Abstract
Study Objectives
Irregular sleep is a major risk factor for adverse health. In a global sample with technology-enabled long-term objective sleep data spanning 3.5 years, we investigated variability in sleep duration and timing over weekdays, months, seasons, and years.
Methods
Registered users of an FDA-cleared under-mattress sleep sensor who had ≥28 nights of sleep recordings and averaged ≥4 nights per/week between January 2020 and September 2023 were included for analyses. Generalized nonlinear fixed effects models were used to assess associations between sleep duration and sleep timing with weekday, month, season, and year. Sub-group analyses were conducted by age, sex, and location.
Results
Data from 116 879 adults (90 333 males, 26 546 females) aged 49 ± 14 years were analyzed. Weekday variation was observed, with 20–35 minutes longer sleep duration on weekends versus weekdays. Time to bed and time out of bed were 30–40 minutes and 60–80 minutes later on weekends, respectively. Seasonal variation in sleep duration was also evident; sleep duration was 15–20 minutes longer during winter in the northern hemisphere, 15–20 minutes shorter during summer in the southern hemisphere, and variations reduced closer to the equator. Sleep duration decreased from 2020 to 2023 but the effect was small (2.5 minutes).
Conclusions
These novel findings underscore the seasonal nature of human sleep, influenced by demographics and geography.
Keywords: big data, aging, sleep technology, sleep health, sleep duration, sleep timing, sleep variability, sleep regularity
Graphical Abstract
Graphical Abstract.
Statement of Significance.
Sleeping at varying times and for varying durations has been associated with poor health in past studies. Sleep patterns likely vary due to a range of factors, including work and study commitments, family responsibilities, and environmental influences. In a large global sample, we found large weekday and seasonal associations with sleep. People on average spent 20–35 minutes longer sleeping in on weekends compared to days during the week, and 15–20 minutes longer sleeping during winter compared to summer. Geographic location influenced this seasonal trend, whereby the season was more impactful on sleep for people further away from the equator. These novel findings highlight the natural variations in human sleep.
Irregular sleep has emerged as a critical risk factor for multiple adverse health and daytime functional outcomes. Variations in sleep timing (bedtimes and wake times) and sleep duration are associated with metabolic disorders, type 2 diabetes, obesity, hypertension, cardiovascular disease, poor mental health, and mortality [1–4]. A recent study of 60 977 people from the UK Biobank dataset found that sleep regularity assessed over 5–7 days was a stronger predictor of mortality risk than sleep duration [3]. Another analysis from the UK Biobank of 103 712 people found irregular sleep duration and delayed sleep onset timing over 5–7 days were associated with cardiovascular disease incidence [5]. We recently showed that irregular sleep is associated with an increased risk of hypertension, even when the average sleep duration is within the recommended healthy range of 7–9 hours per night [6]. For people with obstructive sleep apnea who are already at greater risk of hypertension than healthy sleepers, irregular sleep further increases hypertension risk [7]. Irregular sleep is also associated with adverse mental health outcomes. Indeed, a pooled analysis of eight studies (N = 3053) indicated that irregular sleep duration over 7–14 nights is associated with more depressive and insomnia symptoms in people with insomnia [8]. Given the reported associations between irregular sleep and adverse health outcomes, it is important to determine the degree to which sleep varies across the world in a large, long-term sample which historically has not been feasible due to technological limitations.
Sleep duration and timing vary across weekdays. Weekend catch-up sleep is a prevalent modern phenomenon whereby people often sleep shorter on commitment days (typically weekdays) and longer/later on non-commitment days (typically weekends) [9]. While some studies have associated weekend catch-up sleep with poorer work–academic performance and health outcomes, these effects are complex and subject to ongoing debate [10]. Increasing evidence suggests that weekend catch-up sleep may have beneficial effects on health when weekday sleep duration is low, potentially by mitigating the negative impacts of sleep restriction [11, 12]. It is important to note that weekend catch-up sleep and social jetlag (the difference between sleep midpoints on workdays vs. free days) are related but distinct constructs, each with unique implications for health. A recent review concluded that social jetlag is associated with poor work–academic performance, metabolic health conditions, and a 20% increased risk of obesity [13, 14]. In a seminal cross-sectional study (N = 64 110), at least 1h of social jetlag was observed in 69% of respondents, with ~30% reporting a catch-up of ≥2 hours [15]. A recent analysis of 50 million nights over a year collected with wearables (N = 220 000) showed that people slept 5–25 minutes longer on the weekend versus weekdays [16]. Marked variations were observed across countries with greater weekend sleep extension observed in northern Europe and less weekend sleep extension in countries closer to the equator.
Sleep duration and timing also appear to vary seasonally. In a Swedish cohort of 19 254 people, short (≤6 hours) self-reported sleep duration was more common in summer than in autumn [17]. Studies that examined cohorts of 216 and 500 people in the United States found that objective sleep duration was lowest and wake times were earliest in spring, corresponding with longer day lengths and greater light exposure, respectively [18, 19]. In Japan, a study of 1856 people found longer sleep onset latency (SOL), greater time awake after sleep onset, and reduced sleep efficiency in summer versus winter months [20]. Together, these studies suggest that sleep varies across weekdays and seasons and that social and environmental factors likely contribute to these effects. However, the broader geographical distribution of these seasonal variations in sleep remains unclear.
