Association of accelerometer-measured sleep duration and different intensities of physical activity with incident type 2 diabetes in a population-based cohort study

Xinyi Jin; Yilin Chen; Hongliang Feng; Mingqing Zhou; Joey WY Chan; Yaping Liu; Alice Pik Shan Kong; Xiao Tan; Yun-Kwok Wing; Yannis Yan Liang; Jihui Zhang

doi:10.1016/j.jshs.2023.03.001

. 2023 Mar 3;13(2):222–232. doi: 10.1016/j.jshs.2023.03.001

Association of accelerometer-measured sleep duration and different intensities of physical activity with incident type 2 diabetes in a population-based cohort study

Xinyi Jin ^a,^†, Yilin Chen ^b,^c,^†, Hongliang Feng ^a,^d, Mingqing Zhou ^e,^f, Joey WY Chan ^g,^h, Yaping Liu ^g, Alice Pik Shan Kong ⁱ, Xiao Tan ^j,^k, Yun-Kwok Wing ^g,^h, Yannis Yan Liang ^a,^b,^⁎, Jihui Zhang ^d,^f,^h,^l,^⁎

PMCID: PMC10980868 PMID: 36871624

Highlights

•
Accelerometer-measured short but not long sleep is associated with an increased risk of incident type 2 diabetes among 88,000 participants from the UK Biobank.
•
Accelerometer-measured physical activity (PA) is inversely associated with the risk of incident type 2 diabetes.
•
Both moderate-to-vigorous PA and light-intensity PA potentially ameliorate the negative impact of short sleep duration on incident type 2 diabetes.

Keywords: Light physical activity, Sleep duration, Triaxial accelerometer, Type 2 diabetes, UK Biobank

Abstract

Purpose

The aim of the current study was to investigate the association of accelerometer-measured sleep duration and different intensities of physical activity (PA) with the risk of incident type 2 diabetes in a population-based prospective cohort study.

Methods

Altogether, 88,000 participants (mean age = 62.2 ± 7.9 years, mean ± SD) were included from the UK Biobank. Sleep duration (short: <6 h/day; normal: 6–8 h/day; long: >8 h/day) and PA of different intensities were measured using a wrist-worn accelerometer over a 7-day period between 2013 and 2015. PA was classified according to the median or World Health Organization-recommendation: total volume of PA (high, low), moderate-to-vigorous PA (MVPA) (recommended, not recommended), and light-intensity PA (high, low). Incidence of type 2 diabetes was ascertained using hospital records or death registries.

Results

During a median follow-up of 7.0 years, 1615 incident type 2 diabetes cases were documented. Compared with normal sleep duration, short (hazard ratio (HR) = 1.21, 95% confidence interval (95%CI): 1.03–1.41) but not long sleep duration (HR = 1.01, 95%CI: 0.89–1.15) was associated with excessive type 2 diabetes risk. This increased risk among short sleepers seems to be protected against by PA. Compared with normal sleepers with high or recommended PA, short sleepers with low volume of PA (HR = 1.81, 95%CI: 1.46–2.25), not recommended (below the World Health Organization-recommended level of) MVPA (HR = 1.92, 95%CI: 1.55–2.36), or low light-intensity PA (HR = 1.49, 95%CI: 1.13–1.90) had a higher risk of type 2 diabetes, while short sleepers with a high volume of PA (HR = 1.14, 95%CI: 0.88–1.49), recommended MVPA (HR = 1.02, 95%CI: 0.71–1.48), or high light-intensity PA (HR = 1.14, 95%CI: 0.92–1.41) did not.

Conclusion

Accelerometer-measured short but not long sleep duration was associated with a higher risk of incident type 2 diabetes. A higher level of PA, regardless of intensity, potentially ameliorates this excessive risk.

Graphical abstract

1. Introduction

Type 2 diabetes is one of the leading causes of disability and mortality, and it places a considerable socioeconomic burden on nearly half a billion individuals. The rapid growth of the disease worldwide warrants more effective and practical prevention.¹^,² Short sleep (usually under 6 h) has recently been identified as a potential risk factor for type 2 diabetes;³ likewise, not meeting recommended amounts of sleep has become increasingly prevalent among people with type 2 diabetes.⁴^,⁵ Self-reported short and long sleep durations (those outside the recommended 6–8 h of sleep per day) are repeatedly associated with increased type 2 diabetes risk.⁶ Short and long sleep has been found to adversely impact diabetes development via multiple pathways, including melatonin and cortisol disruption, neural hyperactivation, low-grade inflammation, and energy homeostasis defects.⁷^,⁸ Therefore, researchers have suggested screening for sleep disturbance, particularly for short and long sleep duration, as part a preventive measure for type 2 diabetes.⁹ However, association of device-measured short and long sleep duration with risk of type 2 diabetes remains poorly understood.

While maintaining healthy sleep duration could be a preventive strategy, unhealthy amounts of sleep are sometimes inevitable. In these cases, physical activity (PA) may be considered an alternative for ameliorating the risk of type 2 diabetes. Moderate-to-vigorous PA (MVPA) is recognized and recommended by guidelines as a robust preventive factor for type 2 diabetes.¹⁰ On the other hand, evidence of the protective role of light-intensity PA (LPA) remains limited. This is due to the inability for self-reported measurements, upon which most existing epidemiological studies have relied, to measure LPA accurately.¹¹ Device-based measurements are capable of capturing a wide range of PA intensities, including LPA.¹¹ There are device-measured evidences suggesting that LPA is beneficial to various diabetes-related health outcomes, including obesity and glucose metabolism.¹¹ Considering that for some special groups, such as people with morbid obesity or the elderly, MVPA may not be feasible, the potential of LPA for protecting people from excessive type 2 diabetes risk should be investigated and understood.

Sleep and PA have been widely reported to share a bidirectional relationship and to influence health through overlapping biological pathways.12, 13, 14 Examination of the joint association with type 2 diabetes risks between sleep and PA is therefore necessary. Evidence from cohort studies has suggested that MVPA mitigates the negative effect of poor sleep on important cardiometabolic health outcomes, such as hypertension and cardiovascular mortality.¹⁴^,¹⁵ Regarding diabetes, a cross-sectional study found that among short sleepers, those with a higher level of MVPA tended to have a lower prevalence of diabetes, which suggests a counteractive relationship between unhealthy sleep duration and MVPA.¹⁶ However, it remains unclear whether MVPA and LPA protect short or long sleepers from an elevated risk of type 2 diabetes. Answering this question will help us better understand how to construct the “diabetes-protective lifestyle” promoted by Kolb and Martin.¹³

In the current study, we hypothesized that accelerometer-measured short and long sleep durations are associated with excessive type 2 diabetes risk in the general adult population. We also hypothesized that the elevated risk of type 2 diabetes may be ameliorated by both MVPA and LPA. To verify these hypotheses, we analyzed the independent and joint associations of accelerometer-measured sleep duration and PA (total volume of PA, MVPA, and LPA) with the risk of incident type 2 diabetes using data from a large-scale population-based cohort, the UK Biobank.

2. Materials and methods

2.1. Study population and data source

In a population-based prospective cohort study, the UK Biobank recruited over 500,000 participants aged 37–73 years from 22 centers across the UK between 2006 and 2010. UK Biobank is an open access dataset established for investigations of determinants of a wide range of complex diseases of middle and old age. Data are available upon reasonable application to the UK Biobank (https://www.ukbiobank.ac.uk/) and should be used under the restrictions applied to the data availability. More detailed descriptions of the UK Biobank have been reported.¹⁷ From February 2013 to December 2015, 106,053 participants agreed to wear an AX3 accelerometer (Axivity, Newcastle upon Tyne, UK) on the wrist of the hand they usually write with for 1 week, resulting in 103,670 datasets with acceleration intensity time-series data. The UK Biobank Accelerometer Working Group provided quality metrics. After controlling for data quality, a total number of 88,000 patients were included in the current study. Details of the inclusion and exclusion criteria are depicted in Fig. 1. The UK Biobank study was approved by the North West Multi-centre Research Ethics Committee;¹⁸ all participants provided written informed consent. The present analyses were conducted under UK Biobank application Number 58082.

Fig 1 — Flowchart of participant enrollment. HbA1c = glycated hemoglobin; LPA = light-intensity physical activity; MVPA = moderate-to-vigorous physical activity; PA = physical activity.

2.2. Exposures

Exposures of the current study included accelerometer-measured sleep duration and PA. The raw accelerometer data were processed by the UK Biobank accelerometer working group.¹⁹ The acceleration signals were first calibrated to local gravity.²⁰ The valid data were resampled to 100 Hz. To separate out the activity-related components of these acceleration signals, 1 gravitational unit from the vector magnitude was removed, and the remaining negative values were truncated to 0. PA was calculated from the resampled data combined in 5-s epochs (Field ID 90004). Three categories of PA (total volume of PA, MVPA, and LPA) were included in the current study. The total volume of PA was measured as the weekly average vector magnitude in milli-gravity (mg) units (Field ID 90012).²¹ MVPA was defined as reaching the following standard: (a) within a 5-min period, over 80% of the 5-s epochs having a mean acceleration between 100mg and 400mg (moderate-intensity PA, MPA);²² or (b) any single 5-s epoch having a mean acceleration >400mg (vigorous-intensity PA (VPA)).²¹ LPA was defined as any single epoch with a mean acceleration of 30–100mg.²³ The amounts of MVPA and LPA (min/week) were calculated for each participant. Similarly, considering the expected nonlinear associations with type 2 diabetes risks,²⁴ the total volume of PA and LPA was then divided into high and low groups according to the median, and MVPA was divided into recommended or not recommended groups according to the World Health Organization (WHO) guidelines (≥150 min of MPA, ≥75 min of VPA, or equivalent combinations of MPA and VPA per week).²⁵

Sleep duration was obtained from the returned dataset of UK Biobank (Return ID 1862). The procedures of data processing and analyses were described previously.²⁶ Sleep periods were defined using an automated detection algorithm²⁶^,²⁷ implemented in the GGIR R package²⁸ and validated using polysomnography (PSG) in an external cohort. The algorithm facilitates detection of the sleep period time window (SPT-window) without the use of sleep diaries. The SPT-window refers to the time window starting at sleep onset and ending upon waking from the last sleep episode of the night. Sleep episodes within the SPT-window were defined as periods of ≥5 min with no z-axis changes greater than 5°. Sleep duration for a given SPT-window was calculated by summing the durations of all sleep episodes within the SPT-window. The mean sleep duration across all SPT-windows provided a measurement of average sleep quantity. Due to the expected nonlinear association between sleep duration and type 2 diabetes risks,⁶^,⁹ we further divided the participants into 3 groups with cutoffs based on previous studies²⁹ and the American Academy of Sleep Medicine consensus⁵ suggestion: short (<6 h/day), normal (6–8 h/day), and long sleep duration (>8 h/day).

Participants with at least 3 days of accelerometry data (Field ID 90015) were included in the current study. This inclusion criterion was defined by the UK Biobank accelerometer expert working group, who found that 3 days of wear were needed to be within 10% of a complete 7-day measure after performing missing data simulations on 29,765 participants who had perfect wear time compliance.¹⁹ A total number of 4470 participants were ruled out for not satisfying this inclusion criterion in the current study (Fig. 1). Participants were also excluded if their mean sleep duration was less than 3 or longer than 11 h (Fig. 1). Participants included in the final sample had 6.93 valid days of accelerometer wearing on average.

2.3. Outcomes

The outcome of this study was incident type 2 diabetes, which was defined based on hospital records or death registry (Supplementary Table 1). The data on dates and causes of hospital records for participants from Scotland were obtained from the Scottish Morbidity Records, and those for participants from England and Wales came from health episode statistics. The data on dates and causes of death for participants from England and Wales were obtained from the death registries of the National Health Service Information Center, and those for participants from Scotland came from the National Health Service Central Register Scotland. Further detailed information on linkage procedures is available at http://content.digital.nhs.uk/services. At the time of analysis, hospital record data were available for participants until September 30, 2021, and death registry data were available until November 12, 2021. Therefore, we used November 12, 2021 as the censor date unless hospital admission or death occurred first.

2.4. Covariates

The following variables were considered likely confounding factors: age at the time of accelerometer wearing (continuous, year), sex (male/female), ethnicity (white/others), season of accelerometer wearing (spring/summer/autumn/winter: spring for March to May, summer for June to August, autumn for September to November, and winter for December to February; UK Meteorological Office definitions), recruitment center (England/Wales/Scotland), Townsend Deprivation Index (continuous, a score representing the deprivation of the participant's neighborhood as a reflection of their socioeconomic position), education level (degree or above/any other qualification/no qualification), smoking status (never/previous/current), alcohol consumption (not current/less than 3 times a week/3 or more times a week), healthy diet score (continuous), TV watching time (continuous), obesity status (grouped using body mass index, normal or underweight (<25 kg/m²)/overweight (25–<30 kg/m²)/obese (≥30 kg/m²)), grip strength (continuous), glycated hemoglobin (HbA_1c, continuous), hypertension (yes/no), high cholesterol (yes/no), depression (yes/no), and family history of diabetes (yes/no). Age, sex, recruitment center, and Townsend Deprivation Index were known before arrival at the assessment center. The Townsend Deprivation Index, a composite measure of deprivation, was based on a participant's postcode. Information on ethnicity, education level, smoking status, alcohol consumption, healthy diet score, TV watching time, family history of diabetes, and depression was obtained using touchscreen questionnaires or verbal interviews. Physical measurements, including height, weight, and grip strength, were obtained by training nurses. Body mass index was calculated as weight in kilograms divided by the square of height in meters. HbA_1c levels were measured using high-performance liquid chromatography on a Bio-Rad VARIANT II TURBO analyzer (Bio-Rad, Des Plaines, IL, USA). The prevalence of hypertension and high cholesterol levels was obtained from self-reported questionnaires, hospital records, and death registries. The initial assessments using touchscreen questionnaires or verbal interviews were carried out between 2006 and 2010. Some of the covariates, including education level, smoking status, alcohol consumption, healthy diet score, obesity status, TV watching time, grip strength, and hypertension, were obtained again during 2012 and 2013 and since 2014. For the purposes of this study, these were determined by the time-point closest to the accelerometry. Detailed information is provided in Supplementary Table 2.

2.5. Statistical analyses

Descriptive characteristics were presented as frequencies and percentages if categorical, and as mean or median (interquartile range) if continuous. Missing covariate data were imputed with multiple imputations using chained equations to minimize the potential inferential. The percentage of missing data for each covariate was less than 10% (Supplementary Table 3).

We conducted 3 multivariable Cox proportional hazards regression models, with the time from accelerometer wearing as follow-up time, to estimate the associations of PA (total volume of PA, MVPA, and LPA) and sleep duration with incident type 2 diabetes. Model 1 was adjusted for age and sex. Model 2 was additionally adjusted for ethnicity, season of accelerometer wearing, recruitment center, Townsend Deprivation Index, education level, smoking status, alcohol consumption, healthy diet score, obesity status, TV watching time, grip strength, and HbA_1c. Model 3 was further adjusted for hypertension, high cholesterol levels, depression, and a family history of diabetes. Sleep duration was included as a covariate in the models for PA and vice versa. We also used a restricted cubic spline with 4 knots selected to smoothen the curve and examine the potential linear or nonlinear pattern of the estimated association.

To test the joint association between sleep duration and PA, we first examined the interactions between sleep duration and PA in association with the risk of type 2 diabetes on either an additive or multiplicative scale. Second, we repeated multivariable Cox regression and restricted cubic spline analyses to investigate the association between sleep duration and incident type 2 diabetes in each subset of participants with different levels of PA. Third, we subdivided the overall sample into 6 groups according to PA volume and sleep duration. We then used multivariable Cox models to estimate the joint association between PA and sleep duration with incident type 2 diabetes, with the group having higher PA volume and normal sleep duration, concurrently, used as a reference. Fourth, in the subgroups stratified by sleep duration associated with an increased risk of type 2 diabetes, we repeated multivariable Cox regression and restricted cubic spline analyses to compare the risks of incident type 2 diabetes associated with higher PA volume with those associated with lower PA volume.

Several sensitivity analyses were conducted to examine the robustness of the current findings, including the exclusion of participants with any missing covariate data, the exclusion of type 2 diabetes events occurring in the first 2 years of follow-up, the restriction of participants who wore the accelerometer for the whole 7 days, and the exclusion of participants with a history of night shift work. To minimize the possible bias exerted by the coronavirus disease 2019 pandemic, we repeated the analyses by censoring up to December 31, 2019, which was regarded as the commencement of the pandemic.

Inspection of Schoenfeld residual plots for covariates confirmed the proportional hazards assumption for all variables. Statistical analyses were performed using R software Version 4.1.2 (R Development Core Team, Vienna, Austria). Statistical significance was defined as p < 0.05 (2-sided).

3. Results

3.1. Baseline characteristics

Table 1 shows the baseline characteristics of the study participants stratified by sleep duration. A total of 88,000 participants were included (mean age, 62.2 years; 57.2% females). The percentage of participants with normal, short, and long sleep durations was 73.4%, 6.9%, and 19.7%, respectively. Fewer than half (40.0%) of the participants met recommended MVPA levels. Compared with participants with normal sleep duration, those with short sleep duration were more likely to have socioeconomic deprivation, smoking, obesity, family history of diabetes, and history of depression, hypertension, and dyslipidemia, whereas long sleepers tended to be more physically inactive (having lower level of total volume of PA, MVPA, and LPA) and had a history of depression, hypertension, and dyslipidemia. The baseline characteristics stratified by PA are shown in Supplementary Tables 4–6.

Table 1.

Baseline characteristics of the study participants stratified by sleep duration (n = 88,000), data are presented as mean ± SD or n (%).

Characteristics	Total (n = 88,000)	Short sleep duration (<6 h) (n = 6063)	Normal sleep duration (6–8 h) (n = 64,589)	Long sleep duration (>8 h) (n = 17,348)
Age at accelerometry (year)	62.2 ± 7.9	61.3 ± 8.0	62.0 ± 7.9	63.4 ± 7.6
Female	50,358 (57.2)	2530 (41.7)	37,083 (57.4)	10,745 (61.9)
White ethnicity	85,442 (97.1)	5581 (92.1)	62,737 (97.1)	17,124 (98.7)
Season of accelerometer wearing
Spring	19,898 (22.6)	1359 (22.4)	14,540 (22.5)	3999 (23.1)
Summer	22,986 (26.1)	1803 (29.7)	17,405 (26.9)	3778 (21.8)
Autumn	26,312 (29.9)	1799 (29.7)	19,131 (29.6)	5382 (31.0)
Winter	18,804 (21.4)	1102 (18.2)	13,513 (20.9)	4189 (24.1)
Recruitment center
England	78,967 (89.7)	5495 (90.6)	57,989 (89.8)	15,483 (89.2)
Wales	3262 (3.7)	231 (3.8)	2370 (3.7)	661 (3.8)
Scotland	5771 (6.6)	337 (5.6)	4230 (6.5)	1204 (6.9)
Townsend Deprivation Index (interquartile range)^a	–2.47 (–3.83 to –0.23)	–1.97 (–3.67 to 0.77)	–2.46 (–3.82 to –0.21)	–2.65 (–3.92 to –0.64)
Education level
Degree or above	38,897 (44.2)	2804 (46.2)	29,279 (45.3)	6814 (39.3)
Any other qualification	42,009 (47.7)	2827 (46.6)	30,383 (47.0)	8799 (50.7)
No qualification	7094 (8.1)	432 (7.1)	4927 (7.6)	1735 (10.0)
Smoking status
Never	51,102 (58.1)	3266 (53.9)	37,447 (58.0)	10,389 (59.9)
Previous	31,380 (35.7)	2222 (36.6)	23,107 (35.8)	6051 (34.9)
Current	5518 (6.3)	575 (9.5)	4035 (6.2)	908 (5.2)
Alcohol consumption
Not current	5093 (5.8)	462 (7.6)	3624 (5.6)	1007 (5.8)
<3 times a week	40,444 (46.0)	2747 (45.3)	29,314 (45.4)	8383 (48.3)
≥3 times a week	42,463 (48.3)	2854 (47.1)	31,651 (49.0)	7958 (45.9)
Healthy diet score	2.7 ± 1.2	2.6 ± 1.2	2.7 ± 1.2	2.7 ± 1.2
Obesity status
Normal/underweight (<25 kg/m²)	35,893 (40.8)	1899 (31.3)	26,478 (41.0)	7516 (43.3)
Overweight (25–<30 kg/m²)	36,236 (41.2)	2607 (43.0)	26,628 (41.2)	7001 (40.4)
Obese (≥30 kg/m²)	15,871 (18.0)	1557 (25.7)	11,483 (17.8)	2831 (16.3)
TV watching time (h/day)	1.9 ± 3.3	1.8 ± 3.5	1.8 ± 3.4	2.3 ± 3.0
Grip strength (kg)	30.3 ± 10.6	32.3 ± 11.0	30.3 ± 10.6	29.6 ± 10.5
HbA1c (mmol/mol)	34.7 ± 3.6	35.0 ± 3.8	34.7 ± 3.6	34.6 ± 3.6
HbA1c (%)	5.3 ± 0.3	5.4 ± 0.3	5.3 ± 0.3	5.3 ± 0.3
Hypertension	22,304 (25.3)	1731 (28.6)	15,937 (24.7)	4636 (26.7)
High cholesterol	11,168 (12.7)	801 (13.2)	7967 (12.3)	2400 (13.8)
Depression	7691 (8.7)	548 (9.0)	5528 (8.6)	1615 (9.3)
Family history of diabetes	15,188 (17.3)	1156 (19.1)	11,158 (17.3)	2874 (16.6)
Total volume of PA (milli-gravity)	28.2 ± 8.2	30.2 ± 9.1	28.8 ± 8.1	25.4 ± 7.5
Low volume of PA	44,041 (50.0)	2466 (40.7)	30,312 (46.9)	11,263 (64.9)
High volume of PA	43,959 (50.0)	3597 (59.3)	34,277 (53.1)	6085 (35.1)
MVPA (min/week)	159.5 ± 150.0	164.1 ± 171.0	163.0 ± 150.0	145.0 ± 140.9
Recommended MVPA	35,201 (40.0)	2381 (39.3)	26,537 (41.1)	6283 (36.2)
Not recommended MVPA	52,799 (60.0)	3682 (60.7)	38,052 (58.9)	11,065 (63.8)
LPA (min/week)	1,858.5 ± 409.5	2064.4 ± 470.4	1888.1 ± 394.1	1676.0 ± 378.8
Low LPA	44,000 (50.0)	1903 (31.4)	30163 (46.7)	11,934 (68.8)
High LPA	44,000 (50.0)	4160 (68.6)	34,426 (53.3)	5414 (31.2)

Open in a new tab

Note: Percentages add up not to 100% due to rounding.

Lower income was defined as average total household income before tax less than £1800.

Abbreviations: HbA1c = glycated hemoglobin; LPA = light-intensity physical activity; MVPA = moderate-to-vigorous physical activity; PA = physical activity.

3.2. Independent association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

During a median follow-up period of 7.0 years, 1615 incident type 2 diabetes cases were documented. After full adjustment in Model 3, short sleep duration was associated with a 21% higher risk of incident type 2 diabetes (HR = 1.21, 95% confidence interval (95%CI): 1.03–1.41; Table 2) when compared to normal sleep duration, whereas long sleep duration was not significantly associated with any excessive risk of incident type 2 diabetes (HR = 1.01, 95%CI: 0.89–1.15).

Table 2.

