Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2020 Nov 23;15(11):e0242852. doi: 10.1371/journal.pone.0242852

How do associations between sleep duration and metabolic health differ with age in the UK general population?

Anmol Arora 1, David Pell 2, Esther M F van Sluijs 2, Eleanor M Winpenny 2,*
Editor: Antonio Palazón-Bru3
PMCID: PMC7682906  PMID: 33227026

Abstract

Background

Despite a growing body of evidence suggesting that short sleep duration may be linked to adverse metabolic outcomes, how these associations differ between age groups remains unclear. We use eight years of data from the UK National Diet and Nutritional Survey (NDNS) (2008–2016) to analyse cross-sectional relationships between sleep duration and metabolic risk in participants aged 11–70 years.

Methods

Participants (n = 2008) who provided both metabolic risk and sleep duration data were included. Self-reported sleep duration was standardised by age, to account for differences in age-related sleep requirements. A standardised metabolic risk score was constructed, comprising: waist circumference, blood pressure, serum triglycerides, serum high-density lipoprotein cholesterol, and fasting plasma glucose. Regression models were constructed across four age groups from adolescents to older adults.

Results

Overall, decreased sleep duration (hrs) was associated with an increased metabolic risk (standard deviations) with significant quadratic (B:0.028 [95%CI: 0.007, 0.050]) and linear (B:-0.061 [95%CI: -0.111, -0.011]) sleep duration coefficients. When separated by age group, stronger associations were seen among mid-aged adults (36-50y) (quadratic coefficient: 0.038 [95%CI: 0.002, 0.074]) compared to other age groups (e.g. adolescents (11-18y), quadratic coefficient: -0.009 [95%CI: -0.042, 0.025]). An increased difference between weekend and weekday sleep was only associated with increased metabolic risk in adults aged 51–70 years (B:0.18 [95%CI: 0.005, 0.348]).

Conclusions

Our results indicate that sleep duration is linked to adverse metabolic risk and suggest heterogeneity between age groups. Longitudinal studies with larger sample sizes are required to explore long-term effects of abnormal sleep and potential remedial benefits.

Introduction

Cardiovascular disease and diabetes remain major public health concerns, with global burdens expected to increase further in the next 10 years. Globally, the economic burden is estimated to rise to at least $2.1 trillion for diabetes and to $1 trillion for cardiovascular diseases by 2030 [1, 2]. Indeed, the World Health Organisation has highlighted cardiovascular disease as the leading cause of death worldwide and diabetes as another common cause of death [3]. Diet and exercise are noted as two key risk factors for the development of metabolic diseases such as cardiovascular disease and type 2 diabetes [4]. However, there have been suggestions that other factors may play a modifiable role in the disease aetiologies, ranging from the gut microbiome to obesogenic HIV drugs [5]. One factor that may contribute to the development of metabolic diseases is short sleep duration. Prior research has found associations between sleep duration and both metabolic outcomes and diet quality in adults [6, 7]. As an example, Potter et al., (2017) noted that, in cross-sectional data from UK adults, short sleep duration was associated with increased body mass index (BMI) and several other metabolic markers [8]. Furthermore, experimental studies in adults have suggested that a relationship between sleep duration and metabolic outcomes may be mediated by increased dietary energy intake as well as physical activity [9]. These findings are supported by a systematic review and meta-analysis [10]. However, there have been relatively few studies analysing the impact of sleep on metabolic outcomes in younger age groups or how these relationships may change with age. Early research with relatively small sample sizes has indicated that sleep may influence metabolic health in these age groups, for example by altering insulin sensitivity [11, 12]. Sleep duration requirements are known to differ with age and, similarly, metabolic health is recognised as varying with age [13, 14]. Adolescents are generally recommended 8 to 10 hours sleep per night, decreasing with age to 7 to 8 hours sleep for adults aged over 65 [15–19]. The potentially complex interplay between age, sleep architecture and metabolic health is an understudied area of research. It has been suggested that factors such as diet may exert heterogeneous effects depending on age and this could also be true of sleep health [20].

Two key metrics in describing sleep are quality and duration. Early research has suggested the existence of associations between sleep duration and diet in young people [21, 22]. One of these studies even noted that sleep restriction in childhood was associated with less favourable BMI profiles in older age [23]. The authors concluded that addressing unhealthy sleep patterns in childhood may offer some relief to the increasingly burdensome obesity crisis. However, this remains a contestable area of research with others finding no such relationship in adolescents [24, 25]. Similarly, according to a recent review, the evidence between fluctuations in sleep duration between weekday and weekend nights and metabolic risk in young people is limited and inconclusive [26]. There is, however, a growing body of literature indicating that the difference between weekend and weekday night sleep is associated with unhealthy changes to eating behaviours and diet quality, though this is heavily reliant on cross-sectional data [27, 28]. Eating behaviours and diet quality would likely lie on a causative pathway between sleep and metabolic risk. In order to develop appropriate recommendations, there is a need for better understanding of how sleep affects metabolic risk amongst different age groups, which may call for heterogeneous interventions.

In this study, we focus on sleep duration, which is both modifiable and relatively easy to record [29]. Sleep duration can be self-reported or measured using medical devices including polysomnograms [30]. Although device-measured sleep duration is more precise, self-reported sleep duration is more feasible in large-scale epidemiological research, due to lower cost and expertise requirements [31]. For the purposes of this study, we use self-reported sleep duration as our exposure variable. There are a number of risk factors recognised to increase the risk of developing heart disease, type 2 diabetes and strokes. Central obesity, insulin resistance, hypertension and dyslipidaemia are particularly important and together they are defined as metabolic syndrome [32]. In our analysis, we use a standardised metabolic risk score as an outcome variable which incorporates these risk factors in order to ensure that our outcome variable truly corresponds to risk of future health outcomes, including cardiovascular disease [33].

In this study, we use data from years 1–8 of the UK-representative ‘National Diet and Nutrition Study Rolling Programme (NDNS)’ (2008–2016) to compare associations between sleep duration and metabolic risk score in different age groups. Specifically, we aim to address the research question: ‘How do associations between sleep duration and metabolic health differ with age in the UK general population?’

Methods

Survey design and participants

The NDNS is an annual cross-sectional survey which assesses the diet of the general population of the UK and includes measures on health. The NDNS aims to recruit 1000 participants each year, comprising an equal ratio of adults (aged 19 years and older) and children (aged 1.5 to 18 years). Households were sampled from the U.K. Postcode Address File, a list of all addresses in the U.K., with up to one adult and one child from each household eligible for inclusion in the survey [34]. The survey consisted of two relevant stages, the first of which was a computer-assisted interview and the second was a nurse visit in which blood samples and waist circumference measurements were taken. Only a smaller proportion consent to the nurse visit stage of the study, during which participants provide metabolic data [35]. Written informed consent was obtained from all participants, or from their parents/guardians if aged under 16 years. Ethical approval for the NDNS was obtained from the Oxfordshire A Research Ethics Committee and the Cambridge South NRES Committee (Ref. No. 13/EE/0016). In this analysis, we use data on participants aged from 11 to 70 years, combined from the first eight years (2008–2016) of the NDNS to provide a sufficiently large sample size for age-based subpopulation analysis (Table 1) [36]. Metabolic data was unavailable to be analysed for participants aged below 11 years. An upper age limit of 70 years was set due to the likelihood of unrecorded comorbidities affecting sleep in the elderly population. We excluded participants from the analyses if they reported taking anti-hypertensive or lipid-lowering medication.

Table 1. Descriptive statistics of the weighted sample, grouped by age category, from the NDNS Years 1–8.

Age category
All ages Adolescents Young adults Mid-aged adults Older adults
(n = 2008) 11–18 19–35 36–50 51–70
(n = 431) (n = 423) (n = 632) (n = 522)
Sex (%) Male 47.3 50.7 48.5 47.4 44.6
Female 52.7 49.3 51.5 52.6 55.4
National Statistics socio-economic classification (NS-SEC3) of household reference person (%) Managerial and professional occupations 50.0 41.9 42.2 56.2 54.5
Intermediate occupations 19.8 19.9 21.3 17.5 20.9
Routine and manual occupations 28.1 35.1 34.5 23.8 23.4
Never worked and other 2.1 3.1 2.0 2.5 1.2
Ethnic group (%) White 88.5 88.0 83.6 87.2 96.4
Mixed ethnic group 2.4 3.1 4.1 1.3 1.3
Black or Black British 2.5 2.1 3.9 2.8 0.7
Asian or Asian British 5.2 5.0 7.2 6.4 1.3
Any other group 1.4 1.8 1.2 2.2 0.3
Smoking status (%) Never a regular smoker 65.0 90.0 62.7 61.0 61.1
Ex-smoker 20.4 0.7 16.5 23.6 30.2
Current smoker 14.6 9.4 20.8 15.3 8.7
Screen time (hours/day) (s.d.) Mean screen time at home per day (computer use, TV, DVD or video viewing) 5.0 (2.2) 5.9 (2.2) 5.3 (2.5) 4.7 (2.1) 4.8 (1.9)
Sleep duration (hours/day) (s.d.) Average sleep duration in seven days prior to questioning 7.4 (1.2) 8.6 (1.3) 7.6 (1.1) 7.1 (1.2) 7.0 (1.1)
Average weeknight sleep (Sunday to Thursday) in seven days prior to questioning 7.3 (1.3) 8.4 (1.5) 7.5 (1.1) 7.0 (1.2) 6.9 (1.2)
Average weekday sleep (Friday to Saturday) in seven days prior to questioning 7.7 (1.5) 9.1 (1.8) 7.9 (1.5) 7.5 (1.4) 7.2 (1.2)
Difference between weekend and weekday night sleep (hours) (s.d.) Mean weekend sleep minus mean weekday night sleep (Note: positive value indicates greater sleep at the weekend) 0.4 (1.2) 0.7 (2.0) 0.4 (1.3) 0.5 (1.0) 0.3 (0.8)
Metabolic outcome Mean standardised metabolic risk score (MetZscore) (s.d.) -0.027 (1.02) 0.021 (0.75) 0.035 (0.88) -0.003 (1.15) -0.153 (1.06)
Waist circumference (cm) 88.7 (13.9) 76.0 (10.7) 85.9 (13.6) 91.9 (13.1) 92.8 (12.8)
Triglycerides (mmol/L) 1.2 (0.8) 0.9 (0.5) 1.1 (0.7) 1.4 (1.1) 1.3 (0.7)
HDL (mmol/L) 1.5 (0.4) 1.4 (0.3) 1.4 (0.4) 1.4 (0.4) 1.6 (0.5)
Glucose (mmol/L) 5.1 (1.0) 4.8 (0.5) 4.9 (0.9) 5.2 (1.2) 5.3 (1.0)
Average blood pressure (mm/Hg) 97.2 (12.2) 88.3 (8.2) 93.8 (10.1) 99.0 (12.3) 102.4 (12.9)
Open in a new tab

Sleep duration

In the survey, sleep data were reported in stage 1 (the interviewer stage) of the data collection, in response to two questions which captured both weekend and weekday night sleep duration: ‘Over the last seven days, that is since (date), how long did you usually sleep for on week nights. That is Sunday to Thursday nights?’ and ‘And over the last seven days, how long did you usually sleep for on weekend nights. That is Friday and Saturday nights?’. We calculated sleep duration as the average self-reported quantity of sleep obtained by respondents of the NDNS over the seven days prior to questioning, over both weekday and weekend nights.

To generate an age-adjusted measure of sleep duration that could be compared between participants of different ages, we regressed sleep duration against age to produce a residual sleep duration value for each participant. Since change in sleep with age is non-linear, a restricted cubic spline regression was used, with 5 knots at ages 14, 26, 39, 50 and 65 years, to ensure appropriate adjustment for age-related variation. The resultant ‘age-adjusted sleep’ measure indicates sleep duration in hours relative to the average value for that age. We also subtracted the mean weekday night sleep duration from mean weekend sleep duration for each participant to calculate a measure of difference in sleep duration (in hours) between weekday and weekend nights.

Metabolic risk factors

To provide an overall measure of metabolic risk, we used metabolic risk z-scores (MetZscore) [37–39]. Use of a continuous risk score, standardised for age, facilitates comparison across age groups [40]. The MetZscore was based on the metabolic risk factors included in the definition of metabolic syndrome from the National Cholesterol Education Program—Adult Treatment Program III (NCEP-ATPIII): waist circumference, blood pressure (BP), serum triglycerides, serum high-density lipoprotein (HDL) cholesterol, and fasting plasma glucose [41]. Metabolic risk factors were measured during stage 2 (the nurse visit) of the data collection. Waist circumference was measured at the midpoint between the iliac crest and lower rib, using a tape measure precise to the nearest 0.1cm. Systolic and diastolic BP values were measured in a seated position ensuring that participants had not eaten, consumed alcohol, exercised, or smoked in the preceding 30 minutes. Three BP readings were taken at one-minute intervals. Serum triglycerides, serum HDL cholesterol and plasma glucose were measured from fasted blood samples, Details of each assay and their quality control data are provided in NDNS Appendix Q [35].

The MetZscore was calculated, as has been described previously, by first standardising the individual metabolic risk factors by regressing them onto selected demographic variables (age, sex, ethnicity) [20, 37]. To account for the non-linearity of associations between age and metabolic outcomes, a restricted cubic spline regression was applied to the weighted sample for each outcome, with knots at ages 15, 20, 30, 40 and 50 years, to ensure appropriate adjustment for age-related variation. The resulting residuals were standardised to give a mean of 0 and a standard deviation (SD) of 1. Standardised HDL is inversely related to metabolic risk so it was multiplied by -1. The residuals of systolic and diastolic BP were averaged to give a value of mid-blood pressure. Z-scores for the individual risk factors were summed to create the MetZscore. This score was standardised to have a mean of 0, with a standard deviation of 1, within the analysis population. A higher MetZscore is indicative of a less favourable metabolic syndrome profile.

Covariates

Age, sex, ethnicity, smoking and screen time were self-reported. Ethnicity was classified according to five groups (white, mixed, black, Asian and other). Smoking was categorised into three categories: ‘current smoker’, ‘ex-regular smoker’ and ‘never regular smoker’. Television viewing time has been used as a covariate in comparable studies; the NDNS provided data on television viewing together with other related activities to produce a measure of screen time [23]. Screen time was reported in response to a series of questions on how much time participants spent watching DVDs, TV, or videos and computer use, before 6pm and after 6pm on both weekends and weekday nights. These data were summed to create a single variable of total weekly screen time. National statistics socio-economic classification (NSSEC) of the household reference person was reported by the household reference person (defined as the household member with the highest income), and classified into four groups: Managerial and professional occupations; Intermediate occupations; Routine and manual occupations; and Never worked or other [42]. We did not adjust for diet quality or physical activity as the literature suggests that these are likely to be on the causal pathway between sleep and metabolic outcomes; this was not the case for screen time [9]. We considered inclusion of marital status, alcohol intake and employment status as covariates, but these were not associated with the exposure and outcome variables in our dataset.

Statistical analysis

All the analyses were performed using R (version 3.6.0). The weights (blood sample weights) provided with the NDNS dataset were applied to account for clustering and response biases. In our descriptive analysis, sociodemographic factors, sleep and MetZscore were summarized overall and by age category. 562 participants had missing screen time data, 40 had missing socioeconomic classification data and 13 had missing data regarding smoking habits. Multiple imputations by chained equations were used to impute missing data for covariates, under the missing at random assumption, following recommendations from White et al. [43]. Twenty imputed datasets were generated using the R command ‘mice()’. Each dataset was analysed separately and results combined using Rubin’s rule [43].

We first tested associations between age-adjusted sleep and MetZscore across all age groups combined. We compared linear and quadratic models, by calculating residual sum of squares and Akaike information criterion (AIC), both of which indicated that the quadratic model provided a better fit to the data. Four age groups were constructed: 11-18y, 19-35y, 36-50y and 51-70y and ‘age group’ was included as an interaction term in the regression models to test differences in association by age. Groups were designed such that they included a comparable number of participants whilst still separating adolescents from adults, due to their different sleep duration requirements. Covariates were added in three stages: the first included no covariates to produce an unadjusted model (model 1). Model 2 included sex, ethnicity and socioeconomic classification and was performed with and without age group as a moderating factor. Finally, model 3 included the previous covariates with the addition of screen time and smoking status.

Results

1. Characteristics of study population

Of 12097 participants who took part in years 1–8 of the NDNS, 7860 participants were aged between 11 and 70 years. Of these, 2627 had complete metabolic data and 2365 had complete sleep and metabolic risk data. A further 357 were excluded due to taking anti-lipid or anti-hypertension medication. This left 2008 individuals in the analysis, with 21.5% aged 11 to 18y, 21.1% aged 19 to 35y, 31.5% aged 36 to 50y and 26.0% aged 51 to 70y. Descriptive data on the study population are shown in Table 1.

The data show an average sleep duration of 7.4 hours per night overall, with more sleep recorded in adolescents. Fig 1 illustrates the relationship between age and sleep duration. Sleep duration decreases steeply with age through adolescence then slows but continues to decrease as age increases.

Fig 1. The unadjusted relationship between sleep and age, across all age groups (95% CI shown in grey).

Fig 1

Open in a new tab

2. Relationship between sleep duration and metabolic outcomes

Overall, the results in Table 2 indicate that there is an association between increased sleep (in hours) and reduced metabolic risk (in standard deviations) across all ages (quadratic term: 0.028 [95%CI: 0.007, 0.050, p = 0.01], linear term: -0.061 [95%CI: -0.111, -0.011, p = 0.02]). When age categories are introduced, the association appears only in the 36–50 age group, which suggests that the overall effect may be driven by the effect seen in this group. The coefficients presented in Table 2 refer to regression equation: y = ax2 + bx + cz + ϵ, where y represents change in metabolic risk score, x represents age-adjusted sleep duration, z represents one of the covariate terms and ϵ represents the error term.

Table 2. Cross-sectional associations between sleep duration and metabolic risk score, presenting coefficients for both the quadratic sleep duration term and the linear sleep duration term from the regression model.

Change in metabolic risk score (standard deviations) for sleep duration increase of one hour [95% CI]
Model 1, unadjusted Model 2, adjusted using basic covariates (sex, ethnicity, socioeconomic classification) Model 3, covariates: sex, ethnicity, socioeconomic classification, smoking, screen time
Coefficient of age-adjusted quadratic sleep duration term Coefficient of age-adjusted sleep duration term Coefficient of age-adjusted quadratic sleep duration term Coefficient of age-adjusted sleep duration term Coefficient of age-adjusted quadratic sleep duration term Coefficient of age-adjusted sleep duration term
All ages (n = 2008) 0.031 [0.010, 0.053]* -0.067[-0.118, -0.016]* 0.031 [0.009, 0.053]* -0.069 [-0.120, -0.019]* 0.028[0.007, 0.050]* -0.061 [-0.111, -0.011]*
Adolescents 11–18 years (n = 431) -0.010 [-0.039, 0.020] 0.052 [-0.020, 0.125] -0.006 [-0.039, 0.026] 0.047 [-0.031,0.125] -0.009 [-0.042, 0.025] 0.041 [-0.039, 0.121]
Young adults 19–35 years (n = 423) 0.034[-0.010, 0.078] -0.086 [-0.197, 0.024] 0.032 [-0.013, 0.077] -0.078 [-0.187,0.030] 0.029 [-0.013, 0.072] -0.058 [-0.161, 0.044]
Mid-aged adults 36–50 years (n = 632) 0.046 [0.009, 0.083]* -0.091 [-0.191, 0.009] 0.043[0.006, 0.079]* -0.096 [-0.193, 0.001] 0.038[0.002, 0.074]* -0.092 [-0.189, 0.005]
Older adults 51–70 years (n = 522) 0.019 [-0.017, 0.056] -0.066 [-0.151, 0.020] 0.022 [-0.015, 0.060] -0.070 [-0.156, 0.016] 0.021 [-0.017, 0.060] -0.064 [-0.150, 0.021] 
Open in a new tab

Note

* indicates p<0.05

The positive quadratic term (0.028 [95%CI: 0.007, 0.050, p = 0.01]) in the results for all ages combined indicate that as sleep duration increases away from the mean, its marginal effect on metabolic risk increases. Therefore, the greater the reduction in sleep time, the larger the corresponding increase in metabolic risk, as illustrated in Fig 2. Fig 2 also suggests that sleep durations above the mean may be associated with increasing metabolic risk as sleep duration increases; however, this effect is not significant.

Fig 2. Model predictions of metabolic risk score.

Fig 2

Open in a new tab

Model predictions of metabolic risk score (with 95% confidence intervals) as sleep durations vary above and below the mean (model 3, all age groups).

When we compare the associations between different age groups, we see that for mid-aged adults the coefficient of the quadratic sleep duration term (0.038 [95%CI: 0.002, 0.074, p = 0.04]) is much larger than that seen in adolescents (-0.009 [95%CI: -0.042, 0.025, p = 0.59]), suggesting that associations of metabolic risk with low sleep duration are less marked in a younger population (Fig 3).

Fig 3. Model predictions of metabolic risk score by age group.

Fig 3

Open in a new tab

Model predictions of metabolic risk score (with 95% confidence intervals) as sleep durations vary above and below the mean, separated by age group (model 3, A: age 11 to 18 years, B: 19 to 35 years, C: 36 to 50 years, D: 51 to 70 years).

3. Relationship between intra-week sleep fluctuations and metabolic outcomes

As shown in Table 3, the difference between weekday and weekend nightly sleep only appears to be statistically significantly associated with metabolic outcomes in older adults, where it is associated with increased metabolic risk. For adults aged between 51 and 70y, we see that where the difference between weekday (Sunday to Thursday) and weekend nights (Friday to Saturday) increases by one hour, the MetZscore increases by 0.18 SD [95%CI: 0.005, 0.348, p = 0.04]. The standardised score allows for estimation of centiles within a population and a score of 0.18 corresponds to an increase of 7 percentiles.

Table 3. Cross-sectional associations between the difference between weekend and weekday night sleep and metabolic risk score.

Change in metabolic risk score (s.d.) for sleep duration or sleep duration increase of one hour [95%CI]
Unadjusted models Adjusted models using basic covariates (sex, ethnicity, socioeconomic classification) Adjusted models (covariates: sex, ethnicity, socioeconomic classification, smoking, screen time)
Coefficient of ‘difference between weekday and weekend sleep’ term Coefficient of ‘difference between weekday and weekend sleep’ term Coefficient of ‘difference between weekday and weekend sleep’ term
All ages (n = 2008) 0.052 [-0.003,0.107] 0.048 [-0.008, 0.104] 0.051 [-0.007, 0.108]
Adolescents 11–18 years (n = 431) 0.026 [-0.020, 0.073] 0.021 [-0.030, 0.071] 0.015 [-0.037, 0.068]
Young adults 19–35 years (n = 423) 0.001 [-0.097, 0.098] -0.004 [-0.105, 0.096] 0.003 [-0.101, 0.107]
Mid-aged adults 36–50 years (n = 632) 0.096 [-0.043, 0.236] 0.105 [-0.034, 0.243] 0.104 [-0.033, 0.241]
Older adults 51–70 years (n = 522) 0.176 [0.004, 0.349]* 0.160 [-0.010, 0.330] 0.176 [0.005, 0.348]*
Open in a new tab

Note

* indicates p<0.05

Discussion

1. Summary of key findings

Our results illustrate that when all age groups are combined, longer sleep duration is associated with more favourable metabolic health. When participants were stratified by age group, we found that this effect was only observed in those aged between 36-50y. In the adolescent age group in particular, the association between sleep and metabolic outcomes appears to be much smaller than in mid-aged adults. It is likely that the mid-aged age group is driving the overall effect. Whilst there may be adverse metabolic outcomes associated with prolonged sleep well above average sleep duration, our findings do not indicate this. The difference between weekend and weekday night sleep duration showed associations with metabolic outcomes only in older age groups, with statistically significantly associations seen only in participants aged 51-70y.

2. Strengths and limitations

A major strength of this study was the use of data from the NDNS, a national survey designed to be representative of the UK general population. We used survey weights which are specifically designed to account for non-response to the blood sample collection so that, despite some further participant exclusions specific to this study, our findings can be said to be broadly reflective of the UK population across our included age range. However, the fact that the study population, due to the ethnicity of the UK population, was mostly of white ethnicity is a limitation and complicates the generalisability of the findings. The wide range of data included in the NDNS survey, allow us to test, and where appropriate adjust for, all putative covariates. We were able to exclude participants who were taking blood-pressure or lipid-lowering medications from the analysis, as this would have affected their metabolic risk profile, but unfortunately data was unavailable to allow us to exclude participants taking medication for sleep-related reasons. The NDNS has a well-defined methodology which is consistent between years of the survey, allowing us to combine data across recent years of survey data. Survey data is accompanied with much detail about the interviewer questions and measurement procedures, to ensure that the results may be appropriately compared with those using other data.

Limitations include the self-reported assessment of sleep duration and use of cross-sectional data. Personal experience is highly subjective and previous research has noted that people tend to overestimate sleep duration when self-reporting [44–46]. It has also been noted that psychosocial factors influence sleep reporting and whilst some groups might overestimate their sleep, others may underestimate it [47]. The implication of this is that if the accuracy self-reported sleep duration was associated with certain characteristics this would require correction to mitigate regression dilution bias, though no relevant associations were known at the time of this study. It is possible to measure sleep clinically e.g. by polysomnogram or accelerometer study; however, this was not used in the NDNS. Whilst more precise, the cost and participant burden of polysomnograms reduces their applicability in clinical practice whereas self-reported sleep duration can be routinely assessed in practice. Another limitation to our study is the use of cross-sectional data, which is only capable of elucidating correlation between variables. Whilst the literature generally favours that sleep health is a potential driver of metabolic health, the possibility of reverse causality cannot be ignored.

3. Comparison with previous evidence

Our results indicated a decrease in sleep duration with age, consistent with existing literature. It has been noted that people typically suffer from well-characterised changes to sleep architecture as a function of ageing, with reduced sleep duration requirements [14, 48]. Indeed, sleep duration recommendations tend to reflect this with adults recommended less sleep than adolescents and children; some guidelines have now begun to refine sleep duration recommendations for the elderly [15]. In most cases, guidelines recommend around 8 to 10 hours sleep for adolescents and 7 to 9 hours sleep for adults, with some recommending 7 to 8 hours sleep for adults aged over 65 [15–19]. It is worth noting that the recommendations are designed with a high granularity during adolescence and early childhood, such that they change at frequent intervals with age in these age groups [18]. We were unable to draw conclusions on participants aged over 70 years old, due to the small number of eligible participants in the age group as well as the high prevalence of comorbidities in this age group confounding metabolic health measurements. However, it has been noted that, especially in this age group, additional characteristics such as sleep quality have increased importance [14].

Others have reported that sleep durations both above and below recommendations are associated with unfavourable health outcomes [14]. However, the metabolic health effects of excessively prolonged sleep appear inconclusive in the literature. Our results are, overall, consistent with a recent meta-analysis which found that short sleep duration is associated with increased risk of metabolic syndrome and that prolonged sleep duration was not associated with an increased risk [49]. However, at least one other study has suggested a U-shaped relationship between sleep duration and metabolic risk [50]. This is recognised as a conflicting area of research, with a systematic review noting that in another meta-regression analysis authors found a relationship between long sleep and adverse metabolic outcomes; however, they noted the need for future research exploring potential mechanisms and examining the precise nature of any potential relationship [51]. Future research may also examine any reverse causality and whether metabolic factors exert a causal effect in reducing sleep duration. There is a notable shortage of longitudinal studies exploring the long term effects of sleep deprivation on metabolic health. One prospective cohort study, conducted using New Zealand data, found that short sleep as a child was linked with increased long-term metabolic risk at age 32 [23]. The authors concluded that addressing sleep duration may be a useful tool in stemming the ongoing obesity epidemic. Our results suggest that strengths of associations between sleep duration and metabolic health might be heterogeneous between age groups. The results of the New Zealand study demonstrate long term effects of sleep deprivation in childhood; this is not necessarily inconsistent with our findings, which explored cross-sectional associations. We found that poor sleep was associated with adverse metabolic health effects in mid-aged adults, but it is possible that the individuals driving the effect in this group also slept poorly in childhood, for these effects to only manifest later in life.

The difference between weekend and weekday nightly sleep has been identified as an important avenue for future studies, with a limited body of literature at present [52]. Our descriptive results agree with research noting that there is reduced intra-week sleep variability in older age groups. One study noted that the difference between weekend and weekday sleep may be a healthy phenomenon to compensate for reduced sleep on work/school nights [52]. It may be that younger age groups are more resilient to short-term fluctuations in sleep duration. Indeed, we see that the youngest age group (11–18 years) exhibits the largest mean difference between weekend and weekday sleep, but showed no association of this difference with metabolic risk score, whereas the oldest age group (51–70 years), where associations were seen, exhibits the lowest. Older participants may be more likely to have flexible work times or be retired and so may not have to wake up earlier on weekdays, compared to school or working age participants. In older age groups, whose nightly sleep is typically less variable, any fluctuations may be more likely to be pathological or deleterious to metabolic health. Notably, some research has found differences between weekend and weekday sleep to be associated with increased hunger, but not with increased fat mass [27]. However, conflicts in the literature are apparent with another study finding that such difference was associated with an array of metabolic risk factors, including adiposity, in adults aged 30–54 years [53].

It has been suggested that changes in circadian rhythm may impact dietary behaviours, with a growing body of evidence, this includes a study noting that adults working night shifts are more likely to choose unhealthy breakfast options than those working day shifts [27, 54, 55]. There is also a well-described relationship between shift work and metabolic risk factors, though there is a need for further longitudinal studies in this area with some studies producing discordant results [56, 57]. Overall, the literature suggests the existence of an understudied, but complex, interplay between circadian rhythm shifts and health outcomes. The difference between weekend and weekday sleep represents just one of several potential variables to denote circadian rhythm variability.

4. Implications of findings

Sleep is currently an understated marker of human health at a population level; it is also seldom considered in routine clinical practice as a health marker [58]. In order to warrant widespread assessment of sleeping habits as a diagnostic marker or potential intervention we must consider whether such application would be clinically useful. For this we need to consider feasible interventions, which will depend upon the underlying cause. Screening for poor sleep duration may help in the diagnosis and prevention of mental and physical conditions, which could be either a cause or consequence of poor sleep [59]. There are a range of causes of poor sleep quality, some of which include: head trauma, depression, anxiety, arthritis, chronic pain or hyperthyroidism [60, 61]. Identification that a patient is experiencing poor quality of sleep is inadequate to formulate a diagnosis but it may aid in screening for a number of conditions, including, but by no means limited to, insomnia. Indeed, the growing body of evidence describing the effects of poor sleep has prompted calls for sleep to be routinely assessed in clinical practice [62]. There may be scope to ask questions about sleep in the same way that questions about smoking, alcohol intake and lifestyle are asked. This would take seconds during regular health check-ups, yet would enable access to an otherwise under-utilised health marker. However, it should be noted that despite a growing body of evidence indicating that poor sleep is associated with health outcomes, there is a shortage of studies assessing whether improving sleep may have therapeutic, or public health, benefits in patients with unfavourable metabolic risk factors [63].

Concluding remarks

The data presented in this paper indicate that the relationship between sleep duration and metabolic health exhibits heterogeneity between age groups. This reinforces the need for age-appropriate guidelines when expanding the inclusion of sleep health in health consultations. It also suggests that future research into the complications of poor sleep should not assume that effects would be homogeneous between age groups. There is a growing body of literature suggesting that irregular sleep patterns may act as an independent factor in influencing metabolic outcomes. This remains an active area of research, with several gaps in the literature. Whilst existing literature has separated children and adolescents from adults, our data suggest that further age separation in young adults (19 to 35 years) and older adults may too be important. This study is not able to elucidate a causal relationship between poor sleep and poor metabolic outcomes, which would require longitudinal studies. Further research may include examining other outcomes affected by sleep that may feed into metabolic pathways associated with less favourable metabolic profiles. Technological advancements, including fitness trackers and smart watches capable of monitoring sleep, may soon offer precise measurements of sleep duration and propel future research in the field.

Data Availability

The data underlying the results presented in the study are available from the UK Data Service under the 'National Diet and Nutrition Survey' (https://beta.ukdataservice.ac.uk/datacatalogue/series/series?id=2000033).

Funding Statement

The NDNS RP is funded by Public Health England (PHE), an executive agency of the Department of Health, and the UK Food Standards Agency (FSA). This study was supported by the Centre for Diet and Activity Research (CEDAR), a UKCRC Public Health Research Centre of Excellence (https://www.cedar.iph.cam.ac.uk/). Funding from the British Heart Foundation (https://www.bhf.org.uk/), Department of Health (https://www.gov.uk/government/organisations/department-of-health-and-social-care), Economic and Social Research Council (https://esrc.ukri.org/), Medical Research Council (https://mrc.ukri.org/), and the Wellcome Trust (https://wellcome.ac.uk/), under the auspices of the UK Clinical Research Collaboration, is gratefully acknowledged (087636/Z/08/Z; ES/G007462/1; MR/K023187/1; RES-590-28-0002). EvS and EW are supported by the Medical Research Council (MC_UU_12015/7). EW is funded by a fellowship from the Medical Research Council (MR/T010576/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.WEF_Harvard_HE_GlobalEconomicBurdenNonCommunicableDiseases_2011.pdf [Internet]. [cited 2020 Jul 26]. Available from: http://www3.weforum.org/docs/WEF_Harvard_HE_GlobalEconomicBurdenNonCommunicableDiseases_2011.pdf
  • 2.Bommer C, Sagalova V, Heesemann E, Manne-Goehler J, Atun R, Bärnighausen T, et al. Global Economic Burden of Diabetes in Adults: Projections From 2015 to 2030. Diabetes Care. 2018;41(5):963–70. 10.2337/dc17-1962 [DOI] [PubMed] [Google Scholar]
  • 3.WHO | Top 10 causes of death [Internet]. WHO. World Health Organization; [cited 2020 Jul 26]. Available from: http://www.who.int/gho/mortality_burden_disease/causes_death/top_10/en/
  • 4.Pitsavos C, Panagiotakos D, Weinem M, Stefanadis C. Diet, Exercise and the Metabolic Syndrome. Rev Diabet Stud. 2006;3(3):118–26. 10.1900/RDS.2006.3.118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Saklayen MG. The Global Epidemic of the Metabolic Syndrome. Curr Hypertens Rep. 2018;20(2):12 10.1007/s11906-018-0812-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kanagasabai T, Chaput J-P. Sleep duration and the associated cardiometabolic risk scores in adults. Sleep Health. 2017;3(3):195–203. 10.1016/j.sleh.2017.03.006 [DOI] [PubMed] [Google Scholar]
  • 7.Kim CE, Shin S, Lee H-W, Lim J, Lee J, Shin A, et al. Association between sleep duration and metabolic syndrome: a cross-sectional study. BMC Public Health. 2018. June 13;18(720). 10.1186/s12889-018-5557-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Potter GDM, Cade JE, Hardie LJ. Longer sleep is associated with lower BMI and favorable metabolic profiles in UK adults: Findings from the National Diet and Nutrition Survey. PLoS One. 2017;12(7):e0182195 10.1371/journal.pone.0182195 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Reutrakul S, Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metab Clin Exp. 2018;84:56–66. 10.1016/j.metabol.2018.02.010 [DOI] [PubMed] [Google Scholar]
  • 10.Xi B, He D, Zhang M, Xue J, Zhou D. Short sleep duration predicts risk of metabolic syndrome: a systematic review and meta-analysis. Sleep Med Rev. 2014. August;18(4):293–7. 10.1016/j.smrv.2013.06.001 [DOI] [PubMed] [Google Scholar]
  • 11.Matthews KA, Dahl RE, Owens JF, Lee L, Hall M. Sleep duration and insulin resistance in healthy black and white adolescents. Sleep. 2012. October 1;35(10):1353–8. 10.5665/sleep.2112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Javaheri S, Storfer-Isser A, Rosen CL, Redline S. The Association of Short and Long Sleep Durations with Insulin Sensitivity In Adolescents. J Pediatr. 2011. April;158(4):617–23. 10.1016/j.jpeds.2010.09.080 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Kuk JL, Ardern CI. Age and Sex Differences in the Clustering of Metabolic Syndrome Factors. Diabetes Care. 2010. November;33(11):2457–61. 10.2337/dc10-0942 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chaput J-P, Dutil C, Sampasa-Kanyinga H. Sleeping hours: what is the ideal number and how does age impact this? Nat Sci Sleep. 2018. November 27;10:421–30. 10.2147/NSS.S163071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.How Much Sleep Do We Really Need?—National Sleep Foundation [Internet]. [cited 2020 Apr 24]. Available from: https://www.sleepfoundation.org/articles/how-much-sleep-do-we-really-need
  • 16.How much sleep do we need?—The Sleep Council [Internet]. [cited 2020 Apr 30]. Available from: https://sleepcouncil.org.uk/advice-support/sleep-hub/sleep-matters/how-much-sleep-do-we-need/
  • 17.How much sleep do children need? [Internet]. nhs.uk. 2018 [cited 2020 Apr 30]. Available from: https://www.nhs.uk/live-well/sleep-and-tiredness/how-much-sleep-do-kids-need/
  • 18.Paruthi S, Brooks LJ, D’Ambrosio C, Hall WA, Kotagal S, Lloyd RM, et al. Recommended Amount of Sleep for Pediatric Populations: A Consensus Statement of the American Academy of Sleep Medicine. J Clin Sleep Med. 2016. June 15;12(6):785–6. 10.5664/jcsm.5866 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Watson NF, Badr MS, Belenky G, Bliwise DL, Buxton OM, Buysse D, et al. Recommended Amount of Sleep for a Healthy Adult: A Joint Consensus Statement of the American Academy of Sleep Medicine and Sleep Research Society. Sleep. 2015. June 1;38(6):843–4. 10.5665/sleep.4716 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Winpenny EM, van Sluijs EMF, Forouhi NG. How do short-term associations between diet quality and metabolic risk vary with age? European Journal of Nutrition [Internet]. 2020. 10.1007/s00394-020-02266-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Duraccio KM, Krietsch KN, Chardon ML, Van Dyk TR, Beebe DW. Poor sleep and adolescent obesity risk: a narrative review of potential mechanisms. Adolesc Health Med Ther. 2019;10:117–30. 10.2147/AHMT.S219594 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Chaput J-P, Saunders TJ, Carson V. Interactions between sleep, movement and other non-movement behaviours in the pathogenesis of childhood obesity. Obes Rev. 2017;18 Suppl 1:7–14. 10.1111/obr.12508 [DOI] [PubMed] [Google Scholar]
  • 23.Landhuis CE, Poulton R, Welch D, Hancox RJ. Childhood Sleep Time and Long-Term Risk for Obesity: A 32-Year Prospective Birth Cohort Study. Pediatrics. 2008. November 1;122(5):955–60. 10.1542/peds.2007-3521 [DOI] [PubMed] [Google Scholar]
  • 24.Calamaro CJ, Park S, Mason TBA, Marcus CL, Weaver TE, Pack A, et al. Shortened sleep duration does not predict obesity in adolescents. J Sleep Res. 2010. December;19(4):559–66. 10.1111/j.1365-2869.2010.00840.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Klingenberg L, Chaput J-P, Holmbäck U, Jennum P, Astrup A, Sjödin A. Sleep restriction is not associated with a positive energy balance in adolescent boys. Am J Clin Nutr. 2012. August;96(2):240–8. 10.3945/ajcn.112.038638 [DOI] [PubMed] [Google Scholar]
  • 26.Sun W, Ling J, Zhu X, Lee TM-C, Li SX. Associations of weekday-to-weekend sleep differences with academic performance and health-related outcomes in school-age children and youths. Sleep Med Rev. 2019;46:27–53. 10.1016/j.smrv.2019.04.003 [DOI] [PubMed] [Google Scholar]
  • 27.LeMay-Russell S, Tanofsky-Kraff M, Schvey NA, Kelly NR, Shank LM, Mi SJ, et al. Associations of Weekday and Weekend Sleep with Children’s Reported Eating in the Absence of Hunger. Nutrients. 2019;11(7):1658 10.3390/nu11071658 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Kjeldsen JS, Hjorth MF, Andersen R, Michaelsen KF, Tetens I, Astrup A, et al. Short sleep duration and large variability in sleep duration are independently associated with dietary risk factors for obesity in Danish school children. Int J Obes (Lond). 2014. January;38(1):32–9. 10.1038/ijo.2013.147 [DOI] [PubMed] [Google Scholar]
  • 29.Sharma M, Sawhney JPS, Panda S. Sleep quality and duration–Potentially modifiable risk factors for Coronary Artery Disease? Indian Heart J. 2014;66(6):565–8. 10.1016/j.ihj.2014.10.412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ibáñez V, Silva J, Cauli O. A survey on sleep assessment methods. PeerJ. 2018;6:e4849 10.7717/peerj.4849 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Sleep duration: how well do self-reports reflect objective measures? The CARDIA Sleep Study. Epidemiology. 2008. November;19(6):838–45. 10.1097/EDE.0b013e318187a7b0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.O’Neill S, O’Driscoll L. Metabolic syndrome: a closer look at the growing epidemic and its associated pathologies. Obes Rev. 2015. January;16(1):1–12. 10.1111/obr.12229 [DOI] [PubMed] [Google Scholar]
  • 33.Okosun IS, Lyn R, Davis-Smith M, Eriksen M, Seale P. Validity of a Continuous Metabolic Risk Score as an Index for Modeling Metabolic Syndrome in Adolescents. Annals of Epidemiology. 2010. November 1;20(11):843–51. 10.1016/j.annepidem.2010.08.001 [DOI] [PubMed] [Google Scholar]
  • 34.NDNS: results from years 7 and 8 (combined)—GOV.UK [Internet]. [cited 2019 Sep 15]. Available from: https://www.gov.uk/government/statistics/ndns-results-from-years-7-and-8-combined
  • 35.NDNS: results from years 7 and 8 (combined): appendices—GOV.UK [Internet]. [cited 2019 Sep 15]. Available from: https://www.gov.uk/government/statistics/ndns-results-from-years-7-and-8-combined
  • 36.NatCen Social Research, MRC Elsie Widdowson Laboratory. (2019). National Diet and Nutrition Survey Years 1–9, 2008/09-2016/17 [data collection]. 15th Edition UK Data Service; SN: 6533, 10.5255/UKDA-SN-6533-15 [DOI] [Google Scholar]
  • 37.Eisenmann JC. On the use of a continuous metabolic syndrome score in pediatric research. Cardiovasc Diabetol. 2008. June 5;7(17). 10.1186/1475-2840-7-17 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Brage S, Wedderkopp N, Ekelund U, Franks PW, Wareham NJ, Andersen LB, et al. Features of the Metabolic Syndrome Are Associated With Objectively Measured Physical Activity and Fitness in Danish Children: The European Youth Heart Study (EYHS). Diabetes Care. 2004. September 1;27(9):2141–8. 10.2337/diacare.27.9.2141 [DOI] [PubMed] [Google Scholar]
  • 39.O’Connor L, Imamura F, Brage S, Griffin SJ, Wareham NJ, Forouhi NG. Intakes and sources of dietary sugars and their association with metabolic and inflammatory markers. Clin Nutr. 2018. August;37(4):1313–22. 10.1016/j.clnu.2017.05.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Stavnsbo M, Resaland GK, Anderssen SA, Steene-Johannessen J, Domazet SL, Skrede T, et al. Reference values for cardiometabolic risk scores in children and adolescents: Suggesting a common standard. Atherosclerosis. 2018. November;278:299–306. 10.1016/j.atherosclerosis.2018.10.003 [DOI] [PubMed] [Google Scholar]
  • 41.Huang PL. A comprehensive definition for metabolic syndrome. Dis Model Mech. 2009;2(5–6):231–7. 10.1242/dmm.001180 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.The National Statistics Socio-economic classification (NS-SEC)—Office for National Statistics [Internet]. [cited 2020 Jul 27]. Available from: https://www.ons.gov.uk/methodology/classificationsandstandards/otherclassifications/thenationalstatisticssocioeconomicclassificationnssecrebasedonsoc2010#category-descriptions-and-operational-issues
  • 43.White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Statistics in Medicine. 2011;30(4):377–99. 10.1002/sim.4067 [DOI] [PubMed] [Google Scholar]
  • 44.Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-Reported and Measured Sleep Duration: How Similar Are They? Epidemiology. 2008;19(6):838–45. 10.1097/EDE.0b013e318187a7b0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Combs D, Goodwin JL, Quan SF, Morgan WJ, Hsu CH, Edgin JO, et al. Mother Knows Best? Comparing Child Report and Parent Report of Sleep Parameters With Polysomnography. Journal of Clinical Sleep Medicine. 2019;15(01):111–7. 10.5664/jcsm.7582 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Biddle DJ, Robillard R, Hermens DF, Hickie IB, Glozier N. Accuracy of self-reported sleep parameters compared with actigraphy in young people with mental ill-health. Sleep Health. 2015. September;1(3):214–20. 10.1016/j.sleh.2015.07.006 [DOI] [PubMed] [Google Scholar]
  • 47.Jackowska M, Dockray S, Hendrickx H, Steptoe A. Psychosocial factors and sleep efficiency: discrepancies between subjective and objective evaluations of sleep. Psychosom Med. 2011. December;73(9):810–6. 10.1097/PSY.0b013e3182359e77 [DOI] [PubMed] [Google Scholar]
  • 48.Mander BA, Winer JR, Walker MP. Sleep and Human Aging. Neuron. 2017. April 5;94(1):19–36. 10.1016/j.neuron.2017.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Iftikhar IH, Donley MA, Mindel J, Pleister A, Soriano S, Magalang UJ. Sleep Duration and Metabolic Syndrome. An Updated Dose–Risk Metaanalysis. Ann Am Thorac Soc. 2015. September;12(9):1364–72. 10.1513/AnnalsATS.201504-190OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Fan L, Hao Z, Gao L, Qi M, Feng S, Zhou G. Non-linear relationship between sleep duration and metabolic syndrome. Medicine (Baltimore). 2020. January; 99(2): e18753 10.1097/MD.0000000000018753 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Jike M, Itani O, Watanabe N, Buysse DJ, Kaneita Y. Long sleep duration and health outcomes: A systematic review, meta-analysis and meta-regression. Sleep Medicine Reviews. 2018. June 1;39:25–36. 10.1016/j.smrv.2017.06.011 [DOI] [PubMed] [Google Scholar]
  • 52.Åkerstedt T, Ghilotti F, Grotta A, Zhao H, Adami H, Trolle‐Lagerros Y, et al. Sleep duration and mortality–Does weekend sleep matter? J Sleep Res. 2019;28(1):e12712 10.1111/jsr.12712 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Wong PM, Hasler BP, Kamarck TW, Muldoon MF, Manuck SB. Social Jetlag, Chronotype, and Cardiometabolic Risk. J Clin Endocrinol Metab. 2015. December;100(12):4612–20. 10.1210/jc.2015-2923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Wing YK, Li SX, Li AM, Zhang J, Kong APS. The effect of weekend and holiday sleep compensation on childhood overweight and obesity. Pediatrics. 2009. November;124(5):e994–1000. 10.1542/peds.2008-3602 [DOI] [PubMed] [Google Scholar]
  • 55.Cain SW, Filtness AJ, Phillips CL, Anderson C. Enhanced preference for high-fat foods following a simulated night shift. Scand J Work Environ Health. 2015. May 1;41(3):288–93. 10.5271/sjweh.3486 [DOI] [PubMed] [Google Scholar]
  • 56.Canuto R, Garcez AS, Olinto MTA. Metabolic syndrome and shift work: A systematic review. Sleep Medicine Reviews. 2013. December 1;17(6):425–31. 10.1016/j.smrv.2012.10.004 [DOI] [PubMed] [Google Scholar]
  • 57.Proper KI, Langenberg D van de, Rodenburg W, Vermeulen RCH, Beek AJ van der, Steeg H van, et al. The Relationship Between Shift Work and Metabolic Risk Factors: A Systematic Review of Longitudinal Studies. American Journal of Preventive Medicine. 2016. May 1;50(5):e147–57. 10.1016/j.amepre.2015.11.013 [DOI] [PubMed] [Google Scholar]
  • 58.Reid KJ, Martinovich Z, Finkel S, Statsinger J, Golden R, Harter K, et al. Sleep: a marker of physical and mental health in the elderly. Am J Geriatr Psychiatry. 2006. October;14(10):860–6. 10.1097/01.JGP.0000206164.56404.ba [DOI] [PubMed] [Google Scholar]
  • 59.Mollayeva T, Thurairajah P, Burton K, Mollayeva S, Shapiro CM, Colantonio A. The Pittsburgh sleep quality index as a screening tool for sleep dysfunction in clinical and non-clinical samples: A systematic review and meta-analysis. Sleep Medicine Reviews. 2016. February 1;25:52–73. 10.1016/j.smrv.2015.01.009 [DOI] [PubMed] [Google Scholar]
  • 60.Hanson JA, Huecker MR. Sleep Deprivation In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2020. [cited 2020 Oct 24]. Available from: http://www.ncbi.nlm.nih.gov/books/NBK547676/ [Google Scholar]
  • 61.Insomnia [Internet]. nhs.uk. 2017 [cited 2020 Oct 24]. Available from: https://www.nhs.uk/conditions/insomnia/
  • 62.Perez-Pozuelo I, Zhai B, Palotti J, Mall R, Aupetit M, Garcia-Gomez JM, et al. The future of sleep health: a data-driven revolution in sleep science and medicine. npj Digital Medicine. 2020. March 23;3(1):1–15. 10.1038/s41746-020-0244-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Nedeltcheva AV, Scheer FAJL. Metabolic effects of sleep disruption, links to obesity and diabetes. Curr Opin Endocrinol Diabetes Obes. 2014. August;21(4):293–8. 10.1097/MED.0000000000000082 [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision Letter 0

Antonio Palazón-Bru

22 Oct 2020

PONE-D-20-29508

How do associations between sleep duration and metabolic health differ with age in the UK general population?

PLOS ONE

Dear Dr. Winpenny,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Dec 06 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information, or include a citation if it has been published previously.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: PONE-D-20-29508 “How do associations between sleep duration and metabolic health differ with age in the UK general population?”

In this manuscript, the authors report findings from a cross-sectional study to test whether short sleep duration is related to increased metabolic risk across four age groups from adolescents to older adults. This study has several strengths, including a large representative sample spanning adolescence to older adulthood, objectively-measured metabolic health, MICE to account for missing covariate data, and consideration of both linear and quadratic relationships between sleep and metabolic health. However, there are a number of areas which warrant further description and justification that reduced enthusiasm for the manuscript in its present form:

Introduction:

• Are there not more comprehensive citations for the relationship between sleep and metabolic risk factors from systematic reviews and meta-analyses?

• What is meant by “increased dietary intake” (p.4, lines 17-18)?

• Although it is true that there have been few studies examining relationships between sleep and metabolic risk factors in younger samples, the authors fail to cite any relevant literature (e.g., Matthews et al, 2012; Javaheri et al, 2011).

• Please include citations and current guidelines re: sleep duration requirements across age groups (p.4, line 20).

• p.5, line 26 - there are other dimensions of sleep beyond duration and quality, I would not make this statement so categorical.

• The statements in p.5, lines 32-36 appear somewhat contradictory.

• I am somewhat confused by the statement “In order to develop appropriate recommendations among children and young adults” (p.5, lines 37-39) – your study did not include children below the age of 11, and the authors ultimately did not provide recommendations based on their results.

• p.5, lines 43-45 – citations needed for the statement that “device-measured sleep is more precise”? Furthermore, are you not downplaying the importance of your own study?

• The authors note that the present study focuses on sleep duration because it is “modifiable and relatively easy to record.” That statement needs a citation.

• Citation needed on p.5, lines 50-51 - metabolic syndrome/metabolic risk score is related to risk of what future health outcomes?

Matthews, et al. Sleep duration and insulin resistance in healthy black and white adolescents. Sleep. 2012;35(10):1353-58.

Javaheri, et al. Association of short and long sleep durations with insulin sensitivity in adolescents. J Pediatr. 2011;158:617–23.

Methods:

• Participants taking blood-pressure or lipid-lowering medications were excluded for analysis, which is good. Was data available on the % of participants who were taking medications for sleep purposes?

• Why were participants less than age 11 excluded (p.7, line 75), when the authors state that there is a need to “develop appropriate recommendations among children and young adults” (p.5, lines 37-39)?

• Would be helpful to know the % participants who were excluded due to age (<11 or >70 years), blood pressure or lipid-lowering medication use.

• What % of the sample were currently employed as shift workers?

Results:

• Table 1 – please provide raw data for individual metabolic risk factors, weekday sleep duration, and weekend sleep duration.

• The authors note that “the association appears strongest in the 36-50 age group” (p.12, line 203) – I would think that is better stated as the association was present only in the 36-50 age group, as no association exists in the other groups.

• Tables 2 and 3 – consider using a note to indicate which covariates were included in Models 1-3.

• The Results text and Table 2 uses phrases including “quadratic term”, “squared term,” and coefficient of age-adjusted sleep duration term2” – be consistent.

• p.13, line 213 – consider using the phrase “all ages” vs. “aggregated results” to be consistent with Table 2.

• The authors provide the p value on p.15, line 241 – would be helpful if they provided p values for other significant findings noted in the Results text.

Discussion

• The fact that the sample was largely white should be added as a limitation.

• Not sure what the authors mean by the sentence beginning “the recommendations change frequently with age…” (p.17, lines 294-295). Also provide citations.

• The authors found that the mid-age group appeared to be driving results for the effect of short sleep on metabolic risk but did not find significant results for the difference between weekday and weekend sleep in this age group. This is in contrast to evidence from Wong et al., who found that midlife adults who have greater social jetlag (difference in actigraphy-measured sleep between work and non-work days) showed a worse cardiometabolic profile.

• Provide citations for the sentences beginning “Screening for poor sleep…” and “There are a range of causes…” (p.20, lines 354-358).

• Provide age range for “young adults” (p.21, line 376).

• Need more explanation for the first sentence in the Concluding Remarks section – what, specifically, might be the clinical implications of heterogeneity between age groups?

Wong, et al. Social jetlag, chronotype, and cardiometabolic risk. J Clin Endocrinol Metab 2015;100:4612-20.

Reviewer #2: I think this will be a good addition to existing literature. Generalizability of the data may be difficult in this case since it was predominantly a white population, none the less the study was done thoughtfully.

Authors have highlighted most of the potential limitations which one may think have. The tables are well done. The technique and the stats are well done with no major limitations. I like the fact that it draws more spotlight on weekend and weekday sleep variation which is an area of future research.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: PONE-D-20-29508_ms review _1st submission.docx

Click here for additional data file. (18.9KB, docx)

Decision Letter 1

Antonio Palazón-Bru

11 Nov 2020

How do associations between sleep duration and metabolic health differ with age in the UK general population?

PONE-D-20-29508R1

Dear Dr. Winpenny,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Acceptance letter

Antonio Palazón-Bru

13 Nov 2020

PONE-D-20-29508R1

How do associations between sleep duration and metabolic health differ with age in the UK general population?

Dear Dr. Winpenny:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Antonio Palazón-Bru

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Attachment

    Submitted filename: PONE-D-20-29508_ms review _1st submission.docx

    Click here for additional data file. (18.9KB, docx)
    Attachment

    Submitted filename: Response to Reviewers .docx

    Click here for additional data file. (20.8KB, docx)

    Data Availability Statement

    The data underlying the results presented in the study are available from the UK Data Service under the 'National Diet and Nutrition Survey' (https://beta.ukdataservice.ac.uk/datacatalogue/series/series?id=2000033).

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES