Summary
Objectives
To identify factors affecting early childhood sleep, and investigate the relationship between sleep and overweight/obesity in childhood.
Study Design
Data were collected using parental‐completed questionnaires from N = 10.840 one‐year‐old children in the prospective ABIS‐study (All Babies in Southeast Sweden), followed up until 8 years of age. Chi‐squared test and Pearson Correlation were used to assess the relationship between covariates affecting the children's sleep. Subsequently, longitudinal mixed model analyses were used to predict the effect of different sleep dimensions (bedtime, sleep duration, sleep quality, and the number of awakenings) on BMI Z‐scores.
Results
Children to parents born in Sweden, parents with higher education, non‐single parents, non‐smoking mothers during pregnancy, and children with fewer siblings, were more likely to have appropriate sleep habits at 1 year age. A greater number of awakenings and nocturnal feeds, and particularly later bedtime (β = −0.544, p < 0.0001) were linked to shorter sleep duration. Sleep duration early in life was negatively associated with BMI Z‐scores (adjusted effect estimate [95% CI]: β = −0.09, [(−0.15) – (−0.03)], p = 0.005) later. In addition, higher birth weight, small size for gestational age, unhealthy food habits, children of mothers who smoked during pregnancy, and higher parental BMI resulted in higher BMI Z‐scores.
Conclusion
The child's BMI Z‐score increases by 0.09 units with every hour shorter sleep duration, indicating that short sleep duration might increase the risk of overweight and obesity in children. Parental educational interventions advising appropriate sleeping patterns should be considered when implementing strategies to combat the development of childhood obesity.
Keywords: ABIS, BMI, childhood, obesity, overweight, sleep
1. INTRODUCTION
The increasing prevalence of childhood obesity is a major public health issue. The risk for medical conditions such as cardiovascular diseases and type 2 diabetes, 1 and higher morbidity and mortality in adulthood 2 is elevated among those who were overweight during childhood.
The origin of paediatric adiposity can be heritable, 3 but can also arise from environmental factors and lifestyle factors. Children with obesity are at higher risk to remain obese into adulthood. 4 As treatment of manifest adult obesity and the attendant comorbidity associated with obesity has a low success rate, 5 we should identify early risk factors for the development of obesity.
Although some studies 6 , 7 , 8 , 9 and meta‐analyses 10 , 11 have identified curtailed sleep duration as a risk factor for obesity in childhood and adolescence; other studies contribute to discordant findings. One study with a large, nationally representative dataset of 81.390 children in the United States found no association between parental reported insufficient sleep and obesity. 12 However, most of the studies are based on cross‐sectional observations, thus making it hard to make a conclusion of the long‐term effect and causality. There is a lack of longitudinal studies with large sample size. There is also limited evidence beyond the component sleep duration, 13 , 14 such as bedtime and sleep quality related to obesity risk. Research linking sleep habits to subsequent weight development has received less attention in paediatric populations. 7
In the present study, we aimed to (1) evaluate a set of sleep dimensions (duration, bedtime, quality, and awakenings) in relation to the prospective risk of childhood overweight or obesity measured by body mass index (BMI) Z‐scores and, (2) identify factors that can affect sleep habits at 1 year of age, to elucidate the pathways underlying putative pathophysiological mechanisms linking sleep and obesity, and thereby facilitating prevention strategies.
2. METHODS
2.1. Participants and methods
Data were obtained from a prospective cohort study, the ABIS study (All Babies in Southeast Sweden). The ABIS project was designed to collect information from all children born from October 1997 to October 1999, in Southeast Sweden, mainly to investigate the aetiology of immune‐mediated diseases, in particular, Type 1 diabetes. The 21.700 children were born in the area and 17.055 gave consent to participate in the study. Out of those who received the birth questionnaire, 16.467 (75.9% of the total population) completed it.
Extensive questionnaires administered at birth and at follow‐up check‐ups at 1, 3, 5, and 8 years of age were completed by the caregivers of the children. The structured questionnaires included questions on for example, pregnancy‐ and perinatal period, dietary habits, physical activity, sleep patterns, demographic, socioeconomic, and lifestyle characteristics, as well as morbidity and anthropometric measurements.
The parents of 11.094 children (48% girls) completed the one‐year follow‐up questionnaire where the first questions concerning sleep are found. Data from 8890 children at age three, 7445 at age 5, and 4029 at age 8 were available.
Each subject's parents provided verbal and written informed consent following both oral and written information, to participate in the ABIS study, which was approved by the Research Ethics Committee of the Faculty of Health Sciences, Linköping University, and the Research Ethics Committee of Medical Faculty, Lund University, Sweden.
2.2. Child sleep measurements
Sleeping habits were assessed starting from the 1‐year questionnaire, and consecutively at 3, 5, and 8 years. Sleeping measurements from 10.840, 8773, 7340, and 4004 subjects at 1, 3, 5, and 8 years respectively were available. Categorized sleep variables were used in the bivariate analyses to assess factors that can affect sleep habits at 1 year of age, whereas continuous sleep measurements were used in the longitudinal mixed model analyses.
Bedtime was obtained from the question, ‘At about what time in the evening do you usually put your child to sleep?’ with nine options ranging from 1600 to 2400 or later. The variable was classified into early (17, 18), normal (19–21), and late (22 or later) bedtime Rising time was obtained from the question, ‘At about what time do you usually wake your child up?’ with nine options ranging from 0400 to 1200 or later.
Nocturnal sleep duration was calculated as the difference between bedtime and rising time, classified into short (5–9 h), normal (10–12 h), and long (13–15 h) sleep duration at 1 year of age based on percentage distribution and current sleep duration recommendations. 15 The recommendations for toddlers (1–2 years) are 11–14 h but we set the normal hours of nocturnal sleep as 10–12 h since the recommendations include naps.
The number of wakings was determined with the question, ‘How many times does your child usually wake up during the night?’ with the options never, and six other options ranging from 1 to 6 or more times per night.
Sleep quality was assessed using the question, ‘How do you consider the quality of your child's night sleep?’. A scale of five options was used at the one‐year follow‐up, with values ranging from very good (1) to very bad (5). In subsequent follow‐ups, instead, a scale of six options was used where (6) was very bad. Medium sleep quality was at age one defined as a value of 3, and high and low quality were defined as the two‐scale options above and below medium, respectively.
2.3. BMI measurements
Body Mass Index (BMI; Kg/m2) was computed from parental‐reported length and weight values available in the questionnaires. Age‐ and sex‐specific BMI reference mean‐ and standard deviation scores (BMI‐SDS or Z‐scores) of the children were determined at all available ages (1–8 years of age) using formulas derived from Swedish reference values proposed by Karlberg et al., through a population‐based longitudinal growth study. 16 The mean and ±1, 2, and 3 SD reference ranges were later related to the absolute BMI values computed for each subject [weight (kg)/height (m)2] to categorize the outcome variable in 7 Z‐score categories ranging from severe underweight (BMI < −3SD) to obesity (BMI ≥ 3SD) pursuant to WHO definitions. 17 , 18
2.4. Covariates
Presumed covariates were selected based on previously reported relation to child weight status, and/or plausible prior hypotheses.
Parental factors included parental BMI, ethnicity (born/not born in Sweden), maternal smoking during pregnancy (yes; no), whether mother worked during pregnancy, and the number of months (yes; no, options ranging from 1 to 9 months), and parental relationship status (single; cohabitor; married). Self‐reported height and weight data were used to calculate parental BMI at baseline. In accordance with World Health Organization cut‐off points for BMI, 19 the parents were classified as underweight (<18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), or with obesity (≥30 kg/m2). The parental education level was categorized into none/primary school, high school, and college/university.
Covariates related to family structure and living conditions were: number of siblings, the total number of persons in the household, and the number of rooms in the household. These covariates were collected only at baseline and from the one‐year follow‐up questionnaire.
Birth factors included sex (male; female), birth weight (in kg), gestational age (in weeks), and size for gestational age (SGA, AGA, LGA). Factors relating to the child's nutrition status such as the age at the introduction of formula and canned food, age of cessation of breastfeeding, nocturnal feeds, and cessation of nocturnal feeds were also collected.
A Junk food index was created using several questions concerning the child's food habits available in all questionnaires from 1 to 8 years: ‘How often does your child eat whipped cream/crème Fraiche?’, ‘… eat chocolate?’, ‘… eat candy?’, ‘… eat chips?’, ‘… eat cookies or cake?’, ‘… eat french fries?’. All the questions had the options: every day, 3–5 times a week, 1–2 times a week, or less often. The latter two options were classified as healthy food habits.
Screen time data was available starting from the 3‐year follow‐up. A screen time variable was created using the following questions: ‘How many hours per day on average does your child watch videos/TV?’, ‘… play videogames?’, and ‘… spend on the computer?’. All the questions had options ranging from never to 7 h or more. A total sum was calculated, and screen time was at preschooler age (3 years) classified in Low screen time (≤1 h) and High screen time (>1 h) according to WHO recommendations. 20 For children, 5–8 years screen time was classified in Low screen time (≤2 h) and High screen time (>2 h) according to American Academy of Paediatrics (AAP) guidelines.
2.5. Statistical analyses
Chi‐squared test and Pearson Correlation were used first to examine the bivariate relationship between the sleep variables at 1 year and the covariates, as appropriate. Subsequently, linear mixed model analyses were used to assess the longitudinal relationship between sleep and the outcome variable. All sleep dimensions (bedtime, sleep duration, sleep quality, and the number of awakenings) were included in the analyses and were adjusted for age since age interacts with the child's sleeping habits. Because the variable sleep duration is not independent of the variables bedtime and rising time, rising time was not included in the final analyses. Covariates included in the model were age, sex, birth weight, gestational age, size for gestational age, maternal smoking during pregnancy, parental ethnicity, civil status, parental education level, parental BMI, the introduction of formula and canned food, age of cessation of breastfeeding, age of cessation of nocturnal feeds, junk food index, and screen time.
All statistical tests were performed using SPSS Statistics version 27 (SPSS Inc., Chicago, IL). p values of <0.05 were considered statistically significant.
3. RESULTS
3.1. Sample characteristics
Age‐ and sex‐specific BMI data were available for 10.498, 8194, 6746, and 3038 subjects at 1, 3, 5, and 8 years of age respectively. The majority of the children were in the normal weight category (−2SD ≤ BMI < 2SD) at all ages, with a decreasing tendency (from 88% at 1 year to 76% at 8 years). N = 321 children had obesity as defined by a BMI ≥ 3SD at the age of 8 years. The prevalence of overweight increased from 2.1 to 8.6 percent, and obesity from 0.7% to 10.6% at one and 8 years, respectively.
3.2. Factors affecting child sleep at 1 year of age
The bivariate analyses of selected population characteristics with the sleep variables sleep duration, bedtime, and sleep quality at 1 year, are presented in Tables 1 and 2.
TABLE 1.
Bivariate associations of selected population characteristics and child sleep variables at 1 year of age, with data from 10 840 ABIS‐project participants
| Sleep duration | Bedtime | Sleep quality‡ | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristics | Short (5–9 h) n (%) | Normal (10–12 h) n (%) | Long (13–15 h) n (%) | p value a | Early (5–6 pm) n (%) | Normal (7–9 pm) n (%) | Late (10–12 pm) n (%) | p value a | High n (%) | Medium n (%) | Low n (%) | p value a |
| Parental | ||||||||||||
| Maternal ethnicity | ||||||||||||
| Born in Swe | 227 (2.3) | 9077 (90.8) | 698 (7.0) | 251 (2.5) | 9305 (92.9) | 461 (4.6) | 7316 (73.3) | 1813 (18.2) | 852 (8.5) | |||
| Not born in Swe | 33 (5.5) | 537 (88.9) | 34 (5.6) | <0.0001 | 10 (1.7) | 519 (85.9) | 75 (12.4) | <0.0001 | 386 (65.4) | 113 (19.2) | 91 (15.4) | <0.0001 |
| Paternal ethnicity | ||||||||||||
| Born in Swe | 232 (2.3) | 9026 (90.8) | 688 (6.9) | 244 (2.4) | 9254 (92.9) | 463 (4.6) | 7269 (73.2) | 1794 (18.1) | 863 (8.7) | |||
| Not born in Swe | 27 (4.2) | 570 (88.9) | 44 (6.9) | 0.058 | 17 (2.7) | 553 (86.3) | 71 (11.1) | <0.0001 | 419 (66.9) | 129 (20.6) | 78 (12.5) | 0.004 |
| Maternal smoking during pregnancy | ||||||||||||
| Yes | 29 (3.0) | 865 (88.2) | 87 (8.9) | 40 (4.1) | 886 (90.1) | 57 (5.8) | 725 (74.6) | 169 (17.4) | 78 (8.0) | |||
| No | 232 (2.4) | 8724 (90.9) | 644 (6.7) | 0.020 | 220 (2.3) | 8915 (92.7) | 478 (5.0) | 0.001 | 6962 (72.7) | 1749 (18.3) | 864 (9.0) | 0.409 |
| Mother worked during preg and N of months | ||||||||||||
| Yes | 207 (2.3) | 8088 (90.9) | 600 (6.7) | 192 (2.2) | 8302 (93.2) | 414 (4.6) | 6424 (72.4) | 1657 (18.7) | 791 (8.9) | |||
| No | 53 (3.2) | 1499 (89.1) | 130 (7.7) | 0.041 | 67 (4.0) | 1498 (89.0) | 119 (7.1) | <0.0001 | 1260 (75.4) | 261 (15.6) | 150 (9.0) | 0.011 |
| 1–5 months | 37 (2.7) | 1218 (87.2) | 141 (10.1) | 54 (3.9) | 1268 (90.7) | 76 (5.4) | 995 (71.4) | 256 (18.4) | 143 (10.3) | |||
| 6–7 months | 70 (2.4) | 2628 (90.8) | 196 (6.8) | 64 (2.2) | 2692 (92.8) | 144 (5.0) | 2093 (72.5) | 537 (18.6) | 257 (8.9) | |||
| 8–9 months | 97 (2.1) | 4157 (92.1) | 261 (5.8) | <0.0001 | 75 (1.7) | 4254 (94.1) | 191 (4.2) | <0.0001 | 3263 (72.5) | 847 (18.8) | 392 (8.7) | 0.516 |
| Maternal education level | ||||||||||||
| None/primary school | 38 (5.0) | 648 (84.6) | 80 (10.4) | 37 (4.8) | 670 (87.4) | 60 (7.8) | 546 (72.2) | 144 (19.0) | 66 (8.7) | |||
| High school | 147 (2.3) | 5655 (89.9) | 487 (7.7) | 158 (2.5) | 5812 (92.3) | 330 (5.2) | 4613 (73.5) | 1135 (18.1) | 528 (8.4) | |||
| Higher education | 76 (2.2) | 3284 (93.1) | 166 (4.7) | <0.0001 | 66 (1.9) | 3317 (94.0) | 146 (4.1) | <0.0001 | 2528 (71.9) | 644 (18.3) | 344 (9.8) | 0.200 |
| Paternal education level | ||||||||||||
| None/primary school | 44 (3.2) | 1205 (88.2) | 1366 (8.6) | 44 (3.2) | 1253 (91.6) | 71 (5.2) | 993 (73.0) | 237 (17.4) | 130 (9.6) | |||
| High school | 148 (2.3) | 5870 (90.5) | 470 (7.2) | 165 (2.5) | 6021 (92.7) | 312 (4.8) | 4725 (73.0) | 1189 (18.4) | 559 (8.6) | |||
| College/university | 65 (2.5) | 2382 (92.1) | 2587 (5.4) | <0.0001 | 47 (1.8) | 2399 (92.7) | 143 (5.5) | 0.042 | 1867 (72.4) | 468 (18.2) | 242 (9.4) | 0.650 |
| Relationship status | ||||||||||||
| Single | 6 (3.7) | 139 (86.3) | 16 (9.9) | 6 (3.7) | 141 (87.0) | 15 (9.3) | 117 (71.8) | 33 (20.2) | 13 (8.0) | |||
| Cohabitor | 142 (2.4) | 5270 (89.6) | 468 (8.0) | 145 (2.5) | 5447 (92.5) | 297 (5.0) | 4248 (72.4) | 1085 (18.5) | 535 (9.1) | |||
| Married | 112 (2.5) | 4192 (92.1) | 247 (5.4) | <0.0001 | 110 (2.4) | 4222 (92.7) | 224 (4.9) | 0.115 | 3327 (73.5) | 807 (17.8) | 394 (8.7) | 0.719 |
| Child | ||||||||||||
| Gender | ||||||||||||
| Male | 155 (2.8) | 5125 (91.1) | 348 (6.2) | 130 (2.3) | 5225 (92.7) | 283 (5.0) | 4062 (72.4) | 1015 (18.1) | 535 (9.5) | |||
| Female | 111 (2.1) | 4681 (90.2) | 400 (7.7) | 0.001 | 134 (2.6) | 4800 (92.4) | 263 (5.1) | 0.650 | 3802 (73.5) | 940 (18.2) | 429 (8.3) | 0.078 |
p value/unadjusted r: Chi‐square tests (expressed in frequencies with percentage values) were performed on categorical data, and Pearson correlations were performed on continuous data.
TABLE 2.
Bivariate associations of selected population characteristics and child sleep variables at 1 year of age, with continuous data from 10 840 ABIS‐project participants
| Characteristics | Sleep duration unadjusted (r) a | Bedtime unadjusted (r) a | Sleep quality unadjusted (r) a |
|---|---|---|---|
| Child | |||
| N of nocturnal feeds | −0.070** | 0.133** | 0.358** |
| Cessation of nocturnal feeds | −0.102** | 0.052** | 0.220** |
| N of wakings during the night | −0.112** | 0.108** | 0.689** |
| Bedtime | −0.544** | — | 0.117** |
| Sleep quality | −0.151** | 0.117** | — |
| Family characteristics | |||
| N of siblings | −0.084** | 0.083** | 0.032* |
| N of persons in the household | −0.046** | −0.061** | −0.014 |
| N of rooms in the household | −0.001 | −0.072** | −0.013 |
p value/unadjusted r: Chi‐square tests (expressed in frequencies with percentage values) were performed on categorical data, and Pearson correlations were performed on continuous data.
*p < 0.05.
**p < 0.0001.
Children with immigrant parents had shorter sleep duration, prominently later bedtimes, and lower sleep quality than children whose parents were born in Sweden. Children of non‐smoking mothers, and of mothers having a longer prenatal working period, were more likely to have appropriate sleep duration (11–12 h) and bedtime routines at 1 year. Similarly, among children with parents living together (i.e., cohabitor or married) or having higher educational attainment. Late bedtime was correlated to a higher number of awakenings and number of nocturnal feeds. 36.5% of those parents who reported cause of wakings at 1 year of age believed that the child was hungry.
Details regarding awakenings and nocturnal feeds are published earlier. 21 A higher number of awakenings and nocturnal feeds, later cessation of nocturnal feeds, and particularly later bedtime (β = −0.544, p < 0.0001) were associated with shorter sleep duration. More nocturnal feeds, later cessation of nocturnal feeds, later bedtime, shorter sleep duration, and most outstanding, a higher number of wakings during the night (β = 0.689, p < 0.0001) were strongly associated with lower sleep quality. A later bedtime appeared to be positively correlated with the number of wakings during the night, and the number of nocturnal feeds.
The prevalence of short sleep duration was slightly higher among boys than among girls at the age of one. Children with more siblings were found to have poor sleep in general, while fewer rooms in the household were only related to a later bedtime.
3.3. Longitudinal association between the children's sleep and overweight/obesity
Results from the linear mixed model analyses show that sleep duration was negatively associated with BMI Z‐scores (adjusted effect estimate [95% CI]: β = −0.09, [(−0.15) – (−0.03)], p = 0.005). However, bedtime, sleep quality, and times of nocturnal awakenings were not correlated to the outcome variable, as presented in Supplementary Table S1 and Table S2.
Some covariates remained significant. Boys had lower BMI (β = −0.15, 95% CI: [(−0.19) – (−0.10)], p < 0.0001). Higher birth weight was positively correlated to BMI (β = 0.44, 95% CI: 0.38–0.50, p < 0.0001). Gestational age and size for gestational age was negatively correlated to BMI (β = −0.04, 95% CI: [(−0.05) – (−0.02)], p < 0.0001); respectively (β = −0.15, 95% CI [(−0.26) – (−0.04)], p = 0.008). Unhealthy food habits were associated with higher BMI (β = 0.05, 95% CI: 0.0008–0.09, p = 0.05). Children to smoking mothers during pregnancy had higher BMI (β = 0.26, 95% CI: 0.17–0.35, p < 0.0001). Parental BMI was correlated longitudinally to the child's BMI (maternal β = 0.03, 95% CI: 0.02–0.04, p < 0.0001; respectively paternal β = 0.04, 95% CI: 0.03–0.04, p < 0.0001).
3.4. The children's sleeping trends
The relationships of sleep variables at different ages are displayed in Figure 1, indicating a positive correlation which suggests a strong probability that some children will maintain the same sleep patterns year after year. Effect estimates for bedtime between 1 and 3 years (β = 0.472, p < 0.01), 3 and 5 years (β = 0.455, p < 0.01), and 5 and 8 years (β = 0.453, p < 0.01) suggest that children will be put to bed/go to bed with similar trends later in childhood as when they were at a very early age. Similarly for sleep duration between 1 and 3 years (β = 0.330, p < 0.01), 3 and 5 years (β = 0.345, p < 0.01), and 5 and 8 years (β = 0.345, p < 0.01). There was no interaction between sex and sleeping duration, bedtime, sleeping quality or nocturnal wakenings.
FIGURE 1.
The relation of the children's sleep variables at different ages in the ABIS population of 10 840 children. (A) and (B) sleep duration at ages from one to five in hours. (C) and (D) Bedtime at ages from one to five. Note that the variable values are limited, therefore contributing to overlapping of the points
4. DISCUSSION
4.1. Main findings
Our linear longitudinal analyses assessing the relation between sleep and BMI Z‐scores indicate that poor sleep practices already at 1 year of age may increase the risk of overweight and obesity in children many years later (8 years of age in our study). The child's BMI Z‐score increases on average with 0.09 units with every hour shorter sleep. These results are in agreement with results from some previous studies. 6 , 7 , 8 , 9 , 10 , 11
To be able to distinguish the importance of sleep for the development of obesity it has been necessary to analyse a number of other possible contributing factors. Thus, we found that children born to immigrant parents had in general poor sleep in comparison to children with Swedish parents, which may have different reasons. Thus beside acclimatization difficulties there are cultural differences, environmental (e.g. immigrants often live in poor housing conditions) and occupational factors which may play an important role. Mothers who did not smoke during pregnancy, who worked during pregnancy, parents living together in comparison to single parents, and higher educated parents seemed to have children generally with better sleeping routines. This may, to some extent, reflect that these parents are more aware of the times children are required to be put to bed and rise to get an adequate amount of sleep. Children to smoking mothers during pregnancy were more likely to become overweight or get obesity, which may reflect both physical mechanisms and the impact of an unhealthy lifestyle.
Sleep deprivation may operate to affect obesity risk via several mechanisms such as behavioural, biological, social, cultural, or environmental. Poor sleep may result in daytime fatigue with reduced physical activity and a more sedentary lifestyle. A higher level of screen viewing was associated with obesity among children in this cohort, in agreement with some previous studies. 22 , 23 , 24
Sleep curtailment has been associated with reduced levels of the anorexigenic hormone leptin, and elevated levels of the orexigenic hormone ghrelin. Dysregulated serum levels of these key appetite regulatory hormones are likely to increase hunger and appetite, thereby disturbing energy homeostasis. 25 Nighttime consumption of energy may lead to higher obesity risk. 26 Inadequate sleep duration has also shown to give a slower insulin secretion in response to glucose, resulting in impaired glucose tolerance and reduced insulin sensitivity. 27
It is believed that somatic cells are regulated by intracellular circadian clocks, disruption of which may induce changes in metabolic homeostasis. 28 The strong cross‐sectional association between late bedtime and short sleep duration detected in our analysis could be taken as a support of a mechanism that the child's sleep is regulated by circadian rhythm. However, there are other possible explanations. Children with late bedtimes may still have to be woken up early and then their sleep is cut short, and some parents may have less knowledge on good bedtime habits. Thus many psycho‐social factors may also influence sleep patterns of the children. 21 , 29 As the variable bedtime is used to compute the variable sleep duration it is hard to verify the independence of each sleep dimension when they are analysed in the same model. The strongest variable will rule out the rest. One could assume that those children put to bed late at that early age still sleep the quantity needed, but this does not seem to be the case. Maintaining regular bedtime routines may, therefore, be necessary.
Our data confirm that high birth weight is a strong predictor of overweight and obesity. 30 Children born small for gestational age (SGA) were at higher risk for obesity in agreement with previous studies. 31 , 32
Parental obesity influences obesity risk in their offspring, probably by shared genetics and/or environmental factors forming the child's lifestyle. Our analyses indicate that sleep patterns are established at an early age. Educational interventions advising appropriate sleeping patterns starting from an early age may be important when implementing strategies to combat the development of childhood obesity.
4.2. Strengths and limitations
This study has considerable strengths with the large sample size and the longitudinal approach used. Data were collected at several points permitting us to explore the causal ordering between sleep and overweight/obesity. Another strength relates to testing the association between several sleep dimensions during early childhood rather than sleep duration in isolation, and the risk of overweight or obesity while adjusting for a wide range of factors. However, our results should be interpreted in the context of some limitations.
First it should be noted that the collection of data is done quite many years ago. Society has changed and possibly also sleep patterns and life style. There is a gradual drop‐out, but there is no significant difference in demographics between the original study population and the final samples. We used parental‐reported sleep acquired via the questionnaires, but sleep habits were not reported separately for weekdays versus weekends, and we have no information on naps daytime. Reported sleep may also have been subject to recall and social desirability bias. However, parental‐reported assessments of child sleep are highly consistent with actigraphy and accelerometry. 33 , 34 Bedtime, sleep quality, and times of nocturnal awakenings were not correlated to the development of obesity, but as example, number of nocturnal awakenings was associated to shorter sleep duration it is possible that poor sleep quality further strengthen the risk of short sleep duration.
The height and weight measurements used to calculate the body mass indices were also based on parental reports, but the questionnaires were given in connection with visits at well‐baby clinics where measurements took place and are therefore probably reliable.
Because mixed model analyses are based on repeated measures, physical activity wasn't included in the measures since data was not available at the earlier time points (1 and 3 years).
There is a relatively big dropout but exploratory analyses using a cut‐off point at 5‐years of age where the drop‐out was less significant did not yield different results.
5. CONCLUSION
In conclusion, our longitudinal results from a large prospective general cohort demonstrate that poor sleep may increase adiposity risk. It is crucial to implement reasonable sleep routines already in infancy. Education of parents emphasizing recommended sleep routines, such as consistent and early bedtimes should be considered as an obesity prevention strategy.
AUTHOR CONTRIBUTIONS
Dr. Ban Danial studied the data, performed all the statistical analyses, designed the tables and figures, and drafted the initial manuscript. Prof. Tomas Faresjö planned and supervised the work, assisted in drafting the manuscript, reviewed and revised the manuscript. Prof. Mats Fredriksson conceived the analytical models, supervised Dr. Ban Danial with the analyses, contributed to the interpretation of the results, and reviewed the manuscript. Prof. Johnny Ludvigsson conceptualized and planned the study, and critically reviewed and revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
CONFLICT OF INTEREST
No conflict of interest was declared.
Supporting information
Table S1. Linear mixed model of child sleep variables at age 1–8 years and the children's BMI Z‐scores presented in effect estimates, adjusted for all covariates below.
ACKNOWLEDGEMENTS
Special thanks to the participating families in the ABIS study, and all staff at Obstetric departments and Well‐Baby Clinics. ABIS has been supported by Swedish Research Council (K2005‐72X‐11242‐11A and K2008‐69X‐20826‐01‐4) and the Swedish Child Diabetes Foundation (Barndiabetesfonden), JDRF Wallenberg Foundation (K98‐99D‐12813‐01A), Medical Research Council of Southeast Sweden (FORSS), and the Swedish Council for Working Life and Social Research (FAS2004–1775) and Östgöta Brandstodsbolag.
Danial B, Faresjö T, Fredriksson M, Ludvigsson J. Childhood sleep and obesity risk: A prospective cohort study of 10 000 Swedish children. Pediatric Obesity. 2023;18(2):e12983. doi: 10.1111/ijpo.12983
Funding information ALF Region Östergötland, Grant/Award Number: 0000; Barndiabetesfonden (The Swedish Child Diabetes Foundation); Swedish Council for Working Life and Social Research, Grant/Award Number: FAS2004–1775; Medical Research Council of Southeast Sweden; JDRF Wallenberg Foundation, Grant/Award Number: K98‐99D‐12813‐01A; Swedish Research Council, Grant/Award Numbers: K2008‐69X‐20826‐01‐4, K2005‐72X‐11242‐11A
REFERENCES
- 1. Must A, Strauss RS. Risks and consequences of childhood and adolescent obesity. Int J Obes Relat Metab Disord. 1999;23(Suppl 2):S2‐S11. [DOI] [PubMed] [Google Scholar]
- 2. Dietz WH. Childhood weight affects adult morbidity and mortality. J Nutr. 1998;128:411S‐414S. [DOI] [PubMed] [Google Scholar]
- 3. Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH. Predicting obesity in Young adulthood from childhood and parental obesity. N Engl J Med. 1997;337:869‐873. [DOI] [PubMed] [Google Scholar]
- 4. Serdula MK, Ivery D, Coates RJ, Freedman DS, Williamson DF, Byers T. Do obese children become obese adults? A review of the literature. Prev Med. 1993;22:167‐177. [DOI] [PubMed] [Google Scholar]
- 5. Methods for voluntary weight loss and control . NIH technology assessment conference panel. Consensus development conference, 30 march to 1 April 1992. Ann Intern Med. 1993;119:764‐770. [PubMed] [Google Scholar]
- 6. Halal CSE, Matijasevich A, Howe LD, Santos IS, Barros FC, Nunes ML. Short sleep duration in the first years of life and obesity/overweight at age 4 years: a birth cohort study. J Pediatr. 2016;168:99‐103.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Taveras EM, Rifas‐Shiman SL, Oken E, Gunderson EP, Gillman MW. Short sleep duration in infancy and risk of childhood overweight. Arch Pediatr Adolesc Med. 2008;162:305‐311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Wang F, Liu H, Wan Y, et al. Sleep duration and overweight/obesity in preschool‐aged children: a prospective study of up to 48,922 children of the Jiaxing birth cohort. Sleep. 2016;39:2013‐2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. von Kries R, Toschke AM, Wurmser H, Sauerwald T, Koletzko B. Reduced risk for overweight and obesity in 5‐ and 6‐y‐old children by duration of sleep—a cross‐sectional study. Int J Obes. 2002;26:710‐716. [DOI] [PubMed] [Google Scholar]
- 10. Li L, Zhang S, Huang Y, Chen K. Sleep duration and obesity in children: a systematic review and meta‐analysis of prospective cohort studies. J Paediatr Child Health. 2017;53:378‐385. doi: 10.1111/jpc.13434 [DOI] [PubMed] [Google Scholar]
- 11. Fatima Y, Doi SAR, Mamun AA. 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. [DOI] [PubMed] [Google Scholar]
- 12. Hassan F, Davis MM, Chervin RD. No independent association between insufficient sleep and childhood obesity in the National Survey of children's health. J Clin Sleep Med. 2011;7:153‐157. [PMC free article] [PubMed] [Google Scholar]
- 13. Golley RK, Maher CA, Matricciani L, Olds TS. Sleep duration or bedtime? Exploring the association between sleep timing behaviour, diet and BMI in children and adolescents. Int J Obes. 2013;37:546‐551. [DOI] [PubMed] [Google Scholar]
- 14. Ekstedt M, Darkeh MH, Xiu L, et al. Sleep differences in one‐year‐old children were related to obesity risks based on their parents' weight according to baseline longitudinal study data. Acta Paediatr. 2017;106:304‐311. [DOI] [PubMed] [Google Scholar]
- 15. Paruthi S, Brooks LJ, D'Ambrosio C, et al. Recommended amount of sleep for pediatric populations: a consensus statement of the American Academy of sleep medicine. J Clin Sleep Med. 2016;12:785‐786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Karlberg J, Luo ZC, Albertsson‐Wikland K. Body mass index reference values (mean and SD) for Swedish children. Acta Paediatr. 2001;90:1427‐1434. [DOI] [PubMed] [Google Scholar]
- 17. WHO Multicentre Growth Reference Study Group . WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr. 2006;450:76‐85. [DOI] [PubMed] [Google Scholar]
- 18. Training Course on Child Growth Assessment WHO Child Growth Standards.
- 19. World Health Organization . Obesity: preventing and managing the global epidemic. Report of a WHO Consultation presented at the World Health Organization, June 3–5. [PubMed]
- 20. Guidelines on Physical Activity . Sedentary Behaviour and Sleep for Children under 5 Years of Age. Guidelines on Physical Activity, Sedentary Behaviour and Sleep for Children under 5 Years of Age. 2019.
- 21. Palmstierna P, Sepa A, Ludvigsson J. Parent perceptions of child sleep: a study of 10,000 Swedish children. Acta Paediatr. 2008;97(12):1631‐1639. doi: 10.1111/j.1651-2227.2008.00967.x [DOI] [PubMed] [Google Scholar]
- 22. Marsh S, Ni Mhurchu C, Maddison R. The non‐advertising effects of screen‐based sedentary activities on acute eating behaviours in children, adolescents, and young adults. A systematic review. Appetite. 2013;71:259‐273. [DOI] [PubMed] [Google Scholar]
- 23. Jago R, Baranowski T, Baranowski JC, Thompson D, Greaves KA. BMI from 3‐6y of age is predicted by TV viewing and physical activity, not diet. Int J Obes. 2005;29:557‐564. [DOI] [PubMed] [Google Scholar]
- 24. Ekelund U, Brage S, Froberg K, et al. TV viewing and physical activity are independently associated with metabolic risk in children: the European youth heart study. PLoS Med. 2006;3:2449‐2457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Taheri S, Lin L, Austin D, Young T, Mignot E. Short sleep duration is associated with reduced leptin, elevated ghrelin, and increased body mass index. PLoS Med. 2004;1:e62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Cheng TS, Loy SL, Toh JY, et al. Predominantly nighttime feeding and weight outcomes in infants. Am J Clin Nutr. 2016;104:380‐388. [DOI] [PubMed] [Google Scholar]
- 27. Nedeltcheva AV, Kessler L, Imperial J, Penev PD. Exposure to recurrent sleep restriction in the setting of high caloric intake and physical inactivity results in increased insulin resistance and reduced glucose tolerance. J Clin Endocrinol Metab. 2009;94:3242‐3250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Bray MS, Young ME. Circadian rhythms in the development of obesity: potential role for the circadian clock within the adipocyte. Obes Rev. 2007;8:169‐181. [DOI] [PubMed] [Google Scholar]
- 29. Sadeh A. Evaluating night wakings in sleep‐disturbed infants: a methodological study of parental reports and actigraphy. Sleep. 1996;19(10):757‐762. doi: 10.1093/sleep/19.10.757 [DOI] [PubMed] [Google Scholar]
- 30. Yu ZB, Han SP, Zhu GZ, et al. Birth weight and subsequent risk of obesity: a systematic review and meta‐analysis. Obes Rev. 2011;12:525‐542. [DOI] [PubMed] [Google Scholar]
- 31. Nam HK, Lee KH. Small for gestational age and obesity: epidemiology and general risks. Ann Pediatr Endocrinol Metab. 2018;23:9‐13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ong KKL, Ahmed ML, Dunger DB, Emmett PM, Preece MA. Association between postnatal catch‐up growth and obesity in childhood: prospective cohort study. Br Med J. 2000;320:967‐971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Acebo C, Sadeh A, Seifer R, Tzischinsky O, Hafer A, Carskadon MA. Sleep/wake patterns derived from activity monitoring and maternal report for healthy 1‐ to 5‐year‐old children. Sleep. 2005;28:1568‐1577. [DOI] [PubMed] [Google Scholar]
- 34. Martinez SM, Greenspan LC, Butte NF, et al. Mother‐reported sleep, accelerometer‐estimated sleep and weight status in Mexican American children: sleep duration is associated with increased adiposity and risk for overweight/obese status. J Sleep Res. 2014;23:328‐336. [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
Table S1. Linear mixed model of child sleep variables at age 1–8 years and the children's BMI Z‐scores presented in effect estimates, adjusted for all covariates below.