An Expert Consensus to Standardise Definitions, Diagnosis and Treatment Targets for Anti–fibrotic Stricture Therapies in Crohn's Disease

F. Rieder; D. Bettenworth; C. Ma; C. E. Parker; L. A. Williamson; S. A. Nelson; G. van Assche; A. Di Sabatino; Y. Bouhnik; R. W. Stidham; A. Dignass; G. Rogler; S. A. Taylor; J. Stoker; J. Rimola; M. E. Baker; J. G. Fletcher; J. Panes; W. J. Sandborn; B. G. Feagan; V. Jairath


Aliment Pharmacol Ther. 2018;48(3):347-357. 

In This Article


Systematic Review

The literature search retrieved a total of 2238 citations. After removing duplicates, 1518 citations were screened using predefined eligibility criteria. Of these, 1270 citations were deemed not applicable based on title and abstract review. Ninety studies were excluded during full–text review, leaving a total of 158 included studies (Figure S1).

Data obtained from the systematic review were arranged into 4 tables: (1) radiologic definitions and diagnosis; (2) clinical definitions and diagnosis; (3) endoscopic definitions and diagnosis; and (4) endpoint assessment in pharmacological studies (Tables S1–S4). These data, in addition to other items of potential importance, were subsequently incorporated into a survey and sent to panelists for appropriateness rating (Table S5).

Consensus Process

Panel discussion resulted in minimal edits to the proposed items and the addition of 2 new statements. Item #30 revealed an unexpected disagreement in round two and was re–discussed in a third round.

In the literature, the terms "stenosis" and "stricture" are used interchangeably. In this article, we defined "stricture" based upon the recommendation of the Consensus of the American Gastroenterology Association (AGA).[17] The term stricture encompasses the possibility of the co–existence of inflammatory and fibrotic components.

Appropriateness of Items

Definition of naïve small bowel stricture. The panelists felt that a naïve small bowel stricture (strictures arising in parts of the intestine that do not contain a bowel anastomosis) on cross–sectional imaging is optimally defined by the combination of 3 features: (1) localised luminal narrowing; (2) bowel wall thickening; and (3) pre–stricture dilation. Panelists were uncertain about other combinations of radiologic features for stricture definition.

Specific criteria were generated for each of the radiologic features. With respect to bowel wall thickening, panelists felt that a 25% increase in wall thickness relative to the adjacent nonaffected bowel was an appropriate definition. Relating to the definition of pre–stricture dilation in cross–sectional imaging, a luminal diameter greater than 3 cm was regarded as appropriate. The definition of luminal narrowing as a luminal diameter reduction of at least 50%, measured relative to the normal adjacent bowel loop was considered appropriate. The inability to pass an adult colonoscope through the narrowed area without prior endoscopic dilation with a reasonable amount of pressure applied was felt to be an appropriate definition of stricture on endoscopy.

Obstructive symptoms alone were determined to be insufficient to define a stricture (Table S5; Table 1).

Definition of anastomotic small bowel stricture. Appropriateness ratings were similar for definitions of anastomotic (at site of prior intestinal resection with anastomosis) and naïve small bowel strictures; however, there was uncertainty regarding the definitions of wall thickness (Table S5; Table 1). The authors were cautious about evaluation of small bowel anastomosis, as these definitions only apply to proximal small bowel unaltered by surgical intervention, not enteroenterostomy associated with side–to–side small bowel anastomosis.

Diagnosis of small bowel stricturing CD. Cross–sectional imaging or ileocolonoscopy alone was considered appropriate to diagnose a small bowel stricture. Symptoms alone were considered inappropriate to diagnose a stricture. Moreover, most panelists felt that symptoms are not required to diagnose a stricture. Panelists felt that MR enterography (MRE) is the preferred diagnostic modality (sensitivity 55%–100%; specificity 91%–100%).[18] There was uncertainty about whether CT enterography (CTE) and ultrasound with or without oral contrast are the preferred diagnostic modalities. The high accuracy of both MRE and CTE was considered appropriate for detection of a single or multiple small bowel stricture(s), with CTE and MRE felt to have comparable accuracy. Ultrasound with or without oral contrast was deemed uncertain by the panel for detection of single or multiple small bowel stricture(s). MRE was preferred over CTE due to lack of radiation exposure in nonacutely ill, clinically stable patients (Table S5).

Clinical symptoms of stricturing CD. Clinical symptoms are not highly correlated with the presence of small bowel strictures on cross–sectional imaging or endoscopy and there is a disconnect between clinical symptoms and the severity of small bowel strictures on cross–sectional imaging or endoscopy. Symptoms considered appropriate for collection were acute abdominal distension, cramping, dietary restrictions, nausea, vomiting, abdominal pain (duration and intensity) and post–prandial abdominal pain (duration and intensity) (Table S5).

Detection of inflammation and fibrosis. In advanced small bowel strictures, extensive overlap between fibrotic and inflammatory components can be found on histopathology.[8] To detect the inflammatory component of a small bowel stricture, MRE and CTE were deemed to be highly accurate and clinical symptoms were felt to be highly inaccurate. There was uncertainty about ultrasound, colonoscopy, C–reactive protein (CRP) and faecal calprotectin for detection of the inflammatory component of a small bowel stricture. It was uncertain whether the degree of inflammation should optimally be determined using validated endoscopic scores. Panelists rated the following imaging features, reflecting the inflammatory component of a small bowel stricture on cross–sectional imaging, as appropriate: mural hyperenhancement, presence of ulcers, co–existence with penetrating disease, perienteric fat stranding, comb sign and intramural T2 hyperintensity (for MRE only).[18–23]

It was uncertain whether delayed enhancement MRI, magnetisation transfer MRI, ultrasound elastography, contrast enhanced ultrasound, bowel ultrasound, MRE and CTE are most accurate for confidently quantifying the fibrotic component of a small bowel stricture.[18,21,24–26] Colonoscopy with endoscopic mucosal biopsies was considered inappropriate. Currently, no technique can accurately distinguish the inflammatory from the fibrotic component of a small bowel stricture (Table S5).

Treatment targets for anti–fibrotic treatment of a small bowel stricture. No precedent for a trial of an anti–fibrotic in CD exists. Panelists considered it appropriate that successful treatment of a small bowel stricture requires improvement in clinical symptoms combined with radiologic features, clinical symptoms combined with endoscopic features or radiologic features combined with endoscopic features. Improvement in clinical symptoms alone was considered inappropriate as a clinical trial endpoint. When symptoms are used in combination with radiologic or endoscopic features to indicate successful treatment, then absence of acute abdominal distention, cramping, dietary restrictions, vomiting, abdominal pain and post–prandial abdominal pain, were considered reflective of successful anti–fibrotic treatment of small bowel strictures.

Radiologic features considered to indicate improvement were localised luminal narrowing, wall thickening, pre–stricture dilation and stricture length. Panelists also felt that the following individual radiologic features should improve with successful anti–fibrotic treatment: (1) a greater than 50% improvement in luminal narrowing or luminal diameter reduction of less than 50%; (2) reduction in bowel wall thickening by 50%; (3) reduction in pre–stricture dilation by 50%, a pre–stricture bowel diameter equal to normal bowel or a bowel diameter less than 2.5 cm and (4) improvement in stricture length by 50%. The ability to pass an adult endoscope through the stricture was felt to indicate successful anti–fibrotic treatment of a small bowel obstruction (Table 1).

In terms of time points to evaluate the efficacy of medical therapies for CD stricture on cross–sectional imaging, 24 and 48 weeks were considered appropriate, with 24 weeks chosen as the optimal primary efficacy endpoint for a clinical trial. Twenty–four weeks was also considered the only acceptable time point to evaluate endoscopic treatment success. Twelve weeks was considered the optimal time point to evaluate treatment success based on clinical symptom improvement (Table S5).

Endoscopic treatment of a stricture as a starting point for a clinical trial in Crohn's disease. Endoscopic balloon dilation may be useful for treatment of symptomatic patients with obstruction and may be used to temporise symptoms in an anti–fibrotic trial. The following items were judged to be appropriate: 18 mm as the maximal luminal diameter after dilation in 1 or several sessions; a balloon inflation time of at least 1 minute; and 5 cm as the maximum stricture length that should be dilated. Technical success after dilation is defined as the ability to pass an adult ileocolonoscope through a previously nontraversable stricture with reasonable amount of pressure applied, clinical efficacy for dilation is defined as the relief of clinical symptoms of bowel obstruction after dilation. Comparable items were considered appropriate for anastomotic strictures. Graded–through–the–scope balloons should be the preferred tool for endoscopic dilation (Table S5).

Endpoints for failure of stricture therapy after initial response. There optimal clinical symptoms indicative of treatment failure or re–obstruction of a small bowel stricture are abdominal distention, cramping, vomiting, dietary restrictions, abdominal pain and post–prandial abdominal pain. A combination of pre–stricture dilation, wall thickening and luminal narrowing on radiology and the inability to pass an adult endoscope were felt to be signs of treatment failure or re–obstruction.

Time to re–stricturing on imaging, endoscopic re–dilation or surgery were considered acceptable and optimal long–term endpoints in a clinical trial of an anti–fibrotic drug. There was uncertainty about time to symptom recurrence as an optimal endpoint for failure of stricture therapy after initial response. A trial endpoint for an anti–fibrotic in stricturing Crohn's disease was recommended to include cross–sectional imaging, endoscopy and clinical symptoms (Table S5).

Procedure preparation and reporting of cross–sectional imaging. To standardise procedure preparation and reporting, panelists assessed the appropriateness of cross–sectional imaging procedures. These results can be found in Table S5 and Appendix S1.

Expert consensus–based development of clinical trial prototype. Based on items considered appropriate by the CONSTRICT group, we propose a clinical trial outline to be used in the first anti–fibrotic trial in CD (Figure 1). Primary prevention of a stricture is a large unmet clinical need that novel anti–fibrotic therapies may address. However, the duration from CD diagnosis to stricture formation can be several years[27,28] and there is a lack of validated biomarkers to risk stratify patients.[29,30] Hence, pharmaceutical companies are unlikely to embark on primary prevention trials during the first wave of anti–fibrotic drug development. The panelists therefore felt that eligible patients should be clinically symptomatic, with single, naïve or anastomotic ileal strictures that are in reach of endoscopy and confirmed on cross–sectional imaging (CT or MR enterography). This approach was chosen since the current "gold standard" (surgical specimen analysis) is not feasible in this situation and mucosal biopsy specimen is superficial and will not detect transmural disease. The panel recommended the inclusion of only symptomatic strictures given that new anti–fibrotic agents are likely to have limited safety data. Therefore, patients and investigators would be reluctant to participate in a trial of therapy that did not offer the possibility of improving symptoms. Furthermore, it is highly unlikely any regulatory agency would agree to a trial evaluating asymptomatic patients at this juncture. Stricture presence on cross–sectional imaging should require all 3 identified features: localised luminal narrowing (luminal diameter reduction of at least 50%), bowel wall thickening (increase in wall thickness of 25%) and pre–stenotic dilation (luminal diameter less than 3 cm). Patients with internal penetrating disease should be excluded, as internal penetrating disease associated with stricturing disease is an indication for surgery.[29] On the basis of these 3 criteria, all eligible patients should initially receive optimal anti–inflammatory therapy to control symptoms and treat mucosal healing[31] with or without endoscopic balloon dilation (using graded–through–the–scope balloons). The maximal diameter of balloon should be 18 mm with a minimal inflation time of 1 minute. Strictures longer than 5 cm should not be dilated. Anti–inflammatory therapy optimisation should be performed based on a pre–specified algorithm that reflects optimal standard of care. If patient symptoms improve or subside within 12 weeks, the patients should undergo MRE with inclusion of experimental sequences, such as delayed enhancement or magnetisation transfer. The technical details about preparation for MR can be found in Table S5 and Appendix S1. If symptoms do not improve within the 12 weeks lead–in phase, the patient should be excluded. The 12 weeks mark does not reflect an endpoint, but rather allows selection of patients with symptomatic improvement for inclusion into the trial. The minimum number of symptoms that should be recorded are acute abdominal distension, cramping, dietary restrictions, nausea, vomiting, abdominal pain (duration and intensity) and post–prandial abdominal pain (duration and intensity). Given the current lack of validated tools, the authors recommend using a Likert scale or 100 mm visual analogue scale to quantitate these items. Imaging features representing the inflammatory component of a stricture on cross–sectional imaging are mural hyperenhancement, presence of ulcers, perienteric fat stranding, comb sign and intramural T2 hyperintensity. In addition, an ileocolonoscopy (adult ileocolonoscope only, to standardise the approach in the setting of a clinical trial) should be performed to assess passability of the stricture. This approach allows for direct visualisation of mucosal disease activity and sampling in biomarker studies, while also restricting clinical trial inclusion to patients with the most distal ileal strictures. This should be followed by randomisation to placebo or anti–fibrotic drug, given in combination with optimal anti–inflammatory therapy.

Figure 1.

Proposed approach to early development of anti–fibrotics in stricturing Crohn's disease. It is presumed that strictures in the patient population consist of a mix of inflammation and fibrosis. The optimal primary endpoint is 24 weeks; however, a later timepoint (52 weeks) may also be advantageous. At each endpoint, data relevant to objective assessment of disease activity, such as C–reactive protein and faecal calprotectin, should be collected. No patient–reported outcome (PRO) tool for stricturing Crohn's disease exists and we recommend inclusion of clinical symptoms found appropriate in this consensus statement into the clinical trial until PROs are available. NPO, Nothing per mouth; NG, Nasogastric; expMRe: experimental magnetic resonance enterography (including delayed enhancement and magnetisation transfer sequences)

While there is high accuracy for the detection of inflammation on cross–sectional imaging,[18] currently no imaging technique is able to accurately measure the amount of fibrosis in a stricture.[29] Given that anti–fibrotic therapy approaches may modulate the inflammatory component of a stricture,[1] serial objective parameters of inflammatory activity (ie, serum and faecal biomarkers) throughout the observation should also be collected. This process will facilitate the greatest possible distinction between the anti–fibrotic vs anti–inflammatory effects of an anti–fibrotic drug. This distinction is important as inflammation may be necessary for the development of fibrosis; however, the progression of fibrosis may become independent of inflammation as the disease progresses.[1]

Co–primary endpoints should be recurrence or worsening of clinical symptoms (following randomisation) and documented intestinal obstruction on MRE, with inclusion of experimental sequences signalling the need for endoscopic intervention or surgery. At the end of follow–up, asymptomatic patients should undergo MRE with experimental sequences and ileocolonoscopy. Success of anti–fibrotic treatment should be defined as an asymptomatic patient with reduction in luminal narrowing (greater than 50% improvement and/or luminal diameter reduction of less than 50%), pre–stenotic dilation (reduction in pre–stricture dilation by 50%, bowel diameter equal to normal bowel and/or improvement in pre–stricture dilation to less than 2.5 cm), wall thickening (improvement in bowel wall thickening by 50%) and stricture length (improvement of 50%). On ileocolonoscopy, successful anti–fibrotic treatment should be defined as an increase in luminal diameter or ability to pass an adult endoscope. Twenty–four weeks is considered the optimal time point to evaluate treatment success on cross–sectional imaging and ileocolonoscopy.