Abstract
Interest in sleep and circadian research in inflammatory bowel disease (IBD) (Crohn’s disease and ulcerative colitis) is growing; however, few studies have objectively measured sleep or circadian rhythms in people with these conditions. The purpose was to determine the feasibility of the use of wrist actigraphy, electronic sleep/activity diaries, and participant-collected saliva among adults with both active and inactive IBD. We conducted a 10-day feasibility study of adults aged 18 years to 60 years with IBD. We measured sleep and rest-activity rhythms with wrist actigraphy, self-reported sleep/activity using electronic diaries, and participant-collected saliva to compute salivary dim light melatonin onset. All 37 (100%) participants wore the wrist actigraphs, 91.8% (N = 34) participants completed at least 15 of the 18 electronic diaries, and 34 (91.9%) completed the saliva collection. It is feasible to use wrist actigraphy and electronic sleep/activity diaries in adult participants with inflammatory bowel disease.
Inflammatory bowel disease (IBD) is an immune-mediated disease that includes Crohn’s disease and ulcerative colitis, and follows an unpredictable course of active and inactive disease. Currently, 1.6 million people in the United States have IBD (Ng et al., 2018; Shivashankar et al., 2017). IBD is a highly symptomatic chronic condition, and its symptoms include fatigue, bloody diarrhea, abdominal pain, and nausea/vomiting.
The interest in sleep research in IBD has risen since the first study describing the sleep characteristics of 16 people with inactive IBD using polysomnography (Keefer et al., 2006). Sleep disturbance is common in IBD, with 44% to 88% reporting sleep disturbance (Gilca-Blanariu et al., 2018; Gingold-Belfer et al., 2014; Graff et al., 2011; Habibi et al., 2019). Additionally, sleep disturbance predicts risk of disease relapse from inactive to active disease over six months, but which specific sleep or circadian characteristics predict relapse remains unknown (Ananthakrishnan et al., 2013). Being able to measure and describe specific sleep and circadian characteristics in IBD is needed to answer this question fully.
Sleep latency, the number of minutes it takes to fall asleep, measured by wrist actigraph was longer, and sleep efficiency, the percent of time in bed asleep, was lower in Crohn’s disease compared to healthy controls (van Langenberg et al., 2015). Differences in actigraph measured sleep efficiency were mixed, with Qazi et al. (2018) finding that sleep efficiency was lower in active disease and van Langerberg et al. (2015) finding no difference. People with IBD may also have disrupted circadian rhythms. People with both active and inactive IBD have reduced circadian clock gene expression in both inflamed and noninflamed intestinal mucosa (Palmieri et al., 2015; Weintraub et al., 2020). Additionally, circadian clock gene expression is reduced in peripheral mononuclear cells compared to healthy controls and is correlated with disease activity (Liu et al., 2017). However, few reports have measured sleep or circadian characteristics in the home environment over multiple days.
Measuring sleep characteristics and circadian rhythms in the home environment is needed for an ecologically valid measure that captures night-to-night variations (Ancoli-Israel et al., 2003). Objectively measuring sleep and circadian characteristics in the home environment requires participants to engage in multiple types of data collection, and may include wearing a wrist actigraph to measure sleep and rest-activity characteristics, daily sleep/activity diaries, and saliva collection to measure melatonin. Little is known about the feasibility of using these data collection methods in adults with IBD. Symptoms of nighttime diarrhea, fecal urgency, and nausea/vomiting, which are all common in active IBD, may impair people’s willingness to participate and their ability to complete research that occurs over multiple days with multiple data collection methods. Thus, this study includes adults with both active and inactive IBD to determine if there is a difference in completion rates with disease activity (Abraham & Cho, 2009).
Wrist actigraphy is a valid objective measure of sleep and rest-activity, a measure that reflects both endogenous and exogenous circadian rhythms (Conley et al., 2019; Smith et al., 2018; Wirz-Justice, 2007). A wrist actigraph is a small noninvasive device that measures motor activity, which is used to infer sleep and rest-activity characteristics (Martin & Hakim, 2011; Wirz-Justice, 2007). The use of actigraphy to measure sleep requires the use of a sleep diary to guide in the scoring of the actigraph data (Martin & Hakim, 2011). Diaries are used to elicit activities that occurred during a specific window to limit retrospective recall bias (Bingham et al., 2019). However, the completion of electronic sleep diaries in healthy young adults is poor; one study found that less than 60% completed 75% of the daily diaries (Thurman et al., 2018). Sleep researchers have suggested the use of electronic sleep diaries to limit recall bias; however, actigraphy studies usually do not report completeness of diaries or the type of diary (electronic or paper) used, making it difficult to assess feasibility of these methods from previous studies (Tonetti et al., 2016). Previous studies that used actigraphy in IBD used sleep diaries but did not report completion rates or type of diary used (Qazi et al., 2018; van Langenberg et al., 2015).
Dim light melatonin onset (DLMO) is an accurate measure of endogenous circadian rhythm entrainment (Pandi-Perumal et al., 2007). Although previous researchers found that DLMO is feasible to collect at home, little is known about the feasibility of collecting saliva in the presence of IBD, and if this highly symptomatic population would complete saliva collection at home (Pullman et al., 2012).
Due to the complexity of objectively measuring sleep and circadian rhythm characteristics in the home environment with multiple modes of data collection, we conducted a study to determine the feasibility of this study design in adults with IBD in preparation for a larger study. The purpose of this study was to determine the feasibility of in-clinic recruitment, and the use of wrist actigraphy, email-linked electronic sleep/activity diaries, and participant collected saliva among adults with both active and inactive IBD.
Methods
We conducted a 10-day feasibility study among adults diagnosed with IBD. We previously reported the correlations between gastrointestinal symptoms and rest-activity rhythms and sleep characteristics in this sample (Conley et al., 2020).
Setting/Sample
We included adults between the ages of 18 years and 60 years who were diagnosed with IBD (Crohn’s disease or ulcerative colitis). We excluded those who were currently taking melatonin or other sleep medications, pregnant or breastfeeding, diagnosed with severe psychiatric (bipolar disorder and schizophrenia), blind, or neurological conditions affecting the nondominant hand (due to actigraph monitoring), had untreated previously diagnosed sleep apnea, worked night or rotating shifts, traveled two or more times zones in the week prior to data collection, or had surgery in the previous 4 weeks. We also excluded those who did not read and write English, those without internet access, and without an email address. We did not collect data the week following Daylight Savings Time.
Study Protocol
We obtained institutional review board approval. All participants were recruited from a single academic gastroenterology clinic. Potential participants were introduced to the study by their IBD health care provider during their routine IBD appointment. If interested, they were introduced to a member of the research team following their appointment. The research team member then explained the study in full, screened for inclusion/exclusion criteria, and obtained written informed consent.
We instructed participants to follow their regular schedules and to wear the wrist actigraph on their nondominant wrist continuously for 10 days. Participants collected all the saliva on a single night of their choice. Participants were emailed a link to the baseline survey on the day of their enrollment, and two additional reminder emails were sent as needed. The baseline survey elicited demographic and clinical factors and GI symptoms. Participants were emailed links twice daily to an electronic sleep/activity diary at their regular wakening times and 30 minutes before their regular bedtimes. At the conclusion of the study, we picked up the study materials at a mutually agreeable location.
Variables and Measures
Demographic and clinical characteristics
Demographic variables included age, race/ethnicity, gender, and employment status, which we obtained via self-report. Clinical characteristics included IBD type (Crohn’s disease or ulcerative colitis), duration of IBD diagnosis, age at diagnosis, history of/current bowel abscesses, fistulas, and strictures were obtained from the medical record. We elicited disease activity using the Physician Global Assessment (PGA), a valid measure of disease activity compared to mucosal inflammation (Turner et al., 2010). This was obtained from the medical record from the day of enrollment.
Feasibility
We described feasibility as the percentage of people approached who were eligible; enrolled; and completed the study. Feasibility was defined as 70% completion of data, as reported in similar studies (Pullman et al., 2012). At the conclusion of the study, we asked participants to assess the level of difficulty of wearing the actigraph, the email-linked electronic diaries, and the saliva collection. Each item was scored on a five-point Likert scale from not at all difficult to very difficult (Warms & Belza, 2004).
Rest-activity rhythms and sleep
Wrist actigraphy is a valid measure of rest-activity rhythms and sleep characteristics, and discriminates sleep from wake (Conley et al., 2019; Smith et al., 2018). Data were collected in 30-second epochs. The cleaning and scoring of the actigraph data were reported previously (Conley et al., 2020). In brief, the major sleep period was defined as lights off to lights on. We used three different methods to determine these times. (a) Participants were instructed to depress an event marker on the actigraph when they turned the lights off to go to sleep and again in the morning when they turned the lights on, (b) they were instructed keep sleep diaries, and (c) each Actiwatch Spectrum Plus (Philips Respironics) is fitted with a light meter to capture ambient light. We considered lux = 0 as lights off and lux > 0 as lights on.
Participants completed electronic sleep/activity diaries, which included the time of getting in bed, lights off, falling asleep, number of nighttime awakenings and reason, lights on in the morning, naps, and questions from the Social Rhythm Metric II-5 Item Version. Diaries were closed if not completed within a 6-hour window (to allow for variations in wake/bedtime).
Participants were instructed to refrain from the use of toothpaste, caffeine, alcohol, chocolate or bananas for five hours before the start of saliva collection; refrain from eating or drinking anything during the 15 minutes before collection times; document the time of each sample collection; and store samples in the freezer (Pandi-Perumal et al., 2007). Salivary melatonin is stable at room temperature for 72 hours and is suited for home collection (Nalla et al., 2015). We provided participants freezer bags and ice packs to ensure samples remained frozen during transport. After we collected saliva samples from participants, the saliva samples were stored in a research freezer (-80°C) until they were shipped on dry ice to Salimetrics.
The samples were analyzed at Salimetrics, where samples were thawed to room temperature, vortexed, then centrifuged for 15 minutes at 3,000 RPM (1,500 x g). Melatonin levels were measured using enzyme-linked immunosorbent assay (ELISA) (Cat. No. 1-3402) in duplicate. The Salimetrics assay has a lower limit of sensitivity of 0.58 pg/mL, a standard curve range of 0.78–50 pg/mL, an average intra-assay coefficient of variation of 5.42%, and an average inter-assay coefficient of variation of 8.90%. DLMO phase was defined as the time that salivary melatonin exceeded 3pg/ml, consistent with standard methods (Benloucif et al., 2008).
Data Analysis
Participants entered self-report data into electronic surveys programmed in Qualtrics (Provo, Utah).TM Data extracted from the medical record into an Excel dataset. Data were exported from Qualtrics, Philips Actiware, and Excel and were imported into IBM SPSS version 24 for data management and analysis. We used descriptive statistics to describe the sample and feasibility of each study component. We computed cosinor and non-parametric analysis to describe the rest-activity characteristics using a SAS macro we developed.
Results
The study included 37 participants. All data were collected from October 2017 to April 2018. See Figure 1 for the flow of participants through the study. IBD providers initially screened 101 IBD patients charts for eligibility and excluded 33. We approached 68 potential participants and, after screening for exclusion criteria, we enrolled 61.7% (N = 42) of those approached. The most common reasons potential participants were excluded were that they declined to participate (n = 26) due to being too busy (n = 21) or not interested (n = 5). Of those enrolled, 88% completed the study. All those that did not complete the study after enrolling (n = 5) dropped out before the start of data collection due to being too busy. The sample had a mean age of 38.6 years (SD 13.8), 21 (56.8%) were female, 28 (75.7%) were White, 19 (51.3%) had ulcerative colitis, 14 (37.8%) had active disease, the mean time since diagnosis was 12.2 (9.2) years, and the mean social rhythm metric was 2.9 (SD 0.9). The demographic and clinical characteristics of the sample are reported in Table 1.
Flow of Participants Through Study.
Demographic and Clinical Characteristics of the Sample (N = 37).
All 37 participants wore the wrist actigraph for at least six days, and 35 wore it for at least nine days. The results of the actigraph analysis are presented in Table 2. There was no association between sleep rest-activity characteristics and disease activity, disease duration, age, race/ethnicity, or gender (ps > .05). Most participants, 70.3% (N = 26), pressed the event markers for at least 15 out of the 18 possible events. Of the 11 who depressed less than 15 event markers, 5 had active disease.
Sleep and Activity Rest Characteristics of the Sample (N =37).
Note: data reported in Conley et al., 2020.
Of the possible 18 twice-daily diaries, 91.8% (N = 34) of the participants completed all the diaries with minimal missing data (less than three missing diaries). Of those with incomplete diaries, one participant completed 12 diaries, and two completed less than 5 diaries. Of those with missing or incomplete diaries, one out of the three had active disease.
We were unable to compute DLMO for 29 (75.7%) of the participants. Three participants (8.1%) did not return any saliva. Of the three who did not complete the saliva collection, two had active disease. Twenty-two participants (59.5%) had melatonin levels ≥ 3 pg/ml throughout the saliva collection period; thus, time of DLMO could not be extrapolated with the collected samples. Three participants (8.1%) had melatonin levels that were persistently below 3 pg/ml throughout the saliva collection period; thus, the DLMO threshold was not met. Only nine (24.3%) participants had salivary melatonin levels sufficient to compute DLMO.
Thirty-one participants answered how difficult it was to complete each study component, and 100% (n = 31 out of 31 responses) reported that the wrist actigraph was not difficult to wear. The majority of participants, 97% (30 out of 31 responses), reported low or no difficulty with completing the diaries. Most participants, 93.5% (29 of the 31 responses), reported low to no difficulty with the saliva collection.
Discussion
It is feasible for participants with IBD to use wrist actigraphy, twice daily emailed sleep/activity diaries, and collect saliva, and the rates of completion of study components did not appear to differ for those with and without active IBD. The feasibility of wrist actigraphy in IBD was recently confirmed by Qazi et al., 2018, who found a similar percentage of participants completed 5 days of actigraphy monitoring (90%) compared to the 10 days of monitoring in this study. Thus, using actigraphy for a longer period does not seem to add additional burden. Our study adds to this by reporting the completeness of actigraphy data and adding the feasibility of using multiple modes of data collection, including email-linked electronic diaries and participant collected saliva.
We found it was feasible and acceptable to obtain sleep/activity email-linked electronic diaries twice daily at normal awakening and bedtimes. Over 90% of participants completed the diaries with minimal missing data. Participants reported low levels of difficulty with the completion of the diaries, and the majority were completed with limited recall bias, which is a potential problem in pen/paper diaries (Stone et al., 2003).
Measuring circadian characteristics in people with IBD is challenging because serial melatonin measurement and core body temperature are the gold-standard measures of endogenous circadian rhythms (Edwards et al., 2002). Core body temperature is measured using rectal probes or by ingestible sensors, which are contraindicated in IBD due to possible intestinal strictures and perianal disease. While participants were able and willing to collect serial saliva samples, we were not able to compute DLMO with traditional methods due primarily to the high melatonin levels at the start of saliva collection. This may be due to inaccurate timing of saliva collection in relation to participant reported bedtime, or not having a sufficiently long collection window. Future studies should use 24-hour saliva collection to determine if our findings are due to circadian misalignment in people with IBD, or if salivary melatonin, similar to salivary inflammatory markers does not reflect circulating melatonin in IBD (Nielsen et al., 2005). There are early changes in circadian clock genes expression in intestinal tissue in IBD. In people with a recent IBD diagnosis and not yet on treatment, the expression of circadian clock genes was reduced even in noninflamed intestinal mucosa (Weintraub et al., 2020). While research has yet to connect the changes in mucosal circadian clock gene expression with salivary or gastrointestinal melatonin in people with IBD, it may partially explain why we saw high levels of salivary melatonin, as the mouth is part of the gastrointestinal tract.
In nursing, interest is growing in sleep and circadian rhythm research as the timing of behaviors and quality and quantity of sleep have profound impacts on human health (Khan et al., 2018; Medic et al., 2017). However, the participant burden can be high in sleep and circadian research, which may compete with the demands of managing chronic conditions. While we conducted this study in adults with IBD, we would expect similar feasibility findings in other populations with highly symptomatic chronic conditions. Nurse researchers can be assured that participants with highly symptomatic chronic conditions not only can complete these complex data collection procedures but do not find them difficult.
The primary limitation of this study was that we were unable to determine the reason for missing electronic diaries (e.g., forgot to fill out or timed out). Despite this, we reported high levels of electronic diary completion. We did not screen for undiagnosed sleep apnea; thus, we may have included people with undetected sleep apnea or other sleep disorders. The strengths of this study were that we were able to recruit a diverse IBD sample that matches the US IBD population (Shivashankar et al., 2017). We used a well-validated objective measure of sleep and rest-activity characteristics, and had a high study completion rate.
Our study documented that it is feasible for adults with both active and inactive IBD to complete complex data collection, including wearing a wrist actigraph, email-linked electronic sleep/activity diaries, and serial saliva samples at home. This study will provide the foundation for a fully powered study describing sleep and rest-activity characteristics and how they predict the risk of disease relapse among people with IBD so that precision sleep and circadian interventions can be tested and developed.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by American Nurses Foundation Eastern Nursing Eastern Research Scholar Award (PI Samantha Conley); Self and Family Management T32 (PI Margaret Grey T32NR008346)
