Development of the Harvey-Bradshaw Index-pro (HBI-PRO) Score to Assess Endoscopic Disease Activity in Crohn’s Disease (2024)

,

Article history

Abstract

1. Introduction

Management of Crohn’s disease [CD] has conventionally been based on symptoms and subjective reporting, but this approach frequently fails to adequately reflect the state of disease activity at the mucosal level.^1–4 Studies have shown that traditional clinical indices such as the Harvey-Bradshaw index [HBI] and the Crohn’s disease activity index [CDAI],^5,6 are non-specific and do not correlate well with endoscopic activity.^4,7 A significant proportion of patients in symptomatic remission have evidence of active mucosal inflammation when a colonoscopy is performed.¹

Recently, the approach to improving outcomes in CD has focused on healing the mucosa and avoiding bowel damage. Treatment towards these targets is thought to result in better outcomes and alteration of the natural history of CD.⁸ However, objective assessment of mucosal inflammation and bowel damage typically require invasive and costly procedures such as endoscopy and magnetic resonance/computed tomography enterography [MRE/CTE]. Moreover, there is increasing movement away from reliance on clinical indices such as CDAI and HBI by regulatory authorities in order to improve the ability to assess therapy responses. Of potential increasing value are self-administered questionnaires containing items that are answerable solely by patients, referred to as patient-reported outcomes [PROs], in addition to more objective measures of disease activity such as endoscopy and cross-sectional imaging.^9–12 In contrast to CD, the Mayo score is more reliable and better reflects actual disease activity in ulcerative colitis [UC], in part because of the incorporation of endoscopic scoring and the inclusion of components such as the physician global assessment [PGA] as a physician [clinician]-reported outcome [PRO_c]. Previous studies have evaluated non-invasive scores in UC which included PRO components where patients were asked whether they thought they were in remission at each visit.^13,14

Current integrated scores in CD are complicated, not user-friendly and lack validation. There is an unmet need for improved, less-invasive disease activity indices that provide a representative assessment of endoscopic disease activity which include patient-reported outcomes [PRO_p], clinical symptoms and physician [clinician]-reported outcomes [PRO_c]. In this study, we describe the development of a new clinical score for CD, the HBI-PRO, that incorporates the HBI with PRO_p and PRO_c and we assessed its correlation with endoscopic disease activity in CD using the Simple Endoscopic Score for CD [SES-CD] as the measure of endoscopic disease activity.

2. Methods

2.1. Subjects

A random selection of 88 CD patients between 18 and 75 years of age, attending for a colonoscopy for various indications with confirmed disease based on standard clinical, radiological, endoscopic and histological criteria, and who were managed at Mount Sinai Hospital in Toronto were recruited to this study. We included only patients with colonic CD [Montreal classification¹⁵ L2], terminal ileum CD [L1] or ileo-colonic CD [L3] that could be evaluated by ileo-colonoscopy. Patients with comorbidities such as cancer, acute or chronic enteric infection [e.g. Clostridium difficile], UC, and patients receiving concomitant non-steroidal anti-inflammatory drugs [NSAIDS] were excluded. We also excluded individuals with dominant fibrostenotic disease with obstructive symptoms, CD patients with proximal small bowel disease and/or with upper gastrointestinal CD [L4] and/or perianal disease [P]. Pre-endoscopic assessments were conducted, including collection of HBI components, and the PRO_p and PRO_c questions. The PRO_p question was designed to assess the patients’ perception of their current disease activity: ‘Has your CD been active or inactive during the past two weeks?’ Only ‘active’ or ‘inactive’ were accepted as possible answers. The PRO_c was administered by the treating clinician prior to the colonoscopy based upon the routine pre-endoscopy history in response to the question, ‘If you had to scope your patient today, Do you think your patient would have had endoscopically inactive, mild, moderate or severe disease?’. Blood samples for C-reactive protein [CRP] evaluation were also collected on the day of endoscopy. The study was approved by the research ethics board of Mount Sinai Hospital, and all subjects provided written informed consent to participate.

2.2. Endoscopic assessment and clinical scoring

Colonoscopies were video recorded and two endoscopists experienced in inflammatory bowel disease reviewed each procedure independently to provide disease activity assessment. Only concordant observations were included in the analysis. The tools used were the SES-CD for CD subjects and the Rutgeerts score for post-operative CD subjects. Active endoscopic disease was defined as SES-CD ≥ 3¹⁶ or Rutgeerts¹⁷ > i1. Specifically, an SES-CD score of 0–3 was defined as inactive disease, 4–6 mild disease, 7–16 moderate disease and >17 severe active disease. These categories were used to define disease activity in an ordered logistic regression model.¹⁶ The endoscopists were blinded to the pre-endoscopic assessments by HBI-PRO and CRP levels. Only subjects with complete ileocolonoscopic examinations were included in this study. The categorical HBI score was also used as an independent variable in all analyses. Specifically, HBI < 5 was defined as clinical remission, HBI between 5 and 7 as mild disease, HBI between 8 and 16 as moderate disease, and HBI > 16 as severe disease.^5,7,18

2.3. Statistical analysis

Basic demographic and clinical characteristics of the cohort were presented in the form of mean and standard deviation for continuous variables and frequency and percentage of the total for categorical variables. Receiver operating characteristic [ROC] curves were generated for several models using logistic regression. In each model, a binary endoscopic score was treated as the dependent variable while a few permutations of independent variables were explored. Different models were compared to one another using a chi-squared test. We also applied an ordered logistic regression model in which the full categorical endoscopic score was used as a dependent variable in order to determine optimal cut-offs for the linear predictors for clinical remission, mild/moderate/severe clinical disease according to the HBI-PRO total score. All statistical analyses were completed in STATA 11.1.

