Predicting Relapse of Ulcerative Colitis With Gastrointestinal Ultrasound

INTRODUCTION AND RATIONALE

1.1 Introduction This protocol describes PRELAPSE, an explorative, prospective, longitudinal observational study of diagnostic accuracy comparing trans-abdominal gastrointestinal ultrasound (GIUS) with ileocolonoscopy for the assessment of disease activity in patients with ulcerative colitis (UC) and with regards relapse as an outcome.

1.2 Background Ulcerative colitis is an inflammatory bowel disease characterized by chronic inflammation mostly limited to the colon mucosa. The diagnostic "gold standard" is ileocolonoscopy and histologic examination of the biopsies. The disease has a variable cause and objective diagnostic criteria are needed to evaluate disease activity (1, 2). Although, the disease is mild in many patients some require biologic treatment with tumor necrosis alpha receptor inhibitors (anti-TNFs). These are expensive drugs with potentially serious adverse effects and patients should have repeated examinations to evaluate treatment effect. Although ileocolonoscopy currently is the best method for objectively assessing disease activity both bowel preparation and the examination causes patients considerable discomfort (3, 4). Therefore, GIUS for the evaluation of disease activity could be a good alternative for treatment follow up in these patients.

In ulcerative colitis the term relapse is commonly used for a flare of symptoms in patients previously in clinical remission. It can either occur in patient without ongoing treatment or during maintenance treatment as a loss of treatment effect (5). Although increased stool frequency, rectal bleeding and abnormal mucosa found by sigmoidoscopy were considered important in defining relapse, a consensus was not reached in the definition in the 2nd ECCO guidelines from 2012 (5). However, it was suggested that the criteria for relapse should be predefined in clinical studies. Commonly, a combination of clinical scores and an endoscopic score has been used for assessment of relapse in clinical trials (6).

The rate of relapse after discontinuing anti-TNF treatment varies between 20-50% during the first year (7-9). For patients receiving treatment with anti-TNF the rates of relapse is about 10-30% (10, 11). Thus the overall risk of relapse in both populations is quite high. A strategy for detecting relapse in the patients in an objective manner without discomfort and lower cost and risk is warranted.

GIUS is a rapid, efficient, non-invasive and cheap imaging technique. It can be used for diagnosing IBD and for determining the extent and location of disease (12). Longitudinal studies have shown that GIUS can be used for monitoring the effect of therapy in IBD patients (13-16) and is well accepted by the patients as a method for monitoring (17). However, only a few studies have been performed that compare GIUS with endoscopy (13, 18-21). There is only one prospective study looking at the detection of relapse and this suggest increased flow in the inferior mesenteric artery associated with increased risk of relapse (22). Due to the variability and the difficulty in obtaining these measurements they are clinically of little use, however (23). A study using endoscopic ultrasound showed that oedema in rectal mucosa was associated with increased risk of relapse (24).

In previous studies from our research group we have examined patients with familial GUCY2C diarrhea syndrome (FGDS) (n=35) caused by an activating mutation encoding for guanylate cyclase C (GC-C) receptor on the brush border of the intestinal cells (25, 26). This activation leads to increased intracellular cGMP and to increased secretion of Cl- and HCO3- leading to increased fluid intraluminal in gut. The GC-C receptor is activated by the endogene hormones guanylin (GN) and uroguanylin (UGN), which can be measured as the prohormones ProGN and ProUGN in plasma (27, 28).

The majority of the FGDS patients have diarrhea (median 4 per day) and susceptibility to inflammatory bowel disease (20 % have Crohn's Disease). Unpublished results from our group (von Volkmann et al) show that plasma levels of ProGN and ProUGN are significantly lower in patients with FGDS and CD compared to a healthy age and gender matched group. Furthermore, the investigators have found that the number of diarrhea in CD patients correlated negatively with plasma levels of ProGN while there were no correlation between guanylins and inflammatory markers such as fecal Calprotectin and CRP.

These findings may suggest a link between the endogenous hormone regulation of guanylins and diarrhea. The investigators speculate that decreased secretion of guanylins could counter diarrhea, and therefore this study should evaluate if these peptides may be of importance also in UC patients.

1.3 Rational for the study and purpose The study will include UC patients whom have been on maintenance anti-TNF therapy for at one year or more and in clinical remission for the 3 past months at least. The patients will be included from the same patient population included in the Biostop study (Protocol ID no: HMR2016-0.6 and EudraCT no: 2016-001409-18) in two centers, Haukeland University Hospital and Ålesund Hospital. The infrastructure for recruiting these patients has already been established and the parameters collected in the BIOSTOP study can be accessed in the electronic patient journal and used in the PRELAPSE study. These parameters include endoscopy, repeated simple symptom scoring (6-point Mayo score) (29-32), blood chemistry, faecal calprotectin (33-36) and histology.

The proposed study is explorative and aimed at finding ultrasound parameters that can be used to predict the risk of relapse in patients in endoscopic remission and early predictors of disease relapse during follow up.

As in several previous important studies (3, 11, 37-40) mucosal healing is defined as a Mayo endoscopic subscore (MES) <2 and thus the outcome relapse is defined as MES ≥ 2. As some argue that patients with a MES =0 fare better than those with MES=1 (41, 42) and others that histological remission should be used as a reference (43-45), the investigators will also examine these as exploratory outcomes (6, 35, 45-48).

As mentioned above, monitoring by repeated symptom scoring and faecal calprotectin testing will be evaluation tools before performing unscheduled rectosigmoidoscopy to confirm early stages of true relapse defined as Mayo endoscopic score (MES) ≥2. At the discretion of the responsible physician and in agreement with the patient, anti-inflammatory treatment will then be intensified without delay as described in the BIOSTOP study. In the PRELAPSE study the follow up will stop if a patient reaches the primary end-point in the 2 year follow up period.

In the BIOSTOP study patients will be randomized to either continue or discontinue the anti-TNF treatment at inclusion time. In PRELAPSE all patients will be included regardless of which group they are randomized to. Differences between the groups with regards to ultrasound appearance of relapse will be examined as an exploratory objective.

2. STUDY OBJECTIVES

Primary objective:

To study if GIUS can predict endoscopic relapse at follow up examinations in a group of patients with ulcerative colitis in sustained clinical and endoscopic remission

Secondary objectives:

- To assess if GIUS or individual US parameters at baseline can predict endoscopic and clinical relapse

- To assess inter-observer variation of GIUS

Exploratory objectives:

- To assess if GIUS or individual US parameters at baseline can predict strict endoscopic relapse

- To assess if GIUS or individual US parameters at baseline can predict histological relapse

- To assess if GIUS or individual US parameters at relapse is different in patients with and without maintenance treatment with anti-TNFs.

- To assess if GIUS or individual US parameters correlate with factors indicating relapse such as histology score, CRP, calprotectin

- To assess if prouroguanylin or proguanylin can be used to predict endoscopic relapse

3. STUDY ENDPOINTS

Primary Endpoints:

• Endoscopic relapse defined as a MAYO endoscopic subscore ≥2 during a two year follow up period

Secondary endpoints:

- Endoscopic relapse defined as a MAYO endoscopic subscore ≥2 during a two year follow up period

- Clinical relapse defined as 6. Point MAYO score ≥ 2

- Kappa statistics for inter-observer variation of patient classification into the relapse/no relapse categories

- Bland Altman statistics of individual ultrasound parameters and a combined ultrasound score.

STUDY DESIGN

3.1 Overview of study design The PRELAPSE study is an exploratory, prospective, longitudinal, multicentre trial that investigates the applicability of GIUS for the prediction of disease relapse in UC patients compared to colonoscopy as the reference standard. 40 patients will be included, total inclusion time will be 2 years.

Patients with an established diagnosis of ulcerative colitis that have been treated for a minimum of one year with anti-TNF maintenance therapy and that have been in clinical remission during the last 3 months are potential study patients. Eligibility criteria are described in section 5.3 (Inclusion criteria) and 5.4 (Exclusion criteria) as in the BIOSTOP-study. Finally, only patients who are included in BIOSTOP are eligible for inclusion in PRELAPSE.

