Article Text

Original article
IOIBD technical review on endoscopic indices for Crohn's disease clinical trials
  1. L Vuitton1,2,
  2. P Marteau3,
  3. W J Sandborn4,
  4. B G Levesque4,
  5. B Feagan5,
  6. S Vermeire6,
  7. S Danese7,
  8. G D'Haens8,
  9. M Lowenberg8,
  10. R Khanna5,
  11. G Fiorino7,
  12. S Travis9,
  13. J Y Mary10,
  14. L Peyrin-Biroulet2
  1. 1Department of Gastroenterology and Endoscopy Unit, Besançon University Hospital, Besançon, France
  2. 2Department of Hepato-Gastroenterology, Inserm U954, University Hospital of Nancy-Brabois, Lorraine University, Vandoeuvre-lès-Nancy, France
  3. 3Department of Digestive Diseases, AP-HP, Hôpital Lariboisière and University Paris 7 Denis Diderot, Sorbonne Paris Cité, Paris, France
  4. 4Division of Gastroenterology, University of California-San Diego, La Jolla, California
  5. 5Department of Medicine, Robarts Clinical Trials, Robarts Research Institute, Western University, London, Ontario, Canada
  6. 6Department of Gastroenterology, University of Leuven, University Hospitals Leuven, Leuven, Belgium
  7. 7Department of Gastroenterology, IBD Center, Humanitas Research Hospital, Rozzano, Milan, Italy
  8. 8Department of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, The Netherlands
  9. 9Translational Gastroenterology Unit, John Radcliffe Hospital, Oxford, UK
  10. 10Biostatistics and Clinical Epidemiology, Inserm U717, Hôpital Saint-Louis, Paris, France
  1. Correspondence to Dr Laurent Peyrin-Biroulet, Inserm U954 and Department of Hepato-Gastroenterology, University Hospital of Nancy-Brabois, Lorraine University, 1 Allée du Morvan, Vandoeuvre-lès-Nancy F54511, France; peyrinbiroulet{at}gmail.com

Abstract

Background Crohn's disease (CD) is a chronic disabling and progressive IBD. Only strategies looking beyond symptoms and based on tight monitoring of objective signs of inflammation such as mucosal lesions may have the potential for disease modification. Endoscopic evaluation is currently the gold standard to assess mucosal lesions and has become a major therapeutic endpoint in clinical trials. Several endoscopic indices have been proposed to evaluate disease activity; unvalidated and arbitrary definitions have been used in clinical trials for defining endoscopic response and endoscopic remission in CD.

Methods In these recommendations from the International Organization for the Study of Inflammatory Bowel Disease, we first reviewed all technical aspects of available endoscopic scoring systems in the literature. Second, in order to achieve consensus on endoscopic definitions of remission and response in trials, a two-round vote based on a Delphi method was performed among 14 specialists in the field of IBDs.

Results At the end of the voting process, the investigators ranked first a >50% decrease in Simple Endoscopic Score for Crohn's Disease (SES-CD) or Crohn's Disease Endoscopic Index of Severity for the definition of endoscopic response, and an SES-CD 0–2 for the definition of endoscopic remission in CD. All experts agreed on a Rutgeerts’ score i0–i1 for the definition of endoscopic remission after surgery.

  • CROHN'S DISEASE
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known on this subject?

  • In Crohn's disease (CD) only strategies looking beyond symptoms and targeting signs of inflammation such as mucosal lesions may have the potential for disease modification.

  • Endoscopic evaluation is currently the gold standard to assess mucosal lesions and has become a major therapeutic outcome in clinical trials but various endoscopic endpoints have been used.

  • Consensus definitions on endoscopic definitions of response and remission in CD are needed.

What are the new findings?

  • After reviewing technical aspects on scoring systems for CD, specialists in the field of IBD worked on consensus definitions regarding endoscopic response and remission using a Delphi process involving Organization for the Study of Inflammatory Bowel Disease members.

  • At the end of the voting process, the investigators ranked first a >50% decrease in Simple Endoscopic Score for Crohn's Disease (SES-CD) or Crohn's Disease Endoscopic Index of Severity for the definition of endoscopic response, and an SES-CD 0–2 for the definition of endoscopic remission in CD. Experts agreed on a Rutgeerts' score i0–i1 for the definition of endoscopic remission after surgery.

How might it impact on clinical practice in the foreseeable future?

  • This study is an attempt to establish a current consensus to define endoscopic response and remission in CD for clinical trials.

  • The definitions retained after the Delphi process should be subject to prospective testing in clinical trials of CD.

Introduction

Accumulating evidence indicates that mucosal healing may change the natural course of IBD by decreasing the need for surgery and reducing rates of hospitalisation in Crohn's disease (CD).1 ,2 The absence of endoscopic mucosal healing is associated with an increased risk of dysplasia and colorectal cancer in IBDs.1 As advances in the medical therapy of IBD have made mucosal healing a realistic goal, endoscopic indices are increasingly used in clinical practice as well as trials.

Several scoring systems have been developed in CD to evaluate endoscopic disease activity, but none have yet had their properties fully assessed. In addition, most studies addressing mucosal healing have used an arbitrary definition for endoscopic response and remission. Required qualities of any instrument are the following:

  • Reliability (does the instrument produce the same or similar results in individuals on different occasions or by different observers?)

  • Validity (does the instrument measure what is intended?)

  • Responsiveness (is the instrument able to measure change in an individual when it does occur?).

The primary purpose of this paper was first to review technical aspects on scoring systems for CD (available indices, definitions, construction, strengths and weaknesses). The secondary purpose was to develop a consensus definition among specialists in the field of IBD regarding endoscopic response and remission as well as after surgery in CD, based on a Delphi process involving Organization for the Study of Inflammatory Bowel Disease (IOIBD) members.

Available endoscopic scoring systems for CD

Crohn's Disease Endoscopic Index of Severity

Construction

The Crohn's Disease Endoscopic Index of Severity (CDEIS) was constructed in the late 1980s by the French Groupe d'Etude des Affections Inflammatoires du tube Digestif (GETAID) (table 1).3 Before the advent of video-endoscopy, the endoscopist-pair method was developed, which consisted of two endoscopists performing the endoscopy simultaneously, one handling the instrument, with the other one inspecting through a teaching tube, both recording their findings independently. Bowel examination was divided into five segments—rectum, sigmoid and left colon, transverse colon, right colon and ileum—in order to facilitate endoscopic findings description.4 Various endoscopic features that could be related to endoscopic severity were selected a priori and about hundred endoscopies were performed by various pairs of endoscopists using this method.4 Their results were analysed and discrepancies discussed in plenary sessions, leading to the consensual definition of nine descriptors : pseudopolyp, healed ulceration, frank erythema, frankly swollen mucosa, aphthoid ulceration, superficial ulceration, deep ulceration, ulcerated stenosis and non-ulcerated stenosis.4

Table 1

Endoscopic indices for Crohn's disease: definitions, strengths and validation

An inter-observer variation study on 133 colonoscopies, performed by 32 pairs among 29 endoscopists from 12 centres, demonstrated that it was feasible to collect in a reproducible way, on a standardised form, the presence of the nine descriptors per segment and the segmental surfaces involved by any of the nine descriptors and by ulcerations on a 10 cm analogue scale.4 A global evaluation of lesion severity (GELS) was made on a similar scale.4

In a subsample of 75 patients (65/75 colonoscopies reached the caecum and 26/75 the ileum), using data from one endoscopist randomly selected within each pair, the CDEIS was constructed through multiple linear regression as a linear combination of independent variables (number of segments involved by each lesion cited above, presence of the lesions over all segments, segmental surfaces involved), showing the highest correlation with the dependent variable, the endoscopist's GELS.3 The following independent variables were selected: the number of segments involved by superficial ulcerations and by deep ulcerations only, the sum of the segmental surfaces involved by any of the lesions and by ulcerations, the presence of ulcerated and non-ulcerated stenosis in any of the segments. The CDEIS was highly correlated (Pearson's correlation coefficient 0.83, n=75) to the endoscopist's GELS. Using paired data of 12 endoscopist-pairs from 12 centres, the CDEIS was shown to be highly reproducible (intraclass correlation coefficient (ICC) of 0.96, n=75).3

Strengths and weaknesses

The major strength of CDEIS comes from its validation performed in a series of 103 endoscopies,3 the data of which were not used in the construction phase.3 ,4 Indeed, CDEIS was still highly correlated to the endoscopist's GELS (Pearson's correlation coefficient of 0.81, n=103).3 In addition, in a group of 54 patients with active CD who twice underwent colonoscopy, the index was calculated before and at the end of a course of corticosteroids: CDEIS variations correctly reflected changes in disease severity as evaluated by the endoscopists, which also highly correlated with the change in endoscopist's GELS (Pearson's correlation coefficient of 0.72, n=54).3 In a recent study, Khanna et al5 demonstrated a high reproducibility level of CDEIS components and indices in centralised video readings by four specialists of 50 endoscopies in 49 patients with active CD disease entering a clinical trial (ICCs 0.89 and 0.71 for intra-rater and inter-rater respectively for CDEIS). This good agreement was recently confirmed by Daperno et al.6

The weakness of the CDEIS lies in its apparent lack of practicability, a weakness that is stressed by endoscopists not familiar with the procedure. As stated by the authors,3 the CDEIS requires precise and specific data recording, which needs prior training. In addition, discussion with endoscopists familiar with the CDEIS suggests that this index may underestimate the severity of disease when only one out of five segments is involved, especially the ileum.

Finally, the absence of validated score thresholds associated with specific prognostic values or with endoscopic healing represents a weakness of the CDEIS. It has to be understood that the CDEIS was constructed to quantify lesion severity along the entire range of severity, whereas endoscopic remission is concerned with only one end of the spectrum. This point has methodological consequences. A severity index has to be constructed in relation to the endoscopists’ GELS with one of the multiple regression methods, the simplest being multiple linear regression. On the other hand, endoscopic remission has to be constructed in relation to the endoscopists’ classification of remission using a discrimination method, the simplest being multiple logistic regression. It is therefore not certain that the CDEIS is the most appropriate index to define endoscopic remission.

Simple Endoscopic Score for Crohn's Disease

Construction

In 2002–2003, a new score was developed in an attempt to simplify the CDEIS.7 In the Simple Endoscopic Score for Crohn's Disease (SES-CD) (table 1), the items included were the result of a careful review of the GETAID studies with regard to the importance and the reproducibility of the most relevant endoscopic characteristics of CD.4 Only characteristics that were considered as contributing to clinical symptomatology and showing good reproducibility in the GETAID studies were considered for the SES-CD. The four endoscopic variables selected were ulcers, proportion of the surface covered by ulcers, proportion of the surface with any other lesions and stenosis. Each variable was scored from 0 to 3 in each segment. Ulcers were scored according to size (diameter 0.1–0.5 cm, 0.5–2 cm or >2 cm); proportion of ulcerated surface according to extent (<10%, 10–30% or >30%); proportion of affected surface according to extent (<50%, 50–75% or >75%); and stenosis as single or multiple, and whether the colonoscope could be passed through the narrowed lumen.7 Consecutive patients were enrolled in five centres from three countries. For the development phase of the study, 191 consecutive patients with an established diagnosis of CD were enrolled. The endoscopist completed an endoscopic scoring sheet immediately after colonoscopy. A second endoscopist observed the colonoscopy on the video monitor in some of the examinations (35/70 in the development phase and 36/121 in the validation phase) and recorded the lesions independently.

Endoscopic lesions as defined above were shown to be reproducible in a sample of 71 paired examinations. Then, in the first 70 consecutive patients (60 colonoscopies reaching the ileum), data from one of the endoscopists were randomly selected in the 35 examinations performed by pairs. The new index was constructed by multiple linear regression as a linear combination of independent variables, the sum of the four lesions as described above over the observed segments, the number of observed and affected segments, showing the highest correlation with the dependent variable, the CDEIS. The final choice for SES-CD, based on a balance between simplicity of calculation and high correlation with CDEIS, was the simple sum of the four lesions across the observed segments. The SES-CD was highly correlated with CDEIS (Pearson's correlation coefficient of 0.88, n=70).7

Strengths and weaknesses

SES-CD was validated on a sample of 121 additional patients (103 endoscopies reaching the ileum), with data from one endoscopist selected at random in the 36 examinations performed by a pair. SES-CD remained highly correlated with the CDEIS (Pearson's correlation coefficient of 0.89, n=121).7 Using data from endoscopist-pairs, SES-CD was again shown to be highly reproducible (ICC 0.98, n=71).7

Recently, Sipponen et al8 conducted a prospective, parallel scoring of the CDEIS and SES-CD in 86 patients referred for ileocolonoscopy in a cross-sectional study. After CD therapy, 32 patients underwent a follow-up endoscopy with scoring. The SES-CD significantly correlated with the CDEIS (Spearman’s correlation coefficient 0.94). As expected, changes between baseline and follow-up CDEIS and SES-CD indices also correlated (Pearson's correlation of 0.83).8 As stated before, SES-CD components and index were shown to be highly reproducible among four central readers (ICCs 0.91 and 0.83 for intra-rater and inter-rater respectively for SES-CD).5 This high reproducibility was particularly notable among specialists (ICC 0.93 among specialists vs 0.68 among general gastroenterologists).6

The weaknesses of SES-CD, except for retaining complexity, are similar to those of the CDEIS since SES-CD was constructed from the CDEIS. Nevertheless, in the SES-CD, the cut-offs and scores associated with the separation of continuous variables into discrete items were decided by specialists. Moreover, apart from ulcer surface, both indices characterised ulcers differently, through size in the SES-CD and through depth in the CDEIS. In addition, SES-CD, as defined, does not take into account the number of explored segments. It assumes that unexplored segments, for instance, because of a non-passable stenosis, do not contain lesions.

Endoscopic remission and response in CD

The thresholds of current endoscopic indices for CD to define remission or response remain unvalidated (table 2). Consequently, empirical thresholds using different indices have been proposed in various clinical trials. The GETAID proposed a definition of endoscopic remission and response using data from 562 colonoscopies in 231 patients recorded previously during the CDEIS studies and the clinical studies performed in parallel,3 ,4 ,9 ,10 published as an abstract.11 During this period, in addition to the GELS scored on a linear analogue scale as used to derive CDEIS, a global endoscopic severity was evaluated on a five-point scale, defined as no lesion at all or scarred lesions only; minor; moderate; severe and very severe. By defining endoscopic remission as minor lesions or less on the five-point scale, a CDEIS of <6 or 7 could be chosen to define endoscopic remission with a sensitivity of 86 or 92% (n=155) and a specificity of 88 or 82% (n=72). Similarly, by defining endoscopic response as a decrease of at least one grade on the five-point scale in the study of endoscopic monitoring of corticosteroid treatment (13), a decrease in CDEIS of >4 or 5 could be chosen to define endoscopic response, with a sensitivity of 78 or 70% (n=107) and a specificity of 68 or 84% (n=38).11

Table 2

Definitions of endoscopic remission and endoscopic response in Crohn's disease trials

Definition of endoscopic remission and response in CD trials

Most non-randomised or randomised controlled trials (RCTs) in CD displaying endoscopic endpoints have used unvalidated endoscopic outcome measurements (table 2). Endoscopic healing was originally used as a primary endpoint in an open-label study of methotrexate for CD.12 Following that, several trials that evaluated prednisolone9 ,13 or azathioprine14 ,15 ,19 empirically defined endoscopic remission or response with descriptive items (eg, ulcers, erythema and pseudopolyps). More recently, open-label22 and randomised trials on biological therapies introduced endoscopic items as secondary outcomes with CDEIS,12 ,17 ,18 ,22 ,26 ,28 SES-CD18 ,22 ,25or absence of ulcerations17 ,18 ,20 ,23 ,25 being the most commonly used but unvalidated outcomes.

It is noteworthy that the first prospective, placebo-controlled, double-blind study designed with mucosal healing as the primary endpoint was the Extend the Safety and Efficacy of Adalimumab through Endoscopic Healing (EXTEND) trial.24 First, the SES-CD was used to assess a minimal level of active ileocolonic inflammation to warrant inclusion in the study. The primary efficacy endpoint was mucosal healing, defined as absence of mucosal ulceration at week 12. Secondary efficacy endpoints included mucosal healing at week 52, CDEIS remission (defined as a score of ≤4) and percentage of patients achieving >75% decrease in CDEIS score from baseline at weeks 12 and 52.24 Recently as well, the Endoscopic MUcoSal Improvement in Patients with Active Crohn's Disease Treated with Certolizumab open-label clinical trial evaluated mucosal healing with tumour necrosis factor (TNF) antagonist therapy.26 The primary endpoint was the change in CDEIS at week 10. To reduce measurement bias, endoscopy videos were assessed by a blinded central reader using a prospective observer blinded endpoint study design.

Despite the spread out of currently available endoscopic scores in clinical trials, there was no attempt to further validate the thresholds chosen for response and remission. However, a validation of endoscopic activity scores in patients with CD was performed in a post hoc analysis of data from the Study of Biologic and Immunomodulator Naive Patients in Crohn's Disease (SONIC) trial.25 Data from 172 patients who participated in the SONIC trial were analysed with a central reader to determine the SES-CD and CDEIS. Mucosal healing was defined as absence of ulcers and corticosteroid-free clinical remission (CFREM) at week 50 was used as a binary classifier. Based on analyses of receiver operating characteristic curves, positive likelihood ratios (PLRs) and negative likelihood ratios (NLRs), endoscopic response was defined as a decrease from baseline in SES-CD or CDEIS of at least 50%. Mucosal healing at week 26 was associated with CFREM at week 50, with 56% sensitivity, 65% specificity, a PLR of 1.60 and an NLR of 0.67. Endoscopic response at week 26 was associated with CFREM at week 50, with 74% sensitivity, 48% specificity, a PLR of 1.42 and an NLR of 0.54.

Available endoscopic scores in CD after surgery

The only postoperative endoscopic score is that proposed by Rutgeerts et al (table 1).

Rutgeerts’ score

Definition

The endoscopic appearance of recurrent CD in the postoperative setting was described in the early 1980s in a cross-sectional study of 114 patients after ileocolic resection.28 Among them, 29/114 were examined <1 year after surgery and 72% had recurrent CD, mostly at the neoterminal ileum and anastomosis. Early endoscopic signs of recurrence were small aphthous ulcers in the neoterminal ileum. More advanced lesions described in patients examined 1–3 years after surgery consisted of larger, often serpiginous ulcers and nodular thickening of folds. In patients examined 3–10 years after the operation, the anastomosis was frequently stenosed and rigid, with large ulcers extending from the stenosis into the colon.28

Subsequently between 1979 and 1984, 89 patients were prospectively followed up after surgical resection. Ileocolonoscopy was performed within the first year after surgery.29 A scoring system empirically derived from the observations published earlier28 was used to evaluate the severity of endoscopic lesions. Ileal lesions were scored at the first ileocolonoscopy as follows:

  • i0: no lesions;

  • i1: <5 aphthous lesions;

  • i2: >5 aphthous lesions with normal mucosa between the lesions, or skip areas of larger lesions or lesions confined to the ileocolonic anastomosis (ie,<1 cm in length);

  • i3: diffuse aphthous ileitis with diffusely inflamed mucosa;

  • i4: diffuse inflammation with already larger ulcers, nodules and/or narrowing.29

The endoscopic score of the neoterminal ileum during the first year after surgery proved to be the most important variable predicting symptomatic recurrence during follow-up, even after stratification of preoperative disease activity (p=0.0001). Outcome was then analysed separately for the groups with the different endoscopic scores. The evolution was not different for scores i0 and i1, but between these groups and other groups there was a statistically different evolution. Eighty per cent (28/35) of the patients who had no or very mild lesions (i0 and i1) had unchanged lesions at 3 years. In contrast, 92% (28/35) of patients with severe lesions (i3 or i4) had a progressive or severe evolution at 3 years. A third group consisted of patients with intermediate severity of disease (i2; n=15; 18%). Some of these patients remained asymptomatic, others developed symptoms and showed progression of the lesions at subsequent endoscopy. Difference in symptomatic recurrence Kaplan–Meier curves was borderline significant between i0+i1 and i2, and borderline non-significant between i2 and i3, whereas numerous patients were lost to follow-up between 2 and 4 years (1/3 in group i0+i1 and 1/2 in group i2), when time-to-recurrence curves separated between groups.

Strengths and weaknesses

The score lacks formal validation, evaluation for inter-observer agreement and sensitivity to change. Another limitation of the Rutgeerts’ score is when the anastomosis cannot be reached. Recently, its reproducibility was tested and agreement was fair (κ 0.57) among 14 IBD specialists.6 The pragmatic appeal of the score has been demonstrated by its widespread use in clinical trials and clinical practice.

Definition of endoscopic recurrence and remission in CD trials after surgery

Following the publication of Rutgeerts’ data,29 almost all trials on postoperative recurrence included endoscopic parameters in their study design. In most RCTs evaluating 5-aminosalicylic acid,30–33 thiopurines,34–37 corticosteroids38 ,39 or antibiotics40 ,41 in the postoperative setting, a score of i2 or more was considered as recurrence of CD. In some studies, severe recurrence was defined as a score of i3 or i4.30 ,32 ,35 ,36 ,40 ,42The interleukin-10 study in 200143 and the study by Caprilli et al42 on mesalamine in 2003 defined recurrence as an endoscopy score >i0.

Recently, endoscopic recurrence defined by the Rutgeerts’ score has emerged as the standard primary study outcome in trials for postoperative therapy. In an RCT by Regueiro et al,44 which compared the efficacy of infliximab with placebo after intestinal resection for CD, endoscopic recurrence was defined by an endoscopic score of i2, i3 or i4. Savarino et al45 assessed adalimumab superiority over azathioprine and mesalamine in a postoperative RCT and used endoscopic recurrence (defined as a Rutgeerts’ score ≥i2) as a primary endpoint. However, the Rutgeerts’ score has not yet been accepted as an evaluative instrument by regulatory authorities such as the US Food and Drug Administration or the European Medicines Agency.

IOIBD recommendations for endoscopic definitions of remission and response in CD clinical trials

Methods

In order to establish a current consensus to define endoscopic response and remission in CD, we used a Delphi-like process46 of voting on the wording of proposals. The first step consisted of a careful literature review. A total of 14 specialists in the field of IBD from nine countries from Western Europe, North America and Australia participated in the process, which was based on an internet survey (see online supplementary material for methods).

Results

After the systematic literature review, 12 proposals were retained for the definition of endoscopic remission and 10 for the definition of endoscopic response in CD (table 3; see online supplementary material). For the definition of endoscopic remission in postoperative setting in CD, the only definition retained was Rutgeerts’ score 0 or 1. No definition was found in the literature for endoscopic response.

Table 3

Results of first and second rounds of the vote for endoscopic definitions of remission and response in Crohn's disease trials

First vote: ranking of definitions

The median ranking of the 12 proposed definitions of endoscopic remission in CD identified four definitions with the highest rank (figure 1 and table 3):

  • SES-CD 0–2 (median rank 3)

  • remission: CDEIS<6/complete remission: CDEIS<3 (median rank 4)

  • absence of ulceration (median rank 4.5)

  • CDEIS score of ≤4 (median rank 5).

Figure 1

Summary of the results after the first round of vote in the Delphi process for endoscopic definitions of remission in Crohn's disease (CD). A panel of 14 specialists rated 12 reviewed outcome measurements for endoscopic remission in CD trials. The four definitions with the best median rank were selected and ordered by increasing median rank for the second round of vote. CDEIS, Crohn's Disease Endoscopic Index of Severity; SES-CD, Simple Endoscopic Score for Crohn's Disease.

For the 10 proposed definitions of endoscopic response in CD, 11 out of 12 investigators ranked first a >50% decrease in SES-CD or CDEIS at the first vote, so no second round was organised (table 3).

For postoperative endoscopic remission, all specialists agreed with the proposition of a Rutgeerts’ score of 0 or 1 being the definition.

Second vote: ranking of definitions

At the second round of voting for endoscopic remission in CD, the highest median rank was the definition of an SES-CD 0–2 (median 1, mean 1.33) (table 3), with all specialists ranking this definition first or second.

Discussion

Given the well-known disconnect between clinical symptoms and endoscopic lesions,47 mucosal healing has emerged as a major therapeutic goal in clinical trials of CD, so there has been an urgent need for a consensus on measurement indices for endoscopic outcomes. The CDEIS and the SES-CD are robust validated indices and they show high reproducibility among central readers.5 ,6 The absence of validated score thresholds associated with specific prognostic values or with endoscopic healing represents a weakness of both CDEIS and SES-CD. While most thresholds for remission in trials have been chosen arbitrarily by investigators, specialists in the field of IBD agreed on a SES-CD 0–2 for the definition of endoscopic remission. The only validated thresholds have been published as an abstract for CDEIS.11 It should, however, be noted that SES-CD7 was the result of a review of the GETAID studies and was constructed by multiple linear regression as a linear combination of most reproducible independent variables, CDEIS being the dependent variable. As might be expected, the SES-CD correlated significantly with the CDEIS.8 Indeed, SES-CD was developed to meet the clinical need for a reliable, relatively easy-to-use endoscopic scoring instrument for CD. The practicability of SES-CD over CDEIS probably accounts for the specialists’ preference for SES-CD. The most relevant difference between the endoscopic variables used in CDEIS and SES-CD is the classification of ulcers divided into superficial and deep in CDEIS, but classified according to size of ulcers in SES-CD. The results of the SES-CD study7 confirmed that agreement for the definition of deep ulcers is good to very good, but is weaker for superficial ulcers. Applying a 10 cm visual analogue scale for percentages of surface involved by ulcers and by any lesions represents the most time-consuming step in the calculation of the CDEIS, as well as the one most difficult to understand for the novitiate. To avoid these measurements and conversions, the item ‘percentage of a given segment affected by ulcerations’ was included in the SES-CD. It is interesting that ‘absence of ulceration’, widely used for defining remission in clinical trials, did not reach agreement after the second round of the Delphi process. This is because a binomial (yes/no) definition cannot detect degrees of change and is both ambiguous (regarding aphthous ulcers) and likely to lack accuracy, with unknown reproducibility. Indices such as CDEIS and SES-CD turn out to be more suitable for central reading and a range 0–2 for SES-CD has the advantage of precluding the presence of ulcers.

The definition of endoscopic response >50% decrease in SES-CD or CDEIS (as proposed in the SONIC trial and validated in the post hoc analysis)25 readily achieved consensus. The SONIC trial demonstrated that patients with visible evidence of mucosal inflammation had the best clinical results with anti-TNF therapy.20 Subsequently, endoscopic assessment was used to assess a minimal level of active ileocolonic inflammation (the presence of mucosal ulceration) to warrant inclusion in the EXTEND trial.24 The presence of mucosal ulceration was defined as a minimum score of 2 on the Ulcerated Surface subscore of the SES-CD in at least one of five ileocolonic segments. Defining endoscopic entry criteria in addition to outcome measurements is already a standard practice in UC trials.48 ,49 It seems very likely to be the next step in clinical trials in CD in the era of targeting mucosal healing: CDEIS or SES-CD definitions will require additional investigations for this purpose.