3. Results

3.1. Patient characteristics

A total of 93 subjects were recruited between 2014 and 2015. Five patients were excluded because of non-concordance between the two endoscopists’ SES-CD assessment score. A total of 88 subjects were thus eligible to the study of whom 40 were being followed for possible recurrence after ileal resection [post-operative] [Table 1].

In the non-operated cohort, 33 of the 48 [68.8%] patients were male and the mean age was 29.6 years [SD = 11.1 years]. Mean CRP was 7.8 mg/l [SD = 10.5 mg/l] and 17 [40.5%] patients had CRP levels > 5 mg/l. Thirty-nine [81.3%] patients were in clinical remission with HBI < 5, while nine [18.8%] patients had HBI ≥ 5. Based on the PRO_c, 30 [62.5%] patients were deemed inactive and 18 [37.5%] as having active disease. Based on the PRO_p, 28 [58.3%] were inactive and 20 [41.7%] were active. Within the non-operated group, 27 [56.3%] patients had active endoscopic disease with SES-CD ≥ 3.¹⁶

In the post-operative CD group, 24 [60%] patients were male and the mean age was 38.0 years [SD = 15.3 years]. Mean CRP was 4.1 mg/l [SD = 6.5 mg/l]. Thirty [75%] of the patients in this group were in clinical remission with HBI < 5, while 10 [25%] had HBI ≥ 5. Twenty-four [60%] patients were categorized as inactive based on the PRO_c question. Significant post-operative endoscopic recurrence with Rutgeerts > i1 was found in 21 [52.5%] patients.^17,19

3.2. HBI-PRO integrated score results for CD subjects

In the non-operative CD cohort, the HBI alone did not correlate well with SES-CD score (area under the curve [AUC] = 0.54) [Figure 1]. In addition, PRO_p and PRO_c individually were not reflective of the SES-CD score [AUC = 0.70 and AUC = 0.67, respectively, p = NS]. Combining PRO_c and PRO_p into a single score did not perform better than the HBI alone [AUC = 0.73, p = 0.09]. However, combining PRO_c, PRO_p and the HBI [HBI-PRO] significantly improved the accuracy of the model over the HBI alone [AUC = 0.78, p = 0.002] [Figure 2]. In addition, the HBI-PRO model combined with CRP increased the accuracy even further [AUC = 0.88, p = 0.0001] [Figure 3], but was not significantly better than the HBI-PRO without CRP [p = 0.09] [Table 2]. When a dichotomized endoscopic score was replaced with a categorical classification in the SES-CD as the outcome variable in the HBI-PRO and HBI-PRO combined with CRP models [Table 3], a score range could be determined for inactive, mild, moderate and severe disease. Specifically, for the HBI-PRO model, scores between 0 and 0.47 corresponded to inactive disease, between 0.47 and 1.01 to mild disease, between 1.01 and 3.75 to moderate disease, and greater than 3.75 to severe disease. The respective cut-offs for the HBI-PRO plus CRP model were 0–1.15 for inactive, 1.15–1.77 for mild, 1.77–5.08 for moderate and greater than 5.08 for severe disease.

3.3. HBI-PRO integrated score results for post-op CD subjects

In the post-op CD group, the HBI alone did not correlate well with the Rutgeerts score [AUC = 0.52]. In addition, the PRO_p and PRO_c also correlated poorly with the endoscopic score individually [AUC = 0.53 and AUC = 0.63, respectively]. Combining the PRO_c, PRO_p and HBI in one model did not significantly improve its predictive ability [AUC = 0.64, AUC = 0.65 and AUC = 0.61, respectively]. The addition of CRP to the model also had no tangible effect.

4. Discussion

In this study we developed a new clinical score for CD, the HBI-PRO, which incorporates a PRO_p, a PRO_c and the HBI with the option to include CRP levels. The HBI-PRO is both more specific and more sensitive to identifying endoscopic activity than the HBI, CRP, PRO_p or PRO_c alone. Our study clearly demonstrates that the HBI alone is not a reliable, non-invasive clinical score as a proxy for endoscopic disease activity in a routine clinical setting. In contrast to CD, clinical symptoms correlate better with endoscopic disease activity in UC.^14,20 Furthermore, the Mayo clinical score integrates both clinical symptoms and a physician-reported outcome. Clinical indices in CD, such as the HBI or CDAI, are not wholly reflective of disease activity and mucosal inflammation.^21–23 Unfortunately, most of our disease treatment paradigms and clinical trial outcomes have been based on studies that rely on such clinical indices to measure outcomes. Treatment strategies based solely on clinical manifestations have failed to modify the course of IBD.^1–4