Associations of accelerometer-measured sleep duration and PA with incident type 2 diabetes (n = 88,000)

	Events/n	Incidence rate (%)/per 1000 person-years	Model 1 HR (95%CI)	Model 2 HR (95%CI)	Model 3 HR (95%CI)
Sleep duration
Normal sleep duration	1129/64,589	2.5	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Short sleep duration	181/6063	4.4	1.64 (1.40–1.92)	1.22 (1.04–1.43)	1.21 (1.03–1.41)
Long sleep duration	305/17,348	2.6	0.97 (0.86–1.11)	1.02 (0.90–1.16)	1.01 (0.89–1.15)
Total volume of PA
Low volume of PA	1144/44,041	3.8	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
High volume of PA	471/43,959	1.5	0.47 (0.42–0.52)	0.65 (0.58–0.73)	0.67 (0.60–0.76)
MVPA
Not recommended MVPA	1297/52,799	3.6	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Recommended MVPA	318/35,201	1.3	0.37 (0.33–0.42)	0.59 (0.52–0.68)	0.61 (0.54–0.70)
LPA
Low LPA	955/44,000	3.2	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
High LPA	660/44,000	2.2	0.76 (0.69–0.84)	0.84 (0.76–0.93)	0.85 (0.77–0.95)

Open in a new tab

Notes: Model 1 was adjusted for age and sex. Model 2 was adjusted for Model 1 plus ethnicity, season of accelerometer wearing, recruitment center, Townsend Deprivation Index, education level, smoking status, alcohol consumption, healthy diet score, obesity status, TV watching time, grip strength, and HbA1c. Model 3 was adjusted for Model 2 plus hypertension, high cholesterol, depression, and family history of diabetes.Abbreviations: 95%CI = 95% confidence interval; HR = hazard ratio; LPA = light-intensity physical activity; MVPA = moderate-to-vigorous physical activity; PA = physical activity.

Regarding the association between PA and the risk of type 2 diabetes (Fig. 2), a nonlinear dose–response pattern for MVPA (p_nonlinearity < 0.001) and linear dose–response patterns for total volume of PA (p_nonlinearity = 0.358) and LPA (p_nonlinearity = 0.452) were noted. A higher PA at any intensity was associated with a lower risk of incident type 2 diabetes: high volume of PA (vs. low volume of PA): HR = 0.67, 95%CI: 0.60–0.76; recommended MVPA (vs. not recommended MVPA): HR = 0.61, 95%CI: 0.54–0.70; high LPA (vs. low LPA): HR = 0.85, 95%CI: 0.77–0.95) (Table 2).

3.3. Subgroup analyses and joint association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

Fig. 3 and Supplementary Table 7 illustrate the association between sleep duration and incident type 2 diabetes stratified by the total volume of PA, MVPA, and LPA. A higher hazard for type 2 diabetes with short sleep duration, as compared to that with normal sleep duration, was observed in the low total volume of PA subgroup (HR = 1.26, 95%CI: 1.03–1.54); however, this pattern was not found in the high total volume of PA subgroup (HR = 1.11, 95%CI: 0.85–1.45). No excessive hazard of long sleep duration was found in either high or low total volume of PA subgroups (high: HR = 0.90, 95%CI: 0.66–1.21; low: HR = 1.04, 95%CI: 0.90–1.20). Similar patterns were also observed in MVPA and LPA. For more details see Supplementary Table 7.

Fig. 4 shows the results of the joint analyses of the association of sleep duration and PA with incident type 2 diabetes. Compared with the combination of normal sleep duration and high or recommended PA, a combination of short sleep duration and low volume of PA (HR = 1.81, 95%CI: 1.46–2.25), not recommended MVPA (HR = 1.92, 95%CI: 1.55–2.36), or low LPA (HR = 1.49, 95%CI: 1.16–1.90) consistently had the highest risk of type 2 diabetes. When short sleep duration was combined with a high volume of PA (HR = 1.14, 95%CI: 0.88–1.49), recommended MVPA (HR = 1.02, 95%CI: 0.71–1.48), or high LPA (HR = 1.14, 95%CI: 0.92–1.41), the relative risk of type 2 diabetes was insignificant. The lowest risks of type 2 diabetes in each PA category were found for the combinations of high PA volume and long sleep duration (HR = 0.90, 95%CI: 0.67–1.22), recommended MVPA and normal sleep duration (reference group), and high LPA and long sleep duration (HR = 0.92, 95%CI: 0.71–1.19). None of the multiplicative or additive interactions attained statistical significance (Supplementary Tables 8 and 9); nevertheless, they were consistent in the direction of positive multiplicative and positive additive interactions.

3.4. Sensitivity analyses

The results of most sensitivity analyses remained robust (Supplementary Tables 10–24). The major exception was that in the subsample of participants who completed the 7-day accelerometer wearing (Supplementary Table 17), although similar patterns were observed, the excessive type 2 diabetes risk of the short sleepers combined with a low or not recommended level of PA in the stratified analysis did not attain statistical significance.

4. Discussion

In this large population-based prospective cohort of 88,000 middle-aged UK participants, accelerometer-measured short (<6 h/day) but not long sleep duration (>8 h/day) was associated with an increased risk of incident type 2 diabetes. All intensities of PA (total volume of PA, MVPA, and LPA) were associated with a reduced risk of incident type 2 diabetes. Inactive short sleepers were found to have an increased risk of type 2 diabetes compared to that of active normal sleepers; however, short sleepers with a high total volume of PA (>27.23mg), a recommended level of MVPA (>150 min of MPA, >75 min of VPA, or an equivalent combination of MPA and VPA per week), or a high level of LPA (>1839.69 min/week) did not have this excessive risk. These findings suggest that the detrimental effect of short sleep duration on type 2 diabetes could be mitigated by any intensity of PA that reached a high or recommended level.

4.1. Independent association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

Using accelerometer-measured data, the current findings replicated the widely recognized association between short sleep duration and increased risk of type 2 diabetes based on self-reports, with the strength of this association similar to that of existing findings.⁶ Meanwhile, in contrast to existing findings,⁶ this study found no excessive risk of type 2 diabetes associated with long sleep duration. Similar null associations of long sleep have been observed among several cardiometabolic outcomes, using objective sleep measurements¹⁶^,³⁰ and genetically predicted long sleep.³¹ One possible explanation for the inconsistency between objective and subjective measurements is that self-reported long sleep could have reflected an overestimation of sleep, such as including excessive time in bed, a fragmented sleep process, or a sedentary lifestyle.³² All these factors could contribute to the increased risk of type 2 diabetes in self-reported studies. Indeed, poorer sleep quality and longer sedentary time are associated with a higher risk of type 2 diabetes.³³^,³⁴ This suggests that objective sleep measurements should be involved in the screening and prevention of type 2 diabetes.

The current results replicate solid findings regarding the protective effect of MVPA against type 2 diabetes risk and they corroborate the consensus on adhering to the WHO recommendation of 150 min of MPA or 75 min of VPA per week to prevent type 2 diabetes.²⁴ Interestingly, the current study discovered a weaker but significant inverse prospective association between LPA and incident type 2 diabetes. While MVPA has been well recognized for its significance in reducing type 2 diabetes risk, the effect of LPA is scarcely known because of the insensitivity of self-report measurements adopted by most existing studies.¹¹ To the best of our knowledge, this is the first study to find direct support for the protective role of LPA in the development of type 2 diabetes among the general population. Both MVPA reaching the WHO-recommended level and LPA ≥1840 min/week are associated with a lower risk of type 2 diabetes.

4.2. Joint association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

To our knowledge, this is also the first study to systematically examine the joint association of objectively measured sleep duration and PA with the onset of type 2 diabetes from a prospective perspective. The weak but consistent trend in the current study that short sleep duration was associated with a higher risk of incident type 2 diabetes only when combined with physical inactivity suggests that high or recommended levels of PA have the potential to protect and prevent short sleepers from developing type 2 diabetes. This protective effect of PA against short sleep from developing type 2 diabetes is generally consistent with existing studies. For example, an early experimental study with a small human sample (n = 10) found that PA could partially protect sleep-deprived men from decreased insulin sensitivity.³⁵ Some cross-sectional studies have observed that among individuals with short sleep duration, those with a higher level of PA tend to have a lower prevalence of diabetes or insulin resistance than their peers.¹⁶^,³⁶ In addition, a trial demonstrated that a lifestyle intervention that included increasing 30 min of MPA per day ameliorated the increased risk of type 2 diabetes associated with self-reported long sleep duration; however, the independent contribution of MPA was inseparable from other lifestyle modifications.³⁷ Compared with previous studies, by using a longitudinal design, objective sleep and PA measures, and a large sample size, the current study provided direct and robust evidence supporting the hypothesis that PA reaching a high or recommended level, regardless of intensity, may help diminish the excessive risk of type 2 diabetes among short sleepers.

The current findings highlight the importance of a public health strategy promoting PA of any intensity to combat the risk of type 2 diabetes associated with lack of sleep. The high prevalence of chronic insufficient sleep is a constant health concern.³⁸ Multiple factors may prevent adults from getting an adequate amount of sleep, such as chronic pain³⁹ or work demands.⁴⁰ For these short sleepers, interventions that facilitate sufficient PA should be particularly valued and encouraged. It is worth noting that the relationship between sleep and PA is bidirectional. Previous studies showed that PA improved sleep quantity and quality.¹² These sleep improvements, in turn, protected individuals from PA-induced injuries and diseases.¹² The protective role of the WHO-recommended MVPA level has been well validated up to now. In practical terms, people may achieve the recommended level by including at least 11 min of sports exercises or at least 21 min of walking, cycling, or shopping into their daily schedule.²⁵^,⁴¹ Our study further demonstrated that LPA has a beneficial effect in lowering the risk of type 2 diabetes. This is particularly intriguing because a considerable number of people, usually those at a higher risk for type 2 diabetes, may have difficulty performing MVPA or insist on performing MVPA instead of LPA.⁴² In these cases, performing at least 4–5 h of LPA per day, such as cleaning or slow walking, is considered good alternative protection.

4.3. Strengths and limitations

This study had several strengths. First, the inclusion of a prospective cohort design, a large sample size of 88,000 individuals, and detailed epidemiological profiles allowed for high-quality data analyses and valid estimates of exposure –disease relationships. Second, as stated above, the adoption of accelerometer-based measurements for sleep and PA could minimize the recall bias associated with self-reported measurements and allow for capturing poorly understood LPA.

There are several limitations that need to be addressed. First, in the UK Biobank, participants only began to wear the accelerometer several years after baseline data were collected. This time gap led to difficulties excluding cases who developed type 2 diabetes during this period from the incident dataset. However, the findings remained robust even after removing the cases occurring within the first 2 years. Second, prospective changes in accelerometer-measured sleep duration and PA were unavailable in the UK Biobank cohort. While existing analyses of the UK Biobank have suggested that the patterns of sleep and PA were relatively stable across time overall,⁴³ a dynamic association between sleep, PA, and the development of type 2 diabetes could still exist without being observable. Third, it is unclear from the current results whether one PA intensity was more protective or whether they were all similar. HRs are calculated in the current study through comparisons within the same PA intensity category—to be specific, through comparisons between recommended and not recommended MVPA, or high and low total volume of PA or LPA. This prevented us from drawing any conclusions about between-PA intensity category comparisons.

Fourth, sedentary behavior, which is an important issue closely related to type 2 diabetes, is not included in the current analyses. This is because sedentary behavior is closely associated with PA and sleep, which would not satisfy the premise of Cox proportional hazards regression models. According to the consensus definition of sedentary behavior,⁴⁴ it refers to any waking behaviors characterized by an energy expenditure not reaching the threshold level of PA. Several existing studies have adopted this consensus definition to calculate the accelerometer-measured sedentary time.⁴⁵^,⁴⁶ Under this definition, we calculated sedentary time from the accelerometer data of the current study by defining sedentary as any epoch with a mean acceleration <30 mg outside the SPT-window, and we found that sedentary time was significantly correlated with all intensities of PA (Spearman's correlation: ρ_{total volume of PA} = –0.727, ρ_MVPA = –0.371, ρ_LPA = –0.666). Therefore, we decided not to include sedentary time into the current analyses.

Finally, the lower incidence of type 2 diabetes in the UK Biobank sample (1.8%) compared to that in the general UK population (2.3%)⁴⁷ was consistent with the lower prevalence of self-reported diabetes among UK Biobank participants,⁴⁸ which probably reflects a selection bias favoring healthier participants in the UK Biobank cohort.⁴⁸ A similar selection bias may also occur in the subgroup who agreed with and complied with accelerometer wearing. Sensitivity analyses found the excessive risks of type 2 diabetes for the short sleepers with low or not recommended PA to be insignificant, which likely reflected a floor effect resulting from healthier participants.

5. Conclusion

The current study indicated that accelerometer-measured short but not long sleep duration was associated with an excessive risk of type 2 diabetes. An increased level of PA, regardless of intensity, may help prevent the development of type 2 diabetes among individuals with short sleep duration.

Acknowledgments

This research was conducted using the UK Biobank Resource under application Number 58082. Language editing was performed by Elsevier. The authors want to thank Huachen Xue for his assistance in the preparation of the manuscript. Some of the icons in the graphic abstract were created by andinur, Freepik, geotatah, Kalashnyk, and Raftel Design from www.flaticon.com. The authors also thank them for their contributions. This work was supported by the National Key R&D Program of China (2021YFC2501500) and National Natural Science Foundation of China (82171476). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Authors’ contributions

XJ and YC conceived and designed the study, conducted statistical analyses and data interpretation, and drafted the manuscript with the help of HF; they had full access to all the data in the study and took responsibility for the integrity of the data and accuracy of the data analysis; MZ, JWYC, YL, APSK, and XT helped to draft the manuscript; YKW revised the manuscript; YYL and JZ, as corresponding authors, contributed to the conception and design of the study, had full access to all the data in the study, and took responsibility for the integrity of the data and accuracy of the data analysis. All authors have read and approved the final version of the manuscript, and agree with the order of presentation of the authors.

Competing interests

The authors declare that they have no competing interests.

Footnotes

Peer review under responsibility of Shanghai University of Sport.

Supplementary materials associated with this article can be found in the online version at doi:10.1016/j.jshs.2023.03.001.

Contributor Information

Yannis Yan Liang, Email: liangyan@link.cuhk.edu.hk.

Jihui Zhang, Email: zhangjihui@gzhmu.edu.cn.

Supplementary materials

mmc1.docx^{(307.3KB, docx)}

Download video file^{(10.7MB, mp4)}

References

1.Zheng Y, Ley SH, Hu FB. Global aetiology and epidemiology of type 2 diabetes mellitus and its complications. Nat Rev Endocrinol. 2018;14:88–98. doi: 10.1038/nrendo.2017.151. [DOI] [PubMed] [Google Scholar]
2.Liu J, Bai R, Chai Z, Cooper ME, Zimmet PZ, Zhang L. Low- and middle-income countries demonstrate rapid growth of type 2 diabetes: An analysis based on global burden of disease 1990–2019 data. Diabetologia. 2022;65:1339–1352. doi: 10.1007/s00125-022-05713-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Yuan S, Larsson SC. An atlas on risk factors for type 2 diabetes: A wide-angled mendelian randomisation study. Diabetologia. 2020;63:2359–2371. doi: 10.1007/s00125-020-05253-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Schipper SBJ, Van Veen MM, Elders PJM, et al. Sleep disorders in people with type 2 diabetes and associated health outcomes: A review of the literature. Diabetologia. 2021;64:2367–2377. doi: 10.1007/s00125-021-05541-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Consensus Conference P, Watson NF, Badr MS, et al. Joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society on the recommended amount of sleep for a healthy adult: Methodology and discussion. Sleep. 2015;38:1161–1183. doi: 10.5665/sleep.4886. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Shan Z, Ma H, Xie M, et al. Sleep duration and risk of type 2 diabetes: A meta-analysis of prospective studies. Diabetes Care. 2015;38:529–537. doi: 10.2337/dc14-2073. [DOI] [PubMed] [Google Scholar]
7.Larcher S, Benhamou PY, Pepin JL, Borel AL. Sleep habits and diabetes. Diabetes Metab. 2015;41:263–271. doi: 10.1016/j.diabet.2014.12.004. [DOI] [PubMed] [Google Scholar]
8.Reutrakul S, Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metabolism. 2018;84:56–66. doi: 10.1016/j.metabol.2018.02.010. [DOI] [PubMed] [Google Scholar]
9.Anothaisintawee T, Reutrakul S, Van Cauter E, Thakkinstian A. Sleep disturbances compared to traditional risk factors for diabetes development: Systematic review and meta-analysis. Sleep Med Rev. 2016;30:11–24. doi: 10.1016/j.smrv.2015.10.002. [DOI] [PubMed] [Google Scholar]
10.American Diabetes Association Professional Practice Committee 3. Prevention or delay of type 2 diabetes and associated comorbidities: Standards of medical care in diabetes–2022. Diabetes Care. 2022;45(Suppl.1):S39–S45. doi: 10.2337/dc22-S003. [DOI] [PubMed] [Google Scholar]
11.Fuzeki E, Engeroff T, Banzer W. Health benefits of light-intensity physical activity: A systematic review of accelerometer data of the National Health and Nutrition Examination Survey (NHANES) Sports Med. 2017;47:1769–1793. doi: 10.1007/s40279-017-0724-0. [DOI] [PubMed] [Google Scholar]
12.Chennaoui M, Arnal PJ, Sauvet F, Leger D. Sleep and exercise: A reciprocal issue? Sleep Med Rev. 2015;20:59–72. doi: 10.1016/j.smrv.2014.06.008. [DOI] [PubMed] [Google Scholar]
13.Kolb H, Martin S. Environmental/lifestyle factors in the pathogenesis and prevention of type 2 diabetes. BMC Med. 2017;15:131. doi: 10.1186/s12916-017-0901-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Huang BH, Duncan MJ, Cistulli PA, Nassar N, Hamer M, Stamatakis E. Sleep and physical activity in relation to all-cause, cardiovascular disease and cancer mortality risk. Br J Sports Med. 2022;56:718–724. doi: 10.1136/bjsports-2021-104046. [DOI] [PubMed] [Google Scholar]
15.Chen LJ, Hamer M, Lai YJ, Huang BH, Ku PW, Stamatakis E. Can physical activity eliminate the mortality risk associated with poor sleep? A 15-year follow-up of 341,248 MJ cohort participants. J Sport Health Sci. 2022;11:596–604. doi: 10.1016/j.jshs.2021.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Huang BH, Inan-Eroglu E, Hamer M, Stamatakis E. Joint associations of device-measured physical activity and sleep duration with cardiometabolic health in the 1970 British cohort study. J Sci Med Sport. 2020;23:1191–1196. doi: 10.1016/j.jsams.2020.07.012. [DOI] [PubMed] [Google Scholar]
17.Sudlow C, Gallacher J, Allen N, et al. UK Biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12 doi: 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.National Research Ethics Service. UK Biobank ethical approval. Available at: https://www.ukbiobank.ac.uk/media/cs1h15s3/rtb-nwrec-application-and-approval-2011.pdf. [accessed 17.06.2011].
19.Doherty A, Jackson D, Hammerla N, et al. Large scale population assessment of physical activity using wrist worn accelerometers: The UK Biobank study. PLoS One. 2017;12 doi: 10.1371/journal.pone.0169649. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.van Hees VT, Fang Z, Langford J, et al. Autocalibration of accelerometer data for free-living physical activity assessment using local gravity and temperature: An evaluation on four continents. J Appl Physiol (1985) 2014;117:738–744. doi: 10.1152/japplphysiol.00421.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Ramakrishnan R, Doherty A, Smith-Byrne K, et al. Accelerometer measured physical activity and the incidence of cardiovascular disease: Evidence from the UK Biobank cohort study. PLoS Med. 2021;18 doi: 10.1371/journal.pmed.1003487. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Khurshid S, Weng LC, Al-Alusi MA, et al. Accelerometer-derived physical activity and risk of atrial fibrillation. Eur Heart J. 2021;42:2472–2483. doi: 10.1093/eurheartj/ehab250. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Leroux A, Xu S, Kundu P, et al. Quantifying the predictive performance of objectively measured physical activity on mortality in the UK Biobank. J Gerontol A Biol Sci Med Sci. 2021;76:1486–1494. doi: 10.1093/gerona/glaa250. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Smith AD, Crippa A, Woodcock J, Brage S. Physical activity and incident type 2 diabetes mellitus: A systematic review and dose-response meta-analysis of prospective cohort studies. Diabetologia. 2016;59:2527–2545. doi: 10.1007/s00125-016-4079-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Bull FC, Al-Ansari SS, Biddle S, et al. World health organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54:1451–1462. doi: 10.1136/bjsports-2020-102955. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Jones SE, Lane JM, Wood AR, et al. Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nat Commun. 2019;10:343. doi: 10.1038/s41467-018-08259-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.van Hees VT, Sabia S, Jones SE, et al. Estimating sleep parameters using an accelerometer without sleep diary. Sci Rep. 2018;8:12975. doi: 10.1038/s41598-018-31266-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Migueles JH, Rowlands AV, Huber F, Sabia S, van Hees VT. GGIR: A research community–driven open source R package for generating physical activity and sleep outcomes from multi-day raw accelerometer data. J Meas Phy Behav. 2019;2:188–196. [Google Scholar]
29.Yin J, Jin X, Shan Z, et al. Relationship of sleep duration with all-cause mortality and cardiovascular events: A systematic review and dose–response meta-analysis of prospective cohort studies. J Am Heart Assoc. 2017;6 doi: 10.1161/jaha.117.005947. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Ren R, Covassin N, Yang L, et al. Objective but not subjective short sleep duration is associated with hypertension in obstructive sleep apnea. Hypertension. 2018;72:610–617. doi: 10.1161/HYPERTENSIONAHA.118.11027. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Ai S, Zhang J, Zhao G, et al. Causal associations of short and long sleep durations with 12 cardiovascular diseases: Linear and nonlinear mendelian randomization analyses in UK Biobank. Eur Heart J. 2021;42:3349–3357. doi: 10.1093/eurheartj/ehab170. [DOI] [PubMed] [Google Scholar]
32.Mokhlesi B, Temple KA, Tjaden AH, et al. Association of self-reported sleep and circadian measures with glycemia in adults with prediabetes or recently diagnosed untreated type 2 diabetes. Diabetes Care. 2019;42:1326–1332. doi: 10.2337/dc19-0298. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Sondrup N, Termannsen AD, Eriksen JN, et al. Effects of sleep manipulation on markers of insulin sensitivity: A systematic review and meta-analysis of randomized controlled trials. Sleep Med Rev. 2022;62 doi: 10.1016/j.smrv.2022.101594. [DOI] [PubMed] [Google Scholar]
34.Li X, Zhou T, Ma H, Liang Z, Fonseca VA, Qi L. Replacement of sedentary behavior by various daily-life physical activities and structured exercises: Genetic risk and incident type 2 diabetes. Diabetes Care. 2021;44:2403–2410. doi: 10.2337/dc21-0455. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.VanHelder T, Symons JD, Radomski MW. Effects of sleep deprivation and exercise on glucose tolerance. Aviat Space Environ Med. 1993;64:487–492. [PubMed] [Google Scholar]
36.Zuo H, Shi Z, Yuan B, et al. Interaction between physical activity and sleep duration in relation to insulin resistance among non-diabetic Chinese adults. BMC Public Health. 2012;12:247. doi: 10.1186/1471-2458-12-247. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Tuomilehto H, Peltonen M, Partinen M, et al. Sleep duration, lifestyle intervention, and incidence of type 2 diabetes in impaired glucose tolerance: The Finnish Diabetes Prevention Study. Diabetes Care. 2009;32:1965–1971. doi: 10.2337/dc08-1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Basner M, Dinges DF. Sleep duration in the United States 2003–2016: First signs of success in the fight against sleep deficiency? Sleep. 2018;41:zsy012. doi: 10.1093/sleep/zsy012. [DOI] [PubMed] [Google Scholar]
39.Cohen SP, Vase L, Hooten WM. Chronic pain: An update on burden, best practices, and new advances. The Lancet. 2021;397:2082–2097. doi: 10.1016/S0140-6736(21)00393-7. [DOI] [PubMed] [Google Scholar]
40.Arrona-Palacios A, Gradisar M. Self-reported sleep duration, sleep quality and sleep problems in Mexican adults: Results of the 2016 Mexican National Halfway Health and Nutrition Survey. Sleep Health. 2021;7:246–253. doi: 10.1016/j.sleh.2020.08.006. [DOI] [PubMed] [Google Scholar]
41.UK CMO. UK chief medical officers' physical activity guidelines. 2019. Available at: https://www.gov.uk/government/publications/physical-activity-guidelines-uk-chief-medical-officers-report. [accessed 07.09.2019].
42.Herzig KH, Ahola R, Leppaluoto J, Jokelainen J, Jamsa T, Keinanen-Kiukaanniemi S. Light physical activity determined by a motion sensor decreases insulin resistance, improves lipid homeostasis and reduces visceral fat in high-risk subjects: Prediabex study rct. Int J Obes (Lond) 2014;38:1089–1096. doi: 10.1038/ijo.2013.224. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Fatima Y, Bucks RS, Mamun AA, et al. Sleep trajectories and mediators of poor sleep: Findings from the longitudinal analysis of 41,094 participants of the UK Biobank cohort. Sleep Med. 2020;76:120–127. doi: 10.1016/j.sleep.2020.10.020. [DOI] [PubMed] [Google Scholar]
44.Tremblay MS, Aubert S, Barnes JD, et al. Sedentary Behavior Research Network (SBRN) – Terminology consensus project process and outcome. Int J Behav Nutr Phys Act. 2017;14:75. doi: 10.1186/s12966-017-0525-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Hooker SP, Diaz KM, Blair SN, et al. Association of accelerometer-measured sedentary time and physical activity with risk of stroke among us adults. JAMA Netw Open. 2022;5 doi: 10.1001/jamanetworkopen.2022.15385. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Kandola AA, Del Pozo Cruz B, Osborn DPJ, Stubbs B, Choi KW, Hayes JF. Impact of replacing sedentary behaviour with other movement behaviours on depression and anxiety symptoms: A prospective cohort study in the UK Biobank. BMC Med. 2021;19:133. doi: 10.1186/s12916-021-02007-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Sharma M, Nazareth I, Petersen I. Trends in incidence, prevalence and prescribing in type 2 diabetes mellitus between 2000 and 2013 in primary care: A retrospective cohort study. BMJ Open. 2016;6 doi: 10.1136/bmjopen-2015-010210. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Fry A, Littlejohns TJ, Sudlow C, et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol. 2017;186:1026–1034. doi: 10.1093/aje/kwx246. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

mmc1.docx^{(307.3KB, docx)}

Download video file^{(10.7MB, mp4)}

[bib0001] 1.Zheng Y, Ley SH, Hu FB. Global aetiology and epidemiology of type 2 diabetes mellitus and its complications. Nat Rev Endocrinol. 2018;14:88–98. doi: 10.1038/nrendo.2017.151. [DOI] [PubMed] [Google Scholar]

[bib0002] 2.Liu J, Bai R, Chai Z, Cooper ME, Zimmet PZ, Zhang L. Low- and middle-income countries demonstrate rapid growth of type 2 diabetes: An analysis based on global burden of disease 1990–2019 data. Diabetologia. 2022;65:1339–1352. doi: 10.1007/s00125-022-05713-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0003] 3.Yuan S, Larsson SC. An atlas on risk factors for type 2 diabetes: A wide-angled mendelian randomisation study. Diabetologia. 2020;63:2359–2371. doi: 10.1007/s00125-020-05253-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0004] 4.Schipper SBJ, Van Veen MM, Elders PJM, et al. Sleep disorders in people with type 2 diabetes and associated health outcomes: A review of the literature. Diabetologia. 2021;64:2367–2377. doi: 10.1007/s00125-021-05541-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0005] 5.Consensus Conference P, Watson NF, Badr MS, et al. Joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society on the recommended amount of sleep for a healthy adult: Methodology and discussion. Sleep. 2015;38:1161–1183. doi: 10.5665/sleep.4886. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0006] 6.Shan Z, Ma H, Xie M, et al. Sleep duration and risk of type 2 diabetes: A meta-analysis of prospective studies. Diabetes Care. 2015;38:529–537. doi: 10.2337/dc14-2073. [DOI] [PubMed] [Google Scholar]

[bib0007] 7.Larcher S, Benhamou PY, Pepin JL, Borel AL. Sleep habits and diabetes. Diabetes Metab. 2015;41:263–271. doi: 10.1016/j.diabet.2014.12.004. [DOI] [PubMed] [Google Scholar]

[bib0008] 8.Reutrakul S, Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metabolism. 2018;84:56–66. doi: 10.1016/j.metabol.2018.02.010. [DOI] [PubMed] [Google Scholar]

[bib0009] 9.Anothaisintawee T, Reutrakul S, Van Cauter E, Thakkinstian A. Sleep disturbances compared to traditional risk factors for diabetes development: Systematic review and meta-analysis. Sleep Med Rev. 2016;30:11–24. doi: 10.1016/j.smrv.2015.10.002. [DOI] [PubMed] [Google Scholar]

[bib0010] 10.American Diabetes Association Professional Practice Committee 3. Prevention or delay of type 2 diabetes and associated comorbidities: Standards of medical care in diabetes–2022. Diabetes Care. 2022;45(Suppl.1):S39–S45. doi: 10.2337/dc22-S003. [DOI] [PubMed] [Google Scholar]

[bib0011] 11.Fuzeki E, Engeroff T, Banzer W. Health benefits of light-intensity physical activity: A systematic review of accelerometer data of the National Health and Nutrition Examination Survey (NHANES) Sports Med. 2017;47:1769–1793. doi: 10.1007/s40279-017-0724-0. [DOI] [PubMed] [Google Scholar]

[bib0012] 12.Chennaoui M, Arnal PJ, Sauvet F, Leger D. Sleep and exercise: A reciprocal issue? Sleep Med Rev. 2015;20:59–72. doi: 10.1016/j.smrv.2014.06.008. [DOI] [PubMed] [Google Scholar]

[bib0013] 13.Kolb H, Martin S. Environmental/lifestyle factors in the pathogenesis and prevention of type 2 diabetes. BMC Med. 2017;15:131. doi: 10.1186/s12916-017-0901-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0014] 14.Huang BH, Duncan MJ, Cistulli PA, Nassar N, Hamer M, Stamatakis E. Sleep and physical activity in relation to all-cause, cardiovascular disease and cancer mortality risk. Br J Sports Med. 2022;56:718–724. doi: 10.1136/bjsports-2021-104046. [DOI] [PubMed] [Google Scholar]

[bib0015] 15.Chen LJ, Hamer M, Lai YJ, Huang BH, Ku PW, Stamatakis E. Can physical activity eliminate the mortality risk associated with poor sleep? A 15-year follow-up of 341,248 MJ cohort participants. J Sport Health Sci. 2022;11:596–604. doi: 10.1016/j.jshs.2021.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0016] 16.Huang BH, Inan-Eroglu E, Hamer M, Stamatakis E. Joint associations of device-measured physical activity and sleep duration with cardiometabolic health in the 1970 British cohort study. J Sci Med Sport. 2020;23:1191–1196. doi: 10.1016/j.jsams.2020.07.012. [DOI] [PubMed] [Google Scholar]

[bib0017] 17.Sudlow C, Gallacher J, Allen N, et al. UK Biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12 doi: 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0018] 18.National Research Ethics Service. UK Biobank ethical approval. Available at: https://www.ukbiobank.ac.uk/media/cs1h15s3/rtb-nwrec-application-and-approval-2011.pdf. [accessed 17.06.2011].

[bib0019] 19.Doherty A, Jackson D, Hammerla N, et al. Large scale population assessment of physical activity using wrist worn accelerometers: The UK Biobank study. PLoS One. 2017;12 doi: 10.1371/journal.pone.0169649. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0020] 20.van Hees VT, Fang Z, Langford J, et al. Autocalibration of accelerometer data for free-living physical activity assessment using local gravity and temperature: An evaluation on four continents. J Appl Physiol (1985) 2014;117:738–744. doi: 10.1152/japplphysiol.00421.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0021] 21.Ramakrishnan R, Doherty A, Smith-Byrne K, et al. Accelerometer measured physical activity and the incidence of cardiovascular disease: Evidence from the UK Biobank cohort study. PLoS Med. 2021;18 doi: 10.1371/journal.pmed.1003487. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0022] 22.Khurshid S, Weng LC, Al-Alusi MA, et al. Accelerometer-derived physical activity and risk of atrial fibrillation. Eur Heart J. 2021;42:2472–2483. doi: 10.1093/eurheartj/ehab250. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0023] 23.Leroux A, Xu S, Kundu P, et al. Quantifying the predictive performance of objectively measured physical activity on mortality in the UK Biobank. J Gerontol A Biol Sci Med Sci. 2021;76:1486–1494. doi: 10.1093/gerona/glaa250. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0024] 24.Smith AD, Crippa A, Woodcock J, Brage S. Physical activity and incident type 2 diabetes mellitus: A systematic review and dose-response meta-analysis of prospective cohort studies. Diabetologia. 2016;59:2527–2545. doi: 10.1007/s00125-016-4079-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0025] 25.Bull FC, Al-Ansari SS, Biddle S, et al. World health organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54:1451–1462. doi: 10.1136/bjsports-2020-102955. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0026] 26.Jones SE, Lane JM, Wood AR, et al. Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nat Commun. 2019;10:343. doi: 10.1038/s41467-018-08259-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0027] 27.van Hees VT, Sabia S, Jones SE, et al. Estimating sleep parameters using an accelerometer without sleep diary. Sci Rep. 2018;8:12975. doi: 10.1038/s41598-018-31266-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0028] 28.Migueles JH, Rowlands AV, Huber F, Sabia S, van Hees VT. GGIR: A research community–driven open source R package for generating physical activity and sleep outcomes from multi-day raw accelerometer data. J Meas Phy Behav. 2019;2:188–196. [Google Scholar]

[bib0029] 29.Yin J, Jin X, Shan Z, et al. Relationship of sleep duration with all-cause mortality and cardiovascular events: A systematic review and dose–response meta-analysis of prospective cohort studies. J Am Heart Assoc. 2017;6 doi: 10.1161/jaha.117.005947. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0030] 30.Ren R, Covassin N, Yang L, et al. Objective but not subjective short sleep duration is associated with hypertension in obstructive sleep apnea. Hypertension. 2018;72:610–617. doi: 10.1161/HYPERTENSIONAHA.118.11027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0031] 31.Ai S, Zhang J, Zhao G, et al. Causal associations of short and long sleep durations with 12 cardiovascular diseases: Linear and nonlinear mendelian randomization analyses in UK Biobank. Eur Heart J. 2021;42:3349–3357. doi: 10.1093/eurheartj/ehab170. [DOI] [PubMed] [Google Scholar]

[bib0032] 32.Mokhlesi B, Temple KA, Tjaden AH, et al. Association of self-reported sleep and circadian measures with glycemia in adults with prediabetes or recently diagnosed untreated type 2 diabetes. Diabetes Care. 2019;42:1326–1332. doi: 10.2337/dc19-0298. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0033] 33.Sondrup N, Termannsen AD, Eriksen JN, et al. Effects of sleep manipulation on markers of insulin sensitivity: A systematic review and meta-analysis of randomized controlled trials. Sleep Med Rev. 2022;62 doi: 10.1016/j.smrv.2022.101594. [DOI] [PubMed] [Google Scholar]

[bib0034] 34.Li X, Zhou T, Ma H, Liang Z, Fonseca VA, Qi L. Replacement of sedentary behavior by various daily-life physical activities and structured exercises: Genetic risk and incident type 2 diabetes. Diabetes Care. 2021;44:2403–2410. doi: 10.2337/dc21-0455. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0035] 35.VanHelder T, Symons JD, Radomski MW. Effects of sleep deprivation and exercise on glucose tolerance. Aviat Space Environ Med. 1993;64:487–492. [PubMed] [Google Scholar]

[bib0036] 36.Zuo H, Shi Z, Yuan B, et al. Interaction between physical activity and sleep duration in relation to insulin resistance among non-diabetic Chinese adults. BMC Public Health. 2012;12:247. doi: 10.1186/1471-2458-12-247. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0037] 37.Tuomilehto H, Peltonen M, Partinen M, et al. Sleep duration, lifestyle intervention, and incidence of type 2 diabetes in impaired glucose tolerance: The Finnish Diabetes Prevention Study. Diabetes Care. 2009;32:1965–1971. doi: 10.2337/dc08-1980. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0038] 38.Basner M, Dinges DF. Sleep duration in the United States 2003–2016: First signs of success in the fight against sleep deficiency? Sleep. 2018;41:zsy012. doi: 10.1093/sleep/zsy012. [DOI] [PubMed] [Google Scholar]

[bib0039] 39.Cohen SP, Vase L, Hooten WM. Chronic pain: An update on burden, best practices, and new advances. The Lancet. 2021;397:2082–2097. doi: 10.1016/S0140-6736(21)00393-7. [DOI] [PubMed] [Google Scholar]

[bib0040] 40.Arrona-Palacios A, Gradisar M. Self-reported sleep duration, sleep quality and sleep problems in Mexican adults: Results of the 2016 Mexican National Halfway Health and Nutrition Survey. Sleep Health. 2021;7:246–253. doi: 10.1016/j.sleh.2020.08.006. [DOI] [PubMed] [Google Scholar]

[bib0041] 41.UK CMO. UK chief medical officers' physical activity guidelines. 2019. Available at: https://www.gov.uk/government/publications/physical-activity-guidelines-uk-chief-medical-officers-report. [accessed 07.09.2019].

[bib0042] 42.Herzig KH, Ahola R, Leppaluoto J, Jokelainen J, Jamsa T, Keinanen-Kiukaanniemi S. Light physical activity determined by a motion sensor decreases insulin resistance, improves lipid homeostasis and reduces visceral fat in high-risk subjects: Prediabex study rct. Int J Obes (Lond) 2014;38:1089–1096. doi: 10.1038/ijo.2013.224. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0043] 43.Fatima Y, Bucks RS, Mamun AA, et al. Sleep trajectories and mediators of poor sleep: Findings from the longitudinal analysis of 41,094 participants of the UK Biobank cohort. Sleep Med. 2020;76:120–127. doi: 10.1016/j.sleep.2020.10.020. [DOI] [PubMed] [Google Scholar]

[bib0044] 44.Tremblay MS, Aubert S, Barnes JD, et al. Sedentary Behavior Research Network (SBRN) – Terminology consensus project process and outcome. Int J Behav Nutr Phys Act. 2017;14:75. doi: 10.1186/s12966-017-0525-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0045] 45.Hooker SP, Diaz KM, Blair SN, et al. Association of accelerometer-measured sedentary time and physical activity with risk of stroke among us adults. JAMA Netw Open. 2022;5 doi: 10.1001/jamanetworkopen.2022.15385. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0046] 46.Kandola AA, Del Pozo Cruz B, Osborn DPJ, Stubbs B, Choi KW, Hayes JF. Impact of replacing sedentary behaviour with other movement behaviours on depression and anxiety symptoms: A prospective cohort study in the UK Biobank. BMC Med. 2021;19:133. doi: 10.1186/s12916-021-02007-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0047] 47.Sharma M, Nazareth I, Petersen I. Trends in incidence, prevalence and prescribing in type 2 diabetes mellitus between 2000 and 2013 in primary care: A retrospective cohort study. BMJ Open. 2016;6 doi: 10.1136/bmjopen-2015-010210. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib0048] 48.Fry A, Littlejohns TJ, Sudlow C, et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol. 2017;186:1026–1034. doi: 10.1093/aje/kwx246. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Association of accelerometer-measured sleep duration and different intensities of physical activity with incident type 2 diabetes in a population-based cohort study

Xinyi Jin

Yilin Chen

Hongliang Feng

Mingqing Zhou

Joey WY Chan

Yaping Liu

Alice Pik Shan Kong

Xiao Tan

Yun-Kwok Wing

Yannis Yan Liang

Jihui Zhang

Highlights

Abstract

Purpose

Methods

Results

Conclusion

Graphical abstract

1. Introduction

2. Materials and methods

2.1. Study population and data source

Fig. 1.

2.2. Exposures

2.3. Outcomes

2.4. Covariates

2.5. Statistical analyses

3. Results

3.1. Baseline characteristics

Table 1.

3.2. Independent association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

Table 2.

Fig. 2.

3.3. Subgroup analyses and joint association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

Fig. 3.

Fig. 4.

3.4. Sensitivity analyses

4. Discussion

4.1. Independent association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

4.2. Joint association of accelerometer-measured sleep duration and PA with incident type 2 diabetes

4.3. Strengths and limitations

5. Conclusion

Acknowledgments

Acknowledgments

Authors’ contributions

Competing interests

Footnotes

Contributor Information

Supplementary materials

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases