Long sleep duration predicts a higher risk of obesity in adults: A meta-analysis of prospective cohort studies

Abstract

Background:
The connections between long sleep duration and obesity or weight gain warrant further examination. This meta-analysis aimed to evaluate whether long sleep duration was associated with the risk of obesity, weight gain, body mass index (BMI) change or weight change in adults.

Methods:
PubMed, Embase, Cochrane Library, Elsevier Science Direct, Science Online, MEDLINE and CINAHL were searched for English articles published before May 2017. A total of 16 cohort studies (n = 329 888 participants) from 8 countries were included in the analysis. Pooled relative risks (RR) or regression coefficients (β) with 95% confidence intervals (CI) were estimated. Heterogeneity and publication bias were tested, and sensitivity analysis was also performed.

Results:
We found that long sleep duration was associated with higher risk of obesity (RR [95% CI] = 1.04 [1.00-1.09], P = 0.037), but had no significant associations with weight gain, BMI change or weight change. Long sleep duration increased the risk of weight gain in three situations: among men, in studies with <5 years follow-up, and when sleep duration was 9 or more hours.

Conclusions:
Long sleep duration was associated with risk of obesity in adults. More cohort studies with objective measures are needed to confirm this relationship.

Full text

Journal of Public Health | pp. 1–11 | doi:10.1093/pubmed/fdy135
Long sleep duration predicts a higher risk of obesity in adults:
a meta-analysis of prospective cohort studies
Wenjia Liu
1†
, Rui Zhang
2†
, Anran Tan
1
,BoYe
1
, Xinge Zhang
1
, Yueqiao Wang
1
,
Yuliang Zou
1
,LuMa
1
, Guoxun Chen
3
, Rui Li
1
, Justin B. Moore
4,5,6
1
Department of Healthcare Management, School of Health Sciences, Wuhan University, 115 Donghu Road, Wuhan 430071, China
2
College of Life Sciences, South-Central University for Nationalities, Wuhan 430074, China
3
Department of Nutrition, the University of Tennessee, Knoxville, TN 37996, USA
4
Department of Family & Community Medicine, Wake Forest School of Medicine, Medical Center Boulevard, Winston-Salem, NC 27157, USA
5
Department of Epidemiology & Prevention, Wake Forest School of Medicine, Winston-Salem, NC, USA
6
Department of Implementation Science, Wake Forest School of Medicine, Winston-Salem, NC, USA
Address correspondence to Rui Li, E-mail: rli@whu.edu.cn
ABSTRACT
Background The connections between long sleep duration and obesity or weight gain warrant further examination. This meta-analysis aimed to
evaluate whether long sleep duration was associated with the risk of obesity, weight gain, body mass index (BMI) change or weight change in adults.
Methods PubMed, Embase, Cochrane Library, Elsevier Science Direct, Science Online, MEDLINE and CINAHL were searched for English articles
published before May 2017. A total of 16 cohort studies (n=329 888 participants) from 8 countries were included in the analysis. Pooled
relative risks (RR) or regression coefficients (β) with 95% confidence intervals (CI) were estimated. Heterogeneity and publication bias were
tested, and sensitivity analysis was also performed.
Results We found that long sleep duration was associated with higher risk of obesity (RR [95% CI] =1.04 [1.00–1.09], P=0.037), but had no
significant associations with weight gain, BMI change or weight change. Long sleep duration increased the risk of weight gain in three
situations: among men, in studies with <5 years follow-up, and when sleep duration was 9 or more hours.
Conclusions Long sleep duration was associated with risk of obesity in adults. More cohort studies with objective measures are needed to
confirm this relationship.
Keywords adults, long sleep duration, meta-analysis, obesity, prospective cohort study
Introduction
From 1980 to 2013, the global prevalence of overweight and
obesity has risen by 27.5% for adults.
1
Obesity is associated
with many adverse health outcomes, such as diabetes, cancer
and stroke.
2
Although excess energy intake and reduced phys-
ical activity play a major role in weight gain and obesity,
3
short
or long sleep duration, may also be contributing factors.
4
Sleep plays a crucial role in human’s endocrine, metabolic
and neurologic functions.
5
Among the various sleep mea-
sures such as duration, quality, timing and regularity, duration
is the most frequently studied parameter related to health.
6
A
recent study reported that the prevalence of long sleep dur-
ation (>9 h) increased in Australia, Finland, Sweden, the UK
and the USA while short sleep duration (≤6 h) only increased
in Italy and Norway, indicating long sleep duration was per-
haps more widespread than short sleep duration globally.
7
†
These authors contributed equally to this work.
Wenjia Liu, Graduate Student
Rui Zhang, Research Assistant
Anran Tan, Undergraduate Student
Bo Ye, Graduate Student
Xinge Zhang, Undergraduate Student
Yueqiao Wang, Undergraduate Student
Yuliang Zou, Associate Professor of Healthcare Management
Lu Ma, Associate Professor of Healthcare Management
Guoxun Chen, Associate Professor of Nutrition
Rui Li, Associate Professor of Healthcare Management
Justin B. Moore, Associate Professor of Family and Community Medicine
© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 1
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

Long sleep predicted the risk of higher mortality, multiple
cardiovascular diseases and obesity.
8
Therefore, it warrants
more attention as an indicator of population health.
The epidemiological evidence about long sleep duration
remains equivocal. Present biological mechanisms explaining
the relationship between long sleep duration and obesity are
limited.
9
Moreover, in the existing meta-analyses addressing
this issue,
10,11
BMI change and weight change were not
examined, and some methodologic flaws such as misusing
data of the original study appeared. Therefore, this meta-
analysis was conducted to determine associations between
long sleep duration and obesity, weight gain, BMI change
and weight change in adults.
Materials and methods
Search strategy
The PubMed, Embase, Cochrane Library, Elsevier Science
Direct, Science Online, MEDLINE and CINAHL databases
were searched for English articles published before May
2017, using the key words of (‘long sleep duration’OR
‘sleep hours’) AND (‘obesity’OR ‘BMI’OR ‘weight gain’)
AND ‘adults’. Original studies, reviews and meta-analyses
were retained. Conference abstracts or reports whose full
texts were not accessed were excluded. Titles and abstracts
were examined for relevance. The full texts of related articles
were scrutinized further to identify whether or not they met
the inclusion criteria. Reference lists of the retrieved original
articles and reviews were also hand-searched for additional
relevant studies. Two researchers (W.L. and A.T.) evaluated
obtained articles and selected eligible studies independently.
Consensus was achieved through discussion and consulting
the third researcher (Y.W.).
Inclusion criteria
Studies were included if they met the following criteria: (i) a
longitudinal cohort design; (ii) adult cohorts; (iii) published
in English; (iv) the exposure at baseline was sleep duration
with reference category and long sleep categories; (v) the
outcomes of obesity, BMI, BMI change, weight gain, or
weight change; and (vi) the relative risk (RR), odds ratios
(OR), hazard ratios (HR) or βwith 95% confidence intervals
(CI) were reported. When the necessary data were not pro-
vided in the eligible articles, we contacted the authors to
request the data.
Data extraction
The following information was extracted from each eligible
study: (i) last name of the first author; (ii) year of publication;
(iii) origin country of the cohort; (iv) sample size; (v) gender
and age of study population at baseline; (vi) follow-up years;
(vii) sleep duration categories (criteria for long sleep duration
as well as the reference group); (viii) the duration to measure
sleep (at night or during 24 h); (ix) outcomes (obesity/weight
gain/BMI change/weight change); (x) methods to assess
sleep duration; (xi) variables adjusted in the multivariable
models; (xii) effect size measurements with 95% CI; and
(xiii) findings pertinent to associations among long sleep dur-
ation and the targeted outcomes. If results were reported by
different gender or age groups, they were included as differ-
ent cohorts. If results based on different follow-up periods
were reported, we chose the results with the longest follow-
up period. The quality of each study was assessed using the
Newcastle–Ottawa Scale (NOS) for cohort studies, which
includes criteria in three aspects (selection, comparability and
outcome) and the score can range from 0 to 9, with higher
scores indicating better quality of the study.
12
Statistical analysis
Outcomes of the selected studies included both dichotom-
ous and continuous variables. For dichotomous outcomes,
we calculated the pooled RR with 95% CI using the inverse-
variance method to evaluate the strength of associations
among long sleep duration and risk of obesity or weight
gain. When HR was reported, we regarded HR as RR.
When OR was reported, we converted OR to RR based on
the Zhang–Yu method.
13
For continuous outcomes, we cal-
culated the pooled β. If the estimate values under different
multivariate models were reported, the most stringently con-
trolled estimate was extracted. When multiple comparisons
of long sleep duration categories were included in the same
study, an overall estimate was calculated. In the analysis of
weight gain, we chose weight gain ≥5 kg as the outcome and
adjusted the estimates when weight gain was reported as
≥10 or ≥15 kg using standard methodology.
14
The significance of the pooled RR was determined using
Ztests, and statistical heterogeneity between studies was
examined using the Cochran Qand I
2
statistics.
15
If high
heterogeneity was detected (P<0.1 and I
2
>50%), the ran-
dom effect model was adopted, or the fixed effect model
was employed.
16
We used funnel plots and Egger regression
test to evaluate publication bias
16
and the ‘trim and fill’
method (trim the asymmetric outlying studies, find the true
center of the remaining symmetric funnel, then replace the
trimmed studies and fill the funnel with their ‘missing’coun-
terparts around the center to get the final estimate) to adjust
results.
17
Sensitivity analysis was conducted to identify stud-
ies that significantly contributed to the between-study
2JOURNAL OF PUBLIC HEALTH
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

