Associations between Sleep Pattern and Quality and Cardiovascular Risk Factors among Macao School Students

Abstract

Background:

Despite accumulating evidence showing that sleep duration and cardiometabolic health are correlated, the association of sleep regularity and quality with cardiovascular risk factors in children and adolescents remains inconclusive. Therefore, we examined the relationship between sleep regularity/quality and cardiovascular risk factors in children and adolescents in Macao, China.

Methods:

We conducted a cross-sectional study among primary and secondary school students (age range: 9–18 years) in Macao, China. Body weight, height, waist circumference (WC), and serum lipid levels were measured. Sleep regularity and sleep quality were assessed by using the Pittsburgh Sleep Quality Index. Multivariable logistic regression was conducted to examine the relationship of sleep variables with the increased likelihood of body mass index (BMI)-classified overweight/obesity, WC-classified obesity, and dyslipidemia.

Results:

A total of 1078 school students were included in the analysis. Differences of >2 hours between weekday and weekend bedtimes were associated with the increased risk of BMI-classified overweight/obesity (odds ratio = 2.58, 95% confidence interval = 1.55–4.31, p < 0.01) after being adjusted for sleep variables and lifestyle factors. No statistically significant association was found between poor sleep quality and any other outcome.

Conclusions:

Irregular bedtime was associated with elevated body weight regardless of sleep amount and quality. Thus, regular bedtimes are recommended as a simple but practical strategy for preventing obesity.

Introduction

Although the obesity epidemic has abated in numerous western countries, including the United States and United Kingdom, obesity rates continue to increase among Chinese children and adolescents.¹ Over the period of 1981 to 2010, the prevalence of overweight among children and adolescents increased from 1.8% to 13.1%, whereas that of obesity increased from 0.4% to 7.5%.² Thus, modifiable risk factors for obesity must be identified before developing strategies for its prevention.

Accumulating evidence obtained in recent years suggests that short sleep duration has a causal role in high obesity risks among children.³ Short sleep durations may reduce physical activity and increase dietary consumption by intensifying tiredness and increasing opportunities for eating.⁴ Insufficient sleep may also increase food intake by mediating changes in the levels of several neuropeptides that regulate appetite.⁵ In addition, inadequate sleep is linked with other cardiometabolic risk factors, including dyslipidemia and hypertension.⁶ Although the relationship among sleep duration, obesity, and other cardiometabolic risk factors has been established, the relationships of obesity with other sleep pattern parameters, such as sleep regularity and sleep quality, have not been examined adequately.⁷ Sleep regularity refers to the discrepancy between weekday and weekend sleep durations, waking times, and bedtimes⁸ or large day-to-day variations in sleep durations.⁹ Irregular sleep durations are positively associated with the risk of childhood overweight/obesity⁹ and the increased intake of sugar-sweetened beverages.¹⁰ Considerable delays in weekend bedtimes are associated with high BMI among adolescents.¹¹ Moreover, drastic differences between weekday and weekend wake times are associated with high BMI among female children and adolescents.¹²

In contrast to sleep regularity, sleep quality refers to a collection of sleep measures, such as sleep duration and sleep problems, that can be measured objectively (e.g., polysomnography or actigraphy) or subjectively (e.g., sleep diary or self-reported surveys).¹³ Objective sleep quality is categorized into a set of indexes that include number of awakenings, amount and percentage of sleep stages, latency of rapid eye movement, number of apneas or hypopneas, and occurrence of periodic movements during sleep.¹³ Despite their precision, objective indexes are not consolidated into a global sleep-quality index on the basis of their relative importance.¹⁴

Although subjective sleep measures resemble objective sleep indexes (e.g., number of awakenings during sleep and sleep disturbances), subjective sleep quality refers to the retrospective appraisal of sleep experience as recalled by the individual and can be summarized as a measure of global sleep status.¹³ Poor sleep quality has been associated with obesity in adults.¹⁵ The relationship between poor sleep quality and obesity in adults appears to be mediated by several factors, such as perceived psychosocial stress.¹⁶ Some studies involving adults have shown that poor sleep quality is related to metabolic syndrome.^17,18

The majority of previous research focused on the association between sleep duration and obesity among children or on sleep quality among adults. By contrast, few works have examined the role of sleep regularity and sleep quality in childhood obesity and dyslipidemia. Thus, we conducted the present study to untangle the relationships of sleep regularity and sleep quality with childhood obesity and dyslipidemia in China given the rapidly increasing national burden of childhood obesity.

Method

Participants

The present study was nested within a physical assessment conducted among primary and secondary students 6–18 years of age in Macao over the period of December 2014 to June 2015. A total of 1842 individuals who were subjected to physical assessment, which comprised anthropometric measurement and blood drawing, were randomly selected from each participating school in accordance with the overall gender, grade, and age distributions of students in Macao.¹⁹ Written consent was obtained from the parents of the participants. Only students from primary four (10–11 years of age) to secondary six (18 years of age) were invited to complete the sleep survey because children younger than 10–11 years of age may be unable to comprehend questionnaires, especially the sleep quality questionnaire.²⁰

The inclusion and exclusion criteria of the participants are listed below.

Inclusion criteria:

Primary and secondary students in Macao and who participated in physical assessments. Exclusion criteria: (1)

Students younger than 10–11 years of age or older than 18 years old.

(2)

The students provided completed data on sleep survey.

Sleep Variables

The sleep survey was used to assess the duration, regularity, and quality of sleep. Sleep duration was calculated as the difference between bedtime and wake-up time in hours over the previous month. Items for assessing sleep duration on weekday and weekend nights were separated. Average sleep duration was calculated as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} BMI = { \frac { Weight \; \left( { kg } \right) } { { { \left( { Height \; \left( { cm / 100 } \right) } \right) } ^2 } } } \end{align*} \end{document}

Sleep regularity was defined as the absolute differences between weekday bedtimes, wake times, and sleep durations and weekend bedtimes, wake times, and sleep durations. The midpoint of sleep duration, which is used as a surrogate for circadian rhythm, was obtained by calculating the halfway point between bedtime and wake time.²¹ Sleep variables were categorized into “ < 1 hour,” “1–2 hours,” and “ > 2 hours” in reference to a study that examined the relationship of social jetlag (defined as the difference between weekday and weekend sleep midpoints) with obesity status and metabolic parameters.²²

Sleep quality in the previous month was assessed by applying the 19-item Pittsburgh Sleep Quality Index (PSQI) (possible score range = 0–21 points), which has high internal consistency (α = 0.83) and test–retest reliability (r = 0.85).²³ Individuals with poor sleep quality have high scores on the PSQI.²³ The PSQI has been applied among participants as young as 8 years of age^20,24–26 and has been validated among children as young as 14 years of age.²⁷ The sleep quality of the participants was categorized on the basis of global PSQI scores into “good” (PSQI global score ≤5) and “poor” (PSQI global score >5).²³

Anthropometry and Blood Samples

Physical assessments were carried out on location at participating schools or in hospital (Health Care Center, Kiang Wu Hospital) if necessary. Consent was obtained from subjects' parents before the tests. Body heights and weights were measured with a portable stadiometer SECA 214 (SECA GmbH & Co. KG, Germany) and an electronic scale Tanita TBF300 (Tanita Corporation of America, Inc.). Waist circumference (WC) was measured with a measuring tape at the end of gentle expiration while the subjects were in a standing position in accordance with the guidelines established by the World Health Organization (WHO).²⁸ The average values of two readings for each anthropometric measure were calculated.

Trained research staff and qualified phlebotomists collected anthropometric measurements and blood samples, respectively. Serum samples were tested immediately for lipoprotein levels. All samples were analyzed at an accredited central laboratory in Macao (Health Care Center, Kiang Wu Hospital) commissioned by the research team. Details of the laboratory are provided in their official website (www.kwh.org.mo/FYD/index.html).

BMI was calculated from body weight and height measurements. The WHO growth reference was used to identify overweight/obese participants. Overweight was defined as one standard deviation higher than the median of BMI for age (the median BMI value for children of the same age). Obesity was defined as two standard deviations more than the median of BMI for age.²⁹

Population-based cutoff values were used to identify participants with abdominal obesity. Age- and gender-specific cutoff values were adopted from the percentile charts for the WC values of children³⁰ and adolescents³¹ of Chinese ethnicity to identify participants with abdominal obesity.

Serum lipid profiles, including total cholesterol, triglycerides, and low-density lipoprotein, were classified as normal, borderline, or high. High-density-lipoprotein (HDL) cholesterol levels were classified as normal, borderline, or low. The corresponding cutoff points were set in reference to the recommendations of the Expert Panel on Integrated Guidelines for Cardiovascular Health and Risk Reduction in Children and Adolescents.³² Dyslipidemia outcomes were dichotomized into normal versus borderline or high (normal vs. borderline or low for HDL cholesterol).

Statistics

Results were described and summarized through descriptive statistics. Multiple logistic regression was conducted to examine the associations of sleep duration (per 1 hour increment), sleep midpoint (per 1 hour of delay), sleep regularity (bedtime/wake time/sleep duration difference, categorized as <1 hour/1–2 hours/ > 2 hours), and sleep quality (poor/normal) with the risk of elevated body weight (normal vs. overweight/obese), WC (normal vs. elevated), and dyslipidemia (yes vs. no).

Potential confounding variables included age, gender, time spent watching television, time spent on homework, time spent on computers, and eating 1 hour before going to bed (yes/no).^33–35 Food consumption before bedtime is an indicator of night eating, which is associated with an increased risk of obesity.³⁶ Time spent on television, homework, and computers was dichotomized into ≤2 hours/>2 hours in reference to a cross-sectional study that involved children of a similar age (primary school children) and a focus (examining the relationship between compensation for sleep during weekends and the risk of childhood obesity) similar to those involved in the present study.³³ Outcomes related to serum lipid levels were further adjusted on the basis of BMI because this parameter might distort the relationship between sleep and dyslipidemia.⁶

All variables with p < 0.1 were included in univariate analysis (independent t-test for continuous variables and χ² test for categorical variables) to construct the first model (Model 1). Then, the backward selection procedure was used to select the best model (Model 2).³⁷ All analyses were conducted using SPSS version 23.0 (SPSS, Inc., Chicago, IL) with significance levels (α) set at 0.05. The p-values obtained through logistic regression were adjusted by using Bonferroni correction to avoid false-positive findings due to multiple comparisons. Specifically, p-values were multiplied by the total number of comparisons in all regression models. Ethics approval for the survey and physical assessments was received from the Joint CUHK-NTEC Clinical Research Ethics Committee (Reference number: CRE-2014.175).

Results

Among the 1459 primary four to secondary six students recruited in this study, 1078 (73.9%) completed the sleep survey without missing data. The mean (standard deviation) age of participants was 15.35 (3.12) years. The demographic information, lifestyle factors, and sleep data of the participants are summarized in Tables 1 and 2.

Table 1.

Demographic and Lifestyle Factors of the Participating Students in Mean Values (Standard Deviation) or Number (Percentage)

Sleep variables	Overweight or obese		p-value	Abdominal obesity		p-value	≥1 dyslipidemia		p-value
Sleep variables	No (N = 822)	Yes (N = 256)	p-value	No (N = 741)	Yes (N = 337)	p-value	No (N = 513)	Yes (N = 565)	p-value
Age	14.73 (2.63)	13.60 (2.51)	<0.01^*	14.41 (2.71)	14.57 (2.51)	0.34	14.72 (2.62)	14.23 (2.66)	<0.01^*
BMI	NA	NA		NA	NA		19.70 (3.01)	21.00 (4.53)	<0.01^*
Gender
Female	444 (54.0%)	98 (38.3%)	<0.01^*	365 (49.3%)	177 (52.5%)	0.32	242 (47.2%)	300 (53.1%)	0.05
Eating before going to sleep
Yes	325 (39.5%)	73 (28.5%)	<0.01^*	295 (39.8%)	103 (30.6%)	<0.01^*	201 (39.2%)	197 (34.9%)	0.14
Time spent on homework >2 hours
Yes	253 (30.8%)	82 (32.0%)	0.71	230 (31.0%)	105 (31.2%)	0.97	169 (32.9%)	166 (29.4%)	0.21
Time spent on television >2 hours
Yes	292 (35.5%)	106 (41.4%)	0.09	264 (35.6%)	134 (39.8%)	0.19	192 (37.4%)	206 (36.5%)	0.74
Time spent on computer >2 hours
Yes	453 (55.1%)	142 (55.5%)	0.92	403 (54.4%)	192 (57.0%)	0.43	291 (56.7%)	304 (53.8%)	0.34

p < 0.05.

NA, not applicable to the regression analysis.

Table 2.

Sleep Patterns of the Participating Students in Mean Values (Standard Deviation) or Number (Percentage)

Sleep variables	Overweight or obese		p-value	Abdominal obesity		p-value	≥1 dyslipidemia		p-value
Sleep variables	No (N = 822)	Yes (N = 256)	p-value	No (N = 741)	Yes (N = 337)	p-value	No (N = 513)	Yes (N = 565)	p-value
Poor sleep quality
Poor	251 (30.5%)	65 (25.4%)	0.11	214 (28.9%)	102 (30.3%)	0.64	159 (31.0%)	157 (21.8%)	0.25
Bedtime difference
1–2 hours	325 (39.5%)	119 (46.5%)	<0.01^*	292 (39.4%)	152 (45.1%)	0.01^*	214 (41.7%)	230 (40.7%)	0.90
>2 hours	71 (8.6%)	34 (13.3%)		63 (8.5%)	42 (12.5%)		48 (9.4%)	57 (10.1%)
Wake time difference
1–2 hours	201 (24.5%)	73 (28.5%)	0.43	187 (25.2%)	87 (25.8%)	0.41	126 (24.6%)	148 (26.2%)	0.75
>2 hours	443 (53.9%)	130 (50.8%)		387 (52.2%)	186 (55.2%)		273 (53.2%)	300 (53.1%)
Sleep duration difference
1–2 hours	265 (32.2%)	93 (36.3%)	0.07	245 (33.1%)	113 (33.5%)	0.80	160 (31.2%)	198 (35.0%)	0.40
>2 hours	302 (36.7%)	74 (28.9%)		255 (34.4%)	121 (35.9%)		185 (36.1%)	191 (33.8%)
Sleep midpoint difference
1–2 hours	205 (24.9%)	66 (25.9%)	0.28	193 (26.0%)	78 (23.2%)	<0.01^*	119 (23.2%)	152 (27.0%)	0.33
>2 hours	379 (46.1%)	128 (50.2%)		322 (43.5%)	185 (55.1%)		251 (48.9%)	256 (45.4%)
Average sleep midpoint in hours	3:41 (0:54)	3:30 (0:52)	<0.01^*	3:37 (0:52)	3:43 (0:57)	0.08	3:40 (0:54)	3:38 (0:54)	0.54
Average sleep duration in hours	8.37 (1.02)	8.50 (1.11)	0.09	8.42 (1.01)	8.35 (1.10)	0.31	8.36 (1.04)	8.44 (1.04)	0.19

p < 0.05.

Students with BMI-classified obesity were younger, more likely to be male, less likely to eat before sleep, more likely to have a bedtime difference of >2 hours, and had an earlier average sleep midpoint (all p < 0.05) than other students than those without. Students with WC-classified obesity less tended to eat before going to sleep and had bedtime and sleep midpoint difference of >2 hours (all p < 0.05) than those without. Students with dyslipidemia only showed significant differences in terms of young age and high BMI (p < 0.05) than those without.

The results of logistic regression are presented in Table 3. Poor sleep quality, wake time difference, time spent on homework or computer did not show statistically significant associations with any of the outcomes and were excluded from the regression models. A difference of >2 hours between weekday and weekend bedtimes was significantly associated with an increased likelihood of BMI-classified overweight/obesity (Model 1 odds ratio [OR] = 2.58, 95% confidence interval [CI] = 1.55–4.31, p < 0.01). The magnitude of the association between sleep variables and outcomes maintained statistically significant differences in Model 2 (OR: 2.56, 95% CI = 1.57–4.19, p < 0.01). A difference of >2 hours between weekday and weekend sleep duration did not significantly associate with BMI classified. A difference of >2 hours between weekday and weekend sleep midpoints was significantly associated with the increased likelihood of WC-classified overweight/obesity in Model 2 (OR = 1.83, 95% CI = 1.32, 2.52, p = 0.01) but not in Model 1. Average sleep midpoint did not significantly associate with BMI-classified nor WC-classified overweight/obesity. Average sleep duration did not have significant association with BMI-classified overweight/obesity.

Table 3.

Adjusted Logistic Regression Model Results for the Associations between Sleep Variables, Body Mass Index- or Waist Circumference-Classified Obesity, and Dyslipidemia

Sleep variables	Overweight or obese		Abdominal obesity		≥1 dyslipidemia
			Odds ratio (95% CI)
	Model 1^a	Model 2^a	Model 1^a	Model 2^a	Model 1^a	Model 2^a
Poor sleep quality
Non-poor	—	—	—	—	—	—
Poor
p value
Bedtime difference
<1 hour	Ref	Ref	Ref	—	—	—
1–2 hours	1.74 (1.26–2.40)	1.75 (1.28–2.40)	1.15 (0.81–1.63)
>2 hours	2.58 (1.55–4.31)	2.56 (1.57–4.19)	1.42 (0.84–2.39)
p value	<0.01^*	<0.01^*	1.00
Wake time difference
<1 hour	—	—	—	—	—	—
1–2 hours
>2 hours
p value
Sleep duration difference
<1 hour	Ref	—	—	—	—	—
1–2 hours	1.16 (0.81–1.68)
>2 hours	1.12 (0.74–1.70)
p value	1.00
Sleep midpoint difference
<1 hour	—	—	Ref	Ref	—	—
1–2 hours			1.18 (0.80–1.74)	1.24 (0.86–1.81)
>2 hours			1.51 (0.98–2.35)	1.83 (1.32–2.52)
p value			1.00	0.01^*
Average sleep midpoint^b	0.93 (0.74–1.16)	—	1.05 (0.89–1.24)	—	—	—
p value	1.00		1.00
Average sleep duration^b	0.89 (0.76–1.05)	—	—	—	—	—
p value	1.00
Age	0.82 (0.77–0.88)	0.83 (0.78–0.88)	—	—	0.89 (0.84–0.93)	0.89 (0.84–0.93)
p value	<0.01^*	<0.01^*			<0.01^*	<0.01^*
Gender
Male	Ref	Ref	—	—	Ref	Ref
Female	0.50 (0.37–0.67)	0.50 (0.37–0.68)			1.39 (1.09–1.78)	1.40 (1.09–1.79)
p value	<0.01^*	<0.01^*			0.22	0.20
Eat before sleep
No	Ref	Ref	Ref	Ref	—	—
Yes	0.52 (0.38–0.72)	0.52 (0.37–0.71)	0.63 (0.47–0.83)	0.64 (0.48–0.84)
p value	<0.01^*	<0.01^*	0.02^*	0.04^*
Time spent on homework
≤2 hours	—	—	—	—	—	—
>2 hours
p value
Time spent on television
≤2 hours	Ref	—	—	—	—	—
>2 hours	1.17 (0.86–1.58)
p value	1.00
Time spent on computer
≤2 hours	—	—	—	—	—	—
>2 hours
p value
BMI	NA	NA	NA	NA	1.13 (1.09–1.17)	1.13 (1.09–1.17)
p value					<0.01^*	<0.01^*

Model 1: included all variables, which had p-value <0.1 in the univariate analysis Model 2: applied backward procedure based on Model 1. For the regression of ≥1 dyslipidemia, it was further adjusted for BMI. Significance differences with p < 0.05 are marked with asterisk.

Per 1-hour increment.

—, Not included in the regression model.

BMI, body mass index; NA, not applicable to the regression analysis.

Discussion

To the best of our knowledge, this work is one of the few studies that examined the roles of sleep regularity and quality in obesity and dyslipidemia among school students in the Chinese community. The results obtained after adjusting for age, gender, screening time, and sleep duration suggest that considerable differences between weekday and weekend bedtimes (>2 hours) were associated with the increased likelihood of elevated BMI (OR: 2.58). We included all variables with p < 0.1 in univariate analysis to fit the regression model (Model 1) or used the backward selection procedure in the selection of the best model (Model 2) to ensure the robustness of our results. Then, we adjusted p-values by using Bonferroni correction. The statistical significances of most results, except of those for sleep midpoint difference and WC-classified obesity, were consistent in both models. The significant association between irregular sleep midpoint and WC-classified obesity likely cancelled out because Model 1 failed to exclude related sleep variables that lacked statistical significance. These variables included average sleep midpoint and bedtime difference.

The role of bedtime difference in the etiology of obesity identified in this study was consistent with that found in a previous study among Hong Kong adolescents.¹¹ Irregular bedtimes may be related to circadian disruption, which is prevalent among night-shift workers.³⁸ In addition to sleep deprivation, circadian desynchronization can increase ghrelin levels; this increment, in turn, increases an individual's appetite for energy-dense foods and risk for obesity.³⁹ This explanation is supported by the results of a recent meta-analyses demonstrating that night-shift workers are at high risk of obesity-related diseases, including diabetes⁴⁰ and metabolic syndrome.⁴¹

In the present study, we found that irregular bedtimes may increase the likelihood of obesity before adulthood. Irregular bedtime is also known as social jetlag, a phenomenon that is associated with obesity and metabolic syndromes.²² Nevertheless, few studies have examined the association of social jetlag with abdominal obesity and dyslipidemia.^42–44 This association can be translated to the public health message espoused by the National Sleep Foundation, that is, stick to a sleep schedule of the same bedtime and wake time even on the weekends.⁴⁵ Maintaining regular sleep timing is a major goal for sleep hygiene interventions.⁴⁶ Our research provides guidance for studies aiming to determine whether a regular bedtime improves metabolic health.

Moreover, we examined how sleep timing, that is, the regularity of sleep quality, may influence the risk of dyslipidemia. In this study, we found no statistically significant associations between sleep timing and dyslipidemia risk. However, epidemiological evidence for the relationship between sleep and cardiometabolic risk factors has been inconsistent. Several studies conducted among children and adolescents failed to demonstrate any association between sleep parameters (duration, regularity, or disturbance) and serum lipid levels,^47–50 and one investigation showed that the association between late bedtimes and HDL was attenuated after adjusting for adiposity.⁵¹ Sleep durations of >8 hours were positively associated with elevated total cholesterol (TC) levels among Iranian boys⁵² but were negatively associated with TC levels among Chinese adolescents.⁵³

This discrepancy may be attributed to the adjustment for different confounders and different designs applied in various studies and cannot fully explain the inconsistent relationship between sleep and dyslipidemia. Most previous works were cross-sectional in nature and did not establish the temporal relationship between sleep and dyslipidemia.

Moreover, only a few studies have adjusted for diets and physical activities despite the importance of these two factors as confounders.^54,55 Only one previous work was a longitudinal study.⁵⁶ One investigation demonstrated after confounder adjustment that long sleep duration was associated with a high risk of elevated triglyceride levels among Korean adolescents.⁵⁵ No statistically significant association between long sleep durations and elevated triglyceride levels, however, was found among Canadian adolescents.⁵⁴ A longitudinal study⁵⁶ showed that long sleep durations at <3 years of age was predictive of low HDL at <6 years of age. Previous research revealed inconsistencies among associations between sleep and serum lipid levels despite adjusting for important confounders and/or using prospective study designs, and the results of previous studies might even contradict biological mechanisms of sleep and lipid levels. However, prospective studies on the association between sleep regularity/quality and serum lipid levels in children and adolescents and that included diet and physical activity as confounders remain rare, and their findings must be substantiated by additional investigations.

Delaying the start of school time is a simple but potentially effective policy intervention for promoting healthy sleep among adolescents. In 2014, the Adolescent Sleep Working Group published a policy statement advocating delaying school start times to prevent chronic sleep loss.⁵⁷ Adolescents may experience a shift to late bedtimes due to physiological changes, such as the delayed nocturnal secretion of melatonin⁵⁸ and alterations in brain processes for sleep initiation,⁵⁹ during puberty. Thus, standardizing school starting times across different grade levels may ignore the special needs of adolescents and result in sleep deprivation and poor health status among children.⁶⁰ Intervention studies performed in different parts of the world have shown that delaying school start time can be beneficial. A systematic review involving six intervention studies revealed that delaying school start times by 25–60 minutes can increase sleep times by 25–77 minutes per night⁶¹ and improve daytime sleepiness and depressive symptoms and reduce caffeine usage. In addition, the bedtimes of students remained unchanged despite late waking times on the following day.⁶¹ After the review was published, researchers from New York delayed the start times of high schools by 45 minutes and observed increased sleep durations with regular bedtimes among students.⁶²

A recent study in Hong Kong found that delaying school start time by 15 minutes not only increased time in bed but also improved mental health and prosocial behaviors among students.⁶³ Furthermore, among 375 Singaporean female adolescents, a 45 minutes delay in school start time increased the total time in bed by 23.2 minutes, delayed sleep time by 9 minutes, and reduced sleepiness and improved emotional health.⁶⁴ However, given that delaying the school start time may reduce teaching time, delaying school start times in the Chinese educational system, which is highly competitive from the start of primary school and emphasizes academic performance and intolerance of failure, will be difficult.⁶⁵

The present study has several limitations that must be addressed, and its results must be interpreted with caution. As mentioned above, we could not elucidate the causal relationship between exposure and outcomes given the cross-sectional nature of our study. Second, we did not assess sleep patterns through objective methods, such as polysomnography. Although sleep patterns in the present population were assessed by applying the PSQI, which is a well-established and appropriate tool, the validity of our study will be improved if objective assessment can be conducted simultaneously. Validation studies involving individuals younger than 14 years of age are thus necessary.

Third, we did not assess daytime sleep habits (e.g., napping), which may also contribute to the risk of obesity. Fourth, we were not able to adjust for potential confounders, including dietary patterns, physical activities, screen time before bedtime, pubertal status, mental distress, snoring status, family socioeconomic status, and school-level clustering. Residual confounding effects may account for some of the present findings. Nevertheless, the direction of associations between sleep and obesity identified in the present work agree with those presented in the literature despite the limitations in its design and cofounder adjustment.^6,66 Cohort studies provided inconsistent results for the relationship between sleep and serum lipid levels even after adjusting for diets and physical activities. Therefore, the inverse association between sleep time and dyslipidemia found in the present study is not necessarily a flaw. Nevertheless, the mediating effect of diet and physical activity on the association between sleep and obesity still warrants further investigation.

Conclusion

Considerable discrepancies between weekday and weekend bedtimes are associated with an increased risk of overweight/obesity and abdominal obesity among children and adolescents regardless of sleep duration, bedtime, and sleep quality. Regular daily bedtimes are recommended as a simple but practical strategy for preventing obesity. Large-scale prospective studies with comprehensive confounder assessment are needed to examine the causal relationship between sleep and cardiometabolic risk, particularly the association between sleep quality and dyslipidemia.

Footnotes

Acknowledgments

The authors would also like to thank all the study schools, families, and students for their cooperation and participation.

The study was supported by the Education and Youth Affairs Bureau of Macao Special Administrative Region Government (Ref No: 383/2014).

Author Disclosure Statement

No competing financial interests exist.

References

Rokholm

, Baker

, Sorensen

TIA

. The levelling off of the obesity epidemic since the year 1999-a review of evidence and perspectives. Obes Rev, 2010; 11:835–846.

, Han

, Chu

, et al. Trends in Overweight and Obesity among Children and Adolescents in China from 1981 to 2010: A Meta-Analysis. PLoS One, 2012; 7:e51949.

Cappuccio

, Taggart

, Kandala

, et al. Meta-analysis of short sleep duration and obesity in children and adults. Sleep, 2008; 31:619–626.

Taheri

. The link between short sleep duration and obesity: We should recommend more sleep to prevent obesity. Arch Dis Child, 2006; 91:881–884.

Zheng

, Berthoud

. Neural systems controlling the drive to eat: Mind versus metabolism. Physiology, 2008; 23:75–83.

Quist

, Sjodin

, Chaput

, Hjorth

. Sleep and cardiometabolic risk in children and adolescents. Sleep Med Rev, 2015; 29:76–100.

Nielsen

, Danielsen

, Sorensen

TIA

. Short sleep duration as a possible cause of obesity: Critical analysis of the epidemiological evidence. Obes Rev, 2011; 12:78–92.

Pesonen

, Raikkonen

, Paavonen

, et al. Sleep Duration and Regularity are Associated with Behavioral Problems in 8-year-old Children. Int J Behav Med, 2011; 17:298–305.

Spruyt

, Molfese

, Gozal

. Sleep duration, sleep regularity, body weight, and metabolic homeostasis in school-aged children. Pediatrics, 2011; 127:E345–E352.

10.

Kjeldsen

, Hjorth

, Andersen

, 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; 38:32–39.

11.

Chung

, Kan

KKK

, Yeung

. Sleep duration, sleep–wake schedule regularity, and body weight in Hong Kong Chinese adolescents. Biol Rhythm Res, 2013; 44:169–179.

12.

Chuang

, Fehr

, Ievers-Landis

, et al. Associations of sleep duration and regularity with level of obesity among youth in a weight loss program. Transl Issues Psychol Sci, 2015; 1:45.

