Influence of social jetlag on daytime sleepiness in obstructive sleep apnea

Social jetlag is the discrepancy between socially determined sleep timing on workdays and biologically determined sleep timing on days free of social obligation. Poor circadian timing of sleep may worsen sleep quality and increase daytime sleepiness in obstructive sleep apnea (OSA). We analysed de‐identified data from 2,061 participants (75.2% male, mean [SD] age 48.6 [13.4] years) who completed Sleep Apnea Global Interdisciplinary Consortium (SAGIC) research questionnaires and underwent polysomnography at 11 international sleep clinic sites. Social jetlag was calculated as the absolute difference in the midpoints of sleep between weekdays and weekends. Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS). Linear regression analyses were performed to estimate the association between social jetlag and daytime sleepiness, with consideration of age, sex, body mass index, ethnicity, insomnia, alcohol consumption, and habitual sleep duration as confounders. Of the participants, 61.5% had <1 h of social jetlag, 27.5% had 1 to <2 h, and 11.1% had ≥2 h. Compared to those with <1 h of social jetlag, those with ≥2 h of social jetlag had 2.07 points higher ESS (95% confidence interval [CI] 0.77–3.38, p = 0.002), and those with 1 to <2 h of social jetlag had 0.80 points higher ESS (95% CI 0.04–1.55, p = 0.04) after adjustment for potential confounding. Interaction with OSA severity was observed; social jetlag appeared to have the greatest effect on daytime sleepiness in mild OSA. As social jetlag exacerbates daytime sleepiness in OSA, improving sleep timing may be a simple but novel therapeutic target for reducing the impact of OSA.


Summary
Social jetlag is the discrepancy between socially determined sleep timing on workdays and biologically determined sleep timing on days free of social obligation. Poor circadian timing of sleep may worsen sleep quality and increase daytime sleepiness in obstructive sleep apnea (OSA). We analysed de-identified data from 2,061 participants (75.2% male, mean [SD] age 48.6 [13.4]

years) who completed Sleep Apnea Global
Interdisciplinary Consortium (SAGIC) research questionnaires and underwent polysomnography at 11 international sleep clinic sites. Social jetlag was calculated as the absolute difference in the midpoints of sleep between weekdays and weekends. Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS). Linear regression analyses were performed to estimate the association between social jetlag and daytime sleepiness, with consideration of age, sex, body mass index, ethnicity, insomnia, alcohol consumption, and habitual sleep duration as confounders. Of the participants, 61.5% had <1 h of social jetlag, 27.5% had 1 to <2 h, and 11.1% had ≥2 h. Compared to those with <1 h of social jetlag, those with ≥2 h of social jetlag had 2.07 points higher ESS (95% confidence interval [CI] 0.77-3.38, p = 0.002), and those with 1 to <2 h of social jetlag had 0.80 points higher ESS (95% CI 0.04-1.55, p = 0.04) after adjustment for potential confounding. Interaction with OSA severity was observed; social jetlag appeared to have the greatest effect on daytime sleepiness in mild OSA. As social jetlag exacerbates daytime sleepiness in OSA, improving sleep timing may be a simple but novel therapeutic target for reducing the impact of OSA.
K E Y W O R D S circadian rhythm, cross-sectional study, obstructive sleep apnea, polysomnography, sleep habits, sleepiness

| INTRODUCTION
Obstructive sleep apnea (OSA) is an increasingly prevalent sleep disorder thought to affect almost 1 billion people globally (Benjafield et al., 2019). OSA is characterised by recurring episodes of partial or complete upper airway collapse during sleep. The repetitive obstructions and intermittent hypoxaemia associated with sleep fragmentation can lead to excessive daytime sleepiness (He & Kapur, 2017).
However, there are many factors known to contribute to excessive daytime sleepiness, beyond OSA. Residual sleepiness occurs in a significant proportion of patients with OSA despite seemingly adequate treatment, especially in younger individuals (Pepin et al., 2009).
Circadian disruption may potentially exacerbate the negative health consequences of OSA. However, this has not been previously explored.
Social jetlag is conceptualised as the discrepancy between socially determined sleep timing on workdays and biologically determined sleep timing on free days that is, days free of work or educational commitments (Wittmann et al., 2006). Social jetlag is a highly prevalent form of circadian disruption, with an estimated 30% of Australians experiencing >1 h of social jetlag and 30% of Europeans experiencing >2 h of social jetlag (Lang et al., 2018;Roenneberg et al., 2012). Social jetlag has been associated with depression, obesity, diabetes, and academic underachievement (Diaz-Morales & Escribano, 2015;Islam et al., 2020;Parsons et al., 2015). Sleep-wake timing is regulated by endogenous circadian rhythms, an internal homeostatic process, and exogenous social constraints such as work timing (Borbely, 1982;Foster et al., 2013). Early work schedules may force individuals to sleep and wake at inappropriate times compared to their internally regulated circadian preference. In particular, people who prefer a late sleep onset often suffer from sleep deprivation on weekdays (workdays), which causes them to sleep longer on weekends to compensate (Roenneberg et al., 2003). Social jetlag is a chronic form of circadian misalignment that causes sleep disruption.
Increasing social jetlag decreases sleep quality in shift workers and patients with sleep disorders (Kang et al., 2020;Reis et al., 2020). Poor sleep quality and fragmented sleep both impair daytime function and increase daytime sleepiness (Stepanski, 2002 (Keenan et al., 2018;Qin et al., 2021;Sutherland et al., 2019). Recruited participants complete SAGIC research questionnaires and undergo polysomnography. Sleep studies were scored by technicians at each site to the same standard and high concordance (Magalang et al., 2013;Magalang et al., 2019).

| Ethics
The data were collected with ethics approval from each of the SAGIC sites. The use of the de-identified data for this study was approved by the SAGIC investigators, who act as the data custodians. No further ethics approval was required for this study according to the University of Sydney Human Ethics Committee.

| Social jetlag
Sleep and wake times were obtained from participant responses to questions about usual bed and wake-up times on weekdays and weekends. The midpoint of sleep (mid-sleep) was calculated from usual bed and wake-up times on weekdays and weekends. Social jetlag was calculated as the absolute time difference between mid-sleep on weekdays and mid-sleep on weekends (Wittmann et al., 2006). This definition of social jetlag assumes that participants have regular sleep-wake patterns (which excludes shift workers), weekdays correspond to workdays and weekend days correspond to free days.

| Daytime sleepiness
Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS) (Johns, 1991). This scale consists of eight questions describing eight different activities. Participants rate their chances of falling asleep for each activity, ranging from 'would never doze off' to 'high chance of dozing off'. Each question is scored on a scale from 0 to 3 and the total ESS score ranges from 0 to 24. A higher ESS score represents a greater degree of subjective daytime sleepiness. The reliability and validity of the ESS in patients with OSA has been well established in multiple languages (Beiske et al., 2009;Chen et al., 2002;Chiner et al., 1999;Vignatelli et al., 2003).

| Participants
The participant selection flow chart is shown in Figure 1. The total number of participants in the SAGIC dataset was 5,829 as of January 24, 2022. Only participants with diagnosed OSA (Apnea-Hypopnea Index [AHI] ≥5 events/h) were included in the present study. Participants who were previously treated for sleep apnea (with upper airway surgery, continuous positive airway pressure, mandibular advancement device or other) were excluded because the treatments reduce the severity of OSA and daytime sleepiness. Those who took sleep medications were also excluded due to potential effects on daytime sleepiness. Shift working participants were excluded due to potentially irregular sleep-wake patterns and difficulties in ascertaining social jetlag. Participants with invalid or missing sleep time data, ESS data, or covariate data were further excluded. The remaining 2,061 participants were included in the main analysis on social jetlag and daytime sleepiness.

| Covariates
The AHI was taken from polysomnography and used for categorising OSA severity (i.e., mild, AHI 5 to <15; moderate, AHI 15 to <30; severe, AHI ≥30 events/h). Age and body mass index (BMI) were assessed as part of the overnight sleep study visit. Sex, ethnicity, shift work, insomnia, alcohol consumption, and habitual sleep duration were self-reported as part of the SAGIC Sleep Questionnaire. Ethnicity categories included 'Asian', 'Caucasian', 'Central/South American', and 'Other' (which combined Pacific Islander, African, Other, Do not know, and Multiple). Alcohol consumption was reported as consuming alcohol 0, 1 or ≥2 days/ week. Habitual sleep duration was based on self-report in response to 'How many hours do you sleep on average?', with responses measured in hours and minutes. We note that 558 participants did not report their habitual sleep duration (27.8% of analysis sample). Insomnia was assessed using the Insomnia Symptom Questionnaire (Okun et al., 2009). A diagnosis of insomnia was assigned to participants if they experienced at least one of three sleep symptoms (difficulty initiating sleep, difficulty maintaining sleep, or unrefreshing sleep) at least 3 times/week for ≥4 weeks and at least one aspect of daily life was affected 'quite a bit' or 'extremely' by these symptoms.

| Statistical analysis
The IBM ® Statistical Package for the Social Sciences (SPSS ® ) version 26 was used for statistical analysis. Social jetlag was calculated as the absolute difference in the midpoints of sleep between weekdays and weekends and split into three groups (<1, 1 to <2, and ≥2 h). Associations between social jetlag and other variables were assessed using the chi-square test for categorical variables and one-way analysis of variance for continuous variables.
Age, sex, BMI, ethnicity, insomnia, alcohol consumption, and habitual sleep duration were considered potential confounders. A directed acyclic graph was constructed to consider the causal interrelationships between the exposure, outcome, and potential confounding variables and to evaluate which of the variables was necessary to include in adjusted models in order to estimate the association between social jetlag and sleepiness. The advantage of this approach is that it explicitly considers the network of causal relationships between the exposure and outcome variables and potential covariates. Only age and ethnicity were considered necessary covariates in adjusted models estimating the relationship between social jetlag and daytime sleepiness (Appendix S1, Figure S1).
Linear regression analyses were performed to estimate the effect of the predictor social jetlag (<1, 1 to <2, ≥2 h, categorical) on the outcome of daytime sleepiness (continuous), with adjustment for confounding. Model 0 is a simple regression between social jetlag and ESS. Model 1 adjusted only for OSA severity (AHI, continuous). To explore potential interactions between OSA severity and social jetlag on daytime sleepiness, an interaction term was F I G U R E 1 Flow chart of the analysis of participants. Abbreviations: AHI, Apnea-Hypopnea Index; BMI, body mass index; CPAP, continuous positive airway pressure; ESS, Epworth Sleepiness Scale; OSA, obstructive sleep apnea; SAGIC, Sleep Apnea Global Interdisciplinary Consortium created. Model 2 adjusted for age (years, continuous), ethnicity (categorical), and the interaction term (social jetlag Â AHI). Potential site differences were explored and reported in (Appendix S2, Table S1).
Two sensitivity analyses were performed. The first sensitivity analysis examined whether social jetlag is associated with an alternative definition of excessive daytime sleepiness defined as ESS scores and the question 'Do you feel sleepy during the day?' from the Basic Nordic Sleepiness Questionnaire (BNSQ) (Partinen & Gislason, 1995). This definition of sleepiness may be better for detecting clinically significant sleepiness that has an impact on health consequences and quality of life (Thorarinsdottir et al., 2019;Thorarinsdottir et al., 2021) Table S3).
The second sensitivity analysis was performed to examine whether social jetlag is associated with daytime sleepiness using the traditional approach of adjusting for all covariates that were significantly different between the social jetlag groups (the traditional ' Table 1 approach') (Appendix S4, Table S4). The significance level was defined as p < 0.05 for all analyses.  Table 1. Social jetlag was significantly associated with younger age, higher BMI, ethnicity, shorter habitual sleep durations, and greater rates of insomnia. There were no significant differences in social jetlag by OSA severity or gender groups. There was no significant difference in the mean ESS score by social jetlag categories, although there was a trend towards higher ESS scores with increasing social jetlag (Table 1). When sleepiness was defined as feeling sleepy, having risk of dozing off, or both symptoms, social jetlag was significantly associated with sleepiness (p = 0.01, Appendix S3, Table S2).

| Linear regression results
The results of the regression modelling are presented in Table 2 significantly associated with daytime sleepiness on unadjusted analysis, but moderate social jetlag (1 to <2 h) was not (Table 2). After adjusting for age and ethnicity, moderate and severe social jetlag were significantly associated with daytime sleepiness compared to nil/minimal social jetlag (Table 2) Table S1). While the effect sizes for moderate and severe social jetlag for each site are consistent with the overall findings, there is large uncertainty around the results due to the diminished sample sizes, and as such we cannot make firm conclusions about differences by recruitment site/country (Appendix S2, Table S1).
In the sensitivity analysis using an alternative definition of sleepiness defined as a risk of dozing off (ESS score >10), feeling sleepy, or both symptoms, results showed that the difference in sleepiness among social jetlag categories was largely driven by subjects who were both at risk of dozing off and were feeling sleepy (Appendix S3, However, among those with only one of the two sleepiness symptoms, the ESS score was not significantly associated with social jetlag (Appendix S3, Table S3).
We note that social jetlag was not significantly associated with daytime sleepiness after adjusting for all significant covariates from sleepiness classified as a risk of dozing off, feeling sleepy three or more times during the day, or both symptoms, we found that the association appeared to be driven by patients who reported both a risk of dozing off and feeling sleepy.
The present findings are consistent with the results of a previous study that found a significant association between social jetlag and daytime sleepiness in Japanese adults without OSA (Okajima et al., 2021). In that study, social jetlag significantly contributed to daytime sleepiness in individuals with <2 h of sleep debt. However, social jetlag was no longer significantly associated with daytime sleepiness in individuals with >2 h of sleep debt (Okajima et al., 2021). In the present study, both moderate and severe social jetlag were associated with increased daytime sleepiness compared to nil/minimal social jetlag in OSA. This suggests that regular sleep timing to reduce social jetlag may reduce excessive daytime sleepiness in patients with OSA.
This simple therapeutic approach could be easily trialled in patients with OSA. While our study examined only untreated patients, the impact of social jetlag in treated patients should be explored in the future. This is particularly important given the high prevalence of residual sleepiness in patients with OSA who are treated with positive airway pressure devices. Previous studies have found that 12%-55% of patients with OSA have residual sleepiness despite efficacious treatment (Gasa et al., 2013;Koutsourelakis et al., 2009;Pepin et al., 2009). Understanding the association between social jetlag and residual sleepiness might open new pathways and offer more personalised approaches when treating patients with OSA.
We did not adjust for sleep debt (sleep duration) in our modelling because of the potential for overadjustment according to our causal diagram and the large amount of missing data for habitual sleep duration, which would lead to selection bias. However, habitual sleep durations are on average lower in those with more severe social jetlag (Table 1), and less habitual sleep is also a cause of social jetlag. Future studies should clarify whether social jetlag and other forms of circadian disruption significantly contribute to sleepiness and other OSA symptomology once sleep debt is taken into account.
Another factor in patients with OSA that may contribute to the association between social jetlag and daytime sleepiness is the amount of rapid eye movement (REM) sleep. People with more severe social jetlag typically sleep later and longer on weekends. REM sleep has an inherent circadian variation, with greater amounts of REM sleep in the morning (Czeisler et al., 1980;Endo et al., 1981). REM sleep has been associated with more severe OSA due to an increase in upper airway collapsibility and reduced respiratory drive (Varga & Mokhlesi, 2019 (Koritala et al., 2021;Stenvers et al., 2019). OSA has also been found to alter the circadian rhythmicity of blood pressure (Butler et al., 2020). OSA has been associated with increased risks of cardiovascular disease, metabolic dysfunction, and cognitive impairment (Krysta et al., 2017;Levy et al., 2007), and circadian misalignment has also been linked to a similar range of health conditions (Chen et al., 2020;James et al., 2017;Sudy et al., 2019). Thus, further studies should explore the contribution and impact of circadian disruption in OSA and examine the relationship between social jetlag and other consequences of OSA.

| CONCLUSION
In conclusion, social jetlag is a form of chronic circadian disruption

ACKNOWLEDGMENTS
We would like to acknowledge the participants and investigators of the SAGIC for contributing their data to this study. Open access publishing facilitated by The University of Sydney, as part of the Wiley - The University of Sydney agreement via the Council of Australian University Librarians.

FUNDING INFORMATION
No sources of funding were received for this study.

CONFLICT OF INTEREST
Non-financial disclosure: Peter Cistulli has an appointment to an endowed academic Chair at the University of Sydney that was created All other authors have no conflicts of interest to declare.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the the Sleep Apnoea Global Interdisciplinary Consortium upon reasonable request.