Both patients discontinuing anti-TNF treatment and patients continuing treatment will be included. The patients will follow the same scheduled and unscheduled regime as described in the BIOSTOP protocol (Included)

Study period: Estimated first patient enrolled: September 30th. 2017 Anticipated recruitment period: September 30th. 2017 to 31stDes2019 Estimated date of last patients completed 31stDes 2021

Duration of study follow up: 104 weeks

Endpoint evaluation 104 weeks

3.2 Location The study will be performed at the department medicine in Haukeland University Hospital in Bergen and Ålesund Hospital, Ålesund.

4. STUDY POPULATION

4.1 Population (base) Adult male and female patients with an established diagnosis of ulcerative colitis treated for minimum 1 year with anti-TNF maintenance therapy and in clinical remission during the last 3 months. Patients will be included from two hospitals in Norway (Haukeland University Hospital in Bergen and Ålesund Hospital in Ålesund.).

4.2 Number of patients Forty patients are planned for inclusion in this study. The inclusion period is estimated to last for two years.

5. TREATMENT

Since all the patients in the PRELAPSE study also will participate in the BIOSTOP study they will follow the same treatment regime as they do BIOSTOP study. No treatment interventions will be performed as result of participation in the PRELAPSE trial.

6. STUDY PROCEDURES

6.2 Study visits The regular study visits will be carried out by study nurse and study investigator every 3 months the first year. The study investigator will perform the GIUS while the study nurse performes the other tasks related to the study. After 12 months the frequency of visits will bed reduced to one every 6 months. During the 2 years (104±4 weeks) study period there will be 7 regular study visits, one baseline visit including ileocolonoscopy, and 1 scheduled rectosigmoidoscopy at study end visit at 2 years. The assessments performed at different visits are presented in section 8. Prior to inclusion, assumed eligible patients will receive a letter with information about the study.

6.2.1 Screening evaluation In the BIOPSTOP study patients with ulcerative colitis who fulfill the criteria below and who fulfills the main inclusion criteria, can be scheduled for baseline visit, which should be carried out after two consecutive faecal calprotectin within the inclusion limits. The time frame between screening and baseline visit should not be longer than 4 weeks.

- 6-point MAYO score ≤ 1

- Two consecutive faecal calprotectin tests < 200mg/kg (Calpro)

6.2.2 Baseline visit

After signing the informed consent the following assessments will be done in accordance with the BIOSTOP-study as follows:

1. A "baseline ileocolonoscopy" A. If the Mayo endoscopic score (MES) is 0 or 1, the patient can be included in the study Biopsies from rectum will be sent to a to central laboratory for histological assessment (formalin fixed).

B. If the Mayo endoscopic score (MES) is > 1, the patient can not be randomized. The patients will be treated by responsible physician according to local and international guidelines.

2. Fasting plasma for ProGN and ProUGN as well as serum samples for local laboratory analyses (CRP, Hb, platelet count, leucocyte count, and ferritin will be obtained.)

3. Study nurse/investigator assessments:

i. Demographics (sex, birth date and ethnic origin) ii. Clinical status (medical history including comorbidity, physical examination and vital signs, smoking history) iii. IBD history (age at diagnosis, time from diagnosis to anti-TNF start, Montreal classification, Mayo endoscopic score at diagnosis, previous antiinflammatory UC treatment, need for anti-TNF dose escalation to achieve remission) iv. Concomitant medication (immunomodulator, 5-aminosalisylates, other)

4. Review of inclusion/exclusion criteria

6.2.3 Baseline gastrointestinal ultrasound Performed within 14 days before bowel preparation before ilecolonoscopy starts or between 2-7 days after the ilecolonoscopy by study investigator. Ultrasound cannot be performed in relation to to bowel preparation as this theoretically could cause mucosal edema confounding the results.

6.2.4 Regular visits

Scheduled visit by study nurse in BIOSTOP study:

- 6-point Mayo scoring

- Collection of blood and fecal samples

- Fasting ProGN and ProUGN

- Registration of concomitant medication of any kind according to a patient self-report-diary

Scheduled examination in PRELAPSE study performed on the same day:

• Gastrointestinal ultrasound by study investigator

Laboratory samples for clinical chemistry and faecal analyses will be collected preferably on the same day and not more than 7 days before or after every regular visit .

For patients on infliximab additional blood samples for drug concentration and anti drug antibody measurements will be collected within 3 days before every scheduled treatment dose. For patients on subcutaneously administered anti-TNFs samples for drug concentration and antibody measurements will be collected at regular visits every 3 months.

After two years a scheduled rectosigmoidoscopy with biopsy specimens are performed in all patients still in the study.

6.2.5 Unscheduled visit in case of disease worsening If the patient, study nurse or investigator suspects disease worsening between or at regular visits, an unscheduled visit including an additional faecal calprotectin test should be organized without delay.

6.2.6 End of follow up period and end of study visit The end of the study visit will be performed at or near week 104 ±4 (2 years) or when relapse occurs.

6.2.7 After End Follow-up Further treatment will be decided by the treating physician. 6.3 Criteria Defining Relapse 6.3.1 Indication for Unscheduled GIUS and Endoscopy

- 6-point Mayo score ≥2

- Two consecutive faecal calprotectin measurements >200mg/kg

- At the discretion of investigator or patient In order to verify disease relapse an unscheduled rectosigmoidoscopy will be performed without delay. Disease relapse in this study is defined as Mayo endoscopic (MES) ≥2. GIUS will be performed within 7 days before preparation for the sigmoidoscopy or 2 days after and before a change in medical treatment is performed.

6.4 Procedures for relapse treatment If relapse is found the patient follow up in the PRELAPSE study will stop. Ongoing treatment is dependent if the patient had continued the anti-TNF treatment at the start of the study or not. Patient stopping anti-TNF treatment at study start will follow the treatment algorithm as described in the BIOSTOP study. Patients who continued anti-TNF treatment at study start, who during follow up years of the study period are experiencing an endoscopically verified relapse, are considered to be treatment failures and will be treated on an individual basis at the discretion of the responsible physician.

6.5 Criteria for patient discontinuation

Patients may be discontinued from the study at any time. Specific reasons for discontinuing a patient from this study are:

- Voluntary discontinuation by the patient who is at any time free to discontinue his/her participation in the study, without prejudice to further treatment

- Major protocol deviation

- Incorrect enrolment i.e., the patient does not meet the required inclusion/exclusion criteria for the study

- Patient lost to follow-up

6.6 Procedures for discontinuation 6.6.1 Patient Discontinuation Patients have the right to withdraw from the study at any time for any reason. In the case that a patient decides to prematurely withdraw from the study, he or she should be asked if they can still be contacted for further information, so that a final evaluation can be made with an explanation of why the patient is withdrawing from the study, including assessment of possible adverse events. Although a subject is not obliged to give his or her reason(s) for withdrawing prematurely from a trial, the investigator should make a reasonable effort to ascertain the reason(s), while fully respecting the subject's rights.

If a patient withdraws from the study, the investigators should make an effort (if possible) to do all assessment as if it was "the end of study at 2 years" - see flow trial flow chart for details.

6.6.2 Treatment discontinuation Patients who for some reason (e.g. pregnancy, reactions to treatment, surgical interventions or other safety issues) must stop with anti TNF treatment during the study will be asked to continue follow-up in the study.

6.6.3 Trial discontinuation

The whole trial may be discontinued at the discretion of the PI in the event of any of the following:

- Medical or ethical reasons affecting the continued performance of the trial

- Difficulties in the recruitment of patients The principal investigator will inform all investigators, the relevant Competent Authorities and Ethics Committees of the termination of the trial along with the reasons for such action. If the study is terminated early on grounds of safety, the Competent Authorities and Ethics Committees will be informed within 15 days

6.7 Laboratory tests Blood samples for clinical chemistry analyses will be collected and, depending on availability, analysed at the local laboratory according to hospital procedures. Faecal samples for calprotectin will be referred to a central laboratory for analysis (CALPRO AS). Blood samples for measurements of drug concentrations and anti-drug antibodies, as well of additional blood will be sent to a central laboratory for analysis and biobanking in accordance with the BIOSTOP study. However, copies from the results of these analyses will be sent to both hospitals in PRELAPSE study and made available to the primary investigators in the electronic patient journal.

7. ASSESSMENTS 7.1 Efficacy assessments 7.1.1 General Efficacy Assessments Inflammation assessment by biochemical parameters

The following laboratory tests and faecal analyses will be recorded at all regular visits according to the BIOSTOP study:

- Blood chemistry

- Haemoglobin, CRP, ferritin (analysed at local laboratories)

- Drug levels anti-TNF and antibodies

- Faecal analyses: Calprotectin measurement (analysed at central laboratory, CALPRO AS).

The following laboratory test is exclusive for the PRELAPSE study

- ProGN and ProUGN (analysed at Haukeland University Hospital) 7.1.2 Disease Specific Efficacy Assessment

- Mayo Score The Mayo score (49) consists of four components (rectal bleeding, stool frequency, physician rating of disease activity, and mucosal appearance at endoscopy) rated from 0-3 (Mayo endoscopic subscore, Appendix 14.2) that are summed to give a total score that ranges from 0-12. The non-invasive 6-point Mayo score (29-32) does not require either endoscopy or physician's rating, and thereby ranges from 0-6 (see appendix 14.3).

Clinical remission is defined as a 6-point Mayo score of ≤ 1

7.1.3 Determination of guanylines in EDTA plasma 3 ml EDTA full blood has to be obtained in a fasting state and thereafter centrifuged at 1800 x g, 4 °C for 10 minutes. Plasma samples are then frozen in aliquots of 250 µl and stored at -80 °C until analyzing. Plasma proGN and proUGN concentrations are determined by using sandwich enzyme-linked immunoassay (ELISA) kits (BioVendor, Karasek, Czech Republic).

In Ålesund, the plasma aliquots can be stored temporary at -80°C until an appropriate number of samples are collected and will be sent on dry ice to Haukeland University Hospital in Bergen for long time storage and analyzes.

7.1.4 Assessment with gastrointestinal ultrasound The main study parameter of this study is a simplified ultrasound score based on a combination of bowel wall thickness (BWT) and detection of vessels in the thickest area of the colon (See below and appendix 14.3).

Secondary study parameters of this study are as seen with GIUS:

Colon wall thickness Colon wall vascularity Colon wall layer stratification Colon haustrations Colon visibility Colonic content Relative submucosal thickness Lymph nodes Bowel wall elasticity Free fluid

Changes in these features are typical in ulcerative colitis and have been identified in previous studies (13, 18, 19). The examinations will be performed by all collaboratorating physicians in the study.

A Logiq E9 ultrasound machine (GE, Milwaukee, USA) will be used with the following ultrasound transducers: C1-5 and C1-6 curved array and 9L linear array. The C1-5 will be used for overview of the abdomen and the 9L for detailed examination of the GI wall.

Frequency, focus and gain settings are optimised to get the best images in the patient. The examination will be performed with the patient in the supine position. The large intestine will be scanned by following its course from the terminal ileum to the rectum. All segments (ileum, right colon, transverse colon, descending colon, sigmoid colon and rectum) will be scanned in longitudinal and transverse sections. At least one brief cine loop (<5 seconds) will be obtained from each bowel segment, in each plane and annotated with text and/or external markers.

Colon wall thickness will be measured from, but not including, the central hyperechoic line of the lumen to the end of the outer hypoechoic margin of the wall (representing the muscularis propria). Cut off between normal and pathological bowel is defined as 3 mm except in the rectum where the cut off is 4 mm.

In areas of affected bowel (bowel wall thickness >3 mm) colour Doppler measurements will be made. The rectum will not be included as it is very deeply located. Each area scanned on the B-mode exam is also examined with colour Doppler. Cine loops of 5 seconds are saved. The colour Doppler measurements are performed with standardised presets on the transducer with a velocity scale of 5 cm/s for detection of vessels with slow flow in the GI wall. Vessel density with colour Doppler will be scored as 0=No or only single visible vessel, 1=scattered, spotted vessel signals, 2= scattered, elongated vessel signals, 3 confluent, elongated vessel signals.

The ultrasound score has been adopted from Parente et al (19), but adjusted and is defined as follows:

Ultrasound score: Findings on GIUS 0 BTW<3 mm

1. BWT=3-4.9 mm and colour Doppler score 1 BWT=2-2.9 and colour Doppler score 2-3

2. BWT=3-4.9 mm and colour Doppler score 3 BWT=3-7.9 mm and colour Doppler score 2 BWT≥ 5-7.9 mm and colour Doppler score 1 BWT≥ 8 mm and colour Doppler score 0

3. BWT≥5-7.9 and colour Doppler score 3 BWT≥ 8 if colour Doppler score 1-3

The diagnostic accuracy of the ultrasound score will be investigated both for the cut off ≥1 and ≥2.

Normally five wall layers can be seen in the GI wall with high frequency ultrasound. In the anterior wall of the bowel and labelled from the lumen, the first layer is hyperechoic and the interface echo between the luminal content and the mucosa. The second layer is hypoechoic corresponding partially to the mucosa. The third layer is hyperechoic corresponding partially to the submucosa. The fourth layer is hypoechoic corresponding partially to the proper muscle. Finally, the fifth layer is hyperechoic corresponding to the interface echo between the serosa and whichever structure is adjacent. If all layers are present, there is normal stratification. Focal and diffuse loss of haustrations in the GI wall will be registered per segment.

Presence or absence of haustrations, free fluid in the abdomen, colonic contents and normal (short axis <5mm) or enlarged lymph nodes (short axis ≥5mm) in the mesentery will also be registered per segment.

Elasticity of the bowel wall will be measured in kilopascal using shear wave elastography in the subset of patients at Haukeland University Hospital (50, 51).

All registered parameters are shown in the form "Ultrasound parameters" (Appendix14.3).

7.1.5 Inter-observer assessment of images and cine loops Measurements on recorded images and cine loops will be performed by two observers for 20 patients randomly selected using appropriate software for viewing and editing DICOM files.

7.1.6 Blinding The ultrasound examinations will be performed by an investigator unaware of the endoscopic results. At the time of endoscopy the performing gastroenterologist will be unaware of the results of the ultrasound examination.

7.2 Assessment of Disease worsening A disease worsening in ulcerative colitis is defined as an increase in 6-point Mayo score to a total of ≥ 2 points and/or faecal calprotectin > 200 mg/kg in 2 consecutive measurements regardless of symptoms.

7.3 Biobank samples The plasma obtained for ProGN and ProUGN will be stored in a biobank.

7.4 Other assessments

The following samples will be collected as a part of the BIOSTOP study at inclusion and 2 years, and at unscheduled endoscopy visits for patients that relapse during the study:

Biopsy specimens

- Samples from rectum mucosa (8 specimens per endoscopy)

- Routine white light histologic interpretation and classification by local pathologist

Medical history will be carefully assessed and information on the duration of ulcerative colitis, Montreal classification (Appendix 14.1) concomitant disease, previous medical and surgical treatment, age, sex weight, height will be collected. At baseline and after 2 years pulse rate and blood pressure will be measured.

10. STATISTICAL ANALYSIS AND DATA ANALYSIS

10.1 Determination of sample size: This is an exploratory study and the sample size is based on the primary end-point which is relapse at ileocolonoscopy defines as MES score ≥2 during the study period. Since approximately half of the patients will receive anti TNF treatment and the other half will not approximately 1/3 of the patients are expected to relapse during the follow up. This will be sufficient for examining if the ultrasound score is useful for detecting relapse in this patient group.

10.2 Statistical analysis:

Concordance between the Mayo endoscopic subscore (MES) and the GIUS score at the different time points will be determined using kappa statistics. Sensitivity, specificity and accuracy for predicting relapse will also be calculated for the suggested ultrasound score. To investigate how the individual ultrasound parameters contribute stepwise logistic regression will be performed with the MES a dependent variable.

For inter-observer variability of the measurements, kappa statistics will be performed for categorical data while Bland Altmann analysis will be performed for continuous variables.

Analysis of correlation between the various ultrasound parameters, endoscopic findings as scored with the Mayo endoscopic subscore and biochemical markers will be performed.