13.

Krystal

, Edinger

. Measuring sleep quality. Sleep Med, 2008; 9:S10–S17.

14.

Rosipal

, Lewandowski

, Dorffner

. In search of objective components for sleep quality indexing in normal sleep. Biol Psychol, 2013; 94:210–220.

15.

Hung

, Yang

, Ou

, et al. The association between self-reported sleep quality and overweight in a Chinese population. Obesity, 2013; 21:486–492.

16.

Bidulescu

, Din-Dzietham

, Coverson

, et al. Interaction of sleep quality and psychosocial stress on obesity in African Americans: The Cardiovascular Health Epidemiology Study (CHES). BMC Public Health, 2010; 10:581.

17.

Jennings

, Muldoon

, Hall

, et al. Self-reported sleep quality is associated with the metabolic syndrome. Sleep, 2007; 30:219–223.

18.

Hung

, Yang

, Ou

, et al. The association between self-reported sleep quality and metabolic syndrome. PLoS One, 2013; 8:e54304.

19.

Education-and-Youth-Affairs-Bureau. Overview of Education in Macao 2013/2014. Education-and-Youth-Affairs-Bureau: Macao, 2014.

20.

Tan

. The effects of background music on quality of sleep in elementary school children. J Music Ther, 2004; 41:128–150.

21.

Sato-Mito

, Sasaki

, Murakami

, et al. The midpoint of sleep is associated with dietary intake and dietary behavior among young Japanese women. Sleep Med, 2011; 12:289–294.

22.

Mota

, Silva

, Balieiro

LCT

, et al. Social jetlag and metabolic control in non-communicable chronic diseases: a study addressing different obesity statuses. Sci Rep, 2017; 7:6358.

23.

Buysse

, Reynolds

, Monk

, et al. The Pittsburgh sleep quality index - a new instrument for psychiatric practice and research. Psychiat Res, 1989; 28:193–213.

24.

Ayaki

, Torii

, Tsubota

, Negishi

. Decreased sleep quality in high myopia children. Sci Rep, 2016; 6:33902.

25.

Benhayon

, Youk

, McCarthy

, et al. Characterization of relations among sleep, inflammation, and psychiatric dysfunction in depressed youth with Crohn disease. J Pediatr Gastroenterol Nutr, 2013; 57:335–342.

26.

Uyan

, Turan

, Ay

, et al. Sleep disordered breathing and sleep quality in children with bronchiolitis obliterans. Pediatr Pulmonol, 2016; 51:308–315.

27.

de la Vega

, Tome-Pires

, Sole

, et al. The Pittsburgh Sleep Quality Index: Validity and factor structure in young people. Psychol Assess, 2015; 27:e22–e27.

28.

WHO. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert Consultation, Geneva, 8–11 December 2008. WHO: Geneva, 2011.

29.

de Onis

, Onyango

, Borghi

, et al. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ, 2007; 85:660–667.

30.

Sung

RYT

, Yu

CCW

, Choi

, et al. Waist circumference and body mass index in Chinese children: cutoff values for predicting cardiovascular risk factors. Int J Obes (Lond), 2007; 31:550–558.

31.

VWS

, Kong

APS

, Choi

, et al. BMI and waist circumference in predicting cardiovascular risk factor clustering in Chinese adolescents. Obesity, 2007; 15:494–503.

32.

Expert Panel on Integrated Guidelines for Cardiovascular

, Risk Reduction in

, Adolescents, National Heart

, Blood

. Expert panel on integrated guidelines for cardiovascular health and risk reduction in children and adolescents: Summary report. Pediatrics, 2011; 128 Suppl 5:S213–S256.

33.

Wing

, Li

, et al. The effect of weekend and holiday sleep compensation on childhood overweight and obesity. Pediatrics, 2009; 124:E994–E1000.

34.

Fatima

, Doi

, Mamun

. Longitudinal impact of sleep on overweight and obesity in children and adolescents: A systematic review and bias-adjusted meta-analysis. Obes Rev, 2015; 16:137–149.

35.

Mitchell

, Rodriguez

, Schmitz

, Audrain-McGovern

. Greater Screen Time is Associated with Adolescent Obesity: A Longitudinal Study of the BMI Distribution from Ages 14 to 18. Obesity, 2013; 21:572–575.

36.

Lee

, Kang

, Yang

, et al. Study Time after School and Habitual Eating Are Associated with Risk for Obesity among Overweight Korean Children: A Prospective Study. Obes Facts, 2018; 11:46–55.

37.

Leech

, Barrett

, Morgan

. IBM SPSS for Intermediate Statistics: Use and Interpretation (5th edition). Routledge, NewYork: NY, 2015.

38.

Costa

. The impact of shift and night work on health. Appl Ergon, 1996; 27:9–16.

39.

Qian

, Morris

, Caputo

, et al. Ghrelin is impacted by the endogenous circadian system and by circadian misalignment in humans. Int J Obes (Lond) 2018. [Epub ahead of print; DOI: 10.1038/s41366-018-0208-9.]

40.

Gan

, Yang

, Tong

, et al. Shift work and diabetes mellitus: A meta-analysis of observational studies. Occup Environ Med, 2015; 72:72-U91.

41.

Wang

, Zhang

, et al. Meta-analysis on night shift work and risk of metabolic syndrome. Obes Rev, 2014; 15:709–720.

42.

Arora

, Taheri

. Associations among late chronotype, body mass index and dietary behaviors in young adolescents. Int J Obes (Lond), 2015; 39:39–44.

43.

Parsons

, Moffitt

, Gregory

, et al. Social jetlag, obesity and metabolic disorder: Investigation in a cohort study. Int J Obes (Lond), 2015; 39:842–848.

44.

de Zwart

, Beulens

, Elders

, Rutters

. Pilot data on the association between social jetlag and obesity-related characteristics in Dutch adolescents over 1 year. Sleep Med, 2018; 47:32–35.

45.

National-Sleep-Foundation. Healthy Sleep Tips [cited November 16, 2018]. Available at https://sleepfoundation.org/excessivesleepiness/sleep-tools-tips/healthy-sleep-tips

46.

Irish

, Kline

, Gunn

, et al. The role of sleep hygiene in promoting public health: A review of empirical evidence. Sleep Med Rev, 2015; 22:23–36.

47.

Flint

, Kothare

, Zihlif

, et al. Association between inadequate sleep and insulin resistance in obese children. J Pediatr, 2007; 150:364–369.

48.

Cespedes

, Rifas-Shiman

, Redline

, et al. Longitudinal associations of sleep curtailment with metabolic risk in mid-childhood. Obesity, 2014; 22:2586–2592.

49.

Hitze

, Bosy-Westphal

, Bielfeldt

, et al. Determinants and impact of sleep duration in children and adolescents: Data of the Kiel Obesity Prevention Study. Eur J Clin Nutr, 2009; 63:739–746.

50.

Rey-Lopez

, de Carvalho

, de Moraes

, et al. Sleep time and cardiovascular risk factors in adolescents: The HELENA (Healthy Lifestyle in Europe by Nutrition in Adolescence) study. Sleep Med, 2014; 15:104–110.

51.

Berentzen

, Smit

, Bekkers

, et al. Time in bed, sleep quality and associations with cardiometabolic markers in children: The Prevention and Incidence of Asthma and Mite Allergy birth cohort study. J Sleep Res, 2014; 23:3–12.

52.

Azadbakht

, Kelishadi

, Khodarahmi

, et al. The association of sleep duration and cardiometabolic risk factors in a national sample of children and adolescents: The CASPIAN III study. Nutrition, 2013; 29:1133–1141.

53.

Kong

, Wing

, Choi

, et al. Associations of sleep duration with obesity and serum lipid profile in children and adolescents. Sleep Med, 2011; 12:659–665.

54.

Narang

, Manlhiot

, Davies-Shaw

, et al. Sleep disturbance and cardiovascular risk in adolescents. Can Med Assoc J, 2012; 184:E913–E920.

55.

Lee

, Park

. Relation between sleep duration, overweight, and metabolic syndrome in Korean adolescents. Nutr Metab Cardiovasc Dis, 2014; 24:65–71.

56.

Plumptre

, Anderson

, Chen

, et al. Longitudinal analysis of sleep duration and cardiometabolic risk in young children. Child Obes, 2017; 13:291–299.

57.

Owens

, Allison

, et al. School start times for adolescents. Pediatrics, 2014; 134:642–649.

58.

Carskadon

, Acebo

, Jenni

. Regulation of adolescent sleep - Implications for behavior. Ann NY Acad Sci, 2004; 1021:276–291.

59.

Jenni

, Achermann

, Carskadon

. Homeostatic sleep regulation in adolescents. Sleep, 2005; 28:1446–1454.

60.

Chen

, Wang

, Jeng

. Adequate sleep among adolescents is positively associated with health status and health-related behaviors. BMC Public Health, 2006; 6:59.

61.

Minges

, Redeker

. Delayed school start times and adolescent sleep: A systematic review of the experimental evidence. Sleep Med Rev, 2016; 28:86–95.

62.

Thacher

, Onyper

. Longitudinal outcomes of start time delay on sleep, behavior, and achievement in high school. Sleep, 2016; 39:271–281.

63.

Chan

, Zhang

, Yu

MWM

, et al. Impact of a modest delay in school start time in Hong Kong school adolescents. Sleep Med, 2017; 30:164–170.

64.

, Lee

, et al. Sustained benefits of delaying school start time on adolescent sleep and well-being. Sleep, 2018; 41:1–8.

65.

Hesketh

, Zhen

, Lu

, et al. Stress and psychosomatic symptoms in Chinese school children: Cross-sectional survey. Arch Dis Child, 2010; 95:136–140.

66.

Fobian

, Elliott

, Louie

. A systematic review of sleep, hypertension, and cardiovascular risk in children and adolescents. Curr Hypertens Rep, 2018; 20:42.