Sleep and Obesity –
does the egg come before the chicken?
Francesco P Cappuccio MD MSc FRCP FFPH FAHA
Cephalon Professor of Cardiovascular Medicine & Epidemiology
University of Warwick, Warwick Medical School, Coventry, UK
1
Historical background
Steady decline in total sleep time during the past
100 years, generally attributed to lifestyle changes
(increased shift work, longer working hours in less physically
demanding jobs, increased and extended sedentary leisure time
activities, ...)
Parallel increase in the prevalence of overweight
and obesity, particularly evident in the last 25 years,
also attributable to lifestyle changes
(increased physical
inactivity, energy intake, …)
Recent indication of a possible association between
short sleep duration and obesity
2
BMI and Sleep Duration in US Adults Over Time
Is there a link between the growing obesity epidemic
and the decline in sleeping time?
Average Body Mass Index
in US adults in the last century
28.0
26.9
27.0
26.0
25.2
25.0
24.0
23.0
23.0
22.0
21.0
1910
Average sleep duration (hours)
in US adults in the last century
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
1975
2005
9.0
7.5
6.8
1910
1975
2005
3
Prevalence of overweight and obesity
in children and adults in the USA
40
%
2 to 5
35
6 to 11
30
12 to 19
25
40 to 59
20
20+
Adults
20 to 39
60+
15
10
Children
5
0
70
6
6
19
74
1
7
19
80
6
7
19
94
8
8
19
00
9
9
19
02
1
0
20
04
3
0
20
4
Mean sleep duration in healthy
young adults around the world
Steptoe A et al. Arch Int Med 2006;166:1689-92
5
Sleep duration and self-rated health problems
% poor health
50
Japan
40
Korea
30
Thailand
Taiwan
20
10
0
6
6.5
7
7.5
8
8.5
hours
Steptoe A et al. Arch Int Med 2006;166:1689-92
6
Short duration of sleep and all-cause mortality
Short sleep
Weight
(%)
Relative Risk
(95% CI)
147
1.00%
1.93 (1.12 to 3.35)
2,884
109
1.00%
0.90 (0.50 to 1.61)
2002
6,022
2,303
9.00%
1.00 (0.89 to 1.12)
Heslop (women)
2002
1,006
262
2.00%
0.98 (0.70 to 1.37)
Kripke (men)
2002
480,841
45,200
14.00%
1.11 (1.04 to 1.17)
Kripke (women)
2002
636,095
32,440
14.00%
1.07 (1.01 to 1.14)
Mallon (men)
2002
906
165
0.00%
1.11 (0.32 to 3.80)
Mallon (women)
2002
964
101
1.00%
1.00 (0.58 to 1.73)
Amagai (men)
2004
4,419
289
1.00%
2.41 (1.34 to 4.34)
Amagai (women)
2004
6,906
206
0.00%
0.70 (0.21 to 2.35)
Patel
2004
82,969
5,409
9.00%
1.08 (0.96 to 1.22)
Ferrie
2007
9,871
566
3.00%
1.25 (0.93 to 1.67)
Hublin (men)
2007
9,529
1,850
9.00%
1.26 (1.12 to 1.42)
Hublin (women)
2007
10,265
1,850
8.00%
1.21 (1.05 to 1.39)
Lan (men)
2007
1,748
816
4.00%
0.98 (0.76 to 1.26)
Lan (women)
2007
1,331
522
2.00%
1.14 (0.77 to 1.69)
Gangwisch (32-59) 2008
5,806
273
1.00%
0.67 (0.43 to 1.05)
Gangwisch (60-86) 2008
3,983
1,604
6.00%
1.27 (1.07 to 1.52)
Ikehara (men)
2009
41,489
8,548
4.00%
1.28 (1.01 to 1.62)
Ikehara (women)
2009
57,145
5,992
5.00%
1.28 (1.04 to 1.59)
Stone
2009
8,101
1,922
7.00%
1.02 (0.87 to 1.19)
1,374,718
110,574
100%
1.13 (1.07 to 1.19)
First author
Year
Sample size
Deaths
Kojima (men)
2000
2,438
Kojima (women)
2000
Heslop (men)
Combined effect: p<0.0001
Heterogeneity: I2=44%; p=0.0049
Egger’s test: p=0.48
Reference
0.1
1
10
Relative Risk (log scale)
Cappuccio FP et al. 2009; submitted
7
Obesity Epidemic and Sleep Duration
Association or Causation?
BMI in US adults
28.0
26.9
27.0
26.0
25.2
25.0
24.0
23.0
23.0
22.0
21.0
1910
Sleep duration in US adults 10.0
9.0
1975
2005
9.0
7.5
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
6.8
1910
1975
2005
8
Sleep duration and obesity in 6,862 children
(age 5-6 yrs)
Overweight
%
Obese
16
14
12
10
8
6
4
2
0
<=10h
10.5-11h
=>11.5h
Duration of sleep
von Kries R et al. Int J Obesity 2002;26:710-6
9
Short duration of sleep and obesity in children
age 2 to 20 years
Weight
(%)
Odds Ratio
(95% CI)
1,031
9.00%
2.25 (1.27 to 3.97)
Tunisia
167
6.00%
11.02 (4.75 to 25.60)
2002
Japan
8,941
14.00%
1.19 (0.99 to 1.41)
Von Kries
2002
Germany
6,645
12.00%
2.16 (1.55 to 3.01)
Agras
2004
USA
150
5.00%
1.99 (0.79 to 5.01)
Giugliano
2004
Brazil
165
1.00%
5.64 (0.72 to 44.17)
Padez
2005
Portugal
4,390
14.00%
1.15 (0.93 to 1.43)
Reilly
2005
UK
6,426
14.00%
1.45 (1.19 to 1.76)
Chaput
2006
Canada
422
7.00%
2.64 (1.25 to 5.56)
Chen
2006
Taiwan
656
13.00%
1.75 (1.28 to 2.40)
Seicean
2007
USA
509
5.00%
2.23 (0.87 to 5.70)
29,502
100%
1.89 (1.46 to 2.43)
First author
Year
Country
Sample size
Locard
1992
France
BenSlama
2002
Sekine
Combined effect: p<0.0001
Heterogeneity: Q=46.6; p<0.001
Begg’s test: p=0.12
Short sleep if obese
Short sleep if lean
0.1
1
10
Odds Ratio (log scale)
Sensitivity analysis: from 1.61 (1.33 to 1.96) to 2.07 (1.54 to 2.79)
Cappuccio FP et al. Sleep 2008; 31: 619-26
10
Short Sleep Duration and BMI/Obesity
Significant cross-sectional associations
2
BMI
(kg/m2)
1.5
-0.35 (-0.57;-0.12)
1
P <0.001
0.5
0
-0.5
-1
-1.5
-2
<=5
Odds Ratio
Obesity
6
7
8
9+
6
7
8
9+
2.5
2
1.5
1
1.65
0.5
(1.22-2.24)
0
<=5
Hours of sleep
Stranges S et al. Am J Epidemiol. 2008; 167: 321-9
11
Short duration of sleep and obesity in adults
age 15 to 102 years
First author
Year
Country
Sample size
Vioque
2000
Spain
Shigeta
2001
Kripke
Cournot
Short sleep if obese
Short sleep if lean
Weight
(%)
Odds Ratio
(95% CI)
1,772
3.00%
3.35 (2.22 to 5.06)
Japan
437
1.00%
1.97 (1.03 to 3.77)
2002
USA
497,037
12.00%
1.52 (1.46 to 1.58)
2004
France
3,127
4.00%
1.38 (0.97 to 1.96)
Hasler
2004
Switzerland
457
0.00%
10.8 (0.99 to 118.1)
Bjorkelund
2005
Sweden
1,460
1.00%
1.52 (0.68 to 3.40)
Gangwisch 1
2005
USA
3,682
5.00%
1.84 (1.40 to 2.42)
Gangwisch 2
2005
USA
3,324
6.00%
1.38 (1.07 to 1.78)
Gangwisch 3
2005
USA
2,582
4.00%
0.95 (0.67 to 1.35)
Singh
2005
USA
3,158
5.00%
1.70 (1.27 to 2.28)
Moreno
2006
Brazil
4,878
9.00%
1.22 (1.06 to 1.40)
Vahtera
2006
Finland
26,468
11.00%
1.43 (1.35 to 1.52)
Watari (men)
2006
Japan
19,894
2.00%
1.95 (1.20 to 3.19)
Watari (women)
2006
Japan
5,418
0.00%
2.97 (0.77 to 11.50)
Bjortvan
2007
Norway
8,860
3.00%
1.88 (1.22 to 2.89)
Chaput (men)
2007
Canada
323
1.00%
4.01 (1.73 to 9.33)
Chaput (women)
2007
Canada
417
1.00%
2.64 (1.25 to 5.56)
Ko
2007
Hong Kong
4,793
9.00%
1.30 (1.13 to 1.49)
Tuomilehto
2007
Finland
2,770
7.00%
1.30 (1.05 to 1.61)
Fogelholm (men)
2007
Finland
3,377
6.00%
1.46 (1.13 to 1.89)
Fogelholm (women)
2007
Finland
4,264
6.00%
1.75 (1.36 to 2.26)
Stranges
2008
UK
5,021
6.00%
2.01 (1.56 to 2.60)
603,519
100%
1.55 (1.43 to 1.68)
Combined effect: p<0.0001
Heterogeneity: Q=64.0, p<0.001
Begg’s test: p=0.09
0.1
1
10
Odds Ratio (log scale)
Sensitivity analysis: from 1.50 (1.39 to 1.61) to 1.59 (1.44 to 1.76)
Cappuccio FP et al. Sleep 2008; 31: 619-26
12
Prospective studies
Author
(year)
Hasler
(2004)
Gangwish
(2005)
Biorkelund
(2005)
Country
Switzerland
USA
Sweden
Sample
size
Age - Gender
Study details
Main findings
27 at baseline
Men and women
Prospective single
age cohort
F-up: 13 y
Less weight gain in
those with longer
sleep
3,208
32-49 at baseline
Men and women
Longitudinal
population based
(N-HANES)
F-up: 8-10 y
Highest increase in
BMI in those
sleeping 2-4h pn
498
38-60 at baseline
Women
Prospective
population study
F-up: 32 y
Short sleepers
gained more weight
that long sleepers
457
13
Short Sleep Duration and BMI/Obesity
No prospective associations
BMI
0.6
0.4
0.03 (-0.03; 0.08)
P = 0.36
0.2
0
-0.2
-0.4
-0.6
OR
3
Obesity
2.5
<=5
6
7
8
9+
1.05
(0.60-1.82)
2
1.5
1
0.5
0
<=5
6
7
8
9+
Hours of sleep
Stranges S et al. Am J Epidemiol. 2008; 167: 321-9
14
Possible Mechanisms
Appetite/satiety
Leptin and Ghrelin
Insulin resistance/sensitivity
Insulin, Adiponectin
Other hormonal mechanisms
Cortisol, SNS, Thyroid, GH
Endothelial Function
E-selectin, Soluble intercellular adhesion molecules(sICAM-1)
Inflammation/Immune system
C-reactive Protein (CRP), Interleukin 6 (IL-6), Tumor
necrosis factor  (TNF-), Fibrinogen
15
Sleep and Metabolic functions
A U-shaped association between sleep
duration and BMI.
Short sleep associated with low leptin and
high ghrelin
Leptin
BMI
Ghrelin
S Taheri et al. PLoS Med 2004;1(3):e62;210-7
16
Sleep and Metabolic functions
Spiegel K et al. JCEM 2004;89:5762-71
17
LACK OF SLEEP
less
more
Gale SM et al. J Nutr 2004;134:295-8
18
Vgontzas AN et al. J Int Med 2003;254:32-44
19
Vgontzas AN et al. J Int Med 2003;254:32-44
20
Sleep duration and Incidence of Diabetes
(1987-2004) in men, age 40-70 yrs, in
Massachusetts
6
Risk Ratio
5
4
3
2
1
0
<=5
6
7
Hours of Sleep
8
9+
Adjusted for multiple confounding
Yaggi HK et al. Diabetes Care 2006;29:657-61
21
Short Sleep duration and Diabetes
Elevation of evening cortisol levels
predisposing to insulin resistance
Increase in sympathetic tone, inhibiting
pancreatic function and leading to increased
glucose intolerance
Weight gain and reduction in leptin
Reduction of testosterone levels
22
Inflammation and Sleep
Inflammatory markers are elevated in individuals
undergoing short term sleep deprivation studies.
Short sleep may lead to increased secretion of
inflammatory cytokines, which in turn may lead to an
increase in cardiovascular risk
Miller MA et al. Curr Vasc Pharmacol 2007; 5(2):93-102.
23
Vgontzas AN et al. J Int Med 2003;254:32-44
24
Vgontzas AN et al. J Int Med 2003;254:32-44
25
Activation of the immune system and
inflammatory cytokines may affect sleep
Many inflammatory responses are somnogenic
However, in advance stages of inflammation the
sleep promoting effects are diminished with the result
of reduced NREM and increased wakefulness
(Ann NY Acad Sci 2001;933:201-10)
26
Bi-directional model of the sleep
deprivation-obesity association
Environment
(work, social, physical)
Glucose intolerance/Insulin resistance
Increased appetite ( ↓ leptin ↑ ghrelin)
More time to eat
Increased fatigue
Altered thermoregulation
Activation of inflammatory markers
Sedentary extra time
Short Sleep
(Sleep deprivation)
↑ caloric intake
↓ energy
expenditure
Obesity
27
Bi-directional model of the sleep
deprivation-obesity association
Environment
(work, social, physical)
Glucose intolerance/Insulin resistance
Increased appetite ( ↓ leptin ↑ ghrelin)
More time to eat
Increased fatigue
Altered thermoregulation
Activation of inflammatory markers
Sedentary extra time
Short Sleep
(Sleep deprivation)
↑ caloric intake
↓ energy
expenditure
Obesity
Sleep disordered breathing
Disrupted and short sleep
Inflammatory cytokines and the brain
28
Conclusions
The ‘epidemic’ of obesity is paralleled by a ‘silent epidemic’ of reduced
sleep duration.
These trends are detectable in adults as well as in children as young as 5
years.
Short sleep duration is associated with increased risk of obesity both in
adults and in children.
Evidence from prospective studies does not always confirm a temporal
sequence
A plausible mechanism could be the effect of short sleep on appetite
through the ghrelin-leptin system
Bi-directional effects (obesity causes lack of sleep) might also explain the
association
More research needed to understand the mechanisms by which short sleep
is linked to chronic conditions of affluent societies, such as obesity, diabetes
and hypertension.
29
SLEEP, HEALTH & SOCIETY
University of Warwick, Warwick Medical School
sleepresearch@warwick.ac.uk
Leads:
FP Cappuccio & MA Miller
Warwick Team:
E Peile, O Franco, S Stranges, N-B Kandala, FM Taggart,
C Ji, G Ward, A Bakewell, A Currie, A Lowe, D Cooper
Collaborators:
Warwick
R.C.P.
U.C.L.
Surrey
Boston (Harvard)
Naples (Federico II)
Avellino (CNR)
Buffalo (SUNY)
Funding:
S Williams, D Banejee
R Pounder
MG Marmot, E Brunner, M Kumari, M Shipley, JE Ferrie,
M Kivimaki
D-J Djik, S Archer
C Czeisler, SW Lockley, CP Landrigan, JP Sullivan
P Strazzullo
G Barba
J Dorn, R Donhaue, M Trevisan
Cephalon Inc., Wingate Foundation, Whitehall II,
RDF University of Warwick, NHS Workforce
30