Examination of Sleep and Obesity in Children and Adolescents in the United States

Abstract

Purpose:

This study contributes to the growing literature on the association between sleep and obesity by examining the associations between hours of sleep, consistency of bedtime, and obesity among children in the US.

Design:

Analysis of a nationally representative sample of non-institutionalized children from the 2016-17 National Survey of Children’s Health.

Setting:

US, national.

Subjects:

Children ages 10-17 years (n = 34,640)

Measures:

Parent reported weeknight average hours of sleep and consistency of bedtime. Body mass index classified as underweight, normal, overweight or obesity using parent-reported child height and weight information, classified using CDC BMI-for-Age Growth Charts.

Analysis:

Multivariate logistic regression models were used to estimate associations between measures of sleep and body mass index weight category adjusting for individual, household and neighborhood characteristics.

Results:

An additional hour of sleep was associated with 10.8% lower odds of obesity, net of consistency in bedtime. After controlling for sleep duration, children who usually went to bed at the same time on weeknights had lower odds of obesity (24.8%) relative to children who always went to bed at the same time.

Conclusion:

Sleep duration is predictive of lower odds of obesity in US children and adolescents. Some variability in weeknight bedtime is associated with lower odds of obesity, though there were no additional benefits to extensive variability in bedtime.

Keywords

sleep duration bedtime obesity children adolescents BMI

Purpose

Addressing childhood obesity is a major public health priority in the US.¹ In 2018, approximately a fifth of US children (18.4%) and adolescents (20.6%) were clinically obese, with obesity defined as a body mass index (BMI) at or higher than the 95th percentile of age and sex specific BMI.¹ Obesity during childhood is associated with chronic diseases such as type 2 diabetes, asthma, and sleep apnea.^2-4 Obesity is also associated with poor mental health and problems at school.^5,6 Research exploring the association between sleep and obesity in children shows that poor sleep is an important predictor of obesity in children and adolescents.^7-12 A controlled experimental study that iteratively assigned children to be in bed for 1.5 hours more or less than their typical amount found that, when assigned to spend longer time in bed, children reported lower food consumption and had lower weight.¹³ These findings suggest that improving sleep could be a means to reduce obesity.

There are several mechanisms through which sleep and body weight may be related. Poor sleep could lead to higher energy intake.⁷ Studies on adolescents show an association between inadequate sleep and desire for snacks, especially high-glycemic, high-calorie foods and drinks.^13-16 Increases in consumption may be explained in part by reduced inhibitory control¹⁷ and emotional volatility in sleep-deprived adolescents.¹⁸ Other researchers hypothesize that adolescents who miss or skip breakfast, perhaps in part because of disrupted sleep, may consume more calories later in the day than they would have if they spread their food consumption evenly throughout the day.¹⁹ Poor glucose regulation, known as insulin resistance, in adolescents who experience poor sleep,²⁰ may also increase risk of obesity by encouraging food intake.²¹ Because adolescents are generally prone to higher risk of developing insulin resistance during puberty,²² the risk of obesity due to insulin resistance is exacerbated by poor sleep. In addition to higher energy intake, poor sleep may increase risk of obesity through reductions in energy expenditure;⁷ however, there are few studies about poor sleep and reduced energy expenditure, and the findings are inconsistent.^7-12

Children in the US are not meeting recommended sleep guidelines.²³ The National Sleep Foundation recommends that children aged 6 to 13 years should have between 9 and 11 hours of sleep per night, no less than 7 hours, and no more than 12 hours.²⁴ Adolescents between the ages of 14 and 17 require 8 to 10 hours of sleep per night, and should not get less than 7 hours or more than 11 hours.²⁴ A meta-analysis of US-based studies found that children between 9 and 11 years of age slept on average 8 hours; adolescents between 12 and 18 years of age slept on average 7 hours.²³ Given the simultaneous prevalence of obesity and short sleep duration among children in the US, the 2 could be linked. While much of the literature exploring the associations between sleep and obesity in children study samples of American children, US population-level estimates have received little attention.^25,26

This study presents population-level estimates of the association between sleep and obesity among children based on a representative sample of non-institutionalized children ages 10-17 in the United States collected as part of the 2016-17 National Survey of Children’s Health (NSCH). This analytic sample is about 3 times larger than the samples used in previous studies. The NSCH includes detailed information about children’s sleep patterns, including both sleep duration and bedtime consistency. Earlier literature primarily focused on duration of sleep, but sleep quality and bedtime consistency or disturbances may also be important.^8,27,28 With the NSCH, we explore the associations of multiple measure of sleep with obesity. Lastly, NSCH also includes information on neighborhood characteristics, which researchers have linked to variation in sleep among children;²⁹ neighborhood characteristics were not accounted for in previous research on sleep and obesity, but we include them in this study.

Methods

Survey Design and Sample

The data were from the 2016-2017 NSCH, made available by the Health Resources and Services Administration’s Maternal and Child Health Bureau (HRSA MCHB) and the US Census Bureau.³⁰ The annual survey collects information on a representative sample of non-institutionalized children between infancy and 17 years of age in the United States. The NSCH draws a sample of households from the Census Master Address File across the 50 states and the District of Columbia. Household selection is stratified by state, child presence and neighborhood poverty. Child presence was evaluated using an indicator constructed from administrative data by the Census Bureau. In multiple child households, one child was selected randomly per household;³¹ however, 2 oversamples were also applied in households with multiple children to ensure adequate representation of children with special health care needs and very young children, both of which are priority populations for HRSA MCHB.³¹

Respondents were parents or guardians who lived in the household and were familiar with the health and healthcare needs of the selected child. They were asked about the child’s physical and mental health, insurance coverage, access to healthcare, and family, school, and neighborhood characteristics. The surveys were administered in English or Spanish via email, telephone, or mailed paper copy; additional languages were available by telephone.³¹ Protection of participant data was in the jurisdiction of the U.S. Census Bureau and Office of Management and Budget. Secondary analysis of publicly available de-identified dataset did not require IRB review.

Measures

Children’s BMI was calculated using parent-reported information about their child’s weight and height. CDC BMI-for-Age Growth Charts were used to classify children into weight categories: Children with BMI at the 5th percentile and below for their age and sex were classified as underweight; those between the 5th and 84th percentile were classified as normal weight; those between the 85th to 94th percentile were classified as overweight and those with BMI greater than or equal to the 95th percentile were classified as having obesity.³² For analysis, we also created dichotomous indicator of obesity.

Parents were asked about the child’s average hours of sleep per day during the previous week: less than 6, 6, 7, 8, 9,10, and, 11 or more hours. Children with less than 6 hours or 11 or more hours of sleep were recoded as 6 and 11 hours respectively to create a continuous measure of sleep to allow for comparison with existing literature.²⁶ We explored the sensitivity of our analysis to categorical measures of sleep duration. Parents were also asked how often their child goes to bed at the same time on weeknights: always, usually, sometimes, and rarely or never goes to bed about the same time.

Multivariate analyses were used to account for possible confounders, all based on parent’s reports. Demographic covariates were child sex (male reference), race and ethnicity (Hispanic, Black non-Hispanic or multiracial/other non-Hispanic, White non-Hispanic reference) and age (in years). Since children’s daytime activities, including active time, sedentary time, and electronic devise use may be correlated with sleep and body weight,^7,33-36 we controlled for days active for at least 60 minutes per week (1-3, 4-6, or 7 days, 0 days reference), hours of television watched per day (1, 2-3 or 4 or more hours, 0 hours reference) and hours of screen time other than television per day (1, 2-3, 4 or more hours, 0 hours reference). Socioeconomic variables were included as they may be linked with both obesity and sleep: household income as a percent of the Federal Poverty Line (100-199%, 200-399%, 400% or above, 0-99% reference) and food insufficiency (“we could always afford enough to eat but not always the kinds of foods we should eat”, “sometimes we could not afford enough to eat” and “often we could not afford enough to eat”, “we could always afford to eat good nutritious meals” reference). We accounted for parent–reported neighborhood characteristics: safe neighborhood (“somewhat agree”, and “somewhat or definitely disagree”, “definitely agree” reference), presence of sidewalks (yes reference), and presence of a park (yes reference).

Analytical Sample

Because BMI information was only available for children 10-17 years of age in the NSCH, we included only children ages 10 years and older (n = 37,409).³¹ Among these, 7.4% were missing BMI information and 1.7% were missing sleep duration information, and these were excluded from analysis, leaving an analytical sample of 34,640 children. A small number of children (1-3%) were missing information on covariates, and we retained these in the analyses using missing variable adjustment, a method that avoids dropping observations with missing information and offers an empirical strategy to account for non-random missing values across observations.³⁷ In Table 1, we show prevalence of missing values. We conducted sensitivity analyses using listwise deletion.

Table 1.

Descriptive Statistics for Children 10-17 in the United States.^a

Hours of Sleep, daily		8.187	(8.159,8.214)
BMI	Underweight	0.062	(0.056,0.067)
	Normal	0.629	(0.616,0.641)
	Overweight	0.152	(0.142,0.161)
	Obesity	0.158	(0.148,0.169)
Goes to bed at the same time, days	Always	0.287	(0.275,0.299)
Goes to bed at the same time, days	Usually	0.563	(0.551,0.576)
	Sometimes	0.100	(0.093,0.107)
	Rarely/never	0.046	(0.041,0.052)
	Missing	0.003	(0.002,0.004)
Age		13.54	(13.479,13.593)
Sex	Male	0.511	(0.499,0.524)
	Female	0.489	(0.476,0.501)
Race	White, non-Hispanic	0.525	(0.512,0.538)
	Hispanic	0.241	(0.226,0.255)
	Black, non-Hispanic	0.136	(0.126,0.146)
	Multi-racial/other, non-Hispanic	0.098	(0.092,0.104)
Physical Activity, days	0 days	0.109	(0.101,0.116)
	1 to 3 days	0.398	(0.385,0.410)
	4 to 6 days	0.298	(0.286,0.309)
	Everyday	0.192	(0.182,0.203)
	Missing	0.004	(0.003,0.005)
Electronic devices, hours	Do not use electronic devices	0.029	(0.024,0.034)
	Use less than 1 hour per day	0.108	(0.100,0.116)
	Use 1 hour per day	0.208	(0.198,0.219)
	Use 2 to 3 hours per day	0.448	(0.436,0.461)
	Use 4 or more hours per day	0.200	(0.191,0.210)
	Missing	0.006	(0.003,0.008)
TV, hours	Do not watch TV	0.047	(0.042,0.051)
	Less than 1 hour per day	0.155	(0.147,0.164)
	1 hour per day	0.250	(0.239,0.261)
	2 to 3 hours per day	0.434	(0.421,0.447)
	4 or more hours per day	0.108	(0.100,0.116)
	Missing	0.006	(0.003,0.008)
Poverty level, household	0% to 99% FPL	0.192	(0.180,0.204)
	100% to 199% FPL	0.209	(0.198,0.221)
	200% to 399% FPL	0.273	(0.262,0.284)
	400% FPL or Above	0.326	(0.315,0.336)
Household food insufficiency	Always afford to eat good nutritious meals	0.657	(0.644,0.670)
	Always afford enough to eat but not always the kinds of foods we should eat	0.260	(0.248,0.272)
	Sometimes we could not afford enough to eat	0.056	(0.049,0.064)
	Often we could not afford enough to eat	0.013	(0.009,0.016)
	Missing	0.013	(0.010,0.017)
Neighborhood safe	Definitely agree	0.656	(0.643,0.669)
	Somewhat agree	0.280	(0.268,0.292)
	Somewhat or definitely disagree	0.050	(0.043,0.057)
	Missing	0.014	(0.011,0.017)
Neighborhood sidewalks	Yes	0.727	(0.716,0.738)
	No	0.261	(0.251,0.272)
	Missing	0.012	(0.009,0.015)
Neighborhood park	Yes	0.730	(0.719,0.741)
	No	0.256	(0.245,0.267)
	Missing	0.014	(0.011,0.017)

Notes: Estimates are proportions, except for sleep and age which are means. CI in brackets. Source: Data from The National Survey of Children’s Health (NSCH) 2016 and 2017 reported from the child’s parents or primary caregiver. Estimates reflect the US population of non-institutionalized children aged 10-17 in 2016-17 with no missing BMI or sleep duration data. n =34,640. Household selection in the NSCH is stratified first by state and whether there was a child present and then by in terms of neighborhood poverty. One child is randomly selected per household. Estimates reflect parent or primary caregiver reported data including height and weight information used to calculate BMI.

Statistical Methods

All analysis was survey-adjusted using weights, primary sampling unit and strata variables included in the publicly available NSCH dataset and are thus estimates representative of the US population of non-institutionalized children aged 10-17 in 2016-17. The NSCH primary sampling unit is the household. Within each state, households were randomly selected and then stratified by likelihood of child presence and whether households were located in block groups with poverty rates at or below 30%, or above 30%.³¹ We reported descriptive statistics for our analytical sample and, additionally explored unadjusted differences in average sleep duration by BMI category as well as the bivariate relationship between sleep duration and consistency of weeknight bedtime.

We used unordered multinomial logistic regression to estimate the odds of underweight, overweight, or obesity, relative to normal weight, associated with sleep hours and bedtime consistency. We estimated associations between sleep measures and weight status net of variation in physical and sedentary activity, as well as demographic, household and neighborhood characteristics. The benefit of multinomial logit models is that they produce a series of pairwise odd ratios between each BMI category with normal weight that allows us to explore potential differential associations between BMI category and measures of sleep. As a sensitivity analysis, we also estimated binomial logit models with obesity as the outcome variable. All analyses were completed in Stata version 16.1 using svy commands with the subpop option to focus on children and adolescents 10-17 years of age who comprise our analytical sample.

Results

Descriptive Results

Table 1 shows survey-adjusted means and proportions for children and teens aged 10-17 in the United States. Children with obesity accounted for 15.8% of the population, another 15.2% of the children were overweight, 62.9% were normal weight and 6.2% of the children were underweight. Table 1 also reports population estimates for sleep duration and bedtime consistency. On average, children slept 8.2 hours per night. 28.7% of parents reported that their children always go to sleep at the same time, 56.3% reported that they usually do, 10.0% that they sometimes do, and 4.6% that they rarely or never do. Almost half of children had at least 60 minutes of physical activity 4 or more days per week (49.0%), while approximately 40% did only 1 to 3 days per week. Children most commonly spent 2-3 hours per day watching television.

Consistent with the composition of the US population, 52.5% of children were non-Hispanic White, 24.1% Hispanic, 13.6% non-Hispanic Black and 9.8% other groups. The majority of parents reported that their families have enough money to eat good nutritious meals (65.7%), but 1 in 4 (26.0%) reported that they did not always afford the kind of food they should eat. About a fifth of children (19.2%) lived in households earning below the Federal Poverty Line. The majority of the children lived in neighborhoods that their parents deem to be safe (65.6%), that had sidewalks (72.7%) and that had parks (73.0%).

Figure 1 shows unadjusted average hours slept broken down by children’s BMI category. Children with underweight BMI slept on average the longest (8.5 hours) and children with obesity slept on average the shortest (8.1 hours); however, hours slept did not decrease linearly with BMI category, as overweight children slept on average more than children with normal weight (8.3 versus 8.2 hours). Figure 2 shows the unadjusted bivariate relationship between bedtime consistency and hours of sleep. There was a negative relationship between level of variation in bedtime and hours of sleep; that is, children who always went to bed at the same time more frequently also slept longer.

Figure 1.

Hours of Sleep by BMI Category for Children 10-17 in the United States, 2016-17.

Figure 2.

Hours of Sleep and Bedtime Consistency for Children 10-17 in the United States, 2016-17.

Analytic Results

Table 2 shows the adjusted odds of underweight, overweight, or obesity relative to normal weight associated with sleep duration and variation in bedtime using multinomial logit models. For each additional hour of sleep, children had 10.8% (P < 0.01) lower odds of obesity, holding variation in bedtime constant. Equivalent estimates for odds of overweight BMI associated with an additional hour of sleep were not statistically significant; however, an additional hour of sleep was associated with 16.4% higher odds of underweight BMI (P < 0.01).

Table 2.

Odds of Underweight, Overweight or Obesity (Ref Normal Weight).

		Underweight		Overweight		Obesity
Hours of Sleep, daily		1.164**	(1.047,1.295)	1.018	(0.946,1.094)	0.892**	(0.822,0.968)
Goes to Bed at the Same Time, days	Always	–	–	–	–	–	–
Goes to Bed at the Same Time, days	Usually	1.034	(0.823,1.298)	0.822*	(0.688,0.984)	0.752**	(0.619,0.914)
	Sometimes	1.089	(0.759,1.563)	0.761	(0.579,1.001)	0.857	(0.647,1.135)
	Rarely/Never	0.795	(0.476,1.330)	0.919	(0.608,1.389)	0.776	(0.544,1.108)
	Missing	0.913	(0.324,2.573)	3.119	(0.851,11.427)	0.322*	(0.121,0.855)
Physical Activity, days	0 Days	–	–	–	–	–	–
	1 to 3 Days	0.819	(0.586,1.143)	0.923	(0.735,1.158)	0.638***	(0.506,0.804)
	4 to 6 Days	0.563**	(0.400,0.793)	0.763*	(0.595,0.980)	0.426***	(0.321,0.565)
	Everyday	0.497***	(0.340,0.727)	0.610***	(0.458,0.813)	0.321***	(0.235,0.438)
	Missing	0.339	(0.084,1.370)	0.342	(0.111,1.060)	0.739	(0.292,1.873)
Electronic Devices, hours	Do Not Use Electronic Devices	–	–	–	–	–	–
	Use Less Than 1 Hour Per Day	1.342	(0.790,2.279)	0.956	(0.539,1.694)	0.604	(0.303,1.204)
	Use 1 Hour Per Day	1.382	(0.850,2.245)	0.969	(0.551,1.702)	0.699	(0.357,1.367)
	Use 2 to 3 Hours Per Day	1.193	(0.748,1.902)	1.069	(0.616,1.858)	0.768	(0.398,1.483)
	Use 4 or More Hours Per Day	1.103	(0.671,1.813)	1.136	(0.642,2.011)	0.837	(0.432,1.624)
	Missing	1.193	(0.401,3.548)	0.551	(0.146,2.073)	0.455	(0.147,1.412)
TV, hours	Do Not Watch TV	–	–	–	–	–	–
	Less Than 1 Hour Per Day	1.072	(0.676,1.699)	1.333	(0.934,1.902)	1.102	(0.700,1.735)
	1 Hour Per Day	1.009	(0.650,1.568)	1.169	(0.826,1.655)	1.340	(0.872,2.059)
	2 to 3 Hours Per Day	1.186	(0.763,1.843)	1.738**	(1.232,2.452)	1.821**	(1.193,2.780)
	4 or More Hours Per Day	1.250	(0.752,2.079)	2.101***	(1.399,3.156)	2.328***	(1.448,3.742)
	Missing	0.784	(0.254,2.419)	1.681	(0.650,4.348)	1.256	(0.477,3.307)
Household Food Insufficiency	Always Afford to Eat Good Nutritious Meals	–	–	–	–	–	–
	Always Afford Enough to Eat but Not Always the Kinds of Foods We Should Eat	0.988	(0.762,1.281)	1.240*	(1.025,1.500)	1.422***	(1.162,1.741)
	Sometimes We Could Not Afford Enough to Eat	0.757	(0.477,1.203)	0.859	(0.596,1.238)	1.299	(0.913,1.846)
	Often We Could Not Afford Enough to Eat	0.892	(0.401,1.985)	1.417	(0.702,2.859)	1.998	(0.980,4.077)
	Missing	1.190	(0.498,2.842)	1.457	(0.587,3.617)	1.957	(0.903,4.239)
Poverty Level, Household	0 to 99 FPL	–	–	–	–	–	–
	100 to 199 FPL	0.806	(0.565,1.148)	0.998	(0.763,1.304)	0.824	(0.627,1.083)
	200 to 399 FPL	0.691*	(0.490,0.974)	1.000	(0.773,1.295)	0.748*	(0.582,0.962)
	400 FPL or Above	0.717*	(0.514,0.999)	0.883	(0.694,1.124)	0.489***	(0.384,0.623)
Neighborhood Safe	Definitely Agree	–	–	–	–	–	–
	Somewhat Agree	1.128	(0.888,1.434)	1.135	(0.951,1.356)	0.869	(0.724,1.045)
	Somewhat or Definitely Disagree	0.844	(0.415,1.714)	1.071	(0.742,1.546)	0.755	(0.496,1.149)
	Missing	1.411	(0.739,2.694)	1.546	(0.495,4.834)	1.367	(0.664,2.812)
Neighborhood Sidewalks	Yes	–	–	–	–	–	–
	No	1.073	(0.824,1.398)	1.209*	(1.008,1.449)	1.096	(0.891,1.346)
	Missing	0.889	(0.407,1.939)	0.774	(0.286,2.096)	1.309	(0.565,3.037)
Neighborhood Park	Yes	–	–	–	–	–	–
	No	0.929	(0.721,1.196)	1.203	(0.998,1.450)	1.120	(0.902,1.392)
	Missing	0.540	(0.287,1.016)	1.056	(0.545,2.046)	0.550	(0.254,1.188)
Age	Age	0.895***	(0.851,0.941)	0.909***	(0.877,0.942)	0.904***	(0.869,0.941)
Sex	Male	–	–	–	–	–	–
	Female	0.650***	(0.537,0.786)	0.982	(0.846,1.139)	0.657***	(0.556,0.778)
Race	White, non-Hispanic	–	–	–	–	–	–
	Hispanic	0.920	(0.682,1.241)	1.478***	(1.185,1.844)	1.562***	(1.247,1.956)
	Black, non-Hispanic	0.936	(0.676,1.296)	1.483**	(1.167,1.885)	1.575***	(1.239,2.004)
	Multi-racial/Other, non-Hispanic	1.438*	(1.080,1.913)	1.319**	(1.083,1.606)	0.970	(0.766,1.229)

Notes: Reference category is normal weight. Exponentiated coefficients; ci in brackets * p < 0.05, ** p < 0.01, *** p < 0.001. Data from The National Survey of Children’s Health (NSCH) 2016 and 2017 reported from the child’s parents or primary caregiver. Estimates reflect the US population of non-institutionalized children aged 10-17 in 2016-17 with no missing BMI or sleep duration data. n = 34,640. Household selection in the NSCH is stratified first by state and whether there was a child present and then by in terms of neighborhood poverty. One child is randomly selected per household. Estimates reflect parent or primary caregiver reported data including height and weight information used to calculate BMI.

Compared to children who always went to bed at the same time, children who usually went to bed at the same time had 24.8% (P < 0.01) lower odds of obesity and 17.8% (P < 0.05) lower odds of overweight BMI, holding sleep duration constant. There was no statistical evidence of differential odds of obesity or overweight BMI among children who had even greater levels of variation in bedtimes (i.e., sometimes or rarely/never went to bed at the same time) when compared to children who always went to bed at the same time. There was also no statistical association between any levels of variation in bedtime and odds of underweight BMI.

In sensitivity analyses, we re-estimated models using listwise deletion for variables with missing values on covariates, which eliminated 1,216 observations. There was little change in the point estimates after restricting sample to observations with complete data (see Supplemental 1). We also re-estimated models using categorical levels of sleep duration (7, 8, 9, 10, 11 or more hours, 6 or less hours reference). Estimates for indicators of 8, 9, 10 and 11 or more hours were significantly associated with lower odds of obesity (see Supplemental 2).

Lastly, we estimated survey-adjusted logistic regression models, where the outcome was obesity (yes, no). As shown in Supplement 3, an additional hour of sleep was associated with 12.0% lower odds of obesity (P < 0.01). Children who usually went to bed at the same time, had 22.1% lower odds of obesity (P < 0.01) relative to children who always went to bed at the same time. Again, estimates for indicators of higher levels of variation in bedtime (sometimes or rarely/never went to bed at the same time) were not significant.

Discussion

This study examined associations between sleep and obesity in a nationally representative cross-sectional study of children and adolescents ages 10-17 years in the United States. Application of sleep guidelines²⁴ to population estimates of sleep duration suggest that a third of children ages 10 to 17 years in the United States slept less than the recommended number of hours for their age group. Unadjusted estimates showed that on average children with obesity slept fewer hours than children with lower BMI; however, children with longer sleep duration also had on average greater bedtime consistency, suggesting that a negative association between sleep duration and obesity could be driven by bedtime consistency. In our adjusted analysis we found that children’s sleep duration was inversely associated with obesity even after controlling for children’s bedtime consistency, activity levels and screen time, and demographic characteristics, as well as household socioeconomic and neighborhood characteristics.

Holding hours of sleep constant, we found that children who usually went to bed at the same time had lower odds of obesity or overweight compared with children who always went to bed at the same time. However, reductions in odds of obesity or overweight did not extend to children with even greater levels of bedtime variability; that is, children who sometimes or rarely/never went to bed at the same time compared to children who always went to bed at the same time had no statistically distinct odds of obesity or overweight BMI.

This study was limited by the use of cross-sectional data that do not allow us to test possible causal pathways between obesity and sleep. Multivariate models were adjusted for numerous of individual, household and neighborhood covariates, but there may be additional unmeasured differences that could spuriously account for the observed associations between sleep and weight status. For instance, we were not able to control for when children went to sleep nor children’s own preferences about their bed time. Recent research has shown that earlier bedtimes for adolescents with a preference to go to bed late and wake up late has a smaller effect on reducing caloric intake relative to adolescents with a preference to go to bed early and wake up early.³⁸ Our estimates represent the average associations between sleep duration and obesity and may mask underlying heterogeneity in benefits to additional sleep associated with children’s own preferences. Additionally, analysis of sleep measures and obesity in cross-sectional data does not account for reverse causality; that is, obesity may shorten sleep duration. For instance, obstructive sleep apnea in children with obesity is associated with disrupted and fragmented sleep.³⁹ Lastly, our study uses parent reported data that is vulnerable to measurement issues that can bias estimates. Despite limitations of our data, our finding that additional sleep is associated with lower risk of obesity is largely consistent with the overall literature on sleep and obesity in children and adolescents.^7-12

Among the 2 previous studies that used longitudinal data to estimate population-level associations between sleep duration and obesity, our findings are consistent with one of the 2. A study by from the mid-1990s using data from the National Longitudinal Study of Adolescent Health did not find an association between short sleep duration during the first period and obesity during the second period among adolescents aged 12-18 years.²⁵ The authors report that the study used obesity calculated using self-reported height and weight information because longitudinal objective measures were not consistently available. Another study from the early 2000s using the Early Childhood Longitudinal Study (ECLS) found a positive association between a continuous measure of BMI and hours of sleep but hours of sleep were not associated with growth in BMI overtime by children aged 6-11 years.²⁶ This study also benefited from BMI calculated from researcher measured weight and height information included in the ECLS. Substantive comparisons across the studies are challenging because the NSCH is not a longitudinal sample. It is also unclear how far into the future past sleep duration could affect odds of obesity or changes in BMI.

Our findings that variation in sleep duration and not variation in bedtime was associated with obesity contrast previous studies that report volatility in sleep and not duration of sleep are associated with obesity.^28,40 Since these studies directly measured sleep and body weight information,²⁸ the difference in findings could be due to bias in parent-reported data that could have influenced our estimates. Alternatively, these studies rely on small non-random samples which limits generalizability to broader populations. Studies that collect both parent-reported and direct-measured data can help us understand patterns of bias and of noise in data on weight status, sleep, and other factors relevant to children’s health.

We also note that sleep duration was associated not only with obesity but also with underweight BMI, which is also a serious threat to children’s health.⁴¹ Underweight children had longer sleep duration than normal weight children. The associations between underweight and sleep may be different from those between overweight, obesity and sleep. For instance fatigue and emotional distress can increase both sleepiness and risk of underweight BMI.⁴²

Conclusion

In a nationally representative analysis, sleep duration was negatively associated with obesity among children and adolescents in the United States, even after holding bedtime constant. We did not find evidence that variation in bedtime was associated with increased BMI after holding sleep duration constant. It will be important for future studies to explore whether improvement in sleep patterns could lower the prevalence of unhealthy weight, including obesity, overweight, and underweight. Large-scale public health initiatives attributed with improving sleep among children or natural experiments effecting sleep could allow researchers to estimate the causal effects of changes in sleep on obesity.

So What?

What is already known on this topic?

Poor sleep is associated with obesity in children. Children in the US have both high rates of unhealthy weight and inadequate sleep.

What does this article add?

We estimate the associations between sleep and obesity for US children using 2016/17 data, considering both sleep duration and consistency of bedtime.

What are the implications for health promotion practice or researcher?

Children with obesity have lower average weekly duration of sleep than normal weight children; insufficient sleep and unhealthy weight are both health concerns and may be related to each other, raising the possibility for integrated interventions.

Supplemental Material

Supplemental Material, sj-pdf-1-ahp-10.1177_08901171211029189 - Examination of Sleep and Obesity in Children and Adolescents in the United States

Supplemental Material, sj-pdf-1-ahp-10.1177_08901171211029189 for Examination of Sleep and Obesity in Children and Adolescents in the United States by Puneet Kaur Chehal, Livvy Shafer and Solveig Argeseanu Cunningham in American Journal of Health Promotion

Footnotes

Acknowledgments

We thank Johnathan A. Edwards and Rob O’Reilly for guidance on earlier drafts of this work. This study did not require IRB review.

Author's Note

Livvy Shafer is currently affiliated with National Center for Emerging and Zoonotic Infectious Diseases, Division of Global Migration and Quarantine, Community Interventions for Infection Control Unit, Centers for Disease Control and Prevention, Atlanta, GA, USA.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Solveig Argeseanu Cunningham was partly supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number R01DK115937.

ORCID iD

Puneet Kaur Chehal, PhD

Supplemental Material

Supplemental material for this article is available online.

References

Centers for Disease Control and Prevention. Prevalence of Childhood Obesity in the United States. Published 2019. Accessed July 1, 2020. https://www.cdc.gov/obesity/data/childhood.html

Thompson

Obarzanek

Franko

, et al. Childhood overweight and cardiovascular disease risk factors: the National Heart, Lung, and Blood Institute Growth and Health Study. J Pediatr. 2007;150(1):18–25.

Freedman

Dietz

Srinivasan

Berenson

. The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart Study. Pediatrics. 1999;103(6):1175–1182.

Dietz

. Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics. 1998;101(suppl 2):518–525.

Friedlander

Larkin

Rosen

Palermo

Redline

. Decreased quality of life associated with obesity in school-aged children. Arch Pediatrics Adolesc Med. 2003;157(12):1206–1211.

Storch

Milsom

DeBraganza

Lewin

Geffken

Silverstein

. Peer victimization, psychosocial adjustment, and physical activity in overweight and at-risk-for-overweight youth. J Pediatr Psychol. 2007;32(1):80–89.

Duraccio

Krietsch

Chardon

Van Dyk

Beebe

. Poor sleep and adolescent obesity risk: a narrative review of potential mechanisms. Adolesc Health Med Ther. 2019;10:117.

Fatima

Doi

Mamun

. Sleep quality and obesity in young subjects: a meta-analysis. Obesity Rev. 2016;17(11):1154–1166.

Miller

Lumeng

LeBourgeois

. Sleep patterns and obesity in childhood. Curr Opin Endocrinol Diabetes Obes. 2015;22(1):41.

10.

Patel

. Short sleep duration and weight gain: a systematic review. Obesity. 2008;16(3):643–653.

11.

Thind

Davies

Lewis

Pekmezi

Evans

Baskin

. Does short sleep lead to obesity among children and adolescents? Current understanding and implications. Am J Lifestyle Med. 2015;9(6):428–437.

12.

Hart

Cairns

Jelalian

. Sleep and obesity in children and adolescents. Pediatr Clin North Am. 2011;58(3):715–733.

13.

Hart

Carskadon

Considine

, et al. Changes in children’s sleep duration on food intake, Weight, and leptin. Pediatrics. 2013;132(6):e1473.

14.

Hasler

Dahl

Holm

, et al. Weekend–weekday advances in sleep timing are associated with altered reward-related brain function in healthy adolescents. Biol Psychol. 2012;91(3):334–341.

15.

Simon

Field

Miller

DiFrancesco

Beebe

. Sweet/dessert foods are more appealing to adolescents after sleep restriction. PLoS One. 2015;10(2): e0115434.

16.

Asarnow

Greer

Walker

Harvey

. The impact of sleep improvement on food choices in adolescents with late bedtimes. J Adolesc Health. 2017;60(5):570–576.

17.

Telzer

Fuligni

Lieberman

Galván

. The effects of poor quality sleep on brain function and risk taking in adolescence. Neuroimage. 2013;71:275–283.

18.

Aparicio

Canals

Voltas

Valenzano

Arija

. Emotional symptoms and dietary patterns in early adolescence: a school-based follow-up study. J Nutr Educ Behav. 2017;49(5):405–414. e401.

19.

De Castro

. When, how much and what foods are eaten are related to total daily food intake. Br J Nutr. 2009;102(8):1228–1237.

20.

Kaar

Simon

Schmiege

Nadeau

Kelsey

. Adolescent’s health behaviors and risk for insulin resistance: a review of the literature. Curr Diab Rep. 2017;17(7):49.

21.

Templeman

Skovsø

Page

Lim

Johnson

. A causal role for hyperinsulinemia in obesity. J Endocrinol. 2017;232(3):R173–R183.

22.

Moran

Jacobs

Steinberger

, et al. Insulin resistance during puberty: results from clamp studies in 357 children. Diabetes. 1999;48(10):2039–2044.

23.

Galland

Short

Terrill

, et al. Establishing normal values for pediatrix nighttime sleep measured by actigraphy: a systematic review and meta-analysis. Sleep. 2018;41(4).

24.

National Sleep Foundation. How Much Sleep Do Babies and Kids Need? Accessed December 1, 2020. https://www.sleepfoundation.org/excessive-sleepiness/support/how-much-sleep-do-babies-and-kids-need

25.

Calamaro

Park

Mason

, et al. Shortened sleep duration does not predict obesity in adolescents. J Sleep Res. 2010;19(4):559–566.

26.

Miller

. Associations between the home and school environments and child body mass index. Social Sci Med. 2011;72(5):677–684.

27.

Hayes

Balantekin

Altman

Wilfley

Taylor

Williams

. Sleep patterns and quality are associated with severity of obesity and weight-related behaviors in adolescents with overweight and obesity. Child Obes. 2018;14(1):11–17.

28.

Bixler

Liao

, et al. Habitual sleep variability, mediated by nutrition intake, is associated with abdominal obesity in adolescents. Sleep Med. 2015;16(12):1489–1494.

29.

Troxel

Shih

Ewing

Tucker

Nugroho

D’Amico

. Examination of neighborhood disadvantage and sleep in a multi-ethnic cohort of adolescents. Health Place. 2017;45:39–45.

30.

National Survey of Children’s Health. National Survey of Children’s Health 2016-2017. Health Resources and Services Administration MaCHB, ed2016–2017.

31.

US Census Bureau. 2016 National Survey of Children’s Health Methodology Report Published 2018. Accessed October 1, 2019 https://www.census.gov/content/dam/Census/programs-surveys/nsch/tech-documentation/methodology/2016-NSCH-Methodology-Report.pdf

32.

Centers for Disease Control. Clinical Growth Charts. Promotion NCfHSicwNCfCDPaH, ed2000.

33.

Pesonen

Sjöstén

Matthews

, et al. Temporal associations between daytime physical activity and sleep in children. PLoS One. 2011;6(8):e22958–e22958.

34.

Tambalis

Panagiotakos

Psarra

Sidossis

. Insufficient sleep duration is associated with dietary habits, screen time, and obesity in children. J Clinic Sleep Med. 2018;14(10):1689–1696.

35.

Polos

Bhat

Gupta

, et al. The impact of Sleep Time-Related Information and Communication Technology (STRICT) on sleep patterns and daytime functioning in American adolescents. J Adolesc. 2015;44:232–244.

36.

Carter

Rees

Hale

Bhattacharjee

Paradkar

. Association between portable screen-based media device access or use and sleep outcomes: a systematic review and meta-analysis. JAMA Pediatr. 2016;170(12):1202–1208.

37.

Cunningham

Zavodny

. Does the sale of sweetened beverages at school affect children’s weight? Soc Sci Med. 2011;73(9):1332–1339.

38.

Beebe

Zhou

Rausch

Noe

Simon

. The impact of early bedtimes on adolescent caloric intake varies by chronotype. J Adolesc Health. 2015;57(1):120–122.

39.

Narang

Mathew

. Childhood obesity and obstructive sleep apnea. J Nutr Metab. 2012;2012:134202.

40.

Bixler

Berg

, et al. Habitual sleep variability, not sleep duration, is associated with caloric intake in adolescents. Sleep Med. 2015;16(7):856–861.

41.

Uzogara

. Underweight, the less discussed type of unhealthy weight and its implications: a review. Am J Food Sci Nutr Res. 2016;3(5):126–142.

42.

Tomiyama

Mann

Vinas

Hunger

DeJager

Taylor

. Low calorie dieting increases cortisol. Psychosomat Med. 2010;72(4):357.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB