Identifying a Sleep Screener for Disease Relapse in Children With Inflammatory Bowel Disease

Hypothesis: In children with IBD, high scores for a sleep disturbance screener will be positively associated with IBD relapse

Project Design and Specific Objective: This is a single centre pediatric observational prospective study to develop a non-invasive non-costly tool to screen for relapses in pediatric IBD patients through examining the association between sleep disturbances and disease relapse in children with IBD. The outcomes are

Primary outcome: Cut-off score of a sleep screener in relation to disease relapse (as diagnosed by fecal calprotectin value of >100 microgram/gram of stools) in children with IBD Secondary outcomes

1. Correlation between sleep disturbance scores and clinical disease activity indices (PCDAI and PUCAI), after controlling for symptoms of anxiety and depression
2. Identification of which sleep component such as sleep duration and sleep latency is the best at detecting a disease relapse.
3. Identification of whether sleep disturbance more accurately predicts relapse for CD than for UC, after controlling for symptoms of anxiety and depression.
4. Identifying any differences in intestinal microbiom between those with sleep disturbances versus those without.

Experimental Methods and Analysis

1 Design: Observational prospective study 2. Inclusion Criteria 2.1. Children aged 8 -17 years under the care of the Pediatric IBD Program at the Children's Hospital, Winnipeg and with an established diagnosis of IBD according to the North American Society of Pediatric Gastroenterology (NASPGHAN) 3. Exclusion Criteria 3.1. Children with IBD and known cognitive dysfunction or global developmental delay 3.2. Children with IBD and concurrent gastrointestinal infection 4 Measures

• 4.1. A standardized sleep diary will assess sleep in the past week. Scores will be averaged over the 7-day recording period. The sleep diary will provide information about latency to fall asleep, number of awakenings, duration of awakenings, total sleep time, sleep quality, and sleep efficiency. To complement the sleep diary, the Children's Report of Sleep Patterns, a 62-item questionnaire, will assess self-reported sleep patterns, sleep hygiene, and sleep disturbance in children ages 8-17 years. It has acceptable reliability and validity.
• 4.2. Disease relapse will be measured using fecal calprorectin (F Cal) and the PCDAI for CD and PUCAI for UC.The quantum blue single point of care test (ALPCO) will be used
• 4.3. The disease phenotype, duration, base-line characteristics, patient demographics, smoking habits, medications, and initiation of new medications or other treatments for IBD during the course of the study will be obtained by chart review using an information form
• 4.4. The Spence Children's Anxiety Scale will be used to assess anxiety symptoms (e.g., panic attacks, social phobia, fear of physical injuries, generalized anxiety). With clinical cut-offs, the scale has demonstrated high internal consistency and good test-retest reliability over a six-month period. Evidence of convergent and divergent validity has been reported. As parent and youth report can differ, both self-report and parent-report versions of the SCAS will be administered.
• 4.5. The Child Depression Inventory -II (short-form) is a 12-item scale of depressive symptoms for youth aged 7 to 17 years. It represents one of the most widely used measures of depression in this age group, has excellent reliability, concurrent, construct, and predictive validity. Both self-report and parent-report will be collected.

1. Study procedure: Families with children aged 8 -17 years, who are under the care of the Pediatric IBD Program at the Children's Hospital, Winnipeg, and who have an established diagnosis of IBD will be approached via letters to inform them about the study and study procedures. The letters will give the option to patients and their care givers to declare their lack of interest in participation if they do not want to be approached or participate in the study. Patients and parents/guardians/care givers will choose whether or not to provide "permission to be contacted via phone". This will occur through a phone number provided in the letter, and they will be instructed to call within 1 week of receiving the letter to declare if they decline to be approached. If lack of interest is not declared, the Research Assistant will call the potential participants to explain the study procedure, obtain a verbal consent for participation and ask participants to complete a sleep diary to assess sleep disturbances for one week prior to the clinic visit. During the clinic visit, verbal consent will be confirmed and informed assent/consent will be obtained by the children/parents (guardians) and the completed sleep diary will be collected. During the same visit, fecal samples will be collected for FCAL measurement as a surrogate marker for mucosal inflammation. A fecal sample will be laso sent for microbiome analysis. Other investigations will include blood samples for serum hemoglobin, serum albumin, and inflammatory markers (CRP and ESR) and if participants have diarrhea, stool samples for infection screen will be collected to exclude any possibility for gastrointestinal infection. In addition, participants and their parents will complete the Spence Child Anxiety Scale, the Child Depression Inventory (parent and child report), and the Children's Report of Sleep Patterns (CRSP)(child report). It is expected that administration of these measures during clinic will take 45 minutes. The Research Assistant will help participants who need any clarification on how to respond to any question. Children will also be clinically assessed and clinical disease activity will be measured using PCDAI for CD and PUCAI for UC. Each participant will have all measures repeated at a second elective clinic visit 3 months later during the course of the study period (9 months of recruitment). Patients are normally instructed to call our office if they start developing symptoms of relapse such as diarrhoea and abdominal pain. Patients then will be asked over the phone to complete a sleep diary and to bring a stool sample for fecal calprotectin. Stool samples for infection screen will be also collected and an emergency clinic visit for clinical assessment of the disease and routine blood tests including blood samples for measuring inflammatory markers will be planned within a week of having symptoms of relapse. Completed sleep diary will be collected during the visit.
2. Analysis and sample size To determine whether sleep disturbance is associated with disease relapse, logistic regression analysis will be conducted. Logistic regression will be used as the goal of the analysis is to predict the category of outcome for individual patients (relapsed, not relapsed). As symptoms of anxiety and depression are often comorbid with sleep disturbance, we will use these as covariates in the logistic regression analysis.

To determine the diagnostic performance of the sleep measures, or the level of sleep disturbance associated with disease relapse, ROC will be analyzed using MedCalc. MedCalc is a reliable and valid method of assessing the diagnostic validity of laboratory tests. ROC curve analysis methodology will be used to plot curves representing the true positive (TPR) and false positive rates (FPR) for sleep at different cut-off points. ROC curves will be used in this investigation to determine an optimal cutoff level for the sleep measures and to contrast the relative adequacy of measures to discriminate between participants who relapsed versus those who did not. The MedCalc program assumes an underlying bi-normal distribution and produces a non-parametric estimation of the area under the curve. Results from the PCDAI and PUCAI will be the reference standard for the test. Information pertaining to test, specificity, positive predictive value (PPV, and negative predictive value (NPV will be obtained. Such an analysis will produce a range of sensitivity and specificity values associated with different cut-off points for each of the sleep variables and the MedCalc program chooses a cutoff point that maximizes both sensitivity and specificity. Using a corrected z-test, analogous to a paired t-test, results will show whether there is a statistically significant difference between the sleep predictors in terms of the overall accuracy of measurement.

Based on existing studies of sleep disturbances in IBD, sleep disturbances are assumed to be prevalent in 80% of patients with active IBD compared to 40% of those in remission. To achieve an acceptable power of .80 with alpha = .05, a sample of 90 patients will be recruited. This number is highly feasible to recruit within the proposed clinic's patient flow and where research participation rates exceed 80% in the current and past research projects in the pediatric IBD program. Currently the IBD program cares for over 150 children with IBD with over 130 patients > 8 years old.

Descriptive analyses of the variables will be performed comparing disease type (CD/UC) and disease status (inactive (F Cal < 100 microgram/gram stools /active (F Cal > 100 microgram/gram of stool)). Comparisons of disease subtypes and active/inactive subgroups will be performed using two-tailed t-tests for continuous variables, and chi-square/Fisher exact tests of association for categorical variables.