heterogeneity, and the pooled results were re-estimated after
excluding these studies.
18
To detect potential factors influencing
heterogeneity, subgroup analyses were performed by gender,
region, follow-up period, sleep duration evaluation period, sleep
duration of reference group, and cut-offs of long sleep duration.
P<0.05 (two-sided) indicated statistical significance. R software
(version 3.4.0; R Foundation for Statistical Computing, Vienna,
Austria) was used for all statistical analyses.
Results
Search results and study selection
As shown in Fig. 1, overall 2673 articles were found after
searching the databases. After eliminating duplicate articles
and those reporting results from the same cohort, 43 studies
remained for further review. Consequently, 27 were excluded
for various reasons. Seven were cross-sectional studies or
lacked the longitudinal data; six lacked the required data for
meta-analysis; three only reported estimates of short sleep
duration; five did not differentiate short sleep and long sleep
by categorizing sleep durations into different groups; four
reported the outcomes of BMI or waist circumference but
the number of studies using the same outcome was too little
to conduct a meta-analysis; one was conducted among post-
partum women, not general population; and another one
combined sleep duration and disinhibition eating behavior.
Therefore, 16 cohort studies were included.
Characteristics of included studies
Overall 329 888 participants were involved. Detailed charac-
teristics of included studies are presented in Table 1.Two,
six and eight studies were from the USA, Europe and Asia,
respectively. The follow-up period ranged from 1 to 16
years. Sleep duration either at night or during 24 h was all
self-reported. The referent category was defined as 5–7h,
19
6–9h,
20
7h,
21–29
7–8h
30–33
and 8 h;
34
and the cut-offs for
long sleep duration included >7h,
19
≥8h,
25,26,29,31,32
≥9h
20–23,27,28,30,34
and ≥10 h.
24
Obesity was defined as
BMI ≥30 kg/m
2
in the studies from Europe
20,23,32
and the
USA,
22
and as BMI ≥25 kg/m
2
in studies from
Asia.
19,26,28,30,31,34
Four studies used weight gain ≥5kg as
the outcome,
21,24,27,28
one study adopted weight gain
≥10 kg
20
and another set weight gain ≥15 kg.
22
The total
NOS scores ranged from 6 to 8.
If a study reported results by gender, grouped data were
used in the overall analysis, and the gender-specific data
were regarded as results of two separate cohorts in the sub-
group analysis by gender. If a study only reported results by
gender (i.e. no grouped estimates), the data were treated as
two different cohorts in the overall analysis. Four outcomes
were encompassed: 12 cohorts were included in the group
with obesity as the outcome; eight cohorts were included in
the analysis of weight gain; and five and four cohorts for the
outcome of BMI change and weight change, respectively.
After preliminary statistical analyses, one study was excluded
Fig. 1 Flow chart of literature search and selection.
LONG SLEEP DURATION PREDICTS A HIGHER RISK OF OBESITY IN ADULTS 3
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

Table 1 Characteristics of included cohort studies on long sleep duration and obesity/weight gain/BMI change/weight change
Author year Country Gender Sample
size
Baseline
age
(year)
Follow-up
(year)
Sleep
measurement
duration
Sleep
category
(h)
Outcomes Adjustment
factors
The
Newcastle–
Ottawa
Scale
(NOS)
score
Assessment of sleep
duration
Findings
Patel et al.
22,45
USA F 68 183 39–65 16 24 h Reference:
7,
8,
≥9
Obesity
(BMI≥30 kg/m
2
);
Weight gain
≥15 kg
1,3,4,8,9,13,
14,17,19,23,
26,27,28,29,
32,42,45
6 Participants were asked to
‘indicate total hours of
actual sleep in a 24-h
period.’
Sleeping more than 7 h
was not associated with an
increased risk of obesity or
weight gain.
Stranges
et al.
23,51
England M and F 4378 35–55 4–7 At night Reference:
7,
8,
≥9
Obesity
(BMI≥30 kg/m
2
)
1,2,9,13,14,
17,33,35,38,
46
8 How many hours of sleep
do you have on an average
week night?
Sleep duration was not
associated with significant
changes in BMI or the
incidence of obesity.BMI change 1,2,3,9,13,
14,17,33,35,
38,46
López-García
et al.
24
Spain M 1064 70.7
(7.2)
2 24 h Reference:
7,
8,
9,
≥10
Weight
gain≥5kg
1,2,3,7,12,
13,14,16,
17,19,36,37,
40,41,46
7 How many hours do you
usually sleep per day
(including sleep at night
and during the day)?
Sleeping 8 or 9 h were
associated with weight
gain of ≥5kg in2y in
women, but not in men.
F 1271
Watanabe
et al.
30
Japan M 31 206 40 1 24 h Reference:
7to<8;
8to<9;
≥9
Obesity
(BMI≥25 kg/m
2
)
1,9,13,14,17,46 7 How many hours do you
sleep on weekdays? And on
the weekend?
(weekday sleep duration×5
+weekend sleep
duration×2)/7
Participants who slept ≥
9 h had an increased risk of
developing obesity, but this
was not statistically
significant.
F 3646 38 1,3,9,13,
14,17,46
Sleep duration ≥9 was
significantly associated with
BMI gain among men, but
not in women.
Nishiura and
Hashimoto
25
Japan M 2632 40–59 4 At night Reference:
7–7.9;
≥8
Obesity
(BMI≥25 kg/m
2
)
1,3,13,14,
17,32,39,
48,49,50
7 How many hours on
average do you sleep
during the night?
No significant relationship
between long sleep
duration and incidence of
obesity was found.
Nishiura et al.
31
Japan M 3803 40–59 4 At night Reference:
7
Long: ≥8
BMI change 1,3,13,14,17,
32,39
8 How many hours, on
average, do you sleep each
night?
No significant relationship
between long sleep
duration and BMI change
was found.
4JOURNAL OF PUBLIC HEALTH
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

