Introduction

Inflammatory bowel disease (IBD) is a chronic relapsing remitting immune-mediated disorder that involves a complex interplay of genetic and environmental factors [1]. Symptoms of active IBD, such as abdominal pain and diarrhea, may influence sleep quality and duration. Irritable bowel syndromes like symptoms are common in people with inflammatory bowel disease and can be misinterpreted as active IBD [2].

Short sleep duration, in the general population, has been associated with increased all-cause mortality [3], adverse health effects including cardiovascular disease [4], and metabolic syndrome [5], as well as economic consequences such as lower productivity and greater health care utilisation [6]. Upregulation of pro-inflammatory cytokines, such as IL-1β, IL-6, and TNFα, have been observed in short sleepers—with these same cytokines implicated in the pathogenesis of IBD [7]. Several previous studies have investigated the association between sleep and IBD [8,9,10] and postulated a bi-directional relationship between sleep and IBD activity [11].

Sleep can be assessed subjectively using a measure of perceived sleep via validated survey methods such as the Pittsburgh Sleep Quality Index (PSQI), with standardised cut off values to define likely significant sleep disturbance referred to as “poor sleep” [12]. Meta-analyses show that poor sleep is prevalent in people with IBD [13], worse in people with IBD than healthy controls [14], more common in those with subjectively active IBD [14], and worse in IBD in remission than in healthy controls [15].

However, there have also been several publications [16,17,18,19,20,21,22] suggesting no significant relationship between IBD disease activity and disturbed sleep, with some suggesting that only perceived sleep may be different. Studies incorporating objective sleep measurement have suggested that sleep efficiency is worse in subjectively active IBD and wake after sleep onset is longer in subjectively active IBD [19]. The lack of objective IBD assessment in these studies leads to the possibility of irritable bowel syndrome like symptoms being mistaken for clinically active IBD. The authors aim to address this gap in the literature and compare the gold standard of sleep assessment—polysomnography with objective assessments of IBD activity.

Methods

Ethics approval for this study was obtained from the Southern Adelaide Human Research Ethics Committee (203.20). Informed consent was obtained from all participants. Participants were recruited from a tertiary IBD unit and a private IBD service with the study advertised via email and flyer. Recruitment occurred from 2021 to the end of 2023. Participants were adults (age over 18) with a confirmed diagnosis of IBD from a gastroenterologist. Participants had a home sleep study (polysomnography) within 12 weeks of an objective assessment of IBD activity which was performed as per usual care. Accepted objective assessment(s) of IBD activity included magnetic resonance imaging enterography, colonoscopy, gastrointestinal ultrasound, and fecal calprotectin (> 150ug/g). Current medications including corticosteroids were recorded during the interval between the sleep study and the IBD assessment. Participants were offered $100 AUD for completion of the sleep study. Participants also completed a survey including demographic data, subjective IBD activity, depression, anxiety, and subjective sleep quality. The study aimed to recruit a minimum of twenty participants based on reported differences in polysomnography (sleep efficiency and wake after sleep onset) reported in other studies that incorporated subjective IBD assessment [19], with a goal of forty participants.

Participants were excluded if they had an uncontrolled psychiatric disorder, substance abuse disorder, heavy vehicle license or a known sleep disorder such as obstructive sleep apnea.

Sleep Quality Assessment

Polysomnography consists of the recording of multiple variables during sleep and can be performed in sleep laboratory or home setting. Variables include—electrocardiogram, electromyography to capture muscle movement, electroencephalography to capture brain activity, electrooculogram to capture eye movement, respiratory airflow channels to capture apnoeas or hypopnoeas, and pulse oximetry monitoring and respiratory effort channels to measure movement of chest and abdomen [23]. Sleep staging is divided into REM sleep (rapid eye movement) and NREM sleep—(non-rapid eye movement) which is further divided into stage 1, stage 2, and slow wave sleep (previously stage 3 and 4 sleep). Variables of interest in polysomnography are seen in Table 1. Given the well-established sleep differences due to age and gender [24, 25], polysomnography results were normalized by published values based on age and gender [26].

Table 1 Definition of objective sleep quality parameters from polysomnography
Full size table

Polysomnography was performed using the Compumedics Somte PSG and Somte PSG 2 devices (Compumedicas Limited, Victoria, Australia). All participants were manually setup by a trained technician. The device used does not show impedance values; however, all traces were visually checked via a staff member prior to leaving the sleep laboratory. All sleep studies have electroencephalography, electrooculogram, electromyography, respiratory, ECG, and limb recordings and are directly comparable to gold standard polysomnography. All of the data were manually scored with the exception of limb movements which was predominantly scored by auto analysis by Profusion software (Compumedicas Limited, Victoria, Australia). All the sleep studies were scored by a trained sleep technician with a minimum of 5 years’ experience in sleep scoring. The sleep laboratory used regularly participates in QSleep quality assurance testing.

Questionnaire Data

Clinical IBD activity was assessed using the Harvey Bradshaw Index (HBI) in participants with Crohn’s disease; with HBI > 5 considered active disease [27]. The patient-reported version of the HBI was used in the survey, although a decision was made to maintain the general well-being and abdominal pain score similar to the physician HBI rather than using a ten-point Likert scale [28]. The Simple Clinical Colitis Activity Index (SCCAI) was used in participants with ulcerative colitis; with an SCCAI > 5 considered active disease [29]. The patient-reported form of the SCCAI was used. The use of a self-reported SCCAI has been previously validated with good agreement with physician reported SCCAI [30].

Anxiety was assessed using the generalized anxiety disorder 7-item scale (GAD-7) [31]. The Patient Health Questionnaire 9 (PHQ-9) was used to assess depression [32]. Subjective sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI), a validated tool to assess perceived sleep quality [12].

Statistical Analysis

Statistical analysis was performed using Stata SE 16 (StataCorp, College Station, TX, USA). For normally distributed variables, mean and standard deviation (SD) were reported with comparisons made using the student t-test or paired t-test when appropriate. For non-normally distributed variables, median and interquartile range (IQR) were reported, with comparisons made using the Mann- Whitney U test. For categorical data, Pearsons χ2 test was used or Fisher’s exact test when appropriate. No incomplete survey responses were included in the analysis. Pearson’s or Spearman’s correlation was used as appropriate, with interpretation of coefficients as follows: very weak < 0.19, weak 0.2–0.3, moderate 0.3–0.5, strong 0.5–0.79 and very strong > 0.80 [33]. Any missing survey data result in exclusion of that instrument from analysis.

Individual polysomnography results were matched by age and gender to standardized health adult polysomnography results generated via an online calculator based on a meta-analysis of healthy population values from 169 studies [26] (https://omc.ohri.ca/psgnorms/). Individual polysomnography results were then normalized by matched age and gender polysomnography results. Polysomnography results were reported according to the American Academy of Sleep Medicine manual [23, 34].

This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline [35].

Results

Twenty-nine participants were screened for inclusion. Two were excluded due to an unstable psychiatric disorder. Five were excluded due to a known sleep disorder (insomnia in one case and obstructive sleep apnea in the others). Following screening, 22 participants were included; however, two participants did not undergo the sleep study within twelve weeks of objective IBD assessment and were subsequently excluded. The majority underwent polysomnography within a month of objective IBD activity measurement (65%) (for further details see supplementary figures). Polysomnography was performed prior to the objective IBD assessment in 60%. Objective IBD assessment took the form of calprotectin in 30%, magnetic resonance imaging in 40%, and colonoscopy in 50%. Evidence of active inflammation on any objective assessment was considered as objectively active IBD and was present in 40%.

The median age was 41 years (33–48), 55% male, and the majority had Crohn’s disease (85%) (see Table 2). Previous IBD surgery was reported by 40%, and over half were on advanced therapies such as biologics (65%) with none on small molecules. For all participants, no medication changes, including no steroid use, occurred during the period between the sleep study and IBD assessment. No participant was admitted to hospital during the period between sleep study and IBD assessment. No participant had an ileostomy or colostomy.

Table 2 Demographic and inflammatory bowel disease (IBD) data of study participants
Full size table

Polysomnography

There was similar time available for sleep in both objectively inactive IBD and active IBD groups (see Table 3, Fig. 1). However, total sleep time was less in those with active IBD (p = 0.049) – primarily as decreased NREM sleep duration due to decreased stage 2 sleep time (p = 0.045). No difference in the percentage of time in each sleep stage was seen. Sleep latency was longer in the active IBD group (p = 0.015); i.e., it took a longer time for them to fall asleep. As a result of this, sleep efficiency was lower in the active IBD group (p = 0.014). No difference in wake after sleep onset was seen (p = 0.41). The polysomnography results were normalized by healthy population results matched by age and gender (see Table 3) with similar differences between inactive and active IBD seen—that is those with active IBD had longer sleep latency, shorter total sleep time due to stage 2 sleep, and worse sleep efficiency, although percentage of time in sleep stage was similar.

Table 3 Polysomnography results for active inflammatory bowel disease (IBD) and inactive IBD
Full size table
Fig. 1
figure 1

Sleep stage duration in minutes in those with objectively active inflammatory bowel disease (IBD) and inactive IBD. N1 stage 1 of sleep, N2 stage 2 of sleep, SWS slow wave sleep, WASO awake after sleep onset

Full size image

Questionnaire

Survey results were available for 75% of the cohort within one month of the sleep study. Subjective symptoms of active IBD were present in 45%, with mean HBI 5.6 (4.1) and mean SCCAI 3.3 (2.9). Objective evidence of active IBD was present in 40% of the cohort and was associated with higher depression scores (see supplementary Table 1) but no difference in anxiety or subjective sleep quality scores. Depression scores moderately correlated with stage 2 sleep duration (ρ 0.53 p = 0.046) and strongly correlated with sleep latency (ρ 0.75 p = 0.0013) (see supplementary Table 2). There were no significant correlations between other polysomnography measures and subjective IBD activity, depression scores, or anxiety scores. Subjective sleep quality (PSQI) moderately correlated with depression scores (PHQ9) (ρ 0.57, p = 0.026). A modified PHQ9 score was considered excluding the PHQ9 question concerning sleep with correlations between this score and sleep latency (ρ 0.69, p = 0.0059), and PSQI remaining significant (ρ 0.59, 0.025), with stage 2 sleep duration no longer significant (ρ 0.47, p = 0.088).

Discussion

Herein, we report novel evidence of associations between objective, but not subjective, IBD activity and objectively assessed sleep quality. Prolonged sleep latency in the active IBD group and shortened stage 2 sleep duration led to lower sleep efficiency and shorter total sleep time in the active IBD group. Unlikely previous actigraphy studies, no difference in wake after sleep onset was seen [36, 37]. This may be consistent—noting that actigraphy cannot accurately determine wakefulness while lying still compared to light sleep (unlike EEG measurement as with polysomnography) and therefore may misclassify wake after sleep onset among other parameters [38].

Subjective sleep quality in IBD has been previously associated with depression with similar results obtained in this study [20, 22, 39]. Aspects of polysomnography did correlate with depression scores and not surprisingly depression scores were higher in those with objectively active IBD. Depression is associated with sleep abnormalities such as prolonged sleep latency that was observed in this study, but also other abnormalities such as decreased slow wave sleep and REM sleep abnormalities [40, 41]—which were not seen in the objective active IBD group here. Consequently, we suggest that not all polysomnography differences demonstrated are explained by depression scores alone and may represent an additional influence from active IBD.

Objectively active IBD has been associated with elevated pro-inflammatory cytokines such as TNF-a and IL-6—which are also produced via sleep restriction. In animal studies, elevated pro-inflammatory cytokines drive the need to sleep and correlate with the duration of recovery sleep [42]. The understanding of inflammatory cytokines and sleep in human is less well understood [43]. In rheumatoid arthritis, a small study demonstrated a reciprocal relationship between sleep efficiency and TNFa production [44]. Other studies in rheumatoid arthritis have not demonstrated any differences in polysomnography findings between active and inactive rheumatoid arthritis [45]. A meta-analysis of studies incorporating markers of inflammation and measurement of sleep showed an association between elevated markers of inflammation and sleep disturbance and long sleep duration [46]. Our findings are more consistent with sleep restriction rather than inflammatory cytokine-mediated sleep recovery.

Limitations to this study include the limited sample size which is typically for studies incorporating polysomnography and limits the generalisability of these results. While our use of objective IBD activity allowed a variety of different measures of IBD activity that likely introduces heterogeneity, it did allow for verification of active inflammatory disease. There was some difficulty in arranging sleep studies within an acceptable time of the IBD assessment with two participants not completing the sleep study within the allowed time. This was due to repeated cancelations or non-attendance by the participant. As a post hoc verification of date, we incorporated results from these two further participants and confirmed that it did not change the study outcomes. Also note that the study utilized a single night of polysomnography with others arguing for utilization of multiple nights of polysomnography to increase validity of results and avoid a first night effect [47]. Recruitment for this study proved challenging as with other polysomnography studies because of the time needed for overnight polysomnography.

Future research should consider the prognostic value of objective measurement of sleep as a marker of active IBD with less intrusive methods to measure sleep. Consideration should also be given to performing polysomnography in people with clinical remission but ongoing endoscopic activity in order to observe changes in sleep due to inflammatory cytokines and not IBD-related symptoms.

Conclusion

Specific changes in sleep stages were demonstrated via polysomnography in active IBD compared to inactive IBD. Depression was associated with polysomnography differences along with IBD activity and may explain some but not all of the observed polysomnography changes. Further research should consider quantifying polysomnography differences associated with IBD-related pro-inflammatory cytokines as well as considering the utility of changes in sleep quality as a means of longitudinally assessing objective IBD activity.