Novel biomarker profiles to improve individual diagnosis and prognosis in patients with suspected inflammatory bowel disease: protocol for the Nordic inception cohort study (NORDTREAT)

Introduction

Inflammatory bowel disease (IBD), primarily ulcerative colitis (UC) and Crohn’s disease (CD) are idiopathic disorders characterised by chronic gastrointestinal inflammation and relapsing symptoms like diarrhoea, abdominal pain and fatigue. A large proportion of the patients experience severe disease manifestations like bowel obstruction, fistulising disease, liver disease and colorectal cancer. Additionally, extraintestinal manifestations occur in 15%–40%.1 Consequently, hospitalisations, surgical interventions, recurring sick leave and life-long medication severely impact the lives of patients and families. Even though the incidence of IBD appears to have reached a plateau in most Western European countries, it is rapidly increasing in many newly industrialised countries.2 The prevalence of IBD has increased substantially in the last three decades, exceeding 0.5% in many European countries.2–4 An estimated 1.3 million people in Europe suffer from IBD, imposing a substantial health, social and economic burden.1 There is a need to improve the management of IBD.

Diagnosis of IBD can be cumbersome.5–7 Furthermore, IBD is a heterogeneous disease and its course differs substantially between patients. The lack of vigorous biomarkers for individual disease characterisation potentially leads to delayed diagnosis, organ damage, worsened outcome, increased mortality and an ever-increasing disease burden.8 Personalised medicine (PM) could be a suitable strategy for optimising both diagnostics and treatment of IBD.9 The vision of the International Consortium for Personalised Medicine (ICPerMed) describes ‘an approach to medicine that integrates an individual’s characteristics for early disease diagnosis, prognosis, optimal choice of treatment, accurate disease risk estimation, and targeted prevention’.10 In gastrointestinal diseases, PM is based on advancements in biomarker detection in a variety of omics-platforms, including genomics, transcriptomics, microbiomics, proteomics and metabolomics, as well as body imaging methods, for example, CT scans and MRI scans. Recent advances led to the identification of several potential biomarkers for diagnosis and treatment response in IBD.11–17 However, they are not in use in clinical practice. Even though early diagnosis and efficient treatment strategies for IBD are cornerstones for reducing irreversible organ damage, identifying biomarkers that can efficiently support clinical decision-making is still warranted.

This protocol article introduces a prospective clinical study that aims to identify novel diagnostic biomarkers for more accurate early diagnosis of IBD and assess the utility of novel prognostic biomarkers in clinical outcomes with various disease manifestations. Notably, the study evaluates biomarkers in a relevant clinical setting, that is, among patients referred to the hospital on suspicion of IBD and diagnosed according to the ECCO criteria.5

Aims

This study aims, on the one hand, to improve accuracy in diagnosing IBD in the form of a diagnostic test accuracy study. The main hypothesis is that biochemical markers can differentiate individuals referred to a gastroenterology department with suspected IBD into individuals with (IBD) and without (non-IBD) diagnosis. The primary aim is to identify molecular biomarkers (eg, from blood and stools) for discriminating between individuals diagnosed with IBD (CD, UC and unclassified inflammatory bowel disease (IBD-U)) and non-IBD. Secondary aims include investigating whether clinical, genetic, microbiome and lifestyle information and combinations thereof (eg, gene-environment interaction analyses) are associated with IBD diagnosis.

On the other hand, the main hypothesis of the prognostic utility study is that prognostic factors can distinguish different pathways of disease progression. The primary aim of this study is to identify prognostic biomarkers that can define different IBD disease courses. The secondary aims are to evaluate whether including information on clinical and lifestyle factors and combinations thereof can improve prognostic value.

Methods and analyses

Design

The NORDTREAT cohort study is a Nordic multicentre collaboration on diagnostic factors and prognostic factors of IBD among patients referred to the hospital with suspected IBD. All referred patients will be carefully evaluated to ensure that patients with suspected IBD are invited to participate in the study at their first clinic visit. Informed consent will be obtained at the screening visit (week −4). After study material collection (ie, blood, stool, biopsies, hair, urine, questionnaires and patient-reported outcome measures (PROMs)) and diagnostic evaluation, IBD diagnosis will be established or rejected at the baseline visit (week 0). IBD patients will be followed until week 52 with three additional visits (weeks 12, 26 and 52), whereas non-IBD patients will complete questionnaires at baseline visit and week 52 (figure 1).