It is well known that a significant proportion of patients in symptomatic remission have evidence of active mucosal inflammation when a colonoscopy is performed. Importantly, it has been shown that active mucosal inflammation is one of the best predictors of clinical relapse, disease progression and complications in IBD patients.^8,24 Major caveats of the current approach of using endoscopy to monitor outcome are the difficulty in implementing it in day-to-day clinical practice and research due to high costs, risks, time demand and, last but not least, patient inconvenience. There is a real need for better, less-invasive disease activity indices that provide a representative assessment of endoscopic disease activity. The United States Food and Drug Administration [FDA] is moving away from the CDAI and HBI to PROs and objective measures of disease such as endoscopy.^9,10 Studies in other chronic conditions, such as rheumatoid arthritis [RA],²⁵ multiple sclerosis [MS]^26,27 and myelofibrosis²⁸ have already progressed towards PRO questionnaires designed to be completed by patients to measure their perceptions of functional status and well-being. PROs have also begun to find a role in national audits and registers and there is rapidly growing interest in their potential to enable personalized medicine. PROs are also becoming important end points for drug regulatory approval. Since the early 1980s, many scores have been developed for measuring PROs in RA and MS. In the last decade, a few irritable bowel disease [IBD] research groups have tried to develop PRO questionnaires. For example, the International Programme to Develop New Indexes for Crohn’s Disease [IPNIC] group has developed the IBD disability index, which includes 19 questions related to quality of life, depression, sleep, etc.²⁹ In addition, Budger et al. in the UK have developed and validated the IBD-control questionnaire, which incorporates 14 questions measuring the patient’s perception of disease control, benefit of treatment, quality of life, depression and pain.³⁰ Nevertheless, these PRO questionnaires, among others, are not yet widely accepted and have not been fully validated. Development of an official PRO index and its subsequent validation in IBD according to FDA guidelines will probably take a few years.³¹ In this study we proposed a new, simple, short and non-invasive clinical score, the HBI-PRO, which combines the established HBI clinical score with patient- and physician-reported outcome in one question for each. Although the HBI-PRO score was not developed according to FDA guidelines, it could prove very useful in the next few years in clinical practice and research as a more reliable measure of endoscopic disease activity than traditional scoring systems.

The simplicity of the HBI-PRO score, with or without CRP levels, lies in its non-invasive nature and the ease with which it could be obtained by a simple, short patient interview. Even though neither the HBI nor CRP was able to discriminate between inactive and active disease, combining the HBI and PRO_p with PRO_c greatly improved the model’s fidelity.

Our pilot study had several limitations. Firstly, it is acknowledged that the sample size was small and we lacked a validation cohort separate from the cohort used to develop the HBI-PRO. Moreover, in the post-op CD cohort, the HBI-PRO did not significantly improve the model, probably because we used a Rutgeerts score of >i1 to define endoscopy activity although this cut-off was developed to assess recurrence and prognosis and has not been formally validated. Moreover, the HBI-PRO was used in this study to reflect ileocolonic CD as we excluded proximal small bowel CD and perianal disease and used endoscopic assessment as the comparator.

In summary, notwithstanding the limitations outlined, this study demonstrates that the HBI-PRO is a new tool that can easily be incorporated into clinical practice and serve as a useful surrogate for endoscopic disease activity in ileocolonic CD without prior surgery, and can both be used by practising physicians as well as for clinical research.

Funding

None.

Conflict of Interest

EZ, BK, OBK, RM, none. GCN, ad-hoc advisory board with Janssen and Abbvie. KC received educational grants from Janssen, Abbvie and Takeda and has served on advisory boards for Abbvie and Takeda. AHS received research grants from Abbvie, Amgen and Pfizer, Millennium Honorarium for Educational Event Presentations from Abbvie, Janssen, Takeda, and Shire Advisory Board for Abbvie, Actavis, Janssen, Takeda, Pharmascience, Shire. MSS received research support and consulting fees from Janssen, Abbvie, Takeda and Prometheus.

Author Contributions

EZ, MSS: study concept and design; acquisition of data; analysis; data interpretation; drafting of the manuscript; critical revision of the manuscript for important intellectual content. BK: statistical analysis, critical revision of the manuscript for important intellectual content. OBK, RM: drafting of the manuscript; critical revision of the manuscript for important intellectual content. GCN, KC, AHS: acquisition of data and critical review of the manuscript. EZ and MSS accept full responsibility for the conduct of the study.

References

Author notes

Corresponding author: Eran Zittan, MD, Mount Sinai Hospital, Zane Cohen Centre for Digestive Diseases, Division of Gastroenterology, Department of Medicine, University of Toronto, Toronto, 600 University Avenue, Toronto, Ontario, M5G 1X5, Canada. E-mail: eranzittan@gmail.com

Copyright © 2016 European Crohn’s and Colitis Organisation (ECCO). Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com

Issue Section:

Original Article

Email alerts

Citing articles via

More from Oxford Academic