratification | Preserved | 82% | 67% | 0.43b | 81% | 60% | 0.37b |
| Reduced | 15% | 20% | 13% | 30% | |||
| Absent | 4% | 13% | 6% | 10% | |||
| Submucosal prominence | Absent | 93% | 73% | 0.164c | 94% | 60% | 0.021c |
| Present | 7% | 27% | 6% | 40% | |||
| Peristalsis | Good | 85% | 60% | 0.210b | 87% | 40% | 0.011b |
| Reduced | 8% | 20% | 7% | 30% | |||
| Absent | 8% | 20% | 7% | 30% | |||
| Mesenteric hyper-echogenicity | Absent | 96% | 80% | 0.122c | 97% | 70% | 0.036c |
| Present | 4% | 20% | 3% | 30% | |||
| Lymphadenopathy | Absent | 85% | 53% | 0.034c | 81% | 50% | 0.094c |
| Present | 15% | 47% | 19% | 50% | |||
| Parameter | Pre-pouch ileitis | Severity of pre-pouch inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent | Present | p-value | No or mild | Moderate-severe | p-value | ||
| Number of patients | 27 | 15 | 32 | 10 | |||
| Bowel wall thickness, median [IQR], mm | 2.2 [2–2.4] | 3.0 [2.1–4.6] | 0.075a | 2.2 [2–2.4] | 3.7 [2.2–5] | 0.008a | |
| Hyperaemia | Absent | 89% | 67% | 0.046b | 91% | 50% | 0.002b |
| Mild | 11% | 7% | 9% | 10% | |||
| Moderate | 0% | 13% | 0% | 20% | |||
| Severe | 0% | 13% | 0% | 20% | |||
| Limberg score, median [IQR] | 0[0] | 1[0–3] | 0.013a | 0[0] | 1.5[0–3] | 0.0005a | |
| Stratification | Preserved | 82% | 67% | 0.43b | 81% | 60% | 0.37b |
| Reduced | 15% | 20% | 13% | 30% | |||
| Absent | 4% | 13% | 6% | 10% | |||
| Submucosal prominence | Absent | 93% | 73% | 0.164c | 94% | 60% | 0.021c |
| Present | 7% | 27% | 6% | 40% | |||
| Peristalsis | Good | 85% | 60% | 0.210b | 87% | 40% | 0.011b |
| Reduced | 8% | 20% | 7% | 30% | |||
| Absent | 8% | 20% | 7% | 30% | |||
| Mesenteric hyper-echogenicity | Absent | 96% | 80% | 0.122c | 97% | 70% | 0.036c |
| Present | 4% | 20% | 3% | 30% | |||
| Lymphadenopathy | Absent | 85% | 53% | 0.034c | 81% | 50% | 0.094c |
| Present | 15% | 47% | 19% | 50% | |||
IQR, interquartile range.
aMann‐Whitney U test.
bChi square test.
cFisher’s exact test.
When patients with moderate-severe pre-pouch ileitis were compared with those with no or mild pre-pouch ileitis, bowel wall thickness was greater in patients with moderate-severe pre-pouch ileitis (3.7 [2.2–5] mm; n= 11) compared with those with no or mild pre-pouch ileitis (2.2 [2.0–2.4] mm; n = 32; p = 0.008). The degrees of hyperaemia and Limberg score were also significantly higher in those with moderate-severe pre-pouch ileitis [p <0.005]. As shown in Table 2, additional sonographic features associated with moderate-severe pre-pouch ileitis included submucosal prominence, reduced peristalsis, and mesenteric hyperechogenicity [all p <0.05]. Both bowel wall thickness and Limberg score had good discriminative performance, with AUCs of 0.78 [0.59–0.96] [p = 0.004] and 0.79 [0.60–0.97] [p = 0.003], respectively. A bowel wall thickness cut-off value of 2.0 mm had a 90% sensitivity and a 96% negative predictive value against moderate-severe pre-pouch ileitis, but had a specificity of only 21% and positive predictive value of 26% for the presence of moderate-severe pre-pouch ileitis. Increasing the cut-off value to 3.0 mm increased specificity to 84% and positive predictive value to 58%, at the expense of reducing sensitivity to 70%.
3.4.3. Transperineal gastrointestinal ultrasound assessment of the pouch
Use of the convex probe [1–8 MHz] was associated with a good quality of views of the pouch body in 78%, but only 25% of the cuff. Examples of good quality scans are shown in Figure 2F and G. There was a trend for better quality of views of the pouch in those with BMI <25 kg/m2 compared with those with BMI ≥25 kg/m2 [90% vs 63%; p = 0.06]. Using the microconvex probe [4–11 MHz], 85% of the pouch body views and 97% of the cuff views were considered of good quality;the quality of the views was not affected by BMI status.
Using the convex probe, the pouch wall thickness in the pouchitis group [total PDAI ≥7] was 4.2 [3.2–5.7] mm, which was higher than 2.8 [2.4–3.8] mm in the non-pouchitis group [p = 0.001; Table 3]. There were no significant differences in the presence or absence of hyperaemia, degree of hyperaemia, and the Limberg scores in the pouchitis group compared with those in the no pouchitis group, at cut-off values of 3.0 mm and 4.0 mm for pouch wall thickness [p = 0.244 and p = 0.183, respectively]. Using the microconvex probe [4–11 MHz], none of the GIUS indices of inflammation were different between the pouchitis group and the no pouchitis group.
Gastrointestinal ultrasound findings in the pouch body via transperineal views with convex probe according to inflammatory activity.
| Parameter | Pouchitis | Endoscopic and histological inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent [total PDAI <7] | Present [total PDAI ≥7] | p-value | Absent [subscore <4] | Present [subscore ≥4] | p-value | ||
| Number | 23 | 16 | 26 | 13 | |||
| Pouch wall thickness, median [IQR], mm | 2.8 [2.4–3.8] | 4.2 [3.2–5.7] | 0.001a | 2.8 [2.4–3.7] | 4.1 [3–5.2] | 0.003a | |
| Hyperaemia | Degree, median [IQR] | 0[0–1] | 0.5[0–2] | 0.244a | 0 [0–1] | 0 [0–2] | 0.368a |
| • Absent [0] | 74% | 69% | 0.183b | 80% | 67% | 0.297b | |
| • Mild [1] | 22% | 6% | 20% | 13% | |||
| • Moderate [2] | 4% | 19% | 0% | 17% | |||
| • Severe [3] | 0% | 6% | 0% | 4% | |||
| Limberg score median [IQR] | Pouch wall thickness cut-off = 3.0 mm | 1[0–1] | 1[1-1] | 0.177a | 1 [0–1] | 1 [0–1] | 0.283 a |
| Pouch wall thickness cut-off = 4.0 mm | 0[0-0] | 0.5[0–1] | 0.14a | 0 [0-0] | 1 [0.5–1] | 0.019 | |
| Parameter | Pouchitis | Endoscopic and histological inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent [total PDAI <7] | Present [total PDAI ≥7] | p-value | Absent [subscore <4] | Present [subscore ≥4] | p-value | ||
| Number | 23 | 16 | 26 | 13 | |||
| Pouch wall thickness, median [IQR], mm | 2.8 [2.4–3.8] | 4.2 [3.2–5.7] | 0.001a | 2.8 [2.4–3.7] | 4.1 [3–5.2] | 0.003a | |
| Hyperaemia | Degree, median [IQR] | 0[0–1] | 0.5[0–2] | 0.244a | 0 [0–1] | 0 [0–2] | 0.368a |
| • Absent [0] | 74% | 69% | 0.183b | 80% | 67% | 0.297b | |
| • Mild [1] | 22% | 6% | 20% | 13% | |||
| • Moderate [2] | 4% | 19% | 0% | 17% | |||
| • Severe [3] | 0% | 6% | 0% | 4% | |||
| Limberg score median [IQR] | Pouch wall thickness cut-off = 3.0 mm | 1[0–1] | 1[1-1] | 0.177a | 1 [0–1] | 1 [0–1] | 0.283 a |
| Pouch wall thickness cut-off = 4.0 mm | 0[0-0] | 0.5[0–1] | 0.14a | 0 [0-0] | 1 [0.5–1] | 0.019 | |
IQR, interquartile range.
aMann‐Whitney U test.
bFisher’s exact test.