Figure 1
Figure 1

Schedule for patients with a suspected IBD diagnosis.

The prospective study consists of a cross-sectional part with focus on diagnosis and a longitudinal cohort study (52 weeks) with focus on prognosis (ie, a prognostic utility study). The inclusion of patients into the NORDTREAT cohort study will be reported as a flowchart (figure 2).

Figure 2
Figure 2

Flowchart for patients assessed eligible for inclusion in the NORDTREAT cohort study.

The objective of the diagnostic test accuracy study is to determine the sensitivity and specificity of biomarkers of interest (baseline measurements, see Biomarkers section) to diagnose IBD when applying the ECCO diagnostic criteria as the reference standard.5

The objective of the prognostic utility study is to examine whether the biomarkers of interest predict severe disease outcomes in patients with IBD. Severe IBD is defined as those experiencing IBD-related surgery, IBD-related hospitalisation (except planned procedures) or IBD-related death within 52 weeks of inclusion.

Setting

This Nordic multicentre study is a collaboration between gastroenterology departments in 11 Nordic hospitals (online supplemental appendix 1). The study enrolment will occur between 7 February 2022 and 31 December 2023. The first patient’s first visit was conducted in Oslo University Hospital in February 2022, and the first patient’s last visit was in February 2023. All inclusion sites will conduct their last patient’s first visit and last patient’s last visit in December 2023 and December 2024, respectively.

Supplemental material

Patient characteristics and eligibility criteria

Inclusion criteria: patients≥18 years of age referred to a gastroenterology department with suspected IBD who give written informed consent to participate in the study. Exclusion criteria: patients with a previous diagnosis of CD, UC or IBD-U or patients unable to provide informed consent or comply with protocol requirements, for example, due to recreational drug or alcohol misuse.

The clinical data and lifestyle information collected for the study will be presented in tabular form (see table 1) and include the following.

Clinical interview

Table 1

Overview of clinical and lifestyle information collected according to visits (weeks after diagnosis)

Demographics (sex, age, marital status, educational level, occupational status and smoking), disease history (previous disease including debut of gastrointestinal symptoms) and surgery (abdominal and perianal), occurrence of intestinal complications (defined as strictures and fistulas), occurrence of perianal disease, extraintestinal manifestations, medication (past and present), family history (of IBD and colorectal cancer) and at the week 52 visit: relapses and surgery (abdominal and perianal) are recorded.

Clinical examination

Includes a perianal examination of study subjects with perianal CD (accounting for the number of draining and non-draining perianal fistula).

Vital signs

Include measurement of height, weight, heart rate, blood pressure, and temperature.

Endoscopy

Standardised scoring systems will record macroscopic findings from the ileocolonoscopy: for patients with UC, the Endoscopic Mayo Score per segment is applied, that is, ascending, transverse, descending, sigmoid colon and rectum.18 For patients with CD, the Simple Endoscopic Score for Crohn’s Disease (SES-CD) per segment is applied, that is, terminal ileum, ascending, transverse, left colon and rectum.19 Patients diagnosed with IBD-U at baseline will be assessed with the Endoscopic Mayo Score per segment and the SES-CD per segment. The endoscopy is recorded on video (approximately 20 s per segment). Furthermore, standard routine biopsies from the ileum, ascending, transverse, descending colon, and rectum will be obtained.

Medical records

Routine clinical information, clinical findings and procedures (endoscopic and surgical), test results (ie, histology, radiology (including small bowel MRI, capsule endoscopy, CT scan and ultrasonography), microbiology and biochemistry), diagnosis, treatment and complications including hospitalisations will be retrieved from the medical records.

Montreal classification

Based on results from the routine examinations, the study subjects will be classified according to the internationally accepted IBD classification system, the Montreal Classification.20

Lifestyle information, Disease Activity Index, disease questionnaires and PROMs

Participants will complete validated questionnaires (on paper or electronically) on disease activity, quality of life, including EQ-5D-5L
21 (measure of, eg, mobility, personal care, usual activities, pain/discomfort and anxiety/depression), SF-36
22 (including perception of health, limitations in daily and social activities, general mental health (psychological distress and well-being), and vitality (energy and fatigue) within the last 4 weeks), IBDQ
23 24 (disease impact on patients daily functioning and quality of life, eg, abdominal pain, sleep, incontinence, ability to work and conduct social activities within the last 2 weeks), IBD-F
25 (disease-specific questionnaire comprising five questions about the severity of fatigue (part 1) and ROME IV
26 (The Rome IV Diagnostic Questionnaire for Functional Gastrointestinal Disorders in Adults). The electronic case report form (eCRF) platform Viedoc (Viedoc Technologies, Uppsala Sweden) is applied, and in Denmark REDCap (Vanderbilt University, Nashville, TN, USA) is used for diet questionnaires.

Before each visit, patients with IBD will register clinical disease activity, Harvey-Bradshaw Index (HBI)27 28 (diary for patients with CD including two clinical parameters: abdominal pain and number of liquid or soft stools), or Partial Mayo Score (diary for patients with UC or IBD-U including the two non-invasive components of the full Mayo Score (stool frequency and rectal bleeding)). All participants will perform diet questionnaires at baseline and week 52.29

Biological material

Blood, biopsies, faeces, hair and urine will be collected at the scheduled visits (table 2).

At baseline, all study subjects have blood samples (whole blood (4 mL), plasma (10 mL), serum (20 mL) and Paxgene blood RNA system (2,5 mL)) and biopsies collected (eight biopsies from predefined locations collected in Allprotect reagent and six biopsies from predefined locations snap frozen in liquid nitrogen or FlashFREEZE (Milestone Medical, product no. 100 780 from Axlab A/S, DK), ensuring a standardised freezing procedure at −80°C). Furthermore, the faecal samples collected at the patient’s home will consist of: one stool collection tube with RNALater, DNA stabiliser and a dry sample, respectively. Additionally, a tuft of hair and urine (8mL) will be collected.

Table 2

Overview of the biological material collected in the study

Blood samples will not be collected on the day of the ileocolonoscopy, and faecal samples will not be collected during the colonic cleansing to minimise any potential influence from the standard cleansing procedure conducted before an endoscopy.

On follow-up visits at week 12 and week 26, all patients diagnosed with IBD will have blood samples (plasma (10 mL) and serum (20 mL)) and faecal samples (one stool collection tube with RNALater, DNA stabiliser and a dry sample, respectively) collected. At the week 52 visit for IBD patients, biological material similar to baseline will be collected (except for whole blood). The biological material will be stored at local facilities according to national procedures, and accordingly available for future research projects on approval of the NORDTREAT executive committee. Standard operating procedures with detailed information can be found in online supplemental appendix 2.

Supplemental material

Primary endpoint(s)

The primary endpoint for the cross-sectional diagnostic test accuracy study is diagnosis according to the reference standard from ECCO criteria,5 that is, IBD (including CD, UC, IBD-U) or non-IBD (patients not diagnosed with IBD) at referral.

The primary endpoint for the longitudinal prognostic utility study is a composite outcome measure. We will include multiple (three mutually associated) endpoint components combined as a Boolean function; this will be our primary outcome measure derived from randomised trials while also anticipating that the approach will be associated with increased statistical efficiency (eg, power). The composite outcome consists of the proportion of patients who experience severe IBD over 52 weeks after inclusion, defined as either IBD-related surgery, IBD-related hospitalisation (except planned procedures) or IBD-related death.

Key secondary endpoints in the prognostic utility study

Major secondary outcomes assessed and analysed 52 weeks from baseline will include IBD-related surgery (number (no.) in %), IBD-related hospitalisation (no., %), IBD-related deaths (no., %) as well as non-corticosteroid-free (local acting systemic administered (budesonide) and systemic steroids (prednisolone)) clinical remission at week 12.30 Clinical remission in CD is defined as patient-reported outcome 2 (abdominal pain≤1 and stool frequency≤3) or Harvey Bradshaw Index<5. Clinical remission in UC is defined as patient-reported outcome 2 (rectal bleeding=0 and stool frequency=0) or Partial Mayo score (<3 and no score>1). Non-corticosteroid-free endoscopic remission is assessed at week 52,30 where endoscopic remission in CD is defined as Simple endoscopic score<3 points or absence of ulcerations (eg, SES-CD ulceration subscore=0). Endoscopic remission in UC is defined as Mayo endoscopic score=0 points or UC endoscopic index of severity≤1 point. Disease impact on patient’s life will be assessed by the 32-item version of Inflammatory Bowel Disease Questionnaire (IBDQ) score (number (%) ≤ 170) and the numerical change of the SF36 subscales.

Other secondary endpoints in the prognostic utility study

The composite secondary outcome is the proportion of patients who experience IBD complications within 52 weeks after inclusion, defined as the cumulative number of patients who experience one of the following (1) need for medication (immunomodulators, biologics or steroids) between diagnosis and 1 year, (2) need for surgery between diagnosis and 1 year, (3) need for hospitalisation between diagnosis and 1 year or (4) relapse between diagnosis and 1 year. Furthermore, the cumulative number of flares, adverse drug reactions, numerical change in the Fatigue Questionnaire, and cumulative glucocorticoid exposure will be assessed. Faecal calprotectin>250 µg/g (no., %), time to start of biological therapy, time to IBD-related surgery or hospitalisation and CD-related surgery (stenosis) will also be measured.

Mid-term and long-term complications in the prognostic utility study

IBD-related hospitalisations will be assessed and defined as a composite endpoint combining the total number of IBD-related hospitalisations (including readmissions) and the cumulative length of hospital stay. Bowel damage in patients with CD (Lémann index for patients with an MRI), UC-related surgery (colectomy), CD-related surgery (no., %), IBD-related hospitalisations (no., %) and change in disease extent in UC (with a focus on macroscopic proximal disease extent) will be assessed among the mid-term complications. Among the long-term complications, the following three endpoints will be measured: gastrointestinal and extraintestinal dysplasia (no., %), cancer (gastrointestinal and others in no., %) and mortality (no., %).

Biomarkers

At the time of submission, the NORDTREAT executive committee has approved projects on the following topics in relation to diagnosis and prognosis (including inflammatory intensity, disease extend, tissue damage, tissue repair activity and severe disease course): extracellular matrix proteins, microbiome, mycobiome, faecal markers, genetic risk factors, metabolomics antibodies towards microbial and autoantigens.

Statistical methods

Referring to earlier studies, approximately 40% of a population referred with suspected IBD will have the diagnosis confirmed according to the ECCO criteria (or comparable).5 31 Cost and practical issues will restrict the sample size to approximately 800 patients, with an expected 80% participation rate.

In the diagnostic test accuracy study, we will compare two independent binomial proportions, using the likelihood ratio (LR) statistic with a χ2 approximation and two-sided significance level (0.05). With an expected sample size of 640 participants (assuming a biomarker positivity-to-negativity ratio of 1 to 1), we will reach an approximate statistical power of 100%, assuming that the proportion of patients diagnosed with IBD is 80% in the biomarker positive group and 20% in the biomarker negative group. If we include 640 study participants, we will potentially (‘rule of thumb’) be able to build a multivariable statistical model with as many as 64 covariates. Thus, we will examine a variety of biochemical variables for diagnosing IBD (see Biomarkers section). The expectation is that a proportion of 100–200 preliminary biomarkers will be evaluated in this model to distinguish non-diagnostic markers and preliminary biomarkers that can be replicated. When applying a significance level of 0.05, each biomarker has a 5% risk of being replicated by chance. Thus, in the evaluation of, for example, 100–200 biomarkers, five to 10 biomarkers may be replicated by chance. In contrast, the risk of not replicating ‘true’ biomarkers with low effects is estimated to be approximately 20%. All replicated biomarkers will be built into a prediction model that will be optimised and evaluated by cross-validation. The primary hypothesis is that the biomarker signature will be able to characterise the participant population into an IBD and a non-IBD group.

In the prognostic utility study, we expect a sample size of 256 IBD patients and a 10% event rate for severe IBD within 52 weeks.32 33 If we assume a 15% event rate for the biomarker positive group and a 5% event rate for the biomarker negative group, and anticipate an even 1:1 biomarker positive-to-negative distribution, approximately 80 patients in each group will increase the statistical power above the 90% threshold.

Comparison of measures of diagnostic accuracy

The basic measures of quantifying the diagnostic accuracy of a test will be the diagnostic OR, sensitivity and specificity measures. From these measures, we will calculate the LR, the ratio of the probability of the index test result among patients who have IBD to the probability of the same test among patients who do not have IBD (non-IBD). Since all diagnostic tests will be performed on each patient, paired data results and methods that account for the correlated binary outcomes are necessary (eg, McNemar’s test).

Prognostic factor research (the prognostic utility study)

The primary statistical analysis will be based on the proportion of participants experiencing severe IBD among biomarker-positive versus biomarker-negative patients. The efficacy of the prognostic biomarker(s) will be considered as the proportion of the patients with different disease outcomes among the biomarker-positive patients compared with the biomarker-negative patients. The associations of the suspected important molecular biomarkers with other variables will be tested with non-parametric tests: with Spearman’s rank correlation (rs) for continuous variables and the Wilcoxon rank-sum test or Kruskal-Wallis test, including a Wilcoxon-type test for trend across ordered groups where appropriate, for categorical variables.

Logistic regression models will be utilised with individual markers as the exposure variables and composite clinical severity as the outcome (dependent variable). The analyses will be adjusted simultaneously for sex, age and prescribed therapy. Potential interactions between biomarker status (positive/negative) and specific drug type will be evaluated too. Covariates comprise of various measures within genetics, transcriptomics, microbiomes and proteomics (see Biomarkers section). We will attempt to develop prediction models containing several conventional risk factors (ie, age, sex and type of drug therapy) with or without any novel biomarkers and calculate improvements in predictive ability using risk discrimination and reclassification measures. We will use a two-stage approach that allows for the examination of centre heterogeneity through the calculation of the C-index, a measure of risk discrimination and changes within each centre before pooling results.

We will attempt to develop multivariable prediction models, relating multiple predictors for a particular individual to the probability of or risk for the presence (diagnosis of IBD at baseline) or future occurrence (prognosis) of a particular outcome, such as intestinal surgery within the first year. Predictors such as biomarkers are covariates, explored as prognostic or predictive factors. Models will be built using regularised regression methods, such as sparse Partial Least Squares (PLS) regression or Least Absolute Shrinkage and Selection Operator (LASSO), as well as tree-based machine learning approaches, such as random forest, or non-linear non-symbolic deep learning approaches, capitalising on available published data sets. Also, multiple data sets will be used for integrative analyses, to explore if combinations or interactions of variables of different kinds could aid predictive or prognostic potential of our models.

Patient and public involvement

The NORDTREAT project has evolved in collaboration with patient research partners (PRPs) from the Danish Colitis-Crohn’s Association, the Norwegian Mage-tarmforbundet, the Swedish Mag- och tarmförbundet and from Icelandic patient representatives ensuring implementation of patient perspectives in aspects such as, for example, research aims and outcomes. The PRPs were involved before the formal project start (adding value in the design of the study) and they are continuously informed on study progress. Overall, the PRPs add value in the project conduct through their engagement in the Patient Advisory Council (online supplemental appendix 3) increasing knowledge-sharing, awareness on the project and research comprehensiveness.

Supplemental material

When reporting the diagnostic test accuracy study, we will adhere to the STARD guideline34 and the TRIPOD statement,35 and when reporting the prognostic utility study, we will apply the STROBE36 as well as the TRIPOD statement.35 Before any analyses are initiated, an elaborated statistical analysis plan indicating explicit figures and tables will be provided.

Project organisation

The NORDTREAT partnership builds on extensive professional experience, a strong track record of collaborative clinical projects and participation in a number of European and global consortia (eg, the IBD Character Consortium, the SYSCID Consortium and the International IBD Genetic Consortium).

The NORDTREAT cohort study is conducted in parallel and as part of the NORDTREAT trial (a biomarker-stratified randomised controlled trial (RCT)). The aim of the NORDTREAT RCT is to demonstrate that personalised therapy can be applied to patients with IBD, by assessing if a top-down treatment can improve clinical outcomes in IBD patients with an increased risk of poor disease course (https://clinicaltrials.gov/ct2/show/NCT05180175).

The operational organisation of the NORDTREAT study is highlighted in online supplemental appendix 3. The Steering Committee (SC), whose members are responsible for planning and organising the study and scientific follow-up, includes the leaders of the six predefined work packages (WPs): WP1: Clinical phenotyping and biobanking (Professor Marte Lie Høivik), WP2: NORDTREAT trial (Professor Jonas Halfvarson), WP3: Host and gut microbiome signatures (NORDTREAT cohort) (Professor Vibeke Andersen), WP4: Bioinformatics and integrated analyses (Professor Dirk Repsilber), WP5: Exploitation and dissemination (Professor Johan D Söderholm) and WP6: Project management (Professor Jonas Halfvarson). Elected SC members also include Professor Jens Kjeldsen, Professor Lóa G Davíðsdóttir, Professor Einar S Björnsson, Professor Asle W Medhus, associate Professor Johannes R Hov, consultant PhD Vendel Kristensen, associate Professor Randi Opheim, PhD Joachim Høg Mortensen, PhD fellow David Füchtbauer, and PhD fellow Anja Fejrskov. The Steering Committee conducts regular meetings and has feedback from the External Advisory Board composed of internationally renowned representatives of academia and the biomedical industry: Professor Jack Satsangi, Oxford University (GB), Professor Philip Rosenstiel, Kiel University (DE), Professor Mauro D’Amato, Karolinska Institute (SE), vice president at Q-linea Mats Gullberg (SE) and from PRPs in the national patient organisations (Patient Advisory Council (online supplemental appendix 3)).

Finally, the Executive Committee (members listed in online supplemental appendix 3) is responsible for project coordination and management during the research project’s preparation, operation and finalisation. Biweekly online meetings are conducted, ad-hoc meetings when necessary and face-to-face meetings are held twice a year.

Perspectives

Substantial differences exist between patients suspected of IBD when comparing symptoms, onset, prognosis and treatment, which makes a one-size-fits-all approach insufficient. With current knowledge and clinical methodologies, the diagnosis and prognosis of IBD is challenging without invasive investigations. However, newly identified biomarkers may have implications for future disease management, as they could be used to differentiate patients with IBD and patients with other comparable gastrointestinal symptoms or predict disease outcomes. If we succeed in identifying diagnostic or prognostic biomarkers of IBD, substantial improvement will be given to IBD treatment, including reduced side-effects and complications, increased quality of life for patients and their relatives and a reduced socioeconomic burden on society. Relevant findings with potential clinical value will be evaluated and replicated in other prospective cohorts if possible.

Ethics and dissemination

The study is approved by the Ethics Committee (DK: S-20200051, v1.4, 16.10.2021; IS: VSNb2021070006/03.01, NO: 193064; SE: DNR 2021-05090) and the Danish Data Protecting Agency (20/54594). The procedures followed are in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975 with later amendments.

The target journal for the primary outcome will be a gastroenterology medical journal. Subsequently, other hypotheses will be tested, and manuscripts will be prepared with the intention of submitting to specialised journals in microbiome research, immunology, genetics, nutrition and gastroenterology. In addition to the scientific reporting, key findings with translational implications will be communicated to health professionals, patient organisations, public health policymakers and the public through various media and news activities. When disseminating the results, we will apply the recommendations of the International Committee of Medical Journal Editors (ICMJE).

This post was originally published on https://bmjopen.bmj.com