- Email: [email protected]
Both habitual short sleepers and long sleepers are at greater risk of obesity: a population-based 10-year follow-up in women
Sleep Medicine 15 (2014) 1204–1211
Contents lists available at ScienceDirect
Sleep Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s l e e p
Original Article
Both habitual short sleepers and long sleepers are at greater risk of obesity: a population-based 10-year follow-up in women Jenny Theorell-Haglöw a,*, Lars Berglund b, Christian Berne c, Eva Lindberg a a b c
Department of Medical Sciences, Respiratory Medicine and Allergology, Uppsala University, Sweden Uppsala Clinical Research Center (UCR), Uppsala University, Sweden Department of Medical Sciences, Internal Medicine, Uppsala University, Sweden
A R T I C L E
I N F O
Article history: Received 22 August 2013 Received in revised form 2 February 2014 Accepted 4 February 2014 Available online 12 June 2014 Keywords: Sleep duration Obesity Changed sleep duration Longitudinal Population-based Women
A B S T R A C T
Objective: To assess how change in sleep duration is related to subsequent obesity. Methods: In this 10-year follow-up, 4903 non-pregnant participants answered a questionnaire on sleeping habits, obesity, and lifestyle factors (questions identical to baseline questionnaire). Habitual normal sleepers were defined as sleeping 6–9 h/night at both baseline and follow-up, whereas women sleeping <6 h/night or ≥9 h/night at both occasions were defined as habitual short sleepers and habitual long sleepers, respectively. Logistic regression was used to analyze associations between changes in sleep duration, general obesity (body mass index ≥30 kg/m2), weight gain (≥10 kg) and also, central obesity (waist circumference ≥88 cm), and increase in waist circumference (≥10 cm) at follow-up. Results: Among younger women (aged <40 years) both habitual short sleepers and habitual long sleepers had a higher prevalence of general (short: 31.3%, P < 0.0001; long: 38.1%, P = 0.01) and central obesity (short: 60.5%, P = 0.01; long: 82.4%, P = 0.01) compared with habitual normal sleepers (general obesity: 8.9%; central obesity: 35.9%) at follow-up. Younger women who were short sleepers at baseline but normal sleepers at the follow-up had a higher prevalence of both general (19.3%, P = 0.01) and central obesity (45.4%, P = 0.07) compared with habitual normal sleepers at follow-up. In adjusted analyses, both habitual short [adjusted odds ratio (aOR), 6.78; 95% confidence interval (CI), 2.71–17.0] and long (aOR, 4.64; 95% CI, 1.09–19.8) sleep durations were risk factors for general obesity in younger women. In younger women habitual long sleep duration was a risk factor also for central obesity (aOR, 6.05; 95% CI, 1.19– 30.7) whereas habitual short sleep duration was not (aOR, 1.93; 95% CI, 0.87–4.81). Similar results were seen also for weight gain and increased waist circumference as dependent variables. In addition, decreased sleep duration from normal to short duration was a risk factor for both weight gain (aOR, 1.85; 95% CI, 1.14–3.02) and increased waist circumference (aOR, 1.84; 95% CI, 1.20–2.81). There were no associations between changes in sleep duration and any of the measures of obesity at the follow-up in women aged >40 years at baseline. Conclusion: In younger women, both habitual short and long sleep duration was a risk factor for obesity, whereas no such relationship was seen in older women. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Over the last decade, several cross-sectional studies have shown associations between sleep duration, sleeping habits, and measurements of obesity [1–9]. An inverse relationship between selfreported sleep duration and obesity is found in adult men and women, as well as in children [10]. Short sleep duration is also found to predict weight gain and obesity later in life [2,3].
* Corresponding author at: Department of Medical Sciences, Respiratory Medicine and Allergology, Uppsala University, SE-751 85 Uppsala, Sweden. Tel.: +46 18 611 02 42; fax: +46 18 611 02 28. E-mail address: [email protected] (J. Theorell-Haglöw). http://dx.doi.org/10.1016/j.sleep.2014.02.014 1389-9457/© 2014 Elsevier B.V. All rights reserved.
The results of previous studies on the longitudinal relationship between sleep duration and obesity vary, with some studies only finding a relationship between short sleep duration and obesity [3], some revealing a U-shaped relationship [11] and some showing no relationship [9]. However, sleep duration is not necessarily fixed over time, and studies of the relationship between changes in sleep duration over time and the development of obesity are still sparse and have produced diverse results. This project is part of a 10-year follow-up of the epidemiological population-based study entitled ‘Sleep and Health in women’ (SHE) [12]. The overall aim of the SHE study is to analyze the longterm evolution and consequences of sleeping habits and sleep disturbances. The aim of the present study was to assess how changes in sleep duration over time were related to obesity over a 10-year
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
period, in a population-based sample of women. A second aim was to examine these associations stratified by age. 2. Methods 2.1. SHE questionnaire The SHE study started in 2000, when a questionnaire on sleeping habits and somatic disorders was sent to women aged ≥20 years, randomly selected from the population registry of the City of Uppsala, Sweden. The response rate was 71.6% (n = 7051). In 2010, a followup questionnaire was sent to all the women who had answered the baseline questionnaire and who were still alive. Of the original study population, 8.5% were lost at follow-up due to death (n = 461), emigration (n = 130) and unknown addresses (n = 5). The follow-up questionnaire was therefore sent to 6455 women (91.6% of the initial study population) and completed by 5193 (response rate 80.5%). In the present study, women who were pregnant or who had omitted the question on sleep duration at one or both time-points were excluded, and the final study population comprised a total of 4903 non-pregnant women (Fig. 1). The study was approved by the Ethics Committee at the Medical Faculty at Uppsala University and all the participants gave their informed consent before participating. The follow-up questionnaire included mostly the same questions as the baseline questionnaire, which has been previously described in detail [12]. Briefly, the follow-up questionnaire included questions on sleep duration, snoring habits, insomnia, occupational status, shift work, civil status, physical activity, smoking, alcohol dependence, anxiety and depression. Both at baseline and at follow-up, sleep duration (h/night) was assessed using the question: ‘How many hours do you sleep on average during the night?’ Sleep duration, at baseline and at follow-up, was classified into three categories: short sleepers (<6 h/night), normal sleepers (6– 9 h/night), and long sleepers (≥9 h/night). In addition, a sleep habit variable was constructed where the women were divided into ha-
Fig. 1. Overview of study protocol.
1205
bitual short sleepers (short sleepers at both baseline and followup), habitual long sleepers (long sleepers at both baseline and followup), women with ‘normalized’ sleep duration (short or long sleepers at baseline who were normal sleepers at follow-up), women with decreased sleep duration (normal sleepers at baseline who were short sleepers at follow-up), women with increased sleep duration (normal sleepers at baseline who were long sleepers at followup), habitual normal sleepers (normal sleep duration at both baseline and follow-up), as well as a group of women changing from short to long sleep duration or from long to short sleep duration. Snoring habits were assessed using the question: ‘How often do you snore loudly and disturbingly?’ The response options for this question were ‘never’ (1), ‘seldom’ (2), ‘sometimes’ (3), ‘often’ (4) and ‘very often’ (5). Based on their response to this question, the participants were categorized into two groups: non-snorers (scores 1–3) and snorers (scores 4 and 5). The women further indicated their current height and weight in both questionnaires, and body mass index (BMI, kg/m2) was calculated and rounded off to one decimal point. BMI was categorized as underweight (<19.9), normal weight (20–24.9), overweight (25–29.9), and obese (≥30). In conjunction with the questionnaire, the women were given a tape measure and instructions on how to measure their waist circumference [12]. A waist circumference of ≥88 cm was used to define central obesity according to National Cholesterol Education Program (NCEP) criteria [13]. All the data are self-reported. However, for 400 of the women BMI and waist circumference have been measured by a research nurse 2 years later. In spite of the time lapse between the two measures there was good correlation between measures at the two time-points (BMI: β = 0.80, P < 0.0001; waist circumference: β = 0.85, P < 0.0001). The mean difference between measured and self-reported BMI was 0.87 ± 2.51 and the corresponding difference in waist circumference was −0.39 ± 7.85 cm. The participants’ physical activity was analyzed by four questions adopted from a questionnaire used in a large populationbased study of the correlation between physical activity and mortality [14]. Six questions assessed smoking habits. The participants were categorized as ‘current smokers’ or ‘non-smokers’ (i.e. had quit smoking at least 6 months before answering the questionnaire or had never smoked). Alcohol dependence was assessed by the CAGE (acronym for Cut down, Annoyed by criticism, Guilty about drinking and Eye-opener drinks) alcohol and screening questionnaire [15]. The women were also asked to indicate how many cups of coffee they drank every day and this number was used as a continuous variable in the analyses. Psychological distress was assessed using the Hospital Anxiety and Depression (HAD) scale [16] and was used as a dichotomized score with a cut-off at score 10. The participants were asked at both time-points to indicate how many months they had worked nights or shifts over the past 10 years. A cut-off point of 60 months was chosen when analyzing working nights or shifts. The women were asked whether they were taking any medication on a regular basis. Taking insulin or oral medication, or both, for diabetes were categorized together as ‘diabetes medication’. Compared with the responders, the non-responders at the followup were older (mean ages 45.0 ± 16.0 vs 43.3 ± 15.3 years, P < 0.001), somewhat more obese (mean BMI 24.5 ± 4.3 vs 24.0 ± 4.1, P = 0.0007), and more often smokers (21.9% vs 15.9%, P < 0.0001) at baseline. Responders had somewhat longer sleep duration (mean sleep duration 7.0 vs 6.9 h, P = 0.03) but prevalence of habitual snoring did not differ between the groups (7.6 vs 7.8%, not significant). However, the non-responders were less physically active at baseline as 16.3% reported a high level of physical activity and 23.8% reported a low level of physical activity as compared with 21.1% and 15.7%, respectively for the responders (P < 0.001).
1206
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
2.2. Statistical analyses Statistical analyses were performed using Stata 10 (Stata Corp., College Station, TX, USA). Univariate analyses at both time-points were conducted using the unpaired t-test for continuous variables and the χ2-test for categorical variables to compare the sleep duration groups. The unpaired t-test was used even if the data were not assumed to be normally distributed, as the sample size was large and, likely, the use of a Wilcoxon–Mann–Whitney test would not change the decision about rejecting the null hypothesis. Analyses comparing continuous and categorical variables respectively, between the two time-points, were also made. Variables that differed between groups (P < 0.20) were entered into multiple logistic regression models. The results of the logistic regression models are presented as adjusted odds ratios (ORs) with 95% confidence intervals (CIs) with the dependent variables: BMI ≥30 kg/m2 and weight gain ≥10 kg, respectively, and waist circumference ≥88 cm and increase in waist circumference ≥10 cm, respectively. As only 19 women in total had gone from short to long sleep duration or from long to short sleep duration, they were grouped together and were included in the multiple analyses. In a previous study [17] using a subsample of the same female cohort, we found that short sleep duration is most important with respect to impact on metabolism in younger women (aged <40 years), and there was a significant interaction between age and sleep duration for both general (P < 0.001) and central obesity (P < 0.001). For this reason, in further analyses in the present study the population was grouped according to age ≥40 and <40 years (at baseline). 3. Results Over the 10-year period, the mean sleep duration had decreased and the prevalence of both general and central obesity had
increased significantly. There was an increase in the number of women who reported snoring. Moreover, alcohol dependence increased during the study period. However, smoking and the use of snuff or other nicotine substances decreased significantly. Furthermore, there were fewer shift workers at the 10-year follow-up and there was a significant reduction in the prevalence of anxiety and depression (Table 1). The majority of the women in the study were normal sleepers (i.e. slept 6–9 h/night) at both time-points, 243 were habitual short sleepers (i.e. slept <6 h/night at both time-points), and 60 women were habitual long sleepers (i.e. slept ≥9 h/night at both timepoints). Of the women, 293 had changed from short to normal sleep and 160 had changed from long to normal sleep, making a total of 453 women whose sleep duration had ‘normalized’. In addition, 309 women had changed from normal to short sleep and 123 had changed from normal to long sleep, making a total of 432 women whose sleep duration had either increased or decreased. Finally, eight women had changed from short to long sleep and 11 women had changed from long to short sleep duration during the 10-year period. Moreover, when the women were grouped according to age, there were similar distributions. 3.1. Changes in sleep duration and obesity At the follow-up, all women who had been short or long sleepers either at baseline or at follow-up had a higher prevalence of general obesity (i.e. BMI ≥30 kg/m2) as compared with women with normal sleep duration on both occasions (Fig. 2). When the women were analyzed by age group, both habitual short sleepers and long sleepers in the younger age group (<40 years) had a higher prevalence of general obesity at follow-up compared with habitual normal sleepers in this age group. In addition, short sleepers at baseline whose sleep had ‘normalized’ at follow-up also had a higher
Table 1 Characteristics of the population at baseline and of the 10-year follow-up group. Characteristics
Baseline All (2000) (n = 6843)
Age at baseline (years) Sleep duration (h) Sleep duration Short (<6 h/night) Normal (6–9 h/night) Long (≥9 h/night) Snoring BMI (kg/m2) Obesity (BMI ≥30 kg/m2) Waist circumference (cm) Central obesity (WC ≥88 cm) Smoking/snuff/nicotine use Alcohol dependence (CAGE ≥2) Physical activity High Medium Low Cups of coffee/day Shift work Diabetes medication Oral diabetes medication Insulin Insulin and oral diabetes medication HAD anxiety score ≥10 HAD depression score ≥10
Follow-up Follow-up group (2000) (n = 4903)
46.0 (17.5) 6.9 (1.2)
43.9 (15.2) 7.0 (1.1)
856 (12.7%) 5566 (82.2%) 346 (5.1%) 513 (7.8%) 24.2 (4.2) 595 (8.8%) 84.1 (12.8) 2157 (33.6%) 1352 (19.9%) 385 (5.9%)
P-valuea
(2010) (n = 4903) NA 6.9 (1.1)
NA 0.011
563 (11.3%) 4171 (84.0%) 234 (4.7%) 391 (8.0%) 24.1 (4.1) 425 (8.5%) 83.6 (12.5) 1503 (31.6%) 947 (19.0%) 294 (6.1%)
568 (11.5%) 4175 (84.6%) 192 (3.9%) 457 (9.4%) 25.1 (4.5) 662 (13.4%) 89.7 (12.6) 2523 (51.9%) 634 (12.8%) 360 (7.4%)
0.51 0.51 0.028 0.0013 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0004
1279 (19.0%) 4184 (62.0%) 1285 (19.0%) 4.37 (2.7%) 294 (4.3%)
785 (15.8%) 3164 (63.7%) 1018 (20.5%) 4.39 (2.8%) 234 (4.7%)
1090 (22.0%) 3146 (63.5%) 719 (14.5%) NA 157 (2.3%)
0.059 0.059 0.17 NA <0.0001
77 (1.1%) 32 (0.5%) 12 (0.2%) 936 (14.2%) 291 (4.5%)
33 (0.7%) 18 (0.4%) 4 (0.1%) 655 (13.5%) 203 (4.2%)
109 (2.2%) 9 (0.2%) 11 (0.2%) 567 (11.6%) 186 (3.8%)
<0.0001 0.0495 0.035 <0.0001 <0.0001
Values are mean (SD) or n (%). Abbreviations: BMI, body mass index; WC, waist circumference; CAGE, acronym for ‘Cut down, Annoyed by criticism, Guilty about drinking and Eye-opener drinks’ questionnaire; HAD, Hospital Anxiety and Depression scale. a For test between baseline and follow-up values in the follow-up group.
Age >=40 yrs.
40
**
30
**
20
**
10
** *
NS REF
0
Prevalence (%) of BMI>=30 kg/m2
Age <40 yrs.
1207
50
50
Prevalence (%) of BMI>=30 kg/m2
All women
Prevalence (%)of BMI>=30 kg/m2
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
50
***
40
***
30 20
**
NS
NS
10
NS REF
0
40 30
***
*
20 10
NS NS
NS
NS REF
0
* <0.05; ** <0.01; *** <0.0001 Fig. 2. Prevalence of general obesity (body mass index ≥30 kg/m2) in 2010 in relation to sleep duration in 2000 and 2010, respectively. Data are only presented for groups of ≥15 women. *P < 0.05, **P < 0.01, ***P < 0.0001 as compared with women with normal sleep duration both at baseline and at follow-up (unadjusted comparisons).
prevalence of general obesity at follow-up compared with habitual normal sleepers. In the older age group, the prevalence of general obesity was highest in long sleepers whose sleep had ‘normalized’, but this also applied to normal sleepers who had become long sleepers. The analysis was also performed in women who were nonsnorers (using the snoring question as a proxy for sleep apnea) on both occasions, but this did not significantly change the results (data not shown). There was a greater prevalence of central obesity (i.e. waist circumference ≥88 cm) at the follow-up for women in all categories of sleep duration change (Fig. 3). Again, both habitual short sleepers and long sleepers in the younger age group had a higher prevalence of central obesity at follow-up compared with habitual normal
sleepers. In this group, there was also a higher prevalence of central obesity in short sleepers whose sleep duration had ‘normalized’ over the 10-year period, even if it was not as high as for habitual short or long sleepers. In the older age group, the highest prevalence of central obesity at follow-up was found in women who were long sleepers at baseline and whose sleep had then ‘normalized’. The analyses for central obesity were also performed in women who were non-snorers on both occasions, but this did not significantly change the results (data not shown). Women aged <40 years had in general gained more weight during the 10-year period compared with women aged ≥40 years. In the younger group, the lowest change in weight was found among the habitual normal sleepers with a mean weight gain of 3.5 ± 6.9 kg.
Age <40 yrs.
Age >=40 yrs.
100 90 80 70 60 50 40 30 20 10 0
100 90 80 70 60 50 40 30 20 10 0
Prevalence (%) of WC>=88 cm
All women
Prevalence (%) of WC>=88 cm
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
Prevalence (%) of WC>=88 cm
1208
100 90 80 70 60 50 40 30 20 10 0
*** ***
*
*
NS
NS
REF
* * NS NS
NS
NS
REF
NS NS
NS NS
* NS
REF
* <0.05; ** <0.01; *** <0.0001 Fig. 3. Prevalence of central obesity (waist circumference ≥88 cm) in 2010 in relation to sleep duration in 2000 and 2010 respectively. Data are only presented for groups of ≥15 women. *P < 0.05, **P < 0.01, ***P < 0.0001 as compared with women with normal sleep duration both at baseline and at follow-up (unadjusted comparisons).
Habitual long and short sleepers had gained the most in weight with a mean change of 8.1 ± 8.8 and 7.5 ± 8.3 kg, respectively. In women aged ≥40 years the habitual normal sleepers increased their weight with 1.7 ± 7.4 kg and the corresponding weight changes in habitual long and short sleepers were 1.7 ± 7.7 and 0.5 ± 7.4 kg. Differences between younger and older women were less pronounced when it came to changes in waist circumference over time. In the younger group, habitual normal sleepers reported a waist circumference at the follow-up that was in general 6.9 ± 8.8 cm longer compared with baseline. The corresponding differences for habitual long and short sleepers were 12.0 ± 8.0 and 9.3 ± 12.3 cm. In the older group, the mean changes in waist circumference were 5.5 ± 8.3,
4.8 ± 7.7, and 4.2 ± 9.0 cm for habitual normal, long, and short sleepers, respectively. 3.2. Multiple analyses 3.2.1. General obesity and weight gain In the whole population, unchanged short sleep duration was a risk factor for general obesity also when adjusting for age, snoring, smoking/nicotine use, alcohol dependence, physical activity, caffeine intake, shift work, anxiety, depression, diabetes medication, and baseline weight (OR, 1.71; 95% CI, 1.01–2.87). In younger women (<40 years) both unchanged short (aOR, 6.78; 95% CI, 2.71–17.0) and
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
1209
Table 2 Change in sleep duration in relation to general obesity and weight gain. All women (n = 4903)
Unchanged normal sleep duration (n = 3696) Decreased sleep duration (n = 309)c Increased sleep duration (n = 123)d ‘Normalized’ sleep duration (n = 453) Unchanged short sleep duration (n = 243) Unchanged long sleep duration (n = 60)
Women aged <40 yearsa (n = 2088)
Women aged ≥40 yearsa (n = 2815)
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
Ref. 1.09 (0.68–1.73) 1.59 (0.79–3.21) 1.36 (0.92–2.03) 1.71 (1.01–2.87) 1.52 (0.63–3.67)
Ref. 1.26 (0.87–1.82) 1.36 (075–2.46) 0.80 (0.56–1.13) 1.28 (0.82–1.99) 1.44 (0.67–3.07)
Ref. 1.37 (0.58–3.23) 1.34 (0.30–5.93) 2.23 (0.61–2.49) 6.78 (2.71–17.0) 4.64 (1.09–19.8)
Ref. 1.85 (1.14–3.02) 1.05 (0.39–2.86) 0.67 (0.40–1.10) 2.17 (1.07–4.43) 2.37 (0.88–6.41)
Ref. 0.97 (0.56–1.70) 1.75 (0.78–3.92) 1.41 (0.86–2.31) 0.99 (0.52–1.88) 0.67 (0.18–2.48)
Ref. 0.84 (0.46–1.53) 1.50 (0.71–3.13) 1.02 (0.62–1.68) 0.94 (0.52–1.69) 0.65 (0.15–2.78)
a
Baseline age. Odds ratios (95% confidence intervals) adjusted for baseline age, snoring, smoking/nicotine use, alcohol dependence, physical activity, caffeine intake, shift work, anxiety, depression, diabetes medication, and baseline weight. c Normal sleepers at baseline who were short sleepers at follow-up. d Normal sleepers at baseline who were long sleepers at follow-up. b
unchanged long (aOR, 4.64; 95% CI, 4.64–19.8) sleep duration were risk factors for general obesity after adjustments. In women aged ≥40 years, there were no significant associations between change in sleep duration over time and general obesity after adjustments (Table 2). When using weight gain ≥10 kg as the dependent variable, habitual short sleepers in the younger age group also had a higher risk of weight gain compared with habitual normal sleepers (aOR, 2.17; 95% CI, 1.07–4.43). A decrease in sleep duration was also seen as a risk factor for weight gain ≥10 kg in the younger age group (aOR, 1.85; 95% CI, 1.14–3.02). There were no significant associations in women ≥40 years (Table 2). Also when using weight change as a continuous dependent variable, habitual short sleep duration (adjusted β = 4.09; 95% CI, 1.98–6.20) and habitual long sleep duration (adjusted β = 5.31; 95% CI, 2.20–8.42) were risk factors in younger women. 3.2.2. Central obesity and increased waist circumference Habitual long sleep duration (aOR, 6.05; 95% CI 1.19–30.7) was a risk factor for central obesity in younger women (aged <40 years), and decreased sleep duration was borderline significant for central obesity in this age group (aOR, 1.60; 95% CI, 0.95–2.69) (Table 3). When using an increase in waist circumference by ≥10 cm as the dependent variable, both habitual short sleepers (aOR 2.83; 95% CI, 1.43–5.60) and long sleepers (aOR, 2.80; 1.01–7.76) in the younger age group had a higher risk of increased waist circumference compared with habitual normal sleepers. A decrease in sleep duration was found to be a risk factor for an increase in waist circumference by ≥10 cm in the younger age group (aOR, 1.84; 95% CI, 1.20– 2.81). There were no significant associations in women ≥40 years (Table 3). When using change in waist circumference as a contin-
uous dependent variable, habitual long sleep duration (adjusted β = 5.31; 95% CI, 2.20–8.42) was a risk factor in younger women. There were no significant associations between change in sleep duration and either central obesity measured as waist circumference ≥88 cm or increase of waist circumference ≥10 cm in women ≥40 years. 4. Discussion This 10-year follow-up study in women shows that both habitual short and long sleep duration are risk factors for general and central obesity in women aged <40 years. In younger women, changing from being a normal sleeper to being a short sleeper over the follow-up period was a risk factor for increasing weight and waist circumference. The study also shows that, in women aged <40 years, both short and long sleepers have a higher prevalence of general and central obesity and that, in the same age group, short sleepers whose sleep had ‘normalized’ still had a higher prevalence of general and central obesity compared with habitual normal sleepers. In the older age group, similar results are seen for long sleepers whose sleep duration has ‘normalized’ from baseline to follow-up. However, the prevalence of obesity is not as high as in women who are habitual short or long sleepers, indicating that effects of short or long sleep duration may be present even after sleep duration ‘normalizes’. Over the last decade, longitudinal studies of sleep duration and obesity in adults have produced varying results. Several studies have shown a U-shaped relationship [11,18] or a negative relationship between sleep duration and obesity [1–3,19] over time. In an epidemiological study by Hairston et al. using computed tomography to assess visceral adipose tissue and subcutaneous adipose tissue,
Table 3 Change in sleep duration in relation to central obesity and changes in waist circumference (WC).
Unchanged normal sleep duration (n = 3696) Decreased sleep duration (n = 309)c Increased sleep duration (n = 123)d ‘Normalized’ sleep duration (n = 453) Unchanged short sleep duration (n = 243) Unchanged long sleep duration (n = 60) a
All women (n = 4903)
Women aged <40 yearsa (n = 2088)
Women aged ≥40 yearsa (n = 2815)
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
Ref. 0.81 (0.57–1.15) 1.03 (0.60–1.77) 1.00 (0.75–1.34) 1.26 (0.85–1.87) 2.05 (0.91; 4.59)
Ref. 1.04 (0.78–1.38) 1.23 (0.78–1.95) 1.00 (0.78–1.27) 1.24 (0.89–1.72) 1.46 (0.79–2.71)
Ref. 1.60 (0.95–2.69) 0.81 (0.30–2.20) 0.99 (0.65–1.53) 1.93 (0.87–4.81) 6.05 (1.19–30.7)
Ref. 1.84 (1.20–2.81) 1.17 (0.53–2.58) 1.09 (0.76–1.55) 2.83 (1.43–5.60) 2.80 (1.01–7.76)
Ref. 0.51 (0.33–0.80) 1.25 (0.63–2.44) 1.09 (0.73–1.62) 1.11 (0.71–1.75) 1.33 (0.52–3.43)
Ref. 0.66 (0.43–1.00) 1.23 (0.70–2.18) 0.99 (0.71–1.39) 0.94 (0.63–1.39) 0.94 (0.40–2.18)
Baseline age. Adjusted for baseline age, snoring, smoking/nicotine use, alcohol dependency, physical activity, caffeine intake, shift work, anxiety, depression, diabetic medication and baseline waist circumference. c Normal sleepers at baseline who were short sleepers at follow-up. d Normal sleepers at baseline who were long sleepers at follow-up. b
1210
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
a U-shaped relationship was shown [18]. However, the relationship was only significant for those aged <40 years, similar to the results in the present study. Moreover, in a community-based sample including 278 men and women, Chaput et al. showed that both short and long sleepers gained more weight and were more likely to experience a 5 kg weight gain than those sleeping 7–8 h [11]. However, changes in sleeping habits were not taken into account as they have been in the present study. In addition, inverse linear relationships between sleep duration and the development of obesity have also been shown. In a prospective cohort study, Hasler et al. showed that short sleep duration was a risk factor for the development of obesity in young adults. However, the association between sleep duration and obesity diminished after the age of 34 years. There was also a trend for the average change rate for weight gain to be negatively associated with the average change rate for sleep duration [3]. Furthermore, in a large sample of women, Patel et al. argued that women who slept 5–6 h or less gained more weight over 16 years than those sleeping 7 h. There was also a higher risk of weight gain and incident obesity among women who slept 5 and 6 h, after adjusting for important covariates, including activity or dietary consumption [1]. Also in the present study habitual short sleep as well as decreased sleep duration were independent risk factors for an increase in both weight and waist circumference. Four longitudinal studies examining associations between sleep and obesity in adults have indicated that there is no association between sleep duration and obesity [20–23]. In the study by Lauderdale et al. [21], only linear regression was used and this could have made it possible to miss a U-shaped relationship. In the studies by Stranges et al. [20], the participants’ baseline age was somewhat higher compared with the present study and, in the study by Appelhans et al. [23], the number of participants was somewhat lower, which could explain the discrepancy between these studies and studies showing longitudinal associations between sleep duration and obesity. Marshall et al. [22] assessed changes in sleep duration and subsequent weight gain in a cohort of obese patients, not a population-based cohort, making a comparison with the results in the present study somewhat difficult. There are several possible pathways via which sleep may affect the development of obesity. In previous studies, short sleep duration has been associated with increased hunger [24–26], as well as irregular eating habits, snacking between meals, and the reduced intake of vegetables [27]. In a community-based sample, short sleep has been associated with certain obesity-related behaviors, such as lower physical activity level and lower fruit and vegetable consumption [28], and it has been argued that factors including short sleep duration, high disinhibition eating behavior, high susceptibility to hunger behavior, non-participation in high-intensity physical exercise, high dietary restraint behavior, high dietary lipid intake, and high alcohol intake are all significantly associated with overweight and obesity in the cross-sectional sample [29]. Over the sixyear follow-up period, short-duration sleepers and those with a high disinhibition and restraint eating behavior score were significantly more likely to gain weight and develop obesity [29]. In addition, disrupted eating patterns, such as having a dominance of snacks over meals, a higher intake of fat and sweets and a lower intake of fruit and vegetables, are associated with both short and long sleep duration [30]. Furthermore, in a randomized study of adults, sleep curtailment was shown to undermine dietary efforts to reduce adiposity, as it reduced the fraction of weight loss by 55% and increased the loss of fat-free body mass by 60% [31]. Explanations for the association between long sleep and obesity are sparse. However, it has been argued that long sleepers are characterized by low energy expenditure due to the longer time in bed, but also that long sleepers may overestimate their actual sleep duration because of the long time in bed, and that they may have poor
sleep quality due to sleep disorders or other health issues [11]. In the present study, we adjusted our models for physical activity, snoring, and some health issues; in spite of this, the associations between long sleep duration and obesity remained. In the present study, as well as in some previous studies [18], the association between sleep duration and obesity or weight gain is strongest or only seen in younger subjects. A review article by Magee et al. [32] discusses whether this can be explained by the fact that the possible physiological and behavioral mechanisms underlying this relationship are strongest at younger ages; furthermore, one possible explanation for the lack of significant associations between short sleep and weight gain observed in some studies is that individuals who are short sleepers do not continue to gain weight linearly over the course of their short sleeping. In other words, if a pattern of short sleep duration had started some years before the study and the change in sleep duration resulted in a certain net increase in calories per day, by the time the study began, the subject would have reached a weight threshold and would not have gained any more weight during the study. This may also contribute to studies being less likely to show an association between sleep duration and weight gain in older people [32]. This study was conducted in a large population of women and had a high response rate both at baseline and at follow-up. Nonetheless, there are some limitations to consider when interpreting the results. All the data are self-reported and could therefore be subject to bias. Furthermore, the fact that the study only included women may limit the generalizability of our results. On the other hand, a study comprising only women may also present less of an issue in terms of a post-hoc decision to stratify by gender as a driver for the conclusions. The population representativeness of the sample can also be useful when it comes to generalizing the results to apply to public health in the context of previous observations in cohorts that are not created to be population representative. In summary, both habitual short and long sleep duration as well as decreased sleep duration are risk factors for obesity and increase in weight or waist circumference in younger women, showing that sleep duration is associated with the development of obesity, especially in young women. In addition, short sleepers and long sleepers who improve their sleep (i.e. who become normal sleepers) still show a higher prevalence of general or central obesity compared with habitual normal sleepers, even if it is lower than that for habitual short or long sleepers. These results indicate that effects of short or long sleep duration may be present even after becoming a normal sleeper. It is clinically important to identify younger women who are short and long sleepers, and to advise them accordingly. Funding sources Swedish Heart-Lung Foundation and Uppsala County Association against Heart and Lung Diseases. Conflict of interest None declared. The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2014.02.014. References [1] Patel SR, Malhotra A, White DP, Gottlieb DJ, Hu FB. Association between reduced sleep and weight gain in women. Am J Epidemiol 2006;164:947–54. [2] Gangwisch JE, Malaspina D, Boden-Albala B, Heymsfield SB. Inadequate sleep as a risk factor for obesity: analyses of the NHANES I. Sleep 2005;28:1289–96.
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
[3] Hasler G, Buysse DJ, Klaghofer R, Gamma A, Ajdacic V, Eich D, et al. The association between short sleep duration and obesity in young adults: a 13-year prospective study. Sleep 2004;27:661–6. [4] Kohatsu ND, Tsai R, Young T, Vangilder R, Burmeister LF, Stromquist AM, et al. Sleep duration and body mass index in a rural population. Arch Intern Med 2006;166:1701–5. [5] Vorona RD, Winn MP, Babineau TW, Eng BP, Feldman HR, Ware JC. Overweight and obese patients in a primary care population report less sleep than patients with a normal body mass index. Arch Intern Med 2005;165: 25–30. [6] 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. [7] Singh M, Drake CL, Roehrs T, Hudgel DW, Roth T. The association between obesity and short sleep duration: a population-based study. J Clin Sleep Med 2005;1:357–63. [8] Vgontzas AN, Lin HM, Papaliaga M, Calhoun S, Vela-Bueno A, Chrousos GP, et al. Short sleep duration and obesity: the role of emotional stress and sleep disturbances. Int J Obes (Lond) 2008;32:801–9. [9] Lauderdale DS, Knutson KL, Yan LL, Rathouz PJ, Hulley SB, Sidney S, et al. Objectively measured sleep characteristics among early-middle-aged adults: the CARDIA study. Am J Epidemiol 2006;164:5–16. [10] Chaput JP, Brunet M, Tremblay A. Relationship between short sleeping hours and childhood overweight/obesity: results from the Quebec en Formé Project. Int J Obes (Lond) 2006;30:1080–5. [11] Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between sleep duration and weight gain in adults: a 6-year prospective study from the Quebec Family Study. Sleep 2008;31:517–23. [12] Theorell-Haglöw J, Lindberg E, Janson C. What are the important risk factors for daytime sleepiness and fatigue in women? Sleep 2006;29:751–7. [13] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of The National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III). JAMA 2001;285:2486–97. [14] Lissner L, Bengtsson C, Björkelund C, Wedel H. Physical activity levels and changes in relation to longevity. A prospective study of Swedish women. Am J Epidemiol 1996;143:54–62. [15] Beresford TP, Blow FC, Hill E, Singer K, Lucey MR. Comparison of CAGE questionnaire and computer-assisted laboratory profiles in screening for covert alcoholism. Lancet 1990;336:482–5. [16] Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. [17] Theorell-Haglöw J, Berne C, Janson C, Sahlin C, Lindberg E. Associations between short sleep duration and central obesity in women. Sleep 2010;33:593–8.
1211
[18] Hairston KG, Bryer-Ash M, Norris JM, Haffner S, Bowden DW, Wagenknecht LE. Sleep duration and five-year abdominal fat accumulation in a minority cohort: the IRAS family study. Sleep 2010;33:289–95. [19] Nishiura C, Hashimoto H. A 4-year study of the association between short sleep duration and change in body mass index in Japanese male workers. J Epidemiol 2010;20:385–90. [20] Stranges S, Cappuccio FP, Kandala NB, Miller MA, Taggart FM, Kumari M, et al. Cross-sectional versus prospective associations of sleep duration with changes in relative weight and body fat distribution: the Whitehall II Study. Am J Epidemiol 2008;167:321–9. [21] Lauderdale DS, Knutson KL, Rathouz PJ, Yan LL, Hulley SB, Liu K. Cross-sectional and longitudinal associations between objectively measured sleep duration and body mass index: the CARDIA Sleep Study. Am J Epidemiol 2009;170:805– 13. [22] Marshall NS, Grunstein RR, Peltonen M, Stenlof K, Hedner J, Sjostrom L. Changes in sleep duration and changes in weight in obese patients: the Swedish Obese Subjects Study. Sleep Biol Rhythms 2010;8:63–71. [23] Appelhans BM, Janssen I, Cursio JF, Matthews KA, Hall M, Gold EB, et al. Sleep duration and weight change in midlife women: the SWAN sleep study. Obesity (Silver Spring) 2013;21:77–84. [24] Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med 2004;141:846–50. [25] Schmid SM, Hallschmid M, Jauch-Chara K, Born J, Schultes B. A single night of sleep deprivation increases ghrelin levels and feelings of hunger in normalweight healthy men. J Sleep Res 2008;17:331–4. [26] Spaeth AM, Dinges DF, Goel N. Effects of experimental sleep restriction on weight gain, caloric intake, and meal timing in healthy adults. Sleep 2013;36:981–90. [27] Ohida T, Kamal AM, Uchiyama M, Kim K, Takemura S, Sone T, et al. The influence of lifestyle and health status factors on sleep loss among the Japanese general population. Sleep 2001;24:333–8. [28] Stamatakis KA, Brownson RC. Sleep duration and obesity-related risk factors in the rural Midwest. Prev Med 2008;46:439–44. [29] Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between short sleep duration and weight gain is dependent on disinhibited eating behavior in adults. Sleep 2011;34:1291–7. [30] Kim J, Jo I. Age-dependent association between sleep duration and hypertension in the adult Korean population. Am J Hypertens 2010;23:1286–91. [31] Nedeltcheva AV, Kilkus JM, Imperial J, Schoeller DA, Penev PD. Insufficient sleep undermines dietary efforts to reduce adiposity. Ann Intern Med 2010;153:435–41. [32] Magee L, Hale L. Longitudinal associations between sleep duration and subsequent weight gain: a systematic review. Sleep Med Rev 2012;16:231–41.
Contents lists available at ScienceDirect
Sleep Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s l e e p
Original Article
Both habitual short sleepers and long sleepers are at greater risk of obesity: a population-based 10-year follow-up in women Jenny Theorell-Haglöw a,*, Lars Berglund b, Christian Berne c, Eva Lindberg a a b c
Department of Medical Sciences, Respiratory Medicine and Allergology, Uppsala University, Sweden Uppsala Clinical Research Center (UCR), Uppsala University, Sweden Department of Medical Sciences, Internal Medicine, Uppsala University, Sweden
A R T I C L E
I N F O
Article history: Received 22 August 2013 Received in revised form 2 February 2014 Accepted 4 February 2014 Available online 12 June 2014 Keywords: Sleep duration Obesity Changed sleep duration Longitudinal Population-based Women
A B S T R A C T
Objective: To assess how change in sleep duration is related to subsequent obesity. Methods: In this 10-year follow-up, 4903 non-pregnant participants answered a questionnaire on sleeping habits, obesity, and lifestyle factors (questions identical to baseline questionnaire). Habitual normal sleepers were defined as sleeping 6–9 h/night at both baseline and follow-up, whereas women sleeping <6 h/night or ≥9 h/night at both occasions were defined as habitual short sleepers and habitual long sleepers, respectively. Logistic regression was used to analyze associations between changes in sleep duration, general obesity (body mass index ≥30 kg/m2), weight gain (≥10 kg) and also, central obesity (waist circumference ≥88 cm), and increase in waist circumference (≥10 cm) at follow-up. Results: Among younger women (aged <40 years) both habitual short sleepers and habitual long sleepers had a higher prevalence of general (short: 31.3%, P < 0.0001; long: 38.1%, P = 0.01) and central obesity (short: 60.5%, P = 0.01; long: 82.4%, P = 0.01) compared with habitual normal sleepers (general obesity: 8.9%; central obesity: 35.9%) at follow-up. Younger women who were short sleepers at baseline but normal sleepers at the follow-up had a higher prevalence of both general (19.3%, P = 0.01) and central obesity (45.4%, P = 0.07) compared with habitual normal sleepers at follow-up. In adjusted analyses, both habitual short [adjusted odds ratio (aOR), 6.78; 95% confidence interval (CI), 2.71–17.0] and long (aOR, 4.64; 95% CI, 1.09–19.8) sleep durations were risk factors for general obesity in younger women. In younger women habitual long sleep duration was a risk factor also for central obesity (aOR, 6.05; 95% CI, 1.19– 30.7) whereas habitual short sleep duration was not (aOR, 1.93; 95% CI, 0.87–4.81). Similar results were seen also for weight gain and increased waist circumference as dependent variables. In addition, decreased sleep duration from normal to short duration was a risk factor for both weight gain (aOR, 1.85; 95% CI, 1.14–3.02) and increased waist circumference (aOR, 1.84; 95% CI, 1.20–2.81). There were no associations between changes in sleep duration and any of the measures of obesity at the follow-up in women aged >40 years at baseline. Conclusion: In younger women, both habitual short and long sleep duration was a risk factor for obesity, whereas no such relationship was seen in older women. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Over the last decade, several cross-sectional studies have shown associations between sleep duration, sleeping habits, and measurements of obesity [1–9]. An inverse relationship between selfreported sleep duration and obesity is found in adult men and women, as well as in children [10]. Short sleep duration is also found to predict weight gain and obesity later in life [2,3].
* Corresponding author at: Department of Medical Sciences, Respiratory Medicine and Allergology, Uppsala University, SE-751 85 Uppsala, Sweden. Tel.: +46 18 611 02 42; fax: +46 18 611 02 28. E-mail address: [email protected] (J. Theorell-Haglöw). http://dx.doi.org/10.1016/j.sleep.2014.02.014 1389-9457/© 2014 Elsevier B.V. All rights reserved.
The results of previous studies on the longitudinal relationship between sleep duration and obesity vary, with some studies only finding a relationship between short sleep duration and obesity [3], some revealing a U-shaped relationship [11] and some showing no relationship [9]. However, sleep duration is not necessarily fixed over time, and studies of the relationship between changes in sleep duration over time and the development of obesity are still sparse and have produced diverse results. This project is part of a 10-year follow-up of the epidemiological population-based study entitled ‘Sleep and Health in women’ (SHE) [12]. The overall aim of the SHE study is to analyze the longterm evolution and consequences of sleeping habits and sleep disturbances. The aim of the present study was to assess how changes in sleep duration over time were related to obesity over a 10-year
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
period, in a population-based sample of women. A second aim was to examine these associations stratified by age. 2. Methods 2.1. SHE questionnaire The SHE study started in 2000, when a questionnaire on sleeping habits and somatic disorders was sent to women aged ≥20 years, randomly selected from the population registry of the City of Uppsala, Sweden. The response rate was 71.6% (n = 7051). In 2010, a followup questionnaire was sent to all the women who had answered the baseline questionnaire and who were still alive. Of the original study population, 8.5% were lost at follow-up due to death (n = 461), emigration (n = 130) and unknown addresses (n = 5). The follow-up questionnaire was therefore sent to 6455 women (91.6% of the initial study population) and completed by 5193 (response rate 80.5%). In the present study, women who were pregnant or who had omitted the question on sleep duration at one or both time-points were excluded, and the final study population comprised a total of 4903 non-pregnant women (Fig. 1). The study was approved by the Ethics Committee at the Medical Faculty at Uppsala University and all the participants gave their informed consent before participating. The follow-up questionnaire included mostly the same questions as the baseline questionnaire, which has been previously described in detail [12]. Briefly, the follow-up questionnaire included questions on sleep duration, snoring habits, insomnia, occupational status, shift work, civil status, physical activity, smoking, alcohol dependence, anxiety and depression. Both at baseline and at follow-up, sleep duration (h/night) was assessed using the question: ‘How many hours do you sleep on average during the night?’ Sleep duration, at baseline and at follow-up, was classified into three categories: short sleepers (<6 h/night), normal sleepers (6– 9 h/night), and long sleepers (≥9 h/night). In addition, a sleep habit variable was constructed where the women were divided into ha-
Fig. 1. Overview of study protocol.
1205
bitual short sleepers (short sleepers at both baseline and followup), habitual long sleepers (long sleepers at both baseline and followup), women with ‘normalized’ sleep duration (short or long sleepers at baseline who were normal sleepers at follow-up), women with decreased sleep duration (normal sleepers at baseline who were short sleepers at follow-up), women with increased sleep duration (normal sleepers at baseline who were long sleepers at followup), habitual normal sleepers (normal sleep duration at both baseline and follow-up), as well as a group of women changing from short to long sleep duration or from long to short sleep duration. Snoring habits were assessed using the question: ‘How often do you snore loudly and disturbingly?’ The response options for this question were ‘never’ (1), ‘seldom’ (2), ‘sometimes’ (3), ‘often’ (4) and ‘very often’ (5). Based on their response to this question, the participants were categorized into two groups: non-snorers (scores 1–3) and snorers (scores 4 and 5). The women further indicated their current height and weight in both questionnaires, and body mass index (BMI, kg/m2) was calculated and rounded off to one decimal point. BMI was categorized as underweight (<19.9), normal weight (20–24.9), overweight (25–29.9), and obese (≥30). In conjunction with the questionnaire, the women were given a tape measure and instructions on how to measure their waist circumference [12]. A waist circumference of ≥88 cm was used to define central obesity according to National Cholesterol Education Program (NCEP) criteria [13]. All the data are self-reported. However, for 400 of the women BMI and waist circumference have been measured by a research nurse 2 years later. In spite of the time lapse between the two measures there was good correlation between measures at the two time-points (BMI: β = 0.80, P < 0.0001; waist circumference: β = 0.85, P < 0.0001). The mean difference between measured and self-reported BMI was 0.87 ± 2.51 and the corresponding difference in waist circumference was −0.39 ± 7.85 cm. The participants’ physical activity was analyzed by four questions adopted from a questionnaire used in a large populationbased study of the correlation between physical activity and mortality [14]. Six questions assessed smoking habits. The participants were categorized as ‘current smokers’ or ‘non-smokers’ (i.e. had quit smoking at least 6 months before answering the questionnaire or had never smoked). Alcohol dependence was assessed by the CAGE (acronym for Cut down, Annoyed by criticism, Guilty about drinking and Eye-opener drinks) alcohol and screening questionnaire [15]. The women were also asked to indicate how many cups of coffee they drank every day and this number was used as a continuous variable in the analyses. Psychological distress was assessed using the Hospital Anxiety and Depression (HAD) scale [16] and was used as a dichotomized score with a cut-off at score 10. The participants were asked at both time-points to indicate how many months they had worked nights or shifts over the past 10 years. A cut-off point of 60 months was chosen when analyzing working nights or shifts. The women were asked whether they were taking any medication on a regular basis. Taking insulin or oral medication, or both, for diabetes were categorized together as ‘diabetes medication’. Compared with the responders, the non-responders at the followup were older (mean ages 45.0 ± 16.0 vs 43.3 ± 15.3 years, P < 0.001), somewhat more obese (mean BMI 24.5 ± 4.3 vs 24.0 ± 4.1, P = 0.0007), and more often smokers (21.9% vs 15.9%, P < 0.0001) at baseline. Responders had somewhat longer sleep duration (mean sleep duration 7.0 vs 6.9 h, P = 0.03) but prevalence of habitual snoring did not differ between the groups (7.6 vs 7.8%, not significant). However, the non-responders were less physically active at baseline as 16.3% reported a high level of physical activity and 23.8% reported a low level of physical activity as compared with 21.1% and 15.7%, respectively for the responders (P < 0.001).
1206
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
2.2. Statistical analyses Statistical analyses were performed using Stata 10 (Stata Corp., College Station, TX, USA). Univariate analyses at both time-points were conducted using the unpaired t-test for continuous variables and the χ2-test for categorical variables to compare the sleep duration groups. The unpaired t-test was used even if the data were not assumed to be normally distributed, as the sample size was large and, likely, the use of a Wilcoxon–Mann–Whitney test would not change the decision about rejecting the null hypothesis. Analyses comparing continuous and categorical variables respectively, between the two time-points, were also made. Variables that differed between groups (P < 0.20) were entered into multiple logistic regression models. The results of the logistic regression models are presented as adjusted odds ratios (ORs) with 95% confidence intervals (CIs) with the dependent variables: BMI ≥30 kg/m2 and weight gain ≥10 kg, respectively, and waist circumference ≥88 cm and increase in waist circumference ≥10 cm, respectively. As only 19 women in total had gone from short to long sleep duration or from long to short sleep duration, they were grouped together and were included in the multiple analyses. In a previous study [17] using a subsample of the same female cohort, we found that short sleep duration is most important with respect to impact on metabolism in younger women (aged <40 years), and there was a significant interaction between age and sleep duration for both general (P < 0.001) and central obesity (P < 0.001). For this reason, in further analyses in the present study the population was grouped according to age ≥40 and <40 years (at baseline). 3. Results Over the 10-year period, the mean sleep duration had decreased and the prevalence of both general and central obesity had
increased significantly. There was an increase in the number of women who reported snoring. Moreover, alcohol dependence increased during the study period. However, smoking and the use of snuff or other nicotine substances decreased significantly. Furthermore, there were fewer shift workers at the 10-year follow-up and there was a significant reduction in the prevalence of anxiety and depression (Table 1). The majority of the women in the study were normal sleepers (i.e. slept 6–9 h/night) at both time-points, 243 were habitual short sleepers (i.e. slept <6 h/night at both time-points), and 60 women were habitual long sleepers (i.e. slept ≥9 h/night at both timepoints). Of the women, 293 had changed from short to normal sleep and 160 had changed from long to normal sleep, making a total of 453 women whose sleep duration had ‘normalized’. In addition, 309 women had changed from normal to short sleep and 123 had changed from normal to long sleep, making a total of 432 women whose sleep duration had either increased or decreased. Finally, eight women had changed from short to long sleep and 11 women had changed from long to short sleep duration during the 10-year period. Moreover, when the women were grouped according to age, there were similar distributions. 3.1. Changes in sleep duration and obesity At the follow-up, all women who had been short or long sleepers either at baseline or at follow-up had a higher prevalence of general obesity (i.e. BMI ≥30 kg/m2) as compared with women with normal sleep duration on both occasions (Fig. 2). When the women were analyzed by age group, both habitual short sleepers and long sleepers in the younger age group (<40 years) had a higher prevalence of general obesity at follow-up compared with habitual normal sleepers in this age group. In addition, short sleepers at baseline whose sleep had ‘normalized’ at follow-up also had a higher
Table 1 Characteristics of the population at baseline and of the 10-year follow-up group. Characteristics
Baseline All (2000) (n = 6843)
Age at baseline (years) Sleep duration (h) Sleep duration Short (<6 h/night) Normal (6–9 h/night) Long (≥9 h/night) Snoring BMI (kg/m2) Obesity (BMI ≥30 kg/m2) Waist circumference (cm) Central obesity (WC ≥88 cm) Smoking/snuff/nicotine use Alcohol dependence (CAGE ≥2) Physical activity High Medium Low Cups of coffee/day Shift work Diabetes medication Oral diabetes medication Insulin Insulin and oral diabetes medication HAD anxiety score ≥10 HAD depression score ≥10
Follow-up Follow-up group (2000) (n = 4903)
46.0 (17.5) 6.9 (1.2)
43.9 (15.2) 7.0 (1.1)
856 (12.7%) 5566 (82.2%) 346 (5.1%) 513 (7.8%) 24.2 (4.2) 595 (8.8%) 84.1 (12.8) 2157 (33.6%) 1352 (19.9%) 385 (5.9%)
P-valuea
(2010) (n = 4903) NA 6.9 (1.1)
NA 0.011
563 (11.3%) 4171 (84.0%) 234 (4.7%) 391 (8.0%) 24.1 (4.1) 425 (8.5%) 83.6 (12.5) 1503 (31.6%) 947 (19.0%) 294 (6.1%)
568 (11.5%) 4175 (84.6%) 192 (3.9%) 457 (9.4%) 25.1 (4.5) 662 (13.4%) 89.7 (12.6) 2523 (51.9%) 634 (12.8%) 360 (7.4%)
0.51 0.51 0.028 0.0013 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0004
1279 (19.0%) 4184 (62.0%) 1285 (19.0%) 4.37 (2.7%) 294 (4.3%)
785 (15.8%) 3164 (63.7%) 1018 (20.5%) 4.39 (2.8%) 234 (4.7%)
1090 (22.0%) 3146 (63.5%) 719 (14.5%) NA 157 (2.3%)
0.059 0.059 0.17 NA <0.0001
77 (1.1%) 32 (0.5%) 12 (0.2%) 936 (14.2%) 291 (4.5%)
33 (0.7%) 18 (0.4%) 4 (0.1%) 655 (13.5%) 203 (4.2%)
109 (2.2%) 9 (0.2%) 11 (0.2%) 567 (11.6%) 186 (3.8%)
<0.0001 0.0495 0.035 <0.0001 <0.0001
Values are mean (SD) or n (%). Abbreviations: BMI, body mass index; WC, waist circumference; CAGE, acronym for ‘Cut down, Annoyed by criticism, Guilty about drinking and Eye-opener drinks’ questionnaire; HAD, Hospital Anxiety and Depression scale. a For test between baseline and follow-up values in the follow-up group.
Age >=40 yrs.
40
**
30
**
20
**
10
** *
NS REF
0
Prevalence (%) of BMI>=30 kg/m2
Age <40 yrs.
1207
50
50
Prevalence (%) of BMI>=30 kg/m2
All women
Prevalence (%)of BMI>=30 kg/m2
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
50
***
40
***
30 20
**
NS
NS
10
NS REF
0
40 30
***
*
20 10
NS NS
NS
NS REF
0
* <0.05; ** <0.01; *** <0.0001 Fig. 2. Prevalence of general obesity (body mass index ≥30 kg/m2) in 2010 in relation to sleep duration in 2000 and 2010, respectively. Data are only presented for groups of ≥15 women. *P < 0.05, **P < 0.01, ***P < 0.0001 as compared with women with normal sleep duration both at baseline and at follow-up (unadjusted comparisons).
prevalence of general obesity at follow-up compared with habitual normal sleepers. In the older age group, the prevalence of general obesity was highest in long sleepers whose sleep had ‘normalized’, but this also applied to normal sleepers who had become long sleepers. The analysis was also performed in women who were nonsnorers (using the snoring question as a proxy for sleep apnea) on both occasions, but this did not significantly change the results (data not shown). There was a greater prevalence of central obesity (i.e. waist circumference ≥88 cm) at the follow-up for women in all categories of sleep duration change (Fig. 3). Again, both habitual short sleepers and long sleepers in the younger age group had a higher prevalence of central obesity at follow-up compared with habitual normal
sleepers. In this group, there was also a higher prevalence of central obesity in short sleepers whose sleep duration had ‘normalized’ over the 10-year period, even if it was not as high as for habitual short or long sleepers. In the older age group, the highest prevalence of central obesity at follow-up was found in women who were long sleepers at baseline and whose sleep had then ‘normalized’. The analyses for central obesity were also performed in women who were non-snorers on both occasions, but this did not significantly change the results (data not shown). Women aged <40 years had in general gained more weight during the 10-year period compared with women aged ≥40 years. In the younger group, the lowest change in weight was found among the habitual normal sleepers with a mean weight gain of 3.5 ± 6.9 kg.
Age <40 yrs.
Age >=40 yrs.
100 90 80 70 60 50 40 30 20 10 0
100 90 80 70 60 50 40 30 20 10 0
Prevalence (%) of WC>=88 cm
All women
Prevalence (%) of WC>=88 cm
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
Prevalence (%) of WC>=88 cm
1208
100 90 80 70 60 50 40 30 20 10 0
*** ***
*
*
NS
NS
REF
* * NS NS
NS
NS
REF
NS NS
NS NS
* NS
REF
* <0.05; ** <0.01; *** <0.0001 Fig. 3. Prevalence of central obesity (waist circumference ≥88 cm) in 2010 in relation to sleep duration in 2000 and 2010 respectively. Data are only presented for groups of ≥15 women. *P < 0.05, **P < 0.01, ***P < 0.0001 as compared with women with normal sleep duration both at baseline and at follow-up (unadjusted comparisons).
Habitual long and short sleepers had gained the most in weight with a mean change of 8.1 ± 8.8 and 7.5 ± 8.3 kg, respectively. In women aged ≥40 years the habitual normal sleepers increased their weight with 1.7 ± 7.4 kg and the corresponding weight changes in habitual long and short sleepers were 1.7 ± 7.7 and 0.5 ± 7.4 kg. Differences between younger and older women were less pronounced when it came to changes in waist circumference over time. In the younger group, habitual normal sleepers reported a waist circumference at the follow-up that was in general 6.9 ± 8.8 cm longer compared with baseline. The corresponding differences for habitual long and short sleepers were 12.0 ± 8.0 and 9.3 ± 12.3 cm. In the older group, the mean changes in waist circumference were 5.5 ± 8.3,
4.8 ± 7.7, and 4.2 ± 9.0 cm for habitual normal, long, and short sleepers, respectively. 3.2. Multiple analyses 3.2.1. General obesity and weight gain In the whole population, unchanged short sleep duration was a risk factor for general obesity also when adjusting for age, snoring, smoking/nicotine use, alcohol dependence, physical activity, caffeine intake, shift work, anxiety, depression, diabetes medication, and baseline weight (OR, 1.71; 95% CI, 1.01–2.87). In younger women (<40 years) both unchanged short (aOR, 6.78; 95% CI, 2.71–17.0) and
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
1209
Table 2 Change in sleep duration in relation to general obesity and weight gain. All women (n = 4903)
Unchanged normal sleep duration (n = 3696) Decreased sleep duration (n = 309)c Increased sleep duration (n = 123)d ‘Normalized’ sleep duration (n = 453) Unchanged short sleep duration (n = 243) Unchanged long sleep duration (n = 60)
Women aged <40 yearsa (n = 2088)
Women aged ≥40 yearsa (n = 2815)
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
BMI ≥30 kg/m2 at follow-upb
Weight gain ≥10 kg at follow-upb
Ref. 1.09 (0.68–1.73) 1.59 (0.79–3.21) 1.36 (0.92–2.03) 1.71 (1.01–2.87) 1.52 (0.63–3.67)
Ref. 1.26 (0.87–1.82) 1.36 (075–2.46) 0.80 (0.56–1.13) 1.28 (0.82–1.99) 1.44 (0.67–3.07)
Ref. 1.37 (0.58–3.23) 1.34 (0.30–5.93) 2.23 (0.61–2.49) 6.78 (2.71–17.0) 4.64 (1.09–19.8)
Ref. 1.85 (1.14–3.02) 1.05 (0.39–2.86) 0.67 (0.40–1.10) 2.17 (1.07–4.43) 2.37 (0.88–6.41)
Ref. 0.97 (0.56–1.70) 1.75 (0.78–3.92) 1.41 (0.86–2.31) 0.99 (0.52–1.88) 0.67 (0.18–2.48)
Ref. 0.84 (0.46–1.53) 1.50 (0.71–3.13) 1.02 (0.62–1.68) 0.94 (0.52–1.69) 0.65 (0.15–2.78)
a
Baseline age. Odds ratios (95% confidence intervals) adjusted for baseline age, snoring, smoking/nicotine use, alcohol dependence, physical activity, caffeine intake, shift work, anxiety, depression, diabetes medication, and baseline weight. c Normal sleepers at baseline who were short sleepers at follow-up. d Normal sleepers at baseline who were long sleepers at follow-up. b
unchanged long (aOR, 4.64; 95% CI, 4.64–19.8) sleep duration were risk factors for general obesity after adjustments. In women aged ≥40 years, there were no significant associations between change in sleep duration over time and general obesity after adjustments (Table 2). When using weight gain ≥10 kg as the dependent variable, habitual short sleepers in the younger age group also had a higher risk of weight gain compared with habitual normal sleepers (aOR, 2.17; 95% CI, 1.07–4.43). A decrease in sleep duration was also seen as a risk factor for weight gain ≥10 kg in the younger age group (aOR, 1.85; 95% CI, 1.14–3.02). There were no significant associations in women ≥40 years (Table 2). Also when using weight change as a continuous dependent variable, habitual short sleep duration (adjusted β = 4.09; 95% CI, 1.98–6.20) and habitual long sleep duration (adjusted β = 5.31; 95% CI, 2.20–8.42) were risk factors in younger women. 3.2.2. Central obesity and increased waist circumference Habitual long sleep duration (aOR, 6.05; 95% CI 1.19–30.7) was a risk factor for central obesity in younger women (aged <40 years), and decreased sleep duration was borderline significant for central obesity in this age group (aOR, 1.60; 95% CI, 0.95–2.69) (Table 3). When using an increase in waist circumference by ≥10 cm as the dependent variable, both habitual short sleepers (aOR 2.83; 95% CI, 1.43–5.60) and long sleepers (aOR, 2.80; 1.01–7.76) in the younger age group had a higher risk of increased waist circumference compared with habitual normal sleepers. A decrease in sleep duration was found to be a risk factor for an increase in waist circumference by ≥10 cm in the younger age group (aOR, 1.84; 95% CI, 1.20– 2.81). There were no significant associations in women ≥40 years (Table 3). When using change in waist circumference as a contin-
uous dependent variable, habitual long sleep duration (adjusted β = 5.31; 95% CI, 2.20–8.42) was a risk factor in younger women. There were no significant associations between change in sleep duration and either central obesity measured as waist circumference ≥88 cm or increase of waist circumference ≥10 cm in women ≥40 years. 4. Discussion This 10-year follow-up study in women shows that both habitual short and long sleep duration are risk factors for general and central obesity in women aged <40 years. In younger women, changing from being a normal sleeper to being a short sleeper over the follow-up period was a risk factor for increasing weight and waist circumference. The study also shows that, in women aged <40 years, both short and long sleepers have a higher prevalence of general and central obesity and that, in the same age group, short sleepers whose sleep had ‘normalized’ still had a higher prevalence of general and central obesity compared with habitual normal sleepers. In the older age group, similar results are seen for long sleepers whose sleep duration has ‘normalized’ from baseline to follow-up. However, the prevalence of obesity is not as high as in women who are habitual short or long sleepers, indicating that effects of short or long sleep duration may be present even after sleep duration ‘normalizes’. Over the last decade, longitudinal studies of sleep duration and obesity in adults have produced varying results. Several studies have shown a U-shaped relationship [11,18] or a negative relationship between sleep duration and obesity [1–3,19] over time. In an epidemiological study by Hairston et al. using computed tomography to assess visceral adipose tissue and subcutaneous adipose tissue,
Table 3 Change in sleep duration in relation to central obesity and changes in waist circumference (WC).
Unchanged normal sleep duration (n = 3696) Decreased sleep duration (n = 309)c Increased sleep duration (n = 123)d ‘Normalized’ sleep duration (n = 453) Unchanged short sleep duration (n = 243) Unchanged long sleep duration (n = 60) a
All women (n = 4903)
Women aged <40 yearsa (n = 2088)
Women aged ≥40 yearsa (n = 2815)
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
WC ≥88 cm at follow-upb
WC increase ≥10 cm at follow-upb
Ref. 0.81 (0.57–1.15) 1.03 (0.60–1.77) 1.00 (0.75–1.34) 1.26 (0.85–1.87) 2.05 (0.91; 4.59)
Ref. 1.04 (0.78–1.38) 1.23 (0.78–1.95) 1.00 (0.78–1.27) 1.24 (0.89–1.72) 1.46 (0.79–2.71)
Ref. 1.60 (0.95–2.69) 0.81 (0.30–2.20) 0.99 (0.65–1.53) 1.93 (0.87–4.81) 6.05 (1.19–30.7)
Ref. 1.84 (1.20–2.81) 1.17 (0.53–2.58) 1.09 (0.76–1.55) 2.83 (1.43–5.60) 2.80 (1.01–7.76)
Ref. 0.51 (0.33–0.80) 1.25 (0.63–2.44) 1.09 (0.73–1.62) 1.11 (0.71–1.75) 1.33 (0.52–3.43)
Ref. 0.66 (0.43–1.00) 1.23 (0.70–2.18) 0.99 (0.71–1.39) 0.94 (0.63–1.39) 0.94 (0.40–2.18)
Baseline age. Adjusted for baseline age, snoring, smoking/nicotine use, alcohol dependency, physical activity, caffeine intake, shift work, anxiety, depression, diabetic medication and baseline waist circumference. c Normal sleepers at baseline who were short sleepers at follow-up. d Normal sleepers at baseline who were long sleepers at follow-up. b
1210
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
a U-shaped relationship was shown [18]. However, the relationship was only significant for those aged <40 years, similar to the results in the present study. Moreover, in a community-based sample including 278 men and women, Chaput et al. showed that both short and long sleepers gained more weight and were more likely to experience a 5 kg weight gain than those sleeping 7–8 h [11]. However, changes in sleeping habits were not taken into account as they have been in the present study. In addition, inverse linear relationships between sleep duration and the development of obesity have also been shown. In a prospective cohort study, Hasler et al. showed that short sleep duration was a risk factor for the development of obesity in young adults. However, the association between sleep duration and obesity diminished after the age of 34 years. There was also a trend for the average change rate for weight gain to be negatively associated with the average change rate for sleep duration [3]. Furthermore, in a large sample of women, Patel et al. argued that women who slept 5–6 h or less gained more weight over 16 years than those sleeping 7 h. There was also a higher risk of weight gain and incident obesity among women who slept 5 and 6 h, after adjusting for important covariates, including activity or dietary consumption [1]. Also in the present study habitual short sleep as well as decreased sleep duration were independent risk factors for an increase in both weight and waist circumference. Four longitudinal studies examining associations between sleep and obesity in adults have indicated that there is no association between sleep duration and obesity [20–23]. In the study by Lauderdale et al. [21], only linear regression was used and this could have made it possible to miss a U-shaped relationship. In the studies by Stranges et al. [20], the participants’ baseline age was somewhat higher compared with the present study and, in the study by Appelhans et al. [23], the number of participants was somewhat lower, which could explain the discrepancy between these studies and studies showing longitudinal associations between sleep duration and obesity. Marshall et al. [22] assessed changes in sleep duration and subsequent weight gain in a cohort of obese patients, not a population-based cohort, making a comparison with the results in the present study somewhat difficult. There are several possible pathways via which sleep may affect the development of obesity. In previous studies, short sleep duration has been associated with increased hunger [24–26], as well as irregular eating habits, snacking between meals, and the reduced intake of vegetables [27]. In a community-based sample, short sleep has been associated with certain obesity-related behaviors, such as lower physical activity level and lower fruit and vegetable consumption [28], and it has been argued that factors including short sleep duration, high disinhibition eating behavior, high susceptibility to hunger behavior, non-participation in high-intensity physical exercise, high dietary restraint behavior, high dietary lipid intake, and high alcohol intake are all significantly associated with overweight and obesity in the cross-sectional sample [29]. Over the sixyear follow-up period, short-duration sleepers and those with a high disinhibition and restraint eating behavior score were significantly more likely to gain weight and develop obesity [29]. In addition, disrupted eating patterns, such as having a dominance of snacks over meals, a higher intake of fat and sweets and a lower intake of fruit and vegetables, are associated with both short and long sleep duration [30]. Furthermore, in a randomized study of adults, sleep curtailment was shown to undermine dietary efforts to reduce adiposity, as it reduced the fraction of weight loss by 55% and increased the loss of fat-free body mass by 60% [31]. Explanations for the association between long sleep and obesity are sparse. However, it has been argued that long sleepers are characterized by low energy expenditure due to the longer time in bed, but also that long sleepers may overestimate their actual sleep duration because of the long time in bed, and that they may have poor
sleep quality due to sleep disorders or other health issues [11]. In the present study, we adjusted our models for physical activity, snoring, and some health issues; in spite of this, the associations between long sleep duration and obesity remained. In the present study, as well as in some previous studies [18], the association between sleep duration and obesity or weight gain is strongest or only seen in younger subjects. A review article by Magee et al. [32] discusses whether this can be explained by the fact that the possible physiological and behavioral mechanisms underlying this relationship are strongest at younger ages; furthermore, one possible explanation for the lack of significant associations between short sleep and weight gain observed in some studies is that individuals who are short sleepers do not continue to gain weight linearly over the course of their short sleeping. In other words, if a pattern of short sleep duration had started some years before the study and the change in sleep duration resulted in a certain net increase in calories per day, by the time the study began, the subject would have reached a weight threshold and would not have gained any more weight during the study. This may also contribute to studies being less likely to show an association between sleep duration and weight gain in older people [32]. This study was conducted in a large population of women and had a high response rate both at baseline and at follow-up. Nonetheless, there are some limitations to consider when interpreting the results. All the data are self-reported and could therefore be subject to bias. Furthermore, the fact that the study only included women may limit the generalizability of our results. On the other hand, a study comprising only women may also present less of an issue in terms of a post-hoc decision to stratify by gender as a driver for the conclusions. The population representativeness of the sample can also be useful when it comes to generalizing the results to apply to public health in the context of previous observations in cohorts that are not created to be population representative. In summary, both habitual short and long sleep duration as well as decreased sleep duration are risk factors for obesity and increase in weight or waist circumference in younger women, showing that sleep duration is associated with the development of obesity, especially in young women. In addition, short sleepers and long sleepers who improve their sleep (i.e. who become normal sleepers) still show a higher prevalence of general or central obesity compared with habitual normal sleepers, even if it is lower than that for habitual short or long sleepers. These results indicate that effects of short or long sleep duration may be present even after becoming a normal sleeper. It is clinically important to identify younger women who are short and long sleepers, and to advise them accordingly. Funding sources Swedish Heart-Lung Foundation and Uppsala County Association against Heart and Lung Diseases. Conflict of interest None declared. The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2014.02.014. References [1] Patel SR, Malhotra A, White DP, Gottlieb DJ, Hu FB. Association between reduced sleep and weight gain in women. Am J Epidemiol 2006;164:947–54. [2] Gangwisch JE, Malaspina D, Boden-Albala B, Heymsfield SB. Inadequate sleep as a risk factor for obesity: analyses of the NHANES I. Sleep 2005;28:1289–96.
J. Theorell-Haglöw et al./Sleep Medicine 15 (2014) 1204–1211
[3] Hasler G, Buysse DJ, Klaghofer R, Gamma A, Ajdacic V, Eich D, et al. The association between short sleep duration and obesity in young adults: a 13-year prospective study. Sleep 2004;27:661–6. [4] Kohatsu ND, Tsai R, Young T, Vangilder R, Burmeister LF, Stromquist AM, et al. Sleep duration and body mass index in a rural population. Arch Intern Med 2006;166:1701–5. [5] Vorona RD, Winn MP, Babineau TW, Eng BP, Feldman HR, Ware JC. Overweight and obese patients in a primary care population report less sleep than patients with a normal body mass index. Arch Intern Med 2005;165: 25–30. [6] 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. [7] Singh M, Drake CL, Roehrs T, Hudgel DW, Roth T. The association between obesity and short sleep duration: a population-based study. J Clin Sleep Med 2005;1:357–63. [8] Vgontzas AN, Lin HM, Papaliaga M, Calhoun S, Vela-Bueno A, Chrousos GP, et al. Short sleep duration and obesity: the role of emotional stress and sleep disturbances. Int J Obes (Lond) 2008;32:801–9. [9] Lauderdale DS, Knutson KL, Yan LL, Rathouz PJ, Hulley SB, Sidney S, et al. Objectively measured sleep characteristics among early-middle-aged adults: the CARDIA study. Am J Epidemiol 2006;164:5–16. [10] Chaput JP, Brunet M, Tremblay A. Relationship between short sleeping hours and childhood overweight/obesity: results from the Quebec en Formé Project. Int J Obes (Lond) 2006;30:1080–5. [11] Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between sleep duration and weight gain in adults: a 6-year prospective study from the Quebec Family Study. Sleep 2008;31:517–23. [12] Theorell-Haglöw J, Lindberg E, Janson C. What are the important risk factors for daytime sleepiness and fatigue in women? Sleep 2006;29:751–7. [13] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of The National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III). JAMA 2001;285:2486–97. [14] Lissner L, Bengtsson C, Björkelund C, Wedel H. Physical activity levels and changes in relation to longevity. A prospective study of Swedish women. Am J Epidemiol 1996;143:54–62. [15] Beresford TP, Blow FC, Hill E, Singer K, Lucey MR. Comparison of CAGE questionnaire and computer-assisted laboratory profiles in screening for covert alcoholism. Lancet 1990;336:482–5. [16] Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. [17] Theorell-Haglöw J, Berne C, Janson C, Sahlin C, Lindberg E. Associations between short sleep duration and central obesity in women. Sleep 2010;33:593–8.
1211
[18] Hairston KG, Bryer-Ash M, Norris JM, Haffner S, Bowden DW, Wagenknecht LE. Sleep duration and five-year abdominal fat accumulation in a minority cohort: the IRAS family study. Sleep 2010;33:289–95. [19] Nishiura C, Hashimoto H. A 4-year study of the association between short sleep duration and change in body mass index in Japanese male workers. J Epidemiol 2010;20:385–90. [20] Stranges S, Cappuccio FP, Kandala NB, Miller MA, Taggart FM, Kumari M, et al. Cross-sectional versus prospective associations of sleep duration with changes in relative weight and body fat distribution: the Whitehall II Study. Am J Epidemiol 2008;167:321–9. [21] Lauderdale DS, Knutson KL, Rathouz PJ, Yan LL, Hulley SB, Liu K. Cross-sectional and longitudinal associations between objectively measured sleep duration and body mass index: the CARDIA Sleep Study. Am J Epidemiol 2009;170:805– 13. [22] Marshall NS, Grunstein RR, Peltonen M, Stenlof K, Hedner J, Sjostrom L. Changes in sleep duration and changes in weight in obese patients: the Swedish Obese Subjects Study. Sleep Biol Rhythms 2010;8:63–71. [23] Appelhans BM, Janssen I, Cursio JF, Matthews KA, Hall M, Gold EB, et al. Sleep duration and weight change in midlife women: the SWAN sleep study. Obesity (Silver Spring) 2013;21:77–84. [24] Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med 2004;141:846–50. [25] Schmid SM, Hallschmid M, Jauch-Chara K, Born J, Schultes B. A single night of sleep deprivation increases ghrelin levels and feelings of hunger in normalweight healthy men. J Sleep Res 2008;17:331–4. [26] Spaeth AM, Dinges DF, Goel N. Effects of experimental sleep restriction on weight gain, caloric intake, and meal timing in healthy adults. Sleep 2013;36:981–90. [27] Ohida T, Kamal AM, Uchiyama M, Kim K, Takemura S, Sone T, et al. The influence of lifestyle and health status factors on sleep loss among the Japanese general population. Sleep 2001;24:333–8. [28] Stamatakis KA, Brownson RC. Sleep duration and obesity-related risk factors in the rural Midwest. Prev Med 2008;46:439–44. [29] Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between short sleep duration and weight gain is dependent on disinhibited eating behavior in adults. Sleep 2011;34:1291–7. [30] Kim J, Jo I. Age-dependent association between sleep duration and hypertension in the adult Korean population. Am J Hypertens 2010;23:1286–91. [31] Nedeltcheva AV, Kilkus JM, Imperial J, Schoeller DA, Penev PD. Insufficient sleep undermines dietary efforts to reduce adiposity. Ann Intern Med 2010;153:435–41. [32] Magee L, Hale L. Longitudinal associations between sleep duration and subsequent weight gain: a systematic review. Sleep Med Rev 2012;16:231–41.