Itani et al.
19
Japan M 21 693 <29–
>50
7 24 h Reference:
5–7;
Long: >7
Obesity
(BMI≥25 kg/m
2
)
1,13,14,17,
20,40,44
7 What was your daily
average sleep duration?
Sleep duration was not
associated with new-onset
obesity.
F 2109
Lyytikäinen
et al.
21
Finland M 1298 40–60 5–7 At night Reference:
7,
8,
≥9
Weight
gain≥5kg
1,3,7,9,13,
14,15,17,
41,44
6 How many hours per night
on average they slept
during the week?
Women with long sleep
duration were more likely
to have weight gain. No
association was found in
men.
F 5729
Kobayashi
et al.
26
Japan M 9449 ≥20 3 At night Reference:
7;
Long: ≥8
BMI change;
obesity
(BMI≥25 kg/m
2
)
1,2,3,14,
17,40
8 Self-reported. No significant impact of
long sleep duration on BMI
change or obesity was
found.
F 12 020
Yiengprugsawan
et al.
34
Thai M and F 42 465 20–49 4
(2009–13)
8
(2005–13)
24 h Reference:
8,
≥9
Obesity
(BMI≥25 kg/m
2
)
1,2,3,5,10,
11,14,17,
28,40
6 How many hours per day
do you sleep (including
during the day)?
Long sleep duration was
associated with obesity in
the 4-year follow-up but
not in the 8-year follow-up.
Xiao et al.
27
USA M 35 319 51–72 7.5 At night Reference:
7;
Long: ≥9
Obesity
(BMI≥30 kg/m
2
)
1,3,5,6,7,13,14,
19,37
7 Ask participants to choose
from ‘<5h’,‘5–6h’,
‘7–8h’,‘9 or more hours’
or leave the answer blank.
No increased risk of long
sleep duration on weight
change or weight gain or
obesity was found.
43 176 Weight change;
Weight
gain≥5kg
1,3,6,5,7,13,
14,17,18,19,
23,25,28,29,
37
F 40 201
Nagai et al.
28
Japan M and F 13 629 40–79 11 24 h Reference:
7,
8
Long: ≥9
Weight
gain≥5 kg;
Obesity
(BMI≥25 kg/m
2
)
1,2,3,5,7,9,
13,14,17,
19,24,37,45
7 How many hours on
average do you sleep per
day?
No significant impact of
long sleep duration on
weight gain or obesity was
found.
Sayón-Orea
et al.
32
Spain M and F 10 532 39 (12) 6.5 At night Reference:
7,8;
≥8
Obesity
(BMI≥30 kg/ m
2
)
1,2,3,13,14,
17,18,19,22,
23,30,31,41,
42,43
7 How many hours per day
do you sleep at night
during the weekdays? And
on the weekends?
(weekday sleep duration ×
5+weekend sleep
duration ×2)/7
There was a weak
relationship between long
sleep duration and incident
obesity.
Theorell-Haglöw
et al.
20
Sweden F 4903 43.9
(15.2)
10 At night Reference:
6–9;
Long: ≥9
Obesity
(BMI≥30 kg/
m
2
);
Weight
gain≥10 kg;
Weight change
1,9,13,14,15,
17,19,34,42,
45,46,
7 How many hours do you
sleep on average during the
night?
Long sleep duration was a
risk factor for obesity and
weight gain in younger
women (age<40 y) but not
in older women (age>40 y).
Continued
LONG SLEEP DURATION PREDICTS A HIGHER RISK OF OBESITY IN ADULTS 5
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

Table 1 Continued
Author year Country Gender Sample
size
Baseline
age
(year)
Follow-up
(year)
Sleep
measurement
duration
Sleep
category
(h)
Outcomes Adjustment
factors
The
Newcastle–
Ottawa
Scale
(NOS)
score
Assessment of sleep
duration
Findings
Nishiura and
Hashimoto
29
Japan M 1687 19–39 3 At night Reference:
7,
≥8
BMI change 13,14,17,36,
37,41,46
6 On average, how many
hours did you sleep each
night over the past month?
Participants who slept ≥
8 h per night had
significant gains in BMI.
Kowall et al.
33
Germany M and F 4814 45–75 5.1 At night Reference:
7–7.9;
≥8
Weight change 1,2,5,7,13,14,
15,20,23,37,
47
8 How many hours do you
sleep on average at night?
How many hours do you
normally take a nap? Total
sleep duration was the sum
of nocturnal sleep duration
and mean duration of
daytime napping.
Long (>8) nocturnal sleep
at T0 (2000–03) were
associated with weight
gain<0.5 kg between T1
(2005–8) and T2 (2011–15)
compared with the 6–8h
nocturnal sleep at T0 and
T1 as the reference.
Adjustment factors: 1: age; 2: gender; 3. Baseline BMI; 4: year of follow-up; 5: marital status; 6: race; 7: educational level; 8: spousal level of education; 9: work status; 10: income; 11: rural–urban residence; 12:
number of social links; 13: smoking; 14: alcohol consumption; 15: baseline weight; 16: intentional weight change; 17: physical activity/exercise; 18: time spent sitting; 19: caffeine intake; 20: eating habits; 21:
accordance with dietary guidelines; 22: snacking between meals; 23: total caloric intake; 24: energy consumption/day; 23: metabolic equivalents/week; 25: total fat intake; 26: ratio of polyunsaturated to saturated
fat; 27: trans-fat intake; 28: fruits and vegetables intake; 29: dietary fiber; 30: fast food; 31: sugared soft drinks; 32: medication use; 33: cardiovascular drugs; 34: diabetes medication; 35: hypnotics use; 36: use of
psychiatric medications; 37: perceived health; 38: mental and physical scores (Short Form-36); 39: family history of disease; 40: chronic diseases; 45: menopausal status; 41: sleep problems; 42: snoring status; 43:
siesta hours; 44: mental complaints/disorders; 45: anxiety; 46; depression; 47: stress; 48: preference for fatty food; 49: skipping breakfast; 50: snacking and eating out.
6JOURNAL OF PUBLIC HEALTH
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

from the weight change group due to conspicuous outliers
of the estimates, which were hundreds of times of those in
other included studies.
20
Long sleep duration and obesity risk
Long sleep duration was significantly associated with a 4%
higher obesity risk (RR [95% CI] =1.04 [1.00–1.09], P=
0.037), with no statistically significant heterogeneity between
the 12 studies involved (Q=5.19, P=0.92, I
2
=0.00%)
(Fig. 2). The result was not influenced substantially by sensi-
tivity analysis; exclusion of each study stepwise led to
changes in estimates from 1.039 (95% CI: 0.99–1.09) to
1.071 (95% CI: 1.00–1.15). No publication bias was found
by funnel plot and Egger’s test (P=0.565). In the subgroup
analysis, although the slight risk still existed, it was not sig-
nificant in any of the subgroup (Supplemental Table S1).
Long sleep duration and weight gain
We found no significant association between long sleep
duration and weight gain of ≥5 kg (RR [95% CI] =1.07
[0.98–1.17], P=0.14) (Supplemental Fig. S1). There was sig-
nificant heterogeneity between the included studies (Q=
14.33, P=0.05, I
2
=51.15%), but no individual study was
shown to be a potential outlier in the sensitivity analysis. In
the subgroup analysis, long sleep duration increased the risk
of weight gain of ≥5 kg in males (RR [95% CI] =1.55 [1.00–
1.31], P=0.048), at follow-up of <5 years (RR [95% CI] =
1.70 [1.25–2.30], P<0.001), and with sleep duration ≥9h
(RR [95% CI] =1.08 [1.00–1.17], P=0.045; Table 2).
Publication bias was observed via the funnel plot and
Egger’stest(P=0.047). After using the ‘trim and fill’
method, the combined RR remained unchanged.
Long sleep duration and BMI change
We found no relationship between long sleep duration and
BMI change (β=−0.001, 95% CI: −0.07 to 0.07, P=
0.990) and results were unchanged in all subgroup analyses.
There was no statistically significant heterogeneity between
involved studies (Q=4.64, P=0.20, I
2
=33.78%).
Sensitivity analysis confirmed the stability of the result. No
publication bias was detected by funnel plot and Egger’s test
(P=0.769).
Long sleep duration and weight change
We found no significant association between sleep duration
and weight change (β[95% CI] =0.106 [−0.30–0.51], P=
0.606), with significant heterogeneity between studies (Q=
5.68, P=0.06, I
2
=63.51%). In the sensitivity analysis,
when one study (51) was omitted, the estimates changed sig-
nificantly (β[95% CI] =0.319 [0.01–0.63], P=0.043),
denoting this study a possible source of significant hetero-
geneity. Subgroup analysis could not be performed because
the number of included studies was too small. No significant
publication bias was revealed by funnel plot and Egger’s test
(P=0.70).
Discussion
Main finding of this study
This meta-analysis found that long sleep duration was sig-
nificantly associated with higher risk of obesity, but had no
Fig. 2 Forest plot of the association between long sleep duration and risk of obesity among adults. (The positions of squares and diamonds represent effect
size and their sizes are proportional to the weight assigned to each study; horizontal lines represent 95% CIs.)
LONG SLEEP DURATION PREDICTS A HIGHER RISK OF OBESITY IN ADULTS 7
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

relationship with weight gain, BMI change, or weight change.
Our findings were determined to be stable by heterogeneity
test and sensitivity analysis.
What is already known on this topic?
Compared to the relatively consistent finding that short sleep
is associated with higher risk of obesity and weight gain,
10,35
the evidence about long sleep duration remains inconsistent.
Some studies showed long sleep was significantly associated
with obesity or weight gain;
36,37
some found no relationship
between long sleep duration and obesity/weight gain;
26,38
while some demonstrated long sleep reduced the risk of
obesity.
39,40
For example, Magee et al.
41
found long sleep
duration (≥9 h in each 24-h day) was related to obesity in
55- to 64-year-old individuals, but not in those aged 65 years
and above. Lyytikäinen et al.
21
observed an association
between long sleep duration (≥9 h per night) and major
weight gain in women but not men.
Likewise, while established laboratory evidence has eluci-
dated how short sleep is associated with higher risk of obes-
ity,
35
few present biological mechanisms explained the
relationship between long sleep duration and obesity.
9
It is
still unknown whether long sleep duration is inherently
harmful, or secondary to other comorbidities such as
depression or cardiovascular disease.
42
To our knowledge, there were two previous published
meta-analyses of prospective studies examining the associa-
tions of sleep duration and obesity or weight gain in
adults.
10,11
They found no association of long sleep duration
with future obesity. Wu et al.
10
concluded that long sleep
duration had no effect on incidence of obesity (OR [95%
CI] =1.06 [0.98–1.15]). Zhang et al.
11
reported that long
sleep duration increased the risk of weight gain (RR [95%
CI] =1.12 [1.04–1.20]) but had no contribution to future
obesity (RR [95% CI] =1.07 [0.99–1.16]), although signifi-
cant associations were observed among the subgroup of
high-quality studies.
What this study adds?
In contrast to previous meta-analyses, we found that signifi-
cant associations were present between long sleep duration
and obesity (RR [95% CI] =1.04 [1.00–1.09]), but not
weight gain (RR [95% CI] =1.07 [0.98–1.17]). Although the
effect sizes in our study were quite similar to those in previ-
ous two meta-analyses, the discrepancy was still visible and
could be accounted by some differences in methodology.
First, we included one more study in the analysis of long
sleep duration and obesity by requiring data from the
author
34
and excluded one study in the analysis of weight
gain that had been involved in previous meta-analyses
11
which seemed to have regarded the weight gain data five
years previous to the baseline as subsequent weight gain.
Second, we converted the reported ORs or HRs into RRs
based on a standard method, while the other two meta-
analyses did not. Third, we combined the RRs of all the
long sleep duration categories beyond the reference group as
the final estimates in the overall analysis, which was not
adopted by the other two meta-analyses. Additionally, we
Table 2 Subgroup analysis for long sleep duration and weight gain of ≥5kg
Exposure Subgroup Number of studies QP−QI
2
RR (95% CI) P
Z
Gender Male 3 1.41 0.495 0.00% 1.15 (1.00, 1.31) 0.0475
Female 5 15.42 0.004 74.06% 1.16 (0.94, 1,41) 0.1604
Mixed 1 0.98 (0.90, 1.07) 0.7000
Region USA 3 1.63 0.443 0.00% 1.07 (0.97, 1.17) 0.1673
Europe 4 9.85 0.020 69.54% 1.21 (0.89, 1.63) 0.2250
Follow-up years <51 –– – 1.70 (1.25, 2.30) 0.0007
5–9 4 1.94 0.585 0.00% 1.05 (0.97, 1.14) 0.2008
≥10 3 1.11 0.574 0.00% 0.99 (0.92, 1.07) 0.8777
Sleep duration measurement period At night 5 2.70 0.610 0.00% 1.06 (0.97, 1.14) 0.1845
24 h 3 11.31 0.004 82.31% 1.14 (0.90, 1.44) 0.2640
Sleep duration reference group 7 h 5 11.37 0.023 64.81% 1.07 (0.94, 1.22) 0.3039
7–8 h 2 1.40 0.237 28.42% 1.08 (0.97, 1.20) 0.1538
6–9h 1 –– – 2.07 (0.45, 9.42) –
Cut-off of long sleep duration ≥9 h 7 3.68 0.720 0.00% 1.08 (1.00, 1.17) 0.0450
≥10 h 1 –– – 1.57 (0.91, 2.70) –
8JOURNAL OF PUBLIC HEALTH
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

found long sleep duration had no relationship with BMI
change or weight change, which was not examined in previ-
ous meta-analyses.
One plausible mechanistic explanation for these results
centers on the basic cause of obesity or weight gain. Long
sleepers might have higher caloric intake before bedtime
43
and lower energy expenditures due to more time in bed and
less physical activity.
44
Another potential explanation is self-
reported sleep duration may be inaccurate.
37
Long sleepers
possibly overestimated their actual sleep time due to the
long time spent in bed, or slept longer to compensate for
poor sleep quality caused by sleep disorders or other health
problems.
20,37
In addition, emotional stress, depression,
poor physical health, low socioeconomic status and societal
isolation may also partly explain the association between
long sleep duration and obesity.
45
Limitations of this study
Several other limitations exist. First, all sleep durations in
our included studies were self-reported by a survey question,
as were the weight and height in some studies.
20–22
There is
evidence that subjects are inclined to under-report their
weight, over-report their height
46
and sleep duration.
47
Three longitudinal studies assessing sleep durations with
objective methods all found no longitudinal associations
between sleep duration and obesity or BMI change.
48–50
Thus, our finding needs further confirmation with studies
using objective sleep measures such as actigraphy and
polysomnography.
Second, some confounding variables affecting sleep dur-
ation or obesity or both might be a source of bias. For
example, medical disorders such as chronic pain and depres-
sion might perturb sleep and body weight by limiting phys-
ical activity. Medication use and socioeconomic status could
also confound the sleep–weight relationship.
35
It was also
reported that age, ethnicity, education level, self-reported
health, chronic health conditions, alcohol consumption,
smoking, physical activity and work status were associated
with long sleep duration.
51–53
Therefore, adjusting for these
potential confounders in the statistical analysis was essen-
tial.
42
Many of the included studies in our meta-analysis
adjusted most of the above factors in their model, however,
sleep problems such as obstructive sleep apnea (OSA) were
omitted in most studies, which played an important role in
both sleep duration and body weight. OSA may cause leptin
resistance that predisposed to higher food intake and increase
the appetite-stimulating hormone ghrelin and thus influenced
the weight regulation.
54
Considering that the included studies
adjusted confounding variables inconsistently and certain
critical variables such as depression and physical disorders
were missed in some studies, we acknowledge it is possible
that our results are partially due to residual confounding.
Furthermore, we only focused on sleep duration at base-
line, neglecting the possibility that changes in sleep durations
at various study points may influence the outcomes. For
instance, one study found that short sleepers at three study
points (baseline and two follow-ups) were more prone to be
obese than short sleepers at baseline and at one of the two
follow-ups.
55
Moreover, using BMI as a surrogate anthropo-
metric measure may be insufficient to diagnose obesity or
adiposity.
56
There was a study reporting that BMI failed to
identify half of the people with excess body fat percent.
57
Direct measurement of body fat (e.g. waist circumference)
instead of BMI would be more accurate.
11,56
Conclusion
Results from this meta-analysis showed that long sleep dur-
ation was significantly associated with higher risk of obesity
in adults, but not with risk of weight gain, BMI change or
weight change. Given the detrimental outcomes of long
sleep duration and obesity, the link between long sleep dur-
ation and obesity deserves more attention.
Supplementary data
Supplementary data are available at the Journal of Public
Health online.
Conflicts of interest
None.
Acknowledgements
We acknowledge the editorial assistance of Karen Klein, in
the Wake Forest Clinical and Translational Science Institute
(UL1 TR001420; PI: McClain).
Funding
This work was supported by National Natural Science
Foundation of China [Grant number 81402668].
References
1. Ng M, Fleming T, Robinson M et al. Global, regional, and national
prevalence of overweight and obesity in children and adults during
1980–2013: a systematic analysis for the Global Burden of Disease
Study 2013. Lancet 2014;384(9945):766–81.
LONG SLEEP DURATION PREDICTS A HIGHER RISK OF OBESITY IN ADULTS 9
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

2. Williams EP, Mesidor M, Winters K et al. Overweight and obesity:
prevalence, consequences, and causes of a growing public health
problem. Curr Obes Rep 2015;4(3):363–70.
3. Guh DP, Zhang W, Bansback N et al. The incidence of co-
morbidities related to obesity and overweight: a systematic review
and meta-analysis. BMC Public Health 2009;9:88.
4. Chaput JP, Perusse L, Despres JP et al. Findings from the Quebec
family study on the etiology of obesity: genetics and environmental
highlights. Curr Obes Rep 2014;3:54–66.
5. HealthyPeople. Healthy People 2020 Sleep Health.https://www.
healthypeople.gov/2020/topics-objectives/topic/sleep-health (10
June 2017, data last accessed).
6. Watson NF, Badr MS, Belenky G et al. Recommended amount of
sleep for a healthy adult: a joint consensus statement of the
American Academy of Sleep Medicine and Sleep Research Society.
Sleep 2015;38(6):843–4.
7. Bin YS, Marshall NS, Glozier N. Sleeping at the limits: the chan-
ging prevalence of short and long sleep durations in 10 countries.
Am J Epidemiol 2013;177(8):826–33.
8. Leger D, Beck F, Richard JB et al. The risks of sleeping ‘too
much’. Survey of a National Representative Sample of 24671 adults
(INPES health barometer). PLoS One 2014;9(9):e106950.
9. Marshall NS, Glozier N, Grunstein RR. Is sleep duration related to
obesity? A critical review of the epidemiological evidence. Sleep Med
Rev 2008;12(4):289–98.
10. Wu Y, Zhai L, Zhang D. Sleep duration and obesity among adults:
a meta-analysis of prospective studies. Sleep Med 2014;15(12):
1456–62.
11. Zhang S, Li L, Huang Y et al. Meta-analysis of prospective cohort
studies about sleep duration and risk of weight gain and obesity in
adults. Zhonghua Liu Xing Bing Xue Za Zhi 2015;36(5):519–25.
12 Wells GA, Shea B, O’Connell D et al The Newcastle-Ottawa
Scale (NOS) for Assessing the Quality if Nonrandomized Studies in
Meta-analyses.http://www.ohri.ca/programs/clinical_epidemiology/
oxford.asp (20 May 2017, data last accessed).
13. Zhang J, Yu KF. What’s the relative risk? A method of correcting
the odds ratio in cohort studies of common outcomes. J Am Med
Assoc 1998;280(19):1690–1.
14. Shah ASV, Langrish JP, Nair H et al. Global association of air pol-
lution and heart failure: a systematic review and meta-analysis.
Lancet 2013;382(9897):1039–48.
15. Higgins JPT, Thompson SG, Deeks JJ et al. Measuring inconsist-
ency in meta-analyses. Br Med J 2003;327(7414):557–60.
16. Egger M, Smith GD, Schneider M et al. Bias in meta-analysis
detected by a simple, graphical test. Br Med J 1997;315(7109):
629–34.
17. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based
method of testing and adjusting for publication bias in meta-
analysis. Biometrics 2000;56(2):455–63.
18. Patsopoulos NA, Evangelou E, Ioannidis JP. Sensitivity of
between-study heterogeneity in meta-analysis: proposed metrics
and empirical evaluation. Int J Epidemiol 2008;37(5):1148–57.
19. Itani O, Kaneita Y, Murata A et al. Association of onset of obesity
with sleep duration and shift work among Japanese adults. Sleep
Med 2011;12(4):341–5.
20. Theorell-Haglow J, Berglund L, Berne C et al. Both habitual short
sleepers and long sleepers are at greater risk of obesity: a
population-based 10-year follow-up in women. Sleep Med 2014;
15(10):1204–11.
21. Lyytikainen P, Rahkonen O, Lahelma E et al. Association of sleep
duration with weight and weight gain: a prospective follow-up
study. J Sleep Res 2011;20(2):298–302.
22. Patel SR, Malhotra A, White DP et al. Association between reduced
sleep and weight gain in women. Am J Epidemiol 2006;164(10):
947–54.
23. Stranges S, Cappuccio FP, Kandala NB 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(3):321–9.
24. Lo´pez-García E, Faubel R, Leo´n-Mun˜oz L et al. Sleep duration,
general and abdominal obesity, and weight change among the older
adult population of Spain. Am J Clin Nutr 2008;87(2):310–6.
25. 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(5):385–90.
26. Kobayashi D, Takahashi O, Deshpande GA et al. Association
between weight gain, obesity, and sleep duration: a large-scale 3-
year cohort study. Sleep Breath 2012;16(3):829–33.
27. Xiao Q, Arem H, Moore SC et al. A large prospective investigation
of sleep duration, weight change, and obesity in the NIH-AARP
Diet and Health Study cohort. Am J Epidemiol 2013;178(11):
1600–10.
28. Nagai M, Tomata Y, Watanabe T et al. Association between sleep
duration, weight gain, and obesity for long period. Sleep Med 2013;
14(2):206–10.
29. Nishiura C, Hashimoto H. Sleep duration and weight gain: recon-
sideration by panel data analysis. J Epidemiol 2014;24(5):404–9.
30. Watanabe M, Kikuchi H, Tanaka K et al. Association of short sleep
duration with weight gain and obesity at 1-year follow-up: a large-
scale prospective study. Sleep 2010;33(2):161–7.
31. Nishiura C, Noguchi J, Hashimoto H. Dietary patterns only par-
tially explain the effect of short sleep duration on the incidence of
obesity. Sleep 2010;33(6):753–7.
32. Sayon-Orea C, Bes-Rastrollo M, Carlos S et al. Association between
sleeping hours and siesta and the risk of obesity: the SUN
Mediterranean Cohort. Obes Facts 2013;6(4):337–47.
33. Kowall B, Lehnich AT, Erbel R et al. Associations between sleep
characteristics and weight gain in an older population: results of
the Heinz Nixdorf Recall Study. Nutr Diabetes 2016;6(8):e225.
34. Yiengprugsawan V, Banwell C, Seubsman SA et al. Short sleep and
obesity in a large national cohort of Thai adults. BMJ Open 2012;2
(e000561):1–6.
35. Patel SR, Hu FB. Short sleep duration and weight gain: a system-
atic review. Obesity 2008;16(3):643–53.
10 JOURNAL OF PUBLIC HEALTH
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018

