Introduction

Inflammatory Bowel Disease (IBD) is a chronic and relapsing, remitting inflammatory disorder affecting the gastrointestinal tract. The two major phenotypes of IBD, Ulcerative Colitis (UC) and Crohn’s Disease, affect young adults in their productive years, imposing a significant challenge to an individual’s well-being and a significant burden on the healthcare system [1,2]. Along with disease-related symptoms, IBD is also associated with a higher prevalence of anxiety and depression[3], and decreased quality of life (QoL) [4,5]. Depression and anxiety in those with IBD have been linked to increased mortality [6], decreased quality of life [7], higher likelihood of clinical recurrence, increased hospitalization rates, and decreased treatment compliance[8].

There is inconclusive evidence of the impact physical activity has on IBD disease activity [9,10]. Physical activity is defined as any bodily movement that substantially increases energy expenditure [11]. Some studies have suggested higher levels of physical activity may be associated with more favorable psychological health outcomes in those with IBD [12,13]. A recent narrative review described how greater time spent in physical activity may also be associated with decreased IBD prevalence, decreased IBD symptoms [14], and possibly, maintenance of disease remission [15]. There is evidence that suggests individuals with IBD (both active and in remission) are less physically active than healthy controls [9], with some patients noticing physical activity causing positive changes in disease activity while others say the opposite [10]. One study reported prolonged moderateintensity walking did not have any harmful effects on inflammatory markers in those with IBD [16].

The impact time spent sitting, sleeping, or standing has on IBD disease activity has not been evaluated. However, in adults with type 2 diabetes, accelerometer-measured sedentary time was associated with increased levels of CRP in women but not men (β = 0.24, p = 0.04, β = 0.05, p = 0.54, respectively) [17]. Most studies completed in patients with IBD use self-report assessments of physical activity [18-21]. Self-report measures are more likely to be biased due to recall bias and measurement error. One systematic review looked at 148 studies and found low-moderate correlations (mean r = 0.37, SD = 0.25) between self-reported and objectively measured PA [22]. These self-report studies generally only assess leisure-time physical activity rather than total daily movement, sedentary and sleeping time [23]. To date, only one abstract has reported objective (i.e., device-based) walking and sleep time measurements in 39 patients with IBD [24]. Despite the importance of considering all behaviours occurring throughout a 24-hour day, to our knowledge, none have investigated sitting, lying down, steps, step cadence, or standing time in the context of IBD.

Accelerometers are devices patients can wear that use acceleration and motion to provide a precise, valid, and reliable estimate of daily movement patterns. Different accelerometers can capture a variety of different variables such as total activity, activity intensity, steps, sitting, standing, and lying down[25]. Previous work in individuals with UC has focused mostly on investigating the effect of moderate (i.e., brisk walking) to vigorous physical activity (i.e., running) [15]. However, additional behaviours occurring throughout the day such as steps (and step cadence), sitting, standing, lying down, and their association with mental health, patient reported outcomes, disease course, or inflammatory markers have not been studied.

The objectives of this exploratory study were to examine whether accelerometer measured step count, step cadence, sitting time, time spent lying down, and standing time associated with a) mental health (i.e., depression and anxiety risk), b) health-related quality of life (HRQoL), and c) clinical outcomes (total Mayo score [TMS], fecal calprotectin [FCP] and C-reactive protein [CRP]) in individuals with UC.

Methods

Participants

Thirty individuals aged 18 years or older were recruited from IBD outpatient clinics at the Foothills Medical Centre in Calgary, Alberta, Canada between May 2017, and November 2020. To be eligible, patients required a diagnosis of UC, managed with conventional medical therapies (oral and/or topical 5-ASA, either alone, in combination with biologic and/or immunosuppressant therapies and with or without oral corticosteroid therapy). Exclusion criteria included major medical comorbidities (diabetes, active malignancy within past 5 years, active infection, severe respiratory or cardiac disease, or acute or chronic kidney disease), a history of previous bowel surgery, or was a current smoker. The study protocol was approved by the University of Calgary Conjoint Health Research Ethics Board and all participants provided informed written consent prior to participating in the study.

Physical activity measurement

Time spent lying down, sitting, standing, step count, and step cadence were measured using the activPAL^TM accelerometer (PAL Technologies, Glasgow, Scotland). Participants were instructed to adhere the activPAL^TM device to the front-midline portion of the thigh with stretch tape that was provided. Participants wore the activPAL^TM at all times (except when showering or swimming), for seven days. Time spent sitting, lying down, and step count were calculated using activPAL^TM algorithms (PAL Software version 8). We were also able to calculate the number of steps taken at a higher stepping cadence (≥ 100 steps/min) which was reflective of moderate (100+ steps/min) to vigorous (130+ steps/min) intensity physical activity [26, 27].

Mental health and HRQoL

This study used the Patient Health Questionnaire-8 (PHQ-8) [28] which measures the presence and severity of depression based on the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria [29]. The PHQ-8 consists of eight items measuring symptoms from the past two weeks with a response scale ranging from “0” (not at all) to “3” (nearly every day) and does not ask the question about suicidal ideation included in the PHQ-9. A PHQ-8 score ≥ 10 is used to characterize moderatesevere depression risk in clinical practice and has a sensitivity and specificity of 70% and 98% respectively [28]. The Generalized Anxiety Disorder-7 (GAD-7) [30] is a seven-item clinical tool used to measure severity of generalised anxiety disorder risk and symptoms from the past two weeks. The response scale ranges from “0” (not at all) to “3” (nearly every day). A GAD-7 ≥ 10 characterizes moderate to severe anxiety risk (sensitivity of 89%, specificity of 82%) [30].

HRQoL was measured using the 12-Item Short-Form Health Survey (SF-12) version 1 [31]. The SF-12 is a self-report tool that assesses the impact of mental and physical health on everyday life. The SF-12 measures eight physical, psychological and psychosocial health domains to tabulate a physical (PCS, based on 4 items) and mental health composite score (MCS, based on 4 items). We used PCS and MCS scores > 50 as a cut-off where a score above 50 represents an above average score and a score of 50 or less represents a below average score [31].

UC clinical outcomes

TMS was calculated for participants by the treating physician. Surveys completed at baseline were used to collect information regarding medication use including immunosuppressants, probiotics, antibiotics, steroids, amino salicylates, and biologic therapies. Stool samples were collected at home by the patient to measure FCP levels, and serum samples were taken to measure CRP levels. TMS, FCP, and CRP variables were dichotomized at clinically relevant cut-offs where TMS ≤ 5 indicated mild disease or clinical remission, FCP ≤ 250 mg/g, and CRP < 8 mg/L (biomarker thresholds represent objective markers of remission).

Statistical Analyses

Proportions, means, and standard deviations were calculated for all relevant sociodemographic, clinical and accelerometer characteristics. To analyze relationships between activPAL^TM variables and mental health, HRQoL and clinical outcomes we used both independent sample t-tests and analysis of variance (ANOVA) for categorical cutoffs (n ≥ 7 in each group), and Pearson’s correlations for continuous variables. Due to the small sample size, large variability in physical activity data, and exploratory nature of this study, the significance value was set at p < 0.20.

Categorical data analyses

When a significant independent t-test was present (p < 0.20), an analysis of variance (ANOVA) was used to further evaluate the relationship between activPAL^TM variables and mental health, HRQoL and clinical outcomes, controlling for relevant covariates (i.e., age, gender, steroids, BMI, and valid days activPAL^TM worn). Continuous data analyses

Pearson correlation coefficients were calculated to identify associations between continuous activPAL^TM variables and mental health, HRQoL and clinical outcomes. When a significant correlation was observed (p < 0.20), multivariable linear regression analyses evaluated associations controlling for relevant covariates. Model covariates included activPAL^TM wear time, age, gender, body mass index (BMI), corticosteroid use, TMS, and FCP.

Results

Baseline sociodemographic characteristics and clinical outcomes are presented in (Table 1). Participants (N=30) had a mean age of 38.6 ± 12.1 years and a mean BMI of 26.1 ± 3.2 kg/ m², with 63.3% overweight or obese (i.e., BMI ≥ 25 kg/m²). TMS scores indicated 56.7% (n=17) of participants were in clinical remission or had mild clinical disease activity (TMS ≤ 5). The mean TMS score was 4.5 ± 3.7. Thirteen patients (43.3%) had an

FCP > 250 mg/g. The mean FCP score was 928.2  1410.2 mg/g. Only one patient (3.3%) had a CRP at or above 8 mg/L and the mean CRP score was 2.1 ± 2.5 mg/L.

Table 1: Sociodemographic characteristics and clinical outcomes.

Physical activity characteristics

Accelerometer data are presented in (Table 2)

Table 2: Descriptive statistics for activPAL^TM measured 24-hour activities (N=28); MCS and PCS cutoffs not included due to n < 7 for the cutoff groups.

For the total sample, mean sitting time was 8.7 ± 2.8 hours per day with a mean standing time of 4.0 ± 1.4 hours per day. Mean time spent lying down was 8.5 ± 2.4 hours per day. Participants had an average of 7,869 ± 3,339 steps per day. The mean number of steps patients took at a cadence ≥ 100 steps per minute was 2,319 ± 1,734 while the amount of time spent walking at this pace was 20.59 ± 14.82 minutes per day.

Clinical outcomes

Independent sample t-tests demonstrated a positive association between TMS and standing time (t = 1.04, p = 0.154), step count at a cadence of ≥ 100 steps/min (t = 1.31, p = 0.100), and time spent walking at a cadence of ≥ 100 steps/min (t = 1.13, p = 0.135) (Table 3). There was an inverse association between FCP and step count (t = -1.14, p = 0.133), standing time (t = -1.46, p = 0.077), step count at a cadence of ≥ 100 steps/min (t = -1.05, p = 0.151), and time spent walking at a cadence of ≥ 100 steps/min (t = -1.18, p = 0.123). No associations were found between FCP and the activPAL^TM variables. We did not run a t-test for CRP as there was only one patient who was above the pre-specified cutoff (CRP ≥ 8mg/L) for analysis.

Table 3: Associations between clinical outcomes and activPAL^TM variables (analysis of CRP not completed due to n < 7 patients above cutoff) *p < 0.200; **p < 0.100.

After the t-tests, an ANOVA was conducted on relationships that were significant in the t-tests and covariates were added. In the ANOVA, no significant associations between the clinical outcomes and activPAL^TM variables remained (Table 4).

Table 4: ANOVA tests for significant associations between clinical outcome cutoffs and activPAL^TM variables. Model covariates: age, gender, steroids, BMI, valid days accelerometer worn.

CRP was negatively correlated with the number of steps taken at a cadence ≥ 100 step/min (r = -0.336, p < 0.070) and time spent stepping at this faster cadence (r = -0.343, p = 0.063) (Table 5). There were no significant correlations present between TMS or FCP and activPAL^TM variables.

Table 5: Pearson correlations between clinical outcomes and activPAL^TM variables; *p < 0.100

Mental Health and HRQoL

Overall, 25% (n = 7) of the sample had moderate to severe depression symptoms (PHQ-8 ≥ 10). The mean PHQ-8 score was 6.6 ± 5.2. Meanwhile, 57.1% (n = 16) had moderate to severe anxiety symptoms (GAD-7 ≥ 10). The mean GAD-7 score was 12.0 ± 4.4. Eight patients (28.6%) had both moderate to severe depression and anxiety. Mean MCS scores were 44.2 ± 3.3 and mean PCS scores were 42.7 ± 4.1.

Based on the t-tests, both the PHQ-8 and GAD-7 ≥ 10 were inversely associated with standing time (t = -1.06, p = 0.15, t = -1.26, p = 0.11 respectively; (Table 6).

Table 6: The association between mental health cutoffs and activPAL^TM variables (MCS and PCS removed due to insufficient patients above cutoff);* p < 0.200; **p < 0.100

GAD-7 ≥ 10 scores were also positively associated with greater time spent lying down (t = 0.861, p = 0.198). The ANOVA on the t-test’s significant relationships did not demonstrate any significant associations after including covariates (Table 7). T-tests and ANOVAs were not performed on the MCS or PCS scores as there were insufficient patients above the cutoff of 50 (zero and one patient, respectively).

Table 7: ANOVA tests for significant associations between mental health cutoffs and activPAL^TM variables.

Both MCS and PCS were positively correlated with step count (r = 0.471, p = 0.011; r = 0.265, p = 0.173, respectively), steps at a cadence ≥ 100 steps/min (r = 0.439, p = 0.038; r = 0.336, p = 0.081, respectively), and time spent walking at this faster cadence (r = 0.439, p = 0.019; r = 0.348, p = 0.070, respectively) (Table 8). After controlling for covariates, average daily steps were significantly associated with better MCS (β = 0.700, p = 0.001) (Table 9). No significant correlations were identified between risk of depression and anxiety (PHQ-8 & GAD-7) and the activPAL^TM variables (Table 8). Daily steps at a cadence ≥ 100 steps per minute, and total minutes walking at this higher cadence, were significantly associated with better MCS scores (β = 0.503, p = 0.044 and β = 0.563, p = 0.019, respectively).

Table 8: Correlations between patient reported outcomes and activPAL^TM variables. *p < 0.20; **p < 0.10; *** p < 0.05.

Table 9: Relationships between MCS and PCS scores and activPAL^TM variables; controlling for age, gender, steroids, BMI, total mayo score, baseline FCP, valid days; *p < 0.100; **p < 0.05.