Gastrointestinal ultrasound findings in the pouch body via transperineal views with convex probe according to inflammatory activity.
| Parameter | Pouchitis | Endoscopic and histological inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent [total PDAI <7] | Present [total PDAI ≥7] | p-value | Absent [subscore <4] | Present [subscore ≥4] | p-value | ||
| Number | 23 | 16 | 26 | 13 | |||
| Pouch wall thickness, median [IQR], mm | 2.8 [2.4–3.8] | 4.2 [3.2–5.7] | 0.001a | 2.8 [2.4–3.7] | 4.1 [3–5.2] | 0.003a | |
| Hyperaemia | Degree, median [IQR] | 0[0–1] | 0.5[0–2] | 0.244a | 0 [0–1] | 0 [0–2] | 0.368a |
| • Absent [0] | 74% | 69% | 0.183b | 80% | 67% | 0.297b | |
| • Mild [1] | 22% | 6% | 20% | 13% | |||
| • Moderate [2] | 4% | 19% | 0% | 17% | |||
| • Severe [3] | 0% | 6% | 0% | 4% | |||
| Limberg score median [IQR] | Pouch wall thickness cut-off = 3.0 mm | 1[0–1] | 1[1-1] | 0.177a | 1 [0–1] | 1 [0–1] | 0.283 a |
| Pouch wall thickness cut-off = 4.0 mm | 0[0-0] | 0.5[0–1] | 0.14a | 0 [0-0] | 1 [0.5–1] | 0.019 | |
| Parameter | Pouchitis | Endoscopic and histological inflammation | |||||
|---|---|---|---|---|---|---|---|
| Absent [total PDAI <7] | Present [total PDAI ≥7] | p-value | Absent [subscore <4] | Present [subscore ≥4] | p-value | ||
| Number | 23 | 16 | 26 | 13 | |||
| Pouch wall thickness, median [IQR], mm | 2.8 [2.4–3.8] | 4.2 [3.2–5.7] | 0.001a | 2.8 [2.4–3.7] | 4.1 [3–5.2] | 0.003a | |
| Hyperaemia | Degree, median [IQR] | 0[0–1] | 0.5[0–2] | 0.244a | 0 [0–1] | 0 [0–2] | 0.368a |
| • Absent [0] | 74% | 69% | 0.183b | 80% | 67% | 0.297b | |
| • Mild [1] | 22% | 6% | 20% | 13% | |||
| • Moderate [2] | 4% | 19% | 0% | 17% | |||
| • Severe [3] | 0% | 6% | 0% | 4% | |||
| Limberg score median [IQR] | Pouch wall thickness cut-off = 3.0 mm | 1[0–1] | 1[1-1] | 0.177a | 1 [0–1] | 1 [0–1] | 0.283 a |
| Pouch wall thickness cut-off = 4.0 mm | 0[0-0] | 0.5[0–1] | 0.14a | 0 [0-0] | 1 [0.5–1] | 0.019 | |
IQR, interquartile range.
aMann‐Whitney U test.
bFisher’s exact test.
On ROC analysis, pouch wall thickness, measured with the convex probe, had the highest accuracy for detecting pouchitis AUC of 0.79 [0.66–0.93]; p <0.0001) and for detecting endoscopic and acute histological evidence of pouchitis (AUC 0.78 [0.64–0.92]; p <0.001). The sensitivity and specificity of the different cut-off values of pouch wall thickness using the convex probe are shown in Supplementary Table 1, available as Supplementary data at ECCO-JCC online. A cut-off value of 3.0 mm had a sensitivity of 88% and negative predictive value of 88% in ruling out pouchitis, and a sensitivity of 76% and negative predictive value of 73% in ruling out endoscopic and acute histological evidence of pouchitis. Increasing the cut-off pouch wall thickness to 4.0 mm improved specificity for the diagnosis of pouchitis to 87% and a positive predictive value to 78%, and improved specificity and positive predictive value to for detecting endoscopic and acute histological evidence of pouchitis to 100%, at the expense of sensitivity.
The pouch wall thickness as assessed by convex probe also correlated with the severity of pouchitis, with a strong positive correlation with the total PDAI [r=0.51; p <0.001] and the combined endoscopic and histological subscores [r = 0.52; p <0.001], and a moderate positive correlation with PDAI endoscopic subscore [r = 0.45; p <0.003] and PDAI histological subscore [r = 0.52; p <005].
3.4.4. Transperineal gastrointestinal ultrasound assessment of the cuff
None of the ultrasound features measured using either the convex or microconvex probe showed significant differences between the cuffitis and non-cuffitis group.
3.5. Faecal calprotectin and other laboratory measures
The routine laboratory indices were not statistically significantly different according to their status regarding pouchitis, pre-pouchitis, or cuffitis, or to whether they were classified as predominately microbially mediated or immune mediated [data not shown].
The median faecal calprotectin in the patients without inflammatory disorders of the pouch at pouchoscopy was 58 [IQR 41–120] µg/g, which was lower than 553 [353–1156] µg/g in patients with pouchitis with or without pre-pouch ileitis or cuffitis [p <0.001] and 241 [161–653] µg/g in those with isolated pre-pouch ileitis [p = 0.004], but not different from 123 [IQR 36–302] µg/g in patients with isolated cuffitis [p = 0.469]. Faecal calprotectin was not different between those having pouchitis with or without pre-pouch ileitis and those with isolated pre-pouch ileitis [p = 0.446], but was higher than the calprotectin in patients with isolated cuffitis [p = 0.009]. There was a trend for a higher faecal calprotectin in those with isolated pre-pouch ileitis 241 [161–653] µg/g compared with isolated cuffitis 123 [36–302] µg/g [p = 0.111].
Among patients with different clinical pouchitis phenotypes, calprotectin was 55 [39–120] µg/g in those with a normal pouch, 55 [46–81] µg/g in irritable pouch syndrome, 224 [107–412] µg/g in antibiotic-responsive pouchitis, 356 [98–749] µg/g in chronic antibiotic-dependent pouchitis, and 590 [320–1711] µg/g in chronic antibiotic-refractory pouchitis. When adjusting for presence of current pouchitis, calprotectin was significantly different between patients with recurrent acute antibiotic-responsive pouchitis and patients with chronic pouchitis [p = 0.002]. However, there was no significant difference in calprotectin between the antibiotic-dependent or refractory groups [each p >0.05].
The ROC for faecal calprotectin at detecting any inflammatory pouch condition, pouchitis without or without associated pre-pouch ileitis or cuffitis, and endoscopic/acute histological activity of the pouch are shown in Table 4. Faecal calprotectin had outstanding performance at detecting any inflammatory pouch condition {0.90 (95% confidence interval [CI] 0.81–0.99)}, excellent performance at diagnosing pouchitis [0.85, 0.74–0.97], and good performance for detecting combined endoscopic and histological evidence of pouchitis [0.72, 0.56–0.88]. As also shown in Table 4, calprotectin <100 µg/g ruled out any pouch inflammatory condition [sensitivity 89% and negative predictive value 79%], pouchitis [sensitivity 94% and negative predictive value 93%], and endoscopic/histological evidence of pouchitis [sensitivity 84% and negative predictive value 71%]. Cut-off values between 100 and 349 µg/g had limited utility in diagnosing inflammatory pouch conditions, but ≥350 µg/g was useful in diagnosing pouchitis [specificity 82%, positive predictive value 88%] and endoscopic and acute histological evidence of inflammation [specificity 83%, positive predictive value 83%].
The performance and receiver operating characteristics of faecal calprotectin at detecting any inflammatory pouch condition, total PDAI-defined pouchitis, and endoscopic and endoscopic/histological activity of the pouch at different cut-off values.
| Faecal calprotectin cut-off value | Sensitivity | Specificity | Negative predictive value | Positive predictive value | Receiver operating characteristics AUC [95% CI]; p-value | |
|---|---|---|---|---|---|---|
| Detecting any inflammatory pouch condition | 0.90 [0.81–0.99]; | <0.001 | ||||
| 100 µg/g | 0.89 | 0.73 | 0.79 | 0.86 | ||
| 350 µg/g | 0.64 | 1.00 | 0.60 | 1.00 | ||
| Detecting pouchitis [PDAI score ≥7] | 0.85 [0.74–0.97]; | <0.001 | ||||
| 100 µg/g | 0.94 | 0.50 | 0.93 | 0.55 | ||
| 350 µg/g | 0.82 | 0.84 | 0.78 | 0.88 | ||
| Detecting endoscopic and histological disease activity [subscore ≥4] | 0.72 [0.56–0.88]; | 0.007 | ||||
| 100 µg/g | 0.84 | 0.55 | 0.71 | 0.72 | ||
| 350 µg/g | 0.60 | 0.83 | 0.60 | 0.83 |
| Faecal calprotectin cut-off value | Sensitivity | Specificity | Negative predictive value | Positive predictive value | Receiver operating characteristics AUC [95% CI]; p-value | |
|---|---|---|---|---|---|---|
| Detecting any inflammatory pouch condition | 0.90 [0.81–0.99]; | <0.001 | ||||
| 100 µg/g | 0.89 | 0.73 | 0.79 | 0.86 | ||
| 350 µg/g | 0.64 | 1.00 | 0.60 | 1.00 | ||
| Detecting pouchitis [PDAI score ≥7] | 0.85 [0.74–0.97]; | <0.001 | ||||
| 100 µg/g | 0.94 | 0.50 | 0.93 | 0.55 | ||
| 350 µg/g | 0.82 | 0.84 | 0.78 | 0.88 | ||
| Detecting endoscopic and histological disease activity [subscore ≥4] | 0.72 [0.56–0.88]; | 0.007 | ||||
| 100 µg/g | 0.84 | 0.55 | 0.71 | 0.72 | ||
| 350 µg/g | 0.60 | 0.83 | 0.60 | 0.83 |
AUC, area under the curve; PDAI, Pouchitis Disease Activity Index; CI, confidence interval.
The performance and receiver operating characteristics of faecal calprotectin at detecting any inflammatory pouch condition, total PDAI-defined pouchitis, and endoscopic and endoscopic/histological activity of the pouch at different cut-off values.
| Faecal calprotectin cut-off value | Sensitivity | Specificity | Negative predictive value | Positive predictive value | Receiver operating characteristics AUC [95% CI]; p-value | |
|---|---|---|---|---|---|---|
| Detecting any inflammatory pouch condition | 0.90 [0.81–0.99]; | <0.001 | ||||
| 100 µg/g | 0.89 | 0.73 | 0.79 | 0.86 | ||
| 350 µg/g | 0.64 | 1.00 | 0.60 | 1.00 | ||
| Detecting pouchitis [PDAI score ≥7] | 0.85 [0.74–0.97]; | <0.001 | ||||
| 100 µg/g | 0.94 | 0.50 | 0.93 | 0.55 | ||
| 350 µg/g | 0.82 | 0.84 | 0.78 | 0.88 | ||
| Detecting endoscopic and histological disease activity [subscore ≥4] | 0.72 [0.56–0.88]; | 0.007 | ||||
| 100 µg/g | 0.84 | 0.55 | 0.71 | 0.72 | ||
| 350 µg/g | 0.60 | 0.83 | 0.60 | 0.83 |
| Faecal calprotectin cut-off value | Sensitivity | Specificity | Negative predictive value | Positive predictive value | Receiver operating characteristics AUC [95% CI]; p-value | |
|---|---|---|---|---|---|---|
| Detecting any inflammatory pouch condition | 0.90 [0.81–0.99]; | <0.001 | ||||
| 100 µg/g | 0.89 | 0.73 | 0.79 | 0.86 | ||
| 350 µg/g | 0.64 | 1.00 | 0.60 | 1.00 | ||
| Detecting pouchitis [PDAI score ≥7] | 0.85 [0.74–0.97]; | <0.001 | ||||
| 100 µg/g | 0.94 | 0.50 | 0.93 | 0.55 | ||
| 350 µg/g | 0.82 | 0.84 | 0.78 | 0.88 | ||
| Detecting endoscopic and histological disease activity [subscore ≥4] | 0.72 [0.56–0.88]; | 0.007 | ||||
| 100 µg/g | 0.84 | 0.55 | 0.71 | 0.72 | ||
| 350 µg/g | 0.60 | 0.83 | 0.60 | 0.83 |
AUC, area under the curve; PDAI, Pouchitis Disease Activity Index; CI, confidence interval.
Faecal calprotectin also correlated with pouchitis severity, demonstrating a strong positive correlation with the total PDAI [r = 0.53; p <0.001] and a moderate positive correlation with the PDAI endoscopic subscore [r = 0.44; p = 0.003], PDAI histological subscore [r = 0.44; p = 0.003], and combined endoscopic and histological subscores [r = 0.48; p = 0.001].
3.6. Combination of GIUS and faecal calprotectin
Whether GIUS or faecal calprotectin should be performed first, whether the results are complementary, and whether pouchoscopy could be avoided [acknowledging an acceptable diagnostic inaccuracy] are addressed in Supplementary Tables 2 and 3, available as Supplementary data at ECCO-JCC online. If faecal calprotectin is used first, owing to its higher sensitivity, a level of <100 µg/g would accurately rule out inflammatory pouch disorder with a sensitivity of 89% and a negative predictive value of 79%. Of the 14 patients with low levels, only one had pouchitis and two had isolated cuffitis. GIUS was abnormal in the pouchitis patient and one of the patients with isolated cuffitis. Thus, addition of GIUS to faecal calprotectin increases sensitivity to 93% and negative predictive value to 83%. If faecal calprotectin was ≥350 µg/g it would have accurately identified all 18 patients in our study with an inflammatory pouch disorder. Transabdominal GIUS of pre-pouch ileum bowel wall thickness was abnormal in two patients and equivocal in two others. Transperineal GIUS of the pouch was abnormal in nine patients and equivocal in five. Therefore, all 18 would have had an accurate diagnosis made with GIUS assuming equivocal findings indicated inflammation. Thus, if faecal calprotectin was performed as the initial investigation followed by GIUS, 11 of the 43 did not need a GIUS or pouchoscopy, and 19 would have had an accurate diagnosis made with GIUS assuming equivocal findings indicated inflammation. Only 13 [30%] needed pouchoscopy. ROC analysis of combined FC and GIUS revealed an AUC of 0.85 [0.71–0.99]. If GIUS was performed first, three of the five patients with normal sonographic findings had a faecal calprotectin above 100 µg/g, which would have dictated pouchoscopy for diagnosis. In 17 patients with equivocal sonographic findings, 12 had faecal calprotectin results above 100 µg/g, indicating likely inflammatory disorders to be present and the requirement for pouchoscopy to confirm and localise inflammation. In patients with an abnormal GIUS, faecal calprotectin and pouchoscopy would have provided additional diagnostic information in very few. Hence, of the 39 patients having a successful GIUS first, 24 [62%] would have avoided pouchoscopy, of whom seven needed a faecal calprotectin for accurate diagnosis and 17 were accurately diagnosed by GIUS alone. The caveat to these estimations is that neither test is helpful in diagnosing isolated cuffitis [present in <10% of patients].
4. Discussion
Pouch-related symptoms from inflammatory and non-inflammatory disorders of the ileoanal pouch are non-specific, commonly reported, and negatively impact on health-related quality of life. However, therapeutic strategies differ according to the underlying pathology, which can vary from pouchitis to pre-pouch ileitis to cuffitis to irritable pouch syndrome. Current assessment techniques depend largely upon endoscopic examination of the pouch, which is disliked by patients, particularly when performed repeatedly. Non-invasive, readily repeatable tests are needed. The current study has investigated GIUS as a novel technique that potentially provides accurate information on both inflammation and the anatomical location of that inflammation, in addition to faecal calprotectin which offers information limited to the presence and absence of inflammation. We have shown that the application of GIUS provides with good accuracy the information needed to define the problems in the majority of patients, in terms of both presence/absence of inflammation and anatomical location. It requires the use of two approaches—transabdominal and transperineal—together with different probes. In addition, we have confirmed that faecal calprotectin is an accurate tool to define inflammation as assessed by gold standard techniques, and have defined practical and well-performing cut-off values that can be applied in clinical practice.
Compared with pouchoscopy, GIUS requires no preparation [enema or fasting] and no sedation. With the transperineal approach, a convex probe had good discriminative performance for diagnosing pouchitis and detecting endoscopic and acute histological inflammation of the pouch body. A pouch wall thickness of ≥4.0 mm can be considered equivalent to a PDAI endoscopic and histological score of ≥4 with a specificity of 100%. Furthermore, transperineal GIUS pouch wall thickness showed strong positive correlation with the severity of pouchitis. Thus, transperineal GIUS can be used as a non-invasive substitute for pouchoscopy to diagnose pouchitis and assess for the presence and severity of pouch inflammation. In symptomatic patients [clinical PDAI ≥3], a pouch wall thickness of ≥4.0 mm would be accurate for diagnosing pouchitis with an ultrasound-based PDAI [US-PDAI] of ≥7 (clinical PDAI ≥3 + bowel wallthickness [BWT] of ≥4.0 mm). In asymptomatic patients or patients who achieve symptomatic remission [clinical PDAI <3], GIUS pouch wall thickness of ≥4.0 mm indicates ongoing disease activity, whereas a bowel wall thickness of <3.0 mm indicates no disease activity and potentially ‘sonographic remission’. This is particularly important in patients with chronic pouchitis in whom symptom response alone is not sufficient to objectively assess treatment response.7 Although not assessed in this study, the strong positive correlation of pouch wall thickness with endoscopic and histological pouch inflammation suggests that changes in pouch wall thickness may also correlate with treatment response.
In contrast to the pouch itself, transabdominal GIUS of the pre-pouch ileum using the linear probe offered good quality views in nearly all patients and good discriminatory performance for detecting moderate-severe pre-pouch ileitis. Pre-pouch ileitis, which affects 4.4–11% of patients with ileoanal pouches,20,21 is associated with higher rates of treatment failure and complications, particularly when isolated, extensive, associated with deep ulcerations, and occurring in patients with a history of indeterminate colitis.20–22 Such cases appear to be predominately immune-mediated processes. In our cohort, patients with moderate-severe pre-pouch ileitis [n = 11] were all antibiotic-refractory and/or maintained on immunosuppressive medications; three [27%] had indeterminate colitis and three [27%] were considered likely to have Crohn’s disease. Therefore, it is important to identify this group of patients in order to minimise antibiotic use and direct treatment towards the mucosal immune component early on. Although pouchoscopy can accurately diagnose moderate-severe pre-pouch ileitis,23 GIUS has a few advantages. First, it is non-invasive, less costly, and requires no preparation. Second, since moderate-severe pre-pouch ileitis is associated with high antibiotic failure rates20 and relatively high complications,21 repeated objective disease activity monitoring at short intervals may help inform therapeutic strategy. Third, studies have shown that GIUS is accurate at diagnosing small bowel inflammation, such as small bowel Crohn’s disease, beyond the reach of ileocolonoscopy.24
The application of sonographic assessment of the pouch has its limitations. As discussed above, the transabdominal approach was not accurate in assessing the pouch body. A full urinary bladder and BMI <25 improved the quality of views, but not the accuracy of measurement. This was not surprising, as transabdominal ultrasound has a similar limitation in assessing the rectum in patients with ulcerative colitis.25 The transperineal approach was not accurate for assessing the degree of hyperaemia of the pouch body. One possible explanation is that the Doppler signal may not reach the depth of the pouch needed. Our findings resemble those reported by Sagami et al.,26 where rectal bowel wall thickness in UC patients assessed transperineally was the most accurate ultrasound feature, whereas utility of the Limberg score was hindered by the reduced accuracy of colour Doppler. Finally, the cuff was not accurately assessed by either probe. Therefore, in a symptomatic patient with normal transabdominal and transperineal GIUS, isolated cuffitis could not be ruled out or differentiated from irritable pouch syndrome.
Our study showed that faecal calprotectin at different cut-off values can accurately differentiate inflammatory from non-inflammatory pouch disorders and accurately confirm the diagnosis of pouchitis and reflect its severity. Hence, faecal calprotectin is highly suited to being the initial test in a symptomatic patient. The ‘rule-out’ cut-off value for any inflammatory pouch disorder was 100 µg/g, which was similar to that [56–92.5 µg/g] in previous reports12–14 where pouchitis alone was assessed. Those studies had a high risk of bias as the endoscopic component of the PDAI was assessed by a single endoscopist who was not blinded to the clinical history in one study,12 and the reliability of the endoscopic gold standard was, therefore, likely to be low. Furthermore, they generally had low specificity and, where calculated, low positive predictive values. In the current study, we addressed this by determining a highly accurate ‘rule-in’ cut-off value of 350 µg/g for the diagnosis of pouchitis [with or without associated cuffitis or pre-pouch ileitis]. However, this does not enable localisation of the inflammation. Therefore, following such levels with a GIUS can help localise inflammation and distinguish pouchitis, with or without pre-pouch ileitis, from moderate-severe pre-pouch ileitis. It follows that those with faecal calprotectin levels of 100–349 µg/g would be best to proceed to pouchoscopy [Figure 3a].
![Proposed assessment algorithms of patients with symptomatic ileoanal pouches, [a] starting with faecal calprotectin and [b] starting with gastrointestinal ultrasound. TAC, transabdominal approach using convex probe; TPC, transperineal approach using the convex probe.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/16/1/10.1093_ecco-jcc_jjab125/1/m_jjab125f0003.jpeg?Expires=1749498075&Signature=MpCGvjDWXbchnMLPEbMLSKTnI~oCEZ5D~gL0pqQV~MW4m8sVdSEpTMbbdRo2OHbft3aGzltkwZUNwy7Bh9Z7AXV~IQw1rQ~4eUq8VjEmoowKhsqz1YqavyN~xqgqQYW0g~kiUI9uRVB1Z4DGHkSB765nkJ7yDP8GFsv9rDgTuV57PNv1nRETYr9T4ndllLhRyYdkje~Er8pC0vXonAfgcyzYIrEOZvaWfmOr7O~hHYtJevsTwWXy2iTDI4Y0GD1~glTqqxkwn617BhZUZ~pCIHU9tl~6YSWzHXkrAPPYsfZZXFKguZkpU5DcgZDek0RQf8aGR5bcj9FCKajRtwdSwg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Proposed assessment algorithms of patients with symptomatic ileoanal pouches, [a] starting with faecal calprotectin and [b] starting with gastrointestinal ultrasound. TAC, transabdominal approach using convex probe; TPC, transperineal approach using the convex probe.
As a point-of-care test, GIUS is well-suited to accompany a patient’s clinic review. As discussed above, in those who have had a high recent faecal calprotectin [≥350 µg/g], it may be help localise inflammation. In those who have not had a recent faecal calprotectin, it may serve as the initial test. A normal GIUS with normal pouch wall thickness and pre-pouch ileum rules out pouchitis and moderate-severe pre-pouch ileitis with high sensitivity. Here the addition of faecal calprotectin depends on the patient’s history. In patients with a history or risk factors for isolated pre-pouch ileitis, a low faecal calprotectin of <100 µg/g can be further reassuring. Whereas in those with a history of risk factors for isolated cuffitis, such as long retained cuff, the addition of faecal calprotectin may not be of much value as calprotectin levels in our study were not significantly different between isolated cuffitis and irritable pouch syndrome. In such patients and in patients diagnosed with irritable pouch syndrome with ongoing symptoms despite treatment, a pouchoscopy is still needed. Conversely, if GIUS of the pouch and pre-pouch ileum were abnormal, a diagnosis can be made with high specificity. Here addition of calprotectin is not necessary. When pouch and pre-pouch wall thickness measured on GIUS is in the equivocal zone, calprotectin levels of <100 µg/g help rule out inflammatory pouch disorders without the need for a pouchoscopy [Figure 3b].
Our study has some limitations. GIUS was performed by a single expert gastroenterologist. Thus, inter-reader reliability of our suggested sonographic features of pouchitis and pre-pouch ileitis could not be assessed. However, GIUS studies in ulcerative colitis and Crohn’s disease have reassuringly shown high intra- and inter-rater reliability.27,28 This is particularly important, as it could overcome one of the main inherent limitations of the currently accepted gold standard, the low inter-rater reliability of the PDAI endoscopic subscore. Second, we have not assessed the learning curve for transabdominal and transperineal GIUS assessment of the pre-pouch ileum and pouch. Third, longitudinal assessment following treatment was not reported. One of the main advantages of GIUS is its repeatability and point-of-care capabilities, and this requires further evaluation.
In conclusion, GIUS and faecal calprotectin are accurate and complementary tests for assessing pouch-related symptoms and assessing the degree of inflammation and location of inflammatory pouch disorders. At different cut-off values, faecal calprotectin can help rule in or rule out inflammatory pouch disorders with a high degree of sensitivity, but cannot differentiate inflammatory pouch disorders. GIUS at different approaches and different cut-off values for bowel wall thickness can help confirm and localise inflammation. Transabdominal GIUS can help rule in and rule out moderate-severe pre-pouch ileitis, and transperineal GIUS can help rule in and rule out pouchitis with a high degree of sensitivity and specificity, reserving pouchoscopies for equivocal cases. Either test can be used as the initial non-invasive test to assess a symptomatic pouch, with the second test used in certain circumstances to improve the accuracy of the diagnostic process, limiting and reserving pouchoscopies to cases when both tests are equivocal. In our cohort, the combination of these two non-invasive tests would have reduced the need for pouchoscopy by 60–70%%. Future multicentre prospective studies are needed to validate the proposed GIUS cut-off values of the pouch and pre-pouch ileum, to assess the inter-rater reliability of GIUS in patients with ileoanal pouches, and to assess changes in GIUS features and faecal calprotectin levels with treatment.
The data used in this study are available to download at the following link [https://drive.google.com/file/d/1LXaLsEPYwUK8nrwYeOG4rjDP9NZvqkQf/view?usp=sharing].
Funding
ZSA was in receipt of the RTP Stipend Scholarship from the Monash University.
Conflict of Interest
None of the author have any disclosures
Author Contributions
ZSA: designed the study, initiateded the concepts, wrote the paper. ABF: designed the study, performed the gastrointestinal ultrasound, drafted the paper. DC: helped perform statistical analysis. SC: served as external reviewer of the pouchoscopy videos and drafted the manuscript. DG: reviewed pouchoscopy recorded videos. AP: performed histopathological examination on the biopsy samples. PDC: significant contribution towards recruiting patients. KT: significant contribution towards recruiting patients. SB: significant contribution towards recruiting patients. OR: analysed samples for faecal calprotectin. MPS: designed the study, served as the internal reviewer of the pouchoscopy videos, and drafted the manuscript. PRG: designed the study, critically revised the manuscript for important intellectual content, and drafted the manuscript. All authors approved the final version of the manuscript.