36. Ford ES, Cunningham TJ, Croft JB. Trends in self-reported sleep dur-
ation among US adults from 1985 to 2012. Sleep 2015;38(5):829–32.
37. Chaput J-P, Després J-P, Bouchard C et al. The association between
sleep duration and weight gain in adults: a 6-year prospective study
from the Quebec Family Study. Sleep 2008;31(4):517–23.
38. Gangwisch JE, Malaspina D, Boden-Albala B et al. Inadequate
sleep as a risk factor for obesity: analyses of the NHANES I. Sleep
2005;28(10):1289–96.
39. Hasler G, Buysse DJ, Klaghofer R et al. The association between
short sleep duration and obesity in young adults: a 13-year pro-
spective study. Sleep 2004;27(4):661–6.
40. Lawman HG, D Fryar C, Gu Q et al. The role of prescription
medications in the association of self-reported sleep duration and
obesity in U.S. adults, 2007–2012. Obesity 2016;24(10):2210–6.
41. Magee CA, Caputi P, Iverson DC. Is sleep duration associated with
obesity in older Australian adults? J Aging Health 2010;22(8):1235–55.
42. Coughlin JW, Smith MT. Sleep, obesity, and weight loss in adults:
is there a rationale for providing sleep interventions in the treat-
ment of obesity? Int Rev Psychiatry 2014;26(2):177–88.
43. Mossavar-Rahmani Y, Jung M, Patel SR et al. Eating behavior by
sleep duration in the Hispanic Community Health Study/Study of
Latinos. Appetite 2015;95:275–84.
44. Yan LX, Chen XR, Chen B et al. Gender-specific association of
sleep duration with body mass index, waist circumference, and
body fat in Chinese adults. Biomed Environ Sci 2017;30(3):157–69.
45. Patel SR, Malhotra A, Gottlieb DJ et al. Correlates of long sleep
duration. Sleep 2006;29(7):881–9.
46. Visscher TLS, Viet AL, Kroesbergen HT et al. Underreporting of
BMI in adults and its effect on obesity prevalence estimations in
the period 1998 to 2001. Obesity 2006;14(11):2054–63.
47. Lauderdale DS, Knutson KL, Yan LL et al. Self-reported and mea-
sured sleep duration: how similar are they? Epidemiology 2008;19(6):
838–45.
48. Lauderdale DS, Knutson KL, Rathouz PJ et al. Cross-sectional and
longitudinal associations between objectively measured sleep dur-
ation and body mass index: the CARDIA Sleep Study. Am J
Epidemiol 2009;170(7):805–13.
49. Appelhans BM, Janssen I, Cursio JF et al. Sleep duration and
weight change in midlife women: the SWAN sleep study. Obesity
2013;21(1):77–84.
50. Vgontzas AN, Fernandez-Mendoza J, Miksiewicz T et al. Unveiling
the longitudinal association between short sleep duration and the
incidence of obesity: the Penn State Cohort. Int J Obes 2014;38(6):
825–32.
51. Stranges S, Dorn JM, Shipley MJ et al. Correlates of short and long
sleep duration: cross-cultural comparison between UK and US.
The Whitehall II study and the Western New York Health Study.
Am J Epidemiol 2008;168(12):1353–64.
52. Magee CA, Iverson DC, Caputi P. Factors associated with short
and long sleep. Prev Med 2009;49(6):461–7.
53. Patel SR, Sotres-Alvarez D, Castañeda SF et al. Social and health
correlates of sleep duration in a US Hispanic population: results
from the Hispanic Community Health Study/Study of Latinos.
Sleep 2015;38(10):1515–22.
54. Patel SR. Shared genetic risk factors for obstructive sleep apnea
and obesity. J Appl Physiol 2005;99(4):1600–6.
55. Gutierrez-Repiso C, Soriguer F, Rubio-Martin E et al. Night-time
sleep duration and the incidence of obesity and type 2 diabetes.
Findings from the prospective Pizarra study. Sleep Med 2014;15(11):
1398–404.
56. Prentice AM, Jebb SA. Beyond body mass index. Obes Rev 2001;
2(3):141–7.
57. Okorodudu DO, Jumean MF, Montori VM et al. Diagnostic per-
formance of body mass index to identify obesity as defined by
body adiposity: a systematic review and meta-analysis. Int J Obes
2010;34(5):791–9.
LONG SLEEP DURATION PREDICTS A HIGHER RISK OF OBESITY IN ADULTS 11
Downloaded from https://academic.oup.com/jpubhealth/advance-article-abstract/doi/10.1093/pubmed/fdy135/5071832
by Wake Forest University School of Medicine user
on 16 August 2018