While there is increased focus on sleep regularity in the sleep literature, two reviews [21, 22] have concluded that only 32%–37% of studies have objectively assessed sleep regularity. Most prior studies have relied upon self-reported sleep, which is prone to recall bias. Furthermore, only ~15% have recorded sleep data for >14 days at any time point [21]. No study has objectively monitored sleep variability within individuals continuously over multiple years and in multiple countries. This is essential to accurately elucidate sleep variability across weekdays–weekends, as well as monthly, seasonal, and yearly variations in sleep across geographic locations.
The growing use of consumer sleep trackers in the global community presents an opportunity to analyze sleep patterns over extended recording periods on a much greater scale than previously possible. This longitudinal study reports objective sleep durations and sleep timing collected over ~3.5 years from an FDA-approved, validated under-mattress sleep sensor in 116 879 people across 41 countries worldwide. We determined variability in sleep duration and sleep timing (bedtime, midpoint, wake time) across weekdays–weekends, months, seasons, and several years. These associations were examined across age, sex, and geographical location to describe how these factors are associated with sleep variability across time. A secondary analysis was also conducted to examine changes in sleep regularity across seasons.
Methods
Participants
Data were collected from 125 295 participants who registered to use a Withings Sleep Analyzer under-mattress sleep sensor between January 2020 and September 2023. Inclusion criteria were ≥28 nights of sleep recordings and an average use of ≥4 times per week. The Withings Sleep Analyzer has an associated application for data synchronization and self-reported user information on age, sex, height, and weight. Geographical location was obtained as the closest town–city to the user based on their phone location shared in the app. Participants provided consent for their deidentified data to be used for research by agreeing to the terms and conditions. Ethics approval was obtained from the Flinders University Human Research Ethics Committee (Project number: 4291).
Study measures
The Withings Sleep Analyzer is an inflatable mat positioned under the user’s mattress at chest height. The device partially inflates to record changes in the device’s air pressure to detect body, respiratory, and cardiac motion (i.e. ballistography). Sleep parameters are then estimated from these signals via proprietary algorithms, including sleep duration (in hours), bedtime (time to bed, in clock time), wake time (time out of bed, in clock time), sleep onset (time asleep, in clock time), sleep latency (time taken to fall asleep, in minutes), and sleep offset (time awake, in clock time). The sleep midpoint was calculated as the middle time point between sleep onset and sleep offset. Multiple validation studies have shown that the Withings Sleep Analyzer is comparable in sleep–wake classification accuracy to other consumer sleep monitors and research actigraphy-based devices [23, 24], with sleep duration overestimated by ~30 mins on average compared to polysomnography [25], and confirmed by our in-house validations [26] including a study with >400 nights of comparison data [27]. The Withings Sleep Analyzer can also estimate the apnea–hypopnea-index (AHI) via ballistography and proprietary algorithms. Validation studies have shown that the estimated AHI is comparable to gold-standard polysomnography [28]. In a Supplementary Analysis, we used the average AHI to determine obstructive sleep apnea severity based on clinical cutoffs [29].
A secondary analysis was conducted to test for seasonal effects on sleep regularity. Sleep duration and timing regularity were defined as the standard deviation (SD) of sleep duration and timing per month, respectively. A given month for a participant was included in this additional analysis if ≥15 sleep recordings were available in that month, in addition to the inclusion criteria specified above.
Statistical analysis
Associations between sleep duration and sleep timing (bedtime, midpoint, wake time) with year, month, and day of the week were examined using generalized nonlinear fixed effects models with a random intercept per participant per year. This analysis design is based on the case-time-series methodology [30] and controls for all measurable and unmeasurable time-varying differences between time-stratum (here yearly) as well as subject-specific differences. These time-series models have been used in multiple studies to assess the effects of environmental factors on health [31–33]. We used natural splines (8 degrees of freedom) of the day of the year (1–365) to estimate seasonal variability in the outcomes of interest, and natural splines of the day of the week (1–7; 4 degrees of freedom) to estimate weekly variation. The analysis was undertaken in four geographical areas according to the latitude of the recording location (−90 to −30°, −30 to 0°, 0 to 30°, 30 to 90°). Sub-group analyses were conducted by age group (in 10-year bins) and sex. Estimated marginal means and 95%CI are reported in all figures. Given the large nature of this dataset, 95%CI (or even 99%CI) were often less than 1 min in sleep duration or timing from the mean; hence, our interpretation of the data focuses on the observed mean rather than statistical significance. A similar analytical framework was adopted for secondary analyses. Seasonal components of sleep duration and sleep timing regularity were examined using natural splines (6 degrees of freedom) of the month number (1–12). The models were implemented in the R programming language [34], using the dlnm [35] and gnm packages [36].
Sensitivity analyses
We conducted multiple additional sensitivity analyses to further validate our primary findings. We used a more conservative inclusion–exclusion criteria of ≥26 weeks of data. To understand the potential confounding effect of COVID-19 and associated changes in sleep, especially during lockdowns [37, 38], we reproduce our results by only including data before and after August 2022. Additionally, to avoid the potential confounding effect of OSA and associated change in sleep, we reproduced our main analysis in participants with no sleep apnea (AHI < 5).
Results
Participant characteristics
A total of 116 879 participants (93%) met the inclusion criteria (Figure S1). Participants were mostly male (77%) and middle-aged (Mean ± SD) 49 ± 14 years with a body-mass index (BMI) of 27.6 ± 5.6 kg/m2. As shown in Figure 1, participants were mainly located in Europe and Northern America, especially in the 30-to-90° latitudes (N = 111 006; 96%; Table 1). A total of 41 countries had data from more than 100 users. Participants had a median [IQR] of 585 [289, 912] sleep recordings available for analysis. ~73 million sleep recordings from January 2020 to September 2023 were included in the analyses. The mean sleep duration was 7.1 ± 1 hours, mean bedtime was 23:30 ± 1h30, mean SOL was 31 ± 12 minutes, mean sleep midpoint was 03:30 ± 1h20, mean wake time was 07:55 ± 1h25 (Table 1), and people spent an average of 6 ± 5 minutes in bed awake before getting out of bed.
Figure 1.
Geographical distribution of participants.
Table 1.
Participant Demographics and Sleep Characteristics.
| Latitude categories | ||||||
|---|---|---|---|---|---|---|
| Overall | −90 to −30° | −30 to 0° | 0 to 30° | 30 to 90° | ||
| n | 116 879 | 914 | 2973 | 1986 | 111 006 | |
| Age (y) | 49 (14) | 47 (14) | 45 (14) | 46 (13) | 49 (14) | |
| Sex, n (%) | Male | 90 333 (77.3) | 711 (77.8) | 2319 (78.0) | 1620 (81.6) | 85 683 (77.2) |
| Female | 26 546 (22.7) | 203 (22.2) | 654 (22.0) | 366 (18.4) | 25 323 (22.8) | |
| BMI (kg/m²) | 27.6 (5.6) | 28.3 (5.9) | 28.5 (6.4) | 26.6 (5.0) | 27.6 (5.6) | |
| Average number of nights (median [IQR]) | 585 [289, 912] |
369 [166, 678] |
293 [145,605] |
497 [249, 809] |
597 [300, 921] |
|
| Mean sleep duration (h) | 7.1 (1.0) | 7.1 (1.0) | 7.2 (1.1) | 6.8 (1.0) | 7.1 (1.0) | |
| Mean time to bed (hh:mm) | 23:30 (1:30) | 23h30 (1h35) | 23h05 (1h35) | 23h55 (1h40) | 23h30 (1h30) | |
| Mean sleep onset latency (min) | 31 (12) | 33 (12) | 34 (13) | 32 (11) | 31 (12) | |
| Mean sleep midpoint, (hh:mm) | 03:30 (1:20) | 03:10 (1:30) | 03:20 (1:25) | 03:50 (1:40) | 03:30 (1:20) | |
| Mean time out of bed, (hh:mm) | 7h55 (1:25) | 7:30 (1:30) | 7:50 (1:20) | 8:05 (1:35) | 7:55 (1:25) | |
| Mean time to bed before getting up (min) | 6 (5) | 6 (6) | 7 (6) | 6 (5) | 6 (5) | |
| SD sleep duration (min) | 86 (39) | 91 (44) | 89 (44) | 93 (39) | 85 (39) | |
| SD time to bed (min) | 94 (65) | 107 (73) | 102 (73) | 112 (69) | 93 (65) | |
| SD sleep midpoint (min) | 82 (53) | 92 (61) | 89 (58) | 97 (58) | 82 (53) | |
| SD time out of bed (min) | 90 (45) | 97 (52) | 96 (48) | 102 (53) | 90 (45) | |
| Habitual sleep duration category, n (%) | < 6h | 13 638 (11.7) | 106 (11.6) | 305 (10.3) | 397 (20.0) | 12 830 (11.6) |
| 6 to 7h | 38 220 (32.7) | 290 (31.7) | 872 (29.3) | 775 (39.0) | 36 283 (32.7) | |
| 7 to 8h | 48 434 (41.4) | 381 (41.7) | 1246 (41.9) | 651 (32.8) | 46 156 (41.6) | |
| 8 to 9h | 13 973 (12.0) | 115 (12.6) | 441 (14.8) | 137 (6.9) | 13,280 (12.0) | |
| >9h | 2614 (2.2) | 22 (2.4) | 109 (3.7) | 26 (1.3) | 2457 (2.2) | |
BMI, body mass index; SD, standard deviation; TST, total sleep time.
Seasonal and weekday effects on sleep duration and sleep timing
Seasonal variations in sleep duration were observed (Figure 2A) and were dependent on latitudes. Using the 21st of June as the reference (longest and shortest sunlight day in the northern and southern hemispheres, respectively), sleep duration was 15–25 minutes longer during winter (November–January) in the northern hemisphere (latitudes from 30 to 90°) and 15 to 20 minutes shorter during summer in the southern hemisphere (latitude −90 to −30°; Figure 2A). Northern locations closer to the equator (latitude 0 to 30°) had smaller seasonal variations with approximately 10 minutes difference for winter versus summer. The same was not true for southern locations closer to the equator (latitude −30 to 0°), but the sample sizes were lower for this geographic region (Table 1).
Figure 2.
Seasonal variation in sleep duration and sleep timing. (A) Variation in sleep duration for different latitudes categories (−90 to −30°, −30 to 0°, 0 to 30°, 30 to 90°) across seasons, using the 21st of June as the reference (summer and winter solstice). Seasonal variation in (B) time to bed (C) sleep onset latency (SOL) (D) sleep midpoint and (E) time out of bed were estimated for two latitude categories (−90 to −30° and 30 to 90°). Bolded values on the bottom right of each graph represent the sample average, rounded to the closest 5 minutes.
In the northern hemisphere, time to bed was the latest during both summer and winter months, being ~20 minutes earlier during February–March and October–November (Figure 2B). Time out of bed was also ~25 minutes later during winter versus summer months (Figure 2E). Given the later bedtime and later wake time in winter, the sleep midpoint was also later (~10 to 15 minutes) during winter versus summer months (Figure 2D). Seasonal variation in SOL also occurred (<5 minutes annual variation; Figure 2C).
In the southern hemisphere, seasonal variations in sleep timing were observed in summer (December–March), indicating a reversed trend with respect to the season. Compared to winter months, time to bed was delayed by ~27 minutes (Figure 2B), sleep midpoint was delayed by ~22 minutes (Figure 2D), and time out of bed was delayed by ~10 minutes (Figure 2E) accounting for the reduction in sleep duration of ~17 minutes due to reduced sleep opportunity (Figure 2A).
Weekday variations in sleep duration were detected, with 15–35 minutes longer sleep duration on weekends versus weekdays (Figure 3A). The weekday–weekend (Thursday vs. Saturday) difference (ΔTST) was similar across latitudes, except for a 10 minutes greater difference observed in ΔTST for latitudes 30 to 90° versus latitudes 0 to 30° (Figure 3A). ΔTST was also similar across categories of habitual sleep duration (Figure 3B). However, ΔTST was different across age categories, whereby there was the largest increase in weekend sleep duration in 40 to 60-year-olds, a smaller increase in weekend sleep duration from ≥60 onwards, and less than 10 minutes extra catch-up sleep in the ≥80-year-olds (Figure 3C). Time to bed and time out of bed were much later during weekends versus the weekly average, with 30 to 37 minutes and 62 to 80 minutes differences compared to midweek sleep timings, respectively (Figure 3D).
Figure 3.
Weekly variation in sleep duration and sleep timing. (A Variation in sleep duration for different latitudes categories (−90 to −30°, −30 to 0°, 0 to 30°, 30 to 90°) across days of the week, using Wednesday as the reference. (B) Variation in weekend vs weekday (ΔTST) across habitual sleep duration categories, and (C) across age categories, for the northern hemisphere only. (D) Variation in time to bed and time out of bed duration across the week, for the 30 to 90° latitude category.
Yearly trends were found for sleep duration but with relatively small effects (2023 vs. 2020; −2.5 minutes change in sleep duration; Figure S1). Similarly, we observed ~4 min earlier bedtime and ~2.5 minutes earlier time out of bed in 2023 versus 2020 (Figure S2), respectively. None of the main findings changed during sensitivity analyses (Figures S4 to S8). There were only small differences between sexes in seasonal and weekly variation in sleep timing and duration are available in Figures S11 to S12.
Seasonal effects on sleep duration and timing regularity
There were 2 568 140 person-months of data available for this analysis. Seasonal variations were observed in sleep duration, time to bed, and time out of bed regularity, whereby sleep duration and timing regularity were higher during winter versus summer months (Figure 4, A, C, and E). In the northern hemisphere (latitude 30 to 90°), the seasonal variation effect was larger on sleep timing regularity than on sleep duration regularity. Indeed, time to bed irregularity was 13.7 minutes greater during winter months, which was a ~22% increase compared to the median (62 minutes, Figure 4F). Similarly, the increase in time out of bed regularity was 19% of the median values (13 minutes). However, sleep duration irregularity was 8 minutes higher during winter versus summer, equivalent to a ~12% increase from the median sleep duration regularity (69 minutes, Figure 4B). A similar effect was observed in the southern hemisphere (−30 to −90°), whereby winter months had the largest sleep duration and timing irregularity by 5 to 10 minutes. The seasonal effect also appeared stronger on sleep timing regularity than on sleep duration regularity. We observed similar seasonal effects on sleep duration and timing in northern latitudes closer to the equator (latitude 0 to 30°) but not in southern latitudes closer to the equator (latitude −30 to 0°; Figure S3). Seasonal effects on sleep duration and timing irregularity were similar across the sexes (Figure S13).
Figure 4.
Seasonal variation in monthly sleep duration and timing regularity. (A) Variation in sleep duration regularity (B) Distribution of monthly sleep duration regularity calculated in the dataset. (C–F) are similar to (A) and (B) but for time to bed and time out of bed regularity, respectively. The median (dashed line) and 25th and 75th (dotted line) are highlighted in each distribution.
Discussion
Using a large global sample with 3.5 years of objective, under-mattress sleep sensor data, we showed substantial variation in sleep timing and duration across weekdays and seasons. We found shorter sleep duration in summer versus winter, longer and delayed sleep on weekends versus weekdays, and greater irregularity in sleep duration and timing from December to January These effects were relatively large (in the order of 15 to 35 minutes on sleep duration) supporting that time-related phenomena have a strong effect on sleep duration and timing that warrants consideration. Changes in sleep across years were very small, although these findings are likely impacted by the COVID-19 pandemic to some extent.
Sleep duration was lowest during summer months in both the northern and southern hemispheres by 15 to 20 minutes. This finding is consistent with a large population-based sample (N = 20 000) in Sweden that showed a greater prevalence of short sleep duration during summer [17], but contrasts with a smaller study (N = 216) in the United States that found shorter sleep duration was more common in spring [18]. In the current study, time to bed and time out of bed were later during the respective summer months in both hemispheres, which is consistent with previous reports. Furthermore, time to bed and time out of bed were latest between December and January by 20–27 minutes in both hemispheres, potentially due to holiday periods. The current study also uniquely tested for variations in seasonal effects by geographic location. We found that associations between sleep duration and timing with months and seasons were stronger for people living in −90 to −30° and 30 to 90° latitude categories compared to −30 to 0° and 0 to 30° categories closer to the equator. Collectively, these findings on seasonal variations in sleep duration and timing indicate that environmental factors contribute to seasonal variations in sleep, such as light, temperature, and humidity. This conclusion is supported by past studies that have demonstrated the adverse impacts of harsh environmental conditions on sleep [39, 40] and the profound effect of light on maintaining a stable sleep schedule [41, 42].
Sleep duration varied considerably across the week, such that sleep duration was 20 to 35 minutes longer on weekends than on weekdays. This increased sleep duration was facilitated by a longer time in bed (sleep opportunity): people went to bed about 30 minutes later on the weekend, but they got out of bed 62 to 80 minutes later, resulting in an increase in time in bed. Some past studies have shown self-reported weekend catch-up sleep of 1 hour or more [15], but the current study found a larger effect than the recently reported weekend catch-up sleep of 5–25 minutes from >50 million sleep recordings with an objective sleep ring [43]. However, consistent with the sleep ring study [43], geographical location effects were observed such that weekend catch-up sleep was longest in the 30 to 90° latitude category. The many ways of measuring irregular sleep likely account for some of the differences between studies, with different sleep irregularity variables known to capture different patterns of irregularity [44].
Irregular sleep likely disrupts homeostatic sleep mechanisms and circadian regulation of cellular functions, negatively impacting physiological and psychological functioning and health [45]. The National Sleep Foundation has recently released a consensus statement about the importance of sleep regularity to health, outlining evidence to support that irregular sleep can be harmful [46]. This consensus statement reviewed 63 articles to conclude that irregular sleep (≥1-hour increase in the standard deviations of sleep duration and timing) is associated with adverse cardiovascular, metabolic, mental health, and performance outcomes. However, the authors also concluded that a weekend catch-up sleep may be beneficial to mitigate sleep debt in some instances where longer weekday sleep is not possible. Interestingly, weekend catch-up sleep was not influenced by habitual sleep duration in this study, which is contrary to previous conclusions that weekend sleep extension may be due to insufficient sleep on weeknights. The potential health benefits conferred by compensatory sleep behaviors over the weekend versus potential detriments associated with irregular sleep patterns is a matter of debate [47]. Interventional studies are needed to determine the consequences of this tradeoff between catching up on sleep and increasing sleep irregularity.
A multitude of factors likely contribute to irregular sleep. Social factors such as family responsibilities, work commitments, and socializing, psychological factors such as sleep prioritization and bedtime procrastination, and biological factors such as homeostatic sleep and circadian rhythm drivers likely all contribute to sleep regularity to some degree. Environmental factors also likely contribute to seasonal and weekly variations in sleep, including temperature, humidity, noise, and light [39]. Behavior factors such as alcohol intake or smoking have also been shown to have seasonal and weekly variations [48], which could also explain some of the observed variations in sleep. Given the known importance of light for circadian health, light exposure regularity is also emerging as a predictor of sleep regularity [42]. This study found that both weekday and seasonal-related phenomena influence sleep duration and timing and sleep regularity and that these effects vary by geographic location. Age also had a strong effect, with middle-aged adults showing the longest weekend catch-up sleep. This might suggest that work or family/caregiving responsibilities lead to voluntary sleep restriction during weekdays, resulting in a strong effect on weekday variability. Additionally, the increased homeostatic sleep drive in young and middle-aged adults compared to older adults may also contribute to this pattern [49]. Further observational and experimental studies are needed to examine the potential causal mechanisms underpinning sleep regularity. Irrespective of the cause, sleep, and circadian research studies clearly need to consider weekday and seasonal variations in sleep outcomes in both study designs and statistical analyses.
Given the known associations between irregular sleep and adverse health outcomes, the weekly sleep variations found in the current study warrant consideration. While seasonal variations in sleep were relatively large in this study, more rapid changes across the week are of most concern given that these pose a greater circadian challenge. Although maintaining a consistent sleep schedule is already considered a key aspect of the sleep hygiene recommendations [50], this work has highlighted the marked sleep irregularity worldwide [51]. The importance of sleep regularity needs to be recognized in future public health messaging and interventions to promote sleep health. A degree of variability in sleep duration and timing can be expected across time, so messaging should not be too strict as to restrict natural variations in sleep. This said, what constitutes “unhealthy” sleep irregularity is yet to be established: a gap that is critical to address to inform public messaging. Nonetheless, experimental studies are needed to definitively establish causal mechanisms, and test whether interventions to reduce sleep irregularity produce beneficial effects on health and function.
The use of an under-mattress sensor to objectively assess in-home sleep nightly in a large global sample is a major methodological strength of the current study. This noninvasive monitoring approach has enabled the practical, routine collection of valuable sleep data in ~116 000 people, which would otherwise be infeasible. This approach is particularly advantageous for the assessment of intraindividual sleep regularity, given the ability to monitor sleep for up to 3.5 years within individuals. Nonetheless, there are noteworthy limitations to the current study. In this sample of real-world users of an under-mattress sleep sensor, information about many participant characteristics and health conditions are unknown, and therefore, could not be controlled for in the analyses. The presence of bed partners or pets sleeping on the bed is also unknown, which may impact device accuracy. The sample is also skewed towards individuals sufficiently interested, willing, and able to purchase sleep sensor technology to routinely monitor their sleeping patterns. However, the lack of necessary charging or regular syncing does not preclude disinterested individuals as much as with other consumer devices. Finally, while the geographical locations of participants were diverse, there were fewer users in the southern hemisphere and equatorial regions; hence, latitude differences in sleep duration and timing should be interpreted with caution.
Using data collected with an under-mattress sleep sensor over a 3.5-year period, the current study has expanded our understanding of time-related variations in sleep duration and sleep timing. Relatively large effects of day of the week and seasonal changes were observed on sleep duration and sleep timing and were dependent upon geographic location and age. Given the importance of regular sleep to maintain good health, future research is required to further determine the effects of biological, psychological, and social factors on sleep regularity. Nevertheless, the novel findings of this study highlight that sleep is a behavior profoundly impacted by multiple behavioral and environmental factors, including time-related phenomena.
Supplementary material
Supplementary material is available at SLEEP online.
Acknowledgments
DJE is supported by a National Health and Medical Research Council of Australia Leadership Fellowship (1196261). BL is supported by an NHMRC of Australia Emerging Leadership Fellowship (2025886).
Contributor Information
Hannah Scott, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Bastien Lechat, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Kelly Sansom, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Lucia Pinilla, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Jack Manners, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Andrew J K Phillips, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Duc Phuc Nguyen, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Sebastien Bailly, HP2 Laboratory, University Grenoble Alpes, Grenoble, France.
Jean-Louis Pepin, HP2 Laboratory, University Grenoble Alpes, Grenoble, France.
Pierre Escourrou, Centre Interdisciplinaire du Sommeil, Paris, France.
Ganesh Naik, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Peter Catcheside, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Danny J Eckert, Adelaide Institute for Sleep Health and FHMRI Sleep Health, Flinders University, Australia.
Disclosure Statement
HS reports research support unrelated to this work from Re-Time Pty Ltd, Withings Ltd, Compumedics Ltd, the American Academy of Sleep Medicine Foundation, and Flinders University. AJKP has received research funding from Delos and Versalux, and he is a co-founder of Circadian Health Innovations PTY LTD. Outside the current work, DJE reports research grants from Bayer, Apnimed, Invicta Medical, Withings, Eli Lilly and Takeda and serves on Scientific Advisory Boards for Invicta Medical, Mosanna, SleepRes and Apnimed and as a consultant for Takeda.
Author Contributions
H.S., B.L., K.S., G.N., J.M., and D.J.E. developed the study concepts and aims. B.L. performed data extraction and data analysis. All authors provided important insight into data interpretation and drafting of the manuscript. All authors approve of the final version of the manuscript.
Data Availability
Deidentified data that support the findings of this study, including individual data, are available from the corresponding author upon request subject to ethical and data custodian (Withings) approval.
References
- 1. Parise BK, Santos RB, Mesas AE, et al. Sleep irregularity and the association with hypertension and blood pressure levels: the ELSA-Brasil study. J Hypertens. 2023;41(4):670–677. doi: https://doi.org/ 10.1097/HJH.0000000000003392 [DOI] [PubMed] [Google Scholar]
- 2. Lunsford-Avery JR, Engelhard MM, Navar AM, Kollins SH. Validation of the sleep regularity index in older adults and associations with cardiometabolic risk. Sci Rep. 2018;8(1):14158. doi: https://doi.org/ 10.1038/s41598-018-32402-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Windred DP, Burns AC, Lane JM, et al. Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep. 2024;47(1):zsad253. doi: https://doi.org/ 10.1093/sleep/zsad253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Zuraikat FM, Makarem N, Redline S, Aggarwal B, Jelic S, St-Onge MP. Sleep regularity and cardiometabolic heath: is variability in sleep patterns a risk factor for excess adiposity and glycemic dysregulation? Curr Diab Rep. 2020;20(8):38. doi: https://doi.org/ 10.1007/s11892-020-01324-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Nikbakhtian S, Reed AB, Obika BD, et al. Accelerometer-derived sleep onset timing and cardiovascular disease incidence: a UK Biobank cohort study. Eur Heart J Digit Health. 2021;2(4):658–666. doi: https://doi.org/ 10.1093/ehjdh/ztab088 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Scott H, Lechat B, Guyett A, et al. Sleep irregularity is associated with hypertension: findings from over 2 million nights with a large global population sample. Hypertension. 2023;80:1117–1126. doi: https://doi.org/ 10.1161/hypertensionaha.122.20513 [DOI] [PubMed] [Google Scholar]
- 7. Sansom K, Reynolds A, Windred D, et al. The interrelationships between sleep regularity, obstructive sleep apnea, and hypertension in a middle-aged community population. Sleep. 2024;47(3):zsae001. doi: https://doi.org/ 10.1093/sleep/zsae001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Messman BA, Wiley JF, Feldman E, Dietch JR, Taylor DJ, Slavish DC. Irregular sleep is linked to poorer mental health: a pooled analysis of eight studies. Sleep Health. 2024;10:493–499. doi: https://doi.org/ 10.1016/j.sleh.2024.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Wittmann M, Dinich J, Merrow M, Roenneberg T. Social jetlag: misalignment of biological and social time. Chronobiol Int. 2006;23(1-2):497–509. doi: https://doi.org/ 10.1080/07420520500545979 [DOI] [PubMed] [Google Scholar]
- 10. Chaput J-P, Biswas RK, Ahmadi M, et al. Device-measured weekend catch-up sleep, mortality, and cardiovascular disease incidence in adults. Sleep. 2024;47(11):zsae135. doi: https://doi.org/ 10.1093/sleep/zsae135 [DOI] [PubMed] [Google Scholar]
- 11. Peixoto de Miranda EJF, Parise BK, Santos RB, et al. Weekend sleep extension, social jetlag and incidence of coronary calcium score: the ELSA-Brasil study. Sleep. 2025;zsaf039. doi: https://doi.org/ 10.1093/sleep/zsaf039 [DOI] [PubMed] [Google Scholar]
- 12. Zhu H, Qin S, Wu M. Association between weekend catch-up sleep and cardiovascular disease: evidence from the National Health and Nutrition Examination Surveys 2017-2018. Sleep Health. 2024;10(1):98–103. doi: https://doi.org/ 10.1016/j.sleh.2023.09.006 [DOI] [PubMed] [Google Scholar]
- 13. Caliandro R, Streng AA, van Kerkhof LW, van der Horst GT, Chaves I. Social jetlag and related risks for human health: a timely review. Nutrients. 2021;13(12):4543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Arab A, Karimi E, Garaulet M, Scheer FA. Social jetlag and obesity: a systematic review and meta‐analysis. Obes Rev. 2024;25(3):e13664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Roenneberg T, Allebrandt KV, Merrow M, Vetter C. Social jetlag and obesity. Curr Biol. 2012;22(10):939–943. doi: https://doi.org/ 10.1016/j.cub.2012.03.038 [DOI] [PubMed] [Google Scholar]
- 16. Willoughby AR, Alikhani I, Karsikas M, Chua XY, Chee MWL. Country differences in nocturnal sleep variability: observations from a large-scale, long-term sleep wearable study. Sleep Med. 2023;110:155–165. doi: https://doi.org/ 10.1016/j.sleep.2023.08.010 [DOI] [PubMed] [Google Scholar]
- 17. Titova OE, Lindberg E, Elmståhl S, Lind L, Benedict C. Seasonal variations in sleep duration and sleep complaints: a Swedish cohort study in middle‐aged and older individuals. J Sleep Res. 2022;31(1):e13453. doi: https://doi.org/ 10.1111/jsr.13453 [DOI] [PubMed] [Google Scholar]
- 18. Mattingly SM, Grover T, Martinez GJ, et al. The effects of seasons and weather on sleep patterns measured through longitudinal multimodal sensing. NPJ Digital Med. 2021;4(1):76. doi: https://doi.org/ 10.1038/s41746-021-00435-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Dunster GP, Hua I, Grahe A, et al. Daytime light exposure is a strong predictor of seasonal variation in sleep and circadian timing of university students. J Pineal Res. 2023;74(2):e12843. doi: https://doi.org/ 10.1111/jpi.12843 [DOI] [PubMed] [Google Scholar]
- 20. Hashizaki M, Nakajima H, Shiga T, Tsutsumi M, Kume K. A longitudinal large-scale objective sleep data analysis revealed a seasonal sleep variation in the Japanese population. Chronobiol Int. 2018;35(7):933–945. doi: https://doi.org/ 10.1080/07420528.2018.1443118 [DOI] [PubMed] [Google Scholar]
- 21. Bei B, Wiley JF, Trinder J, Manber R. Beyond the mean: a systematic review on the correlates of daily intraindividual variability of sleep/wake patterns. Sleep Med Rev. 2016;28:108–124. doi: https://doi.org/ 10.1016/j.smrv.2015.06.003 [DOI] [PubMed] [Google Scholar]
- 22. Chaput J-P, Dutil C, Featherstone R, et al. Sleep timing, sleep consistency, and health in adults: a systematic review. Appl Physiol Nutr Metab. 2020;45(10):S232–S247. doi: https://doi.org/ 10.1139/apnm-2020-0032 [DOI] [PubMed] [Google Scholar]
- 23. Lee T, Cho Y, Cha KS, et al. Accuracy of 11 wearable, nearable, and airable consumer sleep trackers: prospective multicenter validation study. JMIR mHealth uHealth. 2023;11:e50983. doi: https://doi.org/ 10.2196/50983 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Chinoy ED, Cuellar JA, Huwa KE, et al. Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep. 2021;44(5):zsaa291. doi: https://doi.org/ 10.1093/sleep/zsaa291 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Edouard P, Campo D, Bartet P, et al. Validation of the Withings Sleep Analyzer, an under-the-mattress device for the detection of moderate-severe sleep apnea syndrome. J Clin Sleep Med. 2021;17:1217–1227. doi: https://doi.org/ 10.5664/jcsm.9168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Scott H, Lechat B, Guyett A, et al. Sleep irregularity is associated with hypertension: findings from over 2 million nights with a large global population sample. Hypertension. 2023;80(5):1117–1126. doi: https://doi.org/ 10.1161/HYPERTENSIONAHA.122.20513 [DOI] [PubMed] [Google Scholar]
- 27. Manners J, Kemps E, Lechat B, Catcheside P, Eckert D, Scott H. Performance evaluation of an under-mattress sleep sensor versus polysomnography in ≥ 400 nights with healthy and unhealthy sleep. medRxiv. 2024;e14480:2024.09.09.24312921. doi: https://doi.org/ 10.1101/2024.09.09.24312921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Edouard P, Campo D, Bartet P, et al. Validation of the Withings Sleep Analyzer, an under-the-mattress device for the detection of moderate-severe sleep apnea syndrome. J Clin Sleep Med. 2021;17(6):1217–1227. doi: https://doi.org/ 10.5664/jcsm.9168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Kapur VK, Auckley DH, Chowdhuri S, et al. Clinical practice guideline for diagnostic testing for adult obstructive sleep apnea: an American Academy of Sleep Medicine clinical practice guideline. J Clin Sleep Med. 2017;13(3):479–504. doi: https://doi.org/ 10.5664/jcsm.6506 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Gasparrini A. The case time series design. Epidemiology. 2021;32(6):829–837. doi: https://doi.org/ 10.1097/EDE.0000000000001410 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Jones PJ, Koolhof IS, Wheeler AJ, et al. Characterising non-linear associations between airborne pollen counts and respiratory symptoms from the AirRater smartphone app in Tasmania, Australia: a case time series approach. Environ Res. 2021;200:111484. doi: https://doi.org/ 10.1016/j.envres.2021.111484 [DOI] [PubMed] [Google Scholar]
- 32. Quilty S, Jupurrurla NF, Lal A, et al. The relative value of sociocultural and infrastructural adaptations to heat in a very hot climate in northern Australia: a case time series of heat-associated mortality. Lancet Planetary Health. 2023;7(8):e684–e693. doi: https://doi.org/ 10.1016/S2542-5196(23)00138-9 [DOI] [PubMed] [Google Scholar]
- 33. Sun Y, Milando CW, Spangler KR, et al. Short term exposure to low level ambient fine particulate matter and natural cause, cardiovascular, and respiratory morbidity among US adults with health insurance: case time series study. BMJ. 2024;384:e076322. doi: https://doi.org/ 10.1136/bmj-2023-076322 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Gasparrini A. Modeling exposure-lag-response associations with distributed lag non-linear models. Stat Med. 2014;33(5):881–899. doi: https://doi.org/ 10.1002/sim.5963 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Gasparrini A. Distributed lag linear and non-linear models in r: the Package dlnm. J Stat Softw. 2011;43(8):1–20. [PMC free article] [PubMed] [Google Scholar]
- 36. gnm: Generalized Nonlinear Models. (Version 1.1-5) [Computer software]. 2023. [Google Scholar]
- 37. Pepin JL, Bailly S, Mignot E, et al. Digital markers of sleep architecture to characterize the impact of different lockdown regimens on sleep health during the COVID-19 pandemic. Sleep. 2022;45(6):zsac074. doi: https://doi.org/ 10.1093/sleep/zsac074 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Pepin JL, Bailly S, Mordret E, et al. Greatest changes in objective sleep architecture during COVID-19 lockdown in night owls with increased REM sleep. Sleep. 2021;44(9):zsab075. doi: https://doi.org/ 10.1093/sleep/zsab075 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Chevance G, Minor K, Vielma C, et al. A systematic review of ambient heat and sleep in a warming climate. Sleep Med Rev. 2024;75:101915. doi: https://doi.org/ 10.1016/j.smrv.2024.101915 [DOI] [PubMed] [Google Scholar]
- 40. Tsuzuki K, Okamoto-Mizuno K, Mizuno K. Effects of humid heat exposure on sleep, thermoregulation, melatonin, and microclimate. J Therm Biol. 2004;29(1):31–36. doi: https://doi.org/ 10.1016/j.jtherbio.2003.10.003 [DOI] [Google Scholar]
- 41. Wallace DA, Qiu X, Schwartz J, et al. Light exposure during sleep is bidirectionally associated with irregular sleep timing: the multi-ethnic study of atherosclerosis (MESA). Environ Pollut. 2024;344:123258. doi: https://doi.org/ 10.1016/j.envpol.2023.123258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Hand AJ, Stone JE, Shen L, et al. Measuring light regularity: sleep regularity is associated with regularity of light exposure in adolescents. Sleep. 2023;46(8):zsad001. doi: https://doi.org/ 10.1093/sleep/zsad001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Willoughby AR, Alikhani I, Karsikas M, Chua XY, Chee MW. Country differences in nocturnal sleep variability: observations from a large-scale, long-term sleep wearable study. Sleep Med. 2023;110:155–165. doi: https://doi.org/ 10.1016/j.sleep.2023.08.010 [DOI] [PubMed] [Google Scholar]
- 44. Fischer D, Klerman EB, Phillips AJ. Measuring sleep regularity: theoretical properties and practical usage of existing metrics. Sleep. 2021;44(10):zsab103. doi: https://doi.org/ 10.1093/sleep/zsab103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Vetter C. Circadian disruption: What do we actually mean? Eur J Neurosci. 2020;51(1):531–550. doi: https://doi.org/ 10.1111/ejn.14255 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Sletten TL, Weaver MD, Foster RG, et al. The importance of sleep regularity: a consensus statement of the National Sleep Foundation sleep timing and variability panel. Sleep Health. 2023;9(6):801–820. doi: https://doi.org/ 10.1016/j.sleh.2023.07.016 [DOI] [PubMed] [Google Scholar]
- 47. Larson O, Gehrman PR. Keeping the balance: the benefits of catch-up sleep versus the risks of sleep irregularity. US: Oxford University Press; 2024. p. zsae046. [DOI] [PubMed] [Google Scholar]
- 48. Sohal A, Bains K, Dhaliwal A, et al. Seasonal variations of hospital admissions for alcohol-related hepatitis in the United States. Gastroenterol Res. 2022;15(2):75–81. doi: https://doi.org/ 10.14740/gr1506 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Dijk DJ, Groeger JA, Stanley N, Deacon S. Age-related reduction in daytime sleep propensity and nocturnal slow wave sleep. Sleep. 2010;33(2):211–223. doi: https://doi.org/ 10.1093/sleep/33.2.211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Irish LA, Kline CE, Gunn HE, Buysse DJ, Hall MH. The role of sleep hygiene in promoting public health: a review of empirical evidence. Sleep Med Rev. 2015;22:23–36. doi: https://doi.org/ 10.1016/j.smrv.2014.10.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Scott H, Naik G, Lechat B, et al. Are we getting enough sleep? Frequent irregular sleep found in an analysis of over 11 million nights of objective in-home sleep data. Sleep Health. 2024;10(1):91–97. doi: https://doi.org/ 10.1016/j.sleh.2023.10.016 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Deidentified data that support the findings of this study, including individual data, are available from the corresponding author upon request subject to ethical and data custodian (Withings) approval.