Regarding endoscopic indices after surgery in CD, the Rutgeerts’ score is the only available endoscopic index but has never been validated. Its widespread use and apparently good prognostic value readily achieved consensus among specialists. As a consequence, there has been no attempt in recent years to develop new indices for postoperative endoscopic assessment for use in clinical trials. Although in the postoperative Crohn's endoscopic recurrence trial50 De Cruz et al have demonstrated that a therapeutic strategy based on the detection of endoscopic recurrence according to Rutgeerts’ score and therapy escalation is more effective than optimum drug therapy alone, the score thresholds for remission and response (≤i1) proposed in this paper still need further investigation. Since the landmark study by Rutgeerts et al,29 i0 and i1 have often been combined in subsequent studies since these patients did very well. Patients with clearly progressive disease (i3 and i4) had a poor prognosis.29 Nevertheless, there is increasing recognition that one of the major issues is whether the score of 2—as originally formulated—really reflects significant relapse or not. Not only the lesions at the anastomosis might simply reflect surgical sequelae—also the separation of individuals at risk based on the number of aphthae appears unrealistic. Indeed, in the princeps study,29 patients belonging to group i2 had a variable prognosis and some progressed, while others did not,29 but the very small number of patients (n=15) at that time render conclusions on that topic weak. The high predictive value of the score, which is its main strength, needs corroboration in the context that nowadays most patients with a score of i2 receive treatment. It is unclear whether such system can be used to evaluate the response to therapy in patients treated after an endoscopic relapse has been demonstrated after surgery. Although the score has been used to assess endoscopic response to drug intensification in clinical trials,51–53 sensitivity to change was not properly assessed and needs to be formally evaluated in a prospective study. Finally, the system might not be used to evaluate relapsing mucosal lesions resulting from therapy interruption after surgery—a scenario that might become frequent in future clinical practice. Further studies are necessary to confirm that anastomotic lesions reflect true disease relapse by histology and bring prospective assessment of whether i2 should be incorporated into score 1 and/or dismiss the system altogether and propose using the traditional ones (CDEIS or SES-CD).

Consequently, after reviewing technical aspects of available indices and using a Delphi process, IOIBD specialists have agreed that the definition of endoscopic response is >50% decrease in SES-CD or CDEIS, endoscopic remission is defined as SES-CD 0–2 and the definition of endoscopic remission after surgery for CD is a Rutgeerts’ score i0–i1. These definitions and recommendations should be subject to prospective testing in clinical trials of CD.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors LV contributed to literature search, acquisition and interpretation of data and drafting of the whole manuscript. PM, WJS, BGL ,BF ,SV, SD, GD, ML, RK, GF and ST contributed to literature search, expert panel for the Delphi process and drafting of the technical review part. JYM contributed to critical review of technical aspects, design of the voting process, analysis and interpretation of data. LP-B supervised the study. All authors contributed to the critical revising and the final approval of the manuscript.

  • Competing interests LV received fees for lectures from Abbvie, Ferring, MSD, Norgine and Takeda. PM received fees for lectures or consultancies from Abbvie, MSD and Ferring. WJS received consulting fees, lecture fees and/or research support from AbbVie, ActoGeniX, AGI Therapeutics, Alba Therapeutics Albireo, Alfa Wasserman, Amgen, AM-Pharma BV, Anaphore, Aptalis, Astellas, Athersys, Atlantic Healthcare, BioBalance, Boehringer-Ingelheim, Bristol-Myers Squibb, Celgene, Celek, Cellerix, Cerimon, ChemoCentryx, CoMentis, Cosmo Technologies, Coronado Biosciences, Cytokine Pharmasciences, Eagle, Eisai Medical Research, Elan, EnGene, Eli Lilly, Enteromedics, Exagen Diagnostics, Ferring, Flexion Therapeutics, Funxional Therapeutics, Genentech, Genzyme, Gilead, Given Imaging, GlaxoSmithKline, Human Genome Sciences, Ironwood, Janssen, KaloBios, Lexicon, Lycera, Meda, Merck & Co., Merck Research Laboratories, MerckSerono, Millennium, Nisshin Kyorin, Novo Nordisk, NPS Pharmaceuticals, Optimer, Orexigen, PDL Biopharma, Pfizer, Procter and Gamble, Prometheus Laboratories, ProtAb, Purgenesis Technologies, Receptos, Relypsa, Salient, Salix, Santarus, Shire Pharmaceuticals, Sigmoid Pharma, Sirtris (a GSK company), S.L.A. Pharma (UK), Targacept, Teva, Therakos, Tillotts, TxCell SA, UCB Pharma, Vascular Biogenics, Viamet and Warner Chilcott UK. BGL: consulting fees: Nestle Health Sciences, Takeda, Abbvie. BF received consulting fees, lecture fees and/or research support from AbbVie, ActogeniX, Albireo Pharma, Amgen, Astra Zeneca, Avaxia Biologics, Avir Pharma, Axcan, Baxter Healthcare Corp., Boehringer Ingelheim, Bristol-Myers Squibb, Calypso Biotech, Celgene, Centocor, Elan/Biogen, EnGene, Ferring Pharmaceuticals, Genentech, GiCare Pharma, Gilead, Given Imaging, GlaxoSmithKline, Ironwood Pharma, Johnson & Johnson/Janssen, Kyowa Kakko, Kirin Co, Lexicon, Lilly, Merck, Millennium, Nektar, Novartis, Novo Nordisk, Pfizer, Prometheus Therapeutics and Diagnostics, Prometheus Laboratories, Receptos, Salix Pharmaceuticals, Santarus, Sanofi, Serono, Shire Pharmaceuticals, Sigmoid Pharma, Synergy Pharma, Takeda, Teva, TiGenix, Tillotts Pharma AG, UCB Pharma, Vertex Pharma, VHsquared, Warner-Chilcott, Wyeth, Zealand Pharma and Zyngenia, and is a member of the Board of Directors of Roberts Clinical Trials. SV received consulting fees and/or grant/research support from, and/or been a speaker for, AbbVie, Centocor, Ferring, Genentech/Roche, MSD, Novartis, Pfizer, Shire Pharmaceuticals, Takeda, and UCB Pharma. Silvio Danese has served as a speaker, consultant and advisory board member for Abbott, Actelion, Alpha Wasserman, AstraZeneca, Cellerix, Cosmo Pharmaceuticals, Ferring, Genentech, Grunenthal, Johnson & Johnson, Merck, Millennium Takeda, Novo Nordisk, Nycomed, Pfizer, Pharmacosmos, Schering-Plough, UCB and Vifor. GD has served as speaker, consultant and/or principal investigator for AbbVie, Centocor, MSD, Pfizer, UCB, Takeda, TEVA, Millenium, Jansen, Boehringer Ingelheim, Elan, Ferring, Dr. Falk Pharma, Shire, Salix, Receptos, Cosmo, AstraZeneca, Vifor, Tillotts, Otsuka, Photopill, Given Imaging, GSK, PDL, Amgen, AM Pharma, Galapagos, Versant, Novonordisk, Norgine, Giuliani, Celgene, Celltrion, Mitsubishi, Setpoint, Hospira, and Tigenix. Mark Löwenberg has served as speaker and/or principal investigator for Abbvie, Dr. Falk, Ferring Pharmaceuticals, Merck Sharp & Dohme, Receptos, Takeda and Tramedico. He has received research grants from AbbVie, Merck Sharp & Dohme and Achmea healthcare. RK has received speaker's fees from Takeda. Gionata Fiorino served as a consultant and member of Advisory Boards for Takeda Pharmaceuticals, AbbVie, MSD, and Janssen Pharmaceuticals. ST received consulting fees from AbbVie, Asahi-Kasei, Bristol-Myers Squibb, Coronado Biosciences, Cosmo Technologies, Ferring Pharmaceuticals, Genentech, Genzyme Corp., GlaxoSmithKline, Janssen, Lexicon Pharmaceuticals, Merck Research Laboratories, Millennium Pharmaceuticals, Nisshin Kyorin Pharmaceutical Co., , Novartis, Novo Nordisk A/S, NPS Pharmaceuticals, PDL BioPharma, Pfizer, Procter and Gamble, Santarus (a wholly owned subsidiary of Salix Pharmaceuticals), Schering Plough, Shire, Sigmoid Pharma, Tillotts Pharma AG, TxCell SA, UCB and Warner Chilcott UK; he has received research grants from AbbVie, Genentech, GlaxoSmithKline, Janssen, Novartis, Pfizer, Procter and Gamble, Shire and UCB; and he has received payments for lectures/speakers bureau participation from AbbVie, Ferring Pharmaceuticals, Janssen and Warner Chilcott UK. JYM has received lecture and consultation fees from AbbVie and Janssen and consultation fees from Pharmacosmo. LP-B has received consulting fees from Abbott, BMS, Boehringer Ingelheim, Celltrion, Ferring, Genentech, Hospira, Janssen, Lilly, Merck, Mitsubishi, Norgine, Pharmacosmos, Pilège, Shire, Takeda, Therakos, Tillots, UCB and Vifor, and has received lecture fees from Abbott, Ferring, HAC-pharma, Janssen, Merck, Norgine, Therakos, Tillots and Vifor.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.