Sleep Medicine 15 (2014) 13981404

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

Night-time sleep duration and the incidence of obesity and type 2

diabetes. Findings from the prospective Pizarra study
Carolina Gutirrez-Repiso a,b,c,*,1, Federico Soriguer a,b,c,1, Elehazara Rubio-Martn a,b,c,
Isabel Esteva de Antonio a,b,c, Mara Soledad Ruiz de Adana a,b,c, Mara Cruz Almaraz a,b,c,
Gabriel Olveira-Fuster a,b,c, Sonsoles Morcillo a,d, Sergio Valds a,b,c,
Ana Mara Lago-Sampedro a,b,c, Eduardo Garca-Fuentes a,b,c, Gemma Rojo-Martnez a,b,c
a

Spanish Biomedical Research Centre in Diabetes and Associated Metabolic Disorders (CIBERDEM), Spain
UGCI de Endocrinologa y Nutricin, Hospital Regional Universitario, Mlaga, Spain
c Instituto de Investigacin Biomdica de Mlaga (IBIMA), Mlaga, Spain
d
Diabetes Research Group, Hospital Universitario de Cruces, UPV-EHU, Baracaldo, Spain
b

A R T I C L E

I N F O

Article history:
Received 13 February 2014
Received in revised form 9 June 2014
Accepted 12 June 2014
Available online 11 August 2014
Keywords:
Night-time sleep duration
Prospective study
Weight gain
Obesity
Type 2 diabetes
Southern Spain

A B S T R A C T

Background: Several recent studies have related short sleep duration with different health problems, though
the results related with the risk of obesity and type 2 diabetes (T2D) are far from conclusive. The aim of
this study was to investigate the association between night-time sleep duration and the incidence of obesity
and T2D in a prospective study with a follow-up of 11 years.
Material and methods: The study comprised 1145 people evaluated in 19971998 and re-evaluated after
6 years and 11 years. At the three study points, subjects without known diabetes mellitus (KDM) were
given an oral glucose tolerance test (OGTT). Anthropometric and biochemical variables were measured.
The subjects were asked about their number of hours of night-time sleep.
Results: After adjustment, the OR of becoming obese was signicantly higher in subjects who slept 7
hours per night, at both the 6-year follow-up (OR = 1.99; 95% CI = 1.123.55) and the 11-year follow-up
(OR = 2.73; 95% CI = 1.475.04). The incidence of T2D at the 6-year follow-up in subjects without T2D at
baseline was higher in those who slept 7 hours per night (OR = 1.96; 95% CI = 1.103.50). However, this
association was not independent of obesity, weight gain or abnormal glucose regulation at baseline. At
the 11-year follow-up however there was no association between night-time sleep duration and the incidence of T2D.
Conclusions: The incidence of obesity over the 11-year follow-up increased in subjects with fewer hours
of night-time sleep. The incidence of T2D according to the hours of night-time sleep depended on obesity
and the carbohydrate metabolism phenotype.
2014 Elsevier B.V. All rights reserved.

1. Introduction
In those countries where sleep duration has been studied, the
number of hours of sleep has decreased over recent years [1,2]. This
decrease has been related to changes in lifestyle [3]. In recent years,
several studies have related short sleep duration with different health
problems, in cross-sectional [47], prospective [810], and intervention studies [11]. However, the particular results related to the

* Corresponding author at: UGCI de Endocrinologa y Nutricin, Hospital Regional

Universitario, Plaza del Hospital Civil, 29009 Mlaga, Spain. Tel.: +34 952286704;
fax: +34 952286704.
E-mail address: [email protected] (C. Gutirrez-Repiso).
1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.sleep.2014.06.014
1389-9457/ 2014 Elsevier B.V. All rights reserved.

risk of obesity and type 2 diabetes (T2D) are far from conclusive
[12].
A large cross-sectional study of 375,653 US adults aged 18 years
(the Behavioral Risk Factor Surveillance System) undertaken in 2009
found a positive association between short sleep duration and the
likelihood of having chronic diseases (coronary heart disease, stroke,
high blood pressure, asthma, arthritis, T2D and obesity). This association became weaker but did not disappear after adjustment
for frequent mental distress (FMD) (FMD 14 days during the past
30 days) [4]. Other recent cross-sectional studies found an association between poor sleep quality and short sleep duration and prediabetes and T2D [7,13].
In a large prospective study, the Nurses Health Study, the authors
concluded that the association between a reduced self-reported sleep
duration and a diagnosis of T2D could be due to confounding by

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

body mass index (BMI), or that sleep restriction may mediate its
effects on T2D through weight gain [8]. A recently published systematic review of 13 prospective studies in adults failed to detect
a clear association between sleep duration and weight gain [14].
The aim, therefore, of this study was to seek a possible association between night-time sleep duration and the incidence of obesity
and T2D at the 6- and 11-year follow-up points in the Pizarra study,
a prospective study in progress since 1995 in southern Spain. We
examined the hypothesis that those subjects with short nighttime sleep duration have a higher risk of becoming obese and/or
diabetic.
2. Methods
2.1. Baseline and follow-up studies
This study formed part of the Pizarra cohort study, the characteristics of which have been published previously [15]. In 1997, 1226
subjects were randomly selected from the adult population of Pizarra,
a village in the province of Malaga (Spain). The inclusion age was
1865 years, and individuals were excluded from the study if they
were institutionalized for any reason, pregnant, or had a severe clinical or psychological disorder. The nal sample distribution, by age
and sex, was not signicantly different from the population
distribution.
The study was approved by the Ethics and Clinical Investigation Committee of Carlos Haya Hospital, and written informed
consent was obtained from all participants.
The night-time sleep duration for 1145 subjects was obtained
from a self-report questionnaire. Phenotyping of the carbohydrate
metabolism was carried out according to the World Health Organization [16] in 1051 of these subjects, performing an oral glucose
tolerance test (OGTT) in those subjects who were unaware of their
diabetic status.
The cohort was re-evaluated in 20032004 (6-year follow-up
study). All those who had completed the baseline study were invited
by letter or by phone to attend for another clinical and anthropometric examination and another OGTT. In total, 968 of these subjects
completed this 6-year follow-up study. In 20092010, 673 subjects were re-evaluated (11-year follow-up study).

1399

asked about night-time sleep duration. The average night-time sleep

duration was assessed by asking On average, how many hours do
you usually sleep at night? Subjects were divided into two groups
according to the night-time sleep duration: those who slept 7 h
per night and those who slept 8 h per night. The cut-off points were
the 25th percentile (7 h) and the 50th percentile (8 h) of the frequency distribution. A food frequency questionnaire was completed
at baseline and at the 6-year follow-up [18]. The transformation to
energy and macronutrients was done by a computer program that
included the composition of local foods based on food composition studies [19].
Baseline measurements were made of leptin and interleukin-6
(IL-6). Additional proinammatory cytokines and adipokines were
measured at the 6-year follow-up study: tumor necrosis factor-
(TNF-) receptors R60KDa (R1) and R80KDa (R2), IL-6, leptin,
adiponectin, fatty acid binding protein 4 (FABP4) and high-sensitivity
C-reactive protein (hs-CRP). At the 6-year follow-up study, ferritin
and resistin were also measured. Ferritin was measured by
immunoturbidimetry (ATOM S.A., Barcelona, Spain) using an A15
autoanalyzer from Biosystems S.A. (Barcelona, Spain). Resistin was
measured using an enzyme-immunoassay commercial kit (SPI Bio
Bertin, York, UK).
Measurements of cytokines were performed using enzymeimmunoassay commercial kits: TNF- receptors R1 and R2 (BLK
Diagnostics, Barcelona, Spain); IL-6 (R&D Systems, Inc., Minneapolis, MN, USA); adiponectin (DRG Diagnostics GmbH, Marburg,
Germany); leptin (Mediagnost, Reutlingen, Germany); FABP4 (SPI Bio,
Montigny le Bretonneux, France); hs-CRP (BLK Diagnostics).
2.3. Statistical analysis
Data are presented as the mean and standard deviation. Differences between means were calculated using Students t-test or oneway or multivariate analysis of variance. The correlation between
variables was measured using Spearmans test. The strength of association between dependent variables and the explanatory variables
was measured using logistic regression analysis, calculating the odds
ratio (OR) and the 95% condence interval (CI). In all cases, a rejection level of = 0.05 was used. Analyses were done with SPSS
v10 (SPSS Inc., Chicago, IL, USA).

2.2. Procedures

3. Results

The protocol was the same at all three study points. The anthropometric study was carried out following a standardized method
[17]. At all three study points, measurements were made of weight
and height, and the BMI was calculated (weight/height2). Subjects
with BMI >30 kg/m2 were considered as obese.
The blood glucose level was measured at the three study points
using the glucose oxidase method (Accu-Chek, Roche Diagnostics,
Barcelona, Spain) at fasting and 120 min after an OGTT with 75 g
of glucose. The fasting serum insulin level was measured at baseline and at the 6-year follow-up by radioimmunoassay (Coat a
Count RIA kit, DPC, Los Angeles, CA, USA). Insulin resistance was
estimated with the HOMA equation (homeostatic model assessment insulin resistance), as follows: HOMA-IR = [fasting insulin
(U/mL) fasting glucose (mmol/L)]/22.5.
At all three study points, the blood pressure was measured twice
with a sphygmomanometer, with an interval of 5 min between measurements, and the average of the two measurements was used in
the analyses. Participants were considered to have hypertension if
their blood pressure was 140/90 mmHg or if they were receiving
antihypertensive treatment.
At all three study points, information on physical activity level,
smoking habit, educational level, and other lifestyle habits was obtained using a self-administered questionnaire. Subjects were also

3.1. Population variables according to night-time sleep

The number of hours of night-time sleep decreased over the 11year follow-up: 8.23 1.30 h (median, 8 h) at baseline, 8.01 1.38 h
(median, 7.9 h) at the 6-year follow-up, and 7.18 1.48 h (median,
7 h) at the 11-year follow-up.
Sex, educational level, smoking habit, daily energy consumed,
intake of carbohydrates, proteins and lipids, proportion of fatty acids
in the diet, hours watching television (TV), dietary habits watching TV, number of meals per day, snacking habits, and alcohol intake
(all variables studied on at least one of the three study points) did
not vary according to the number of hours of night-time sleep
(Table 1).
Subjects who rested 7 h per night practised sport in their leisure
time less often at both the 6-year (P = 0.01) and the 11-year followup (P = 0.02) (Table 1). On the other hand, daily work activity was
described as more intense in those subjects who rested 7 h per night
at both the 6-year (P = 0.01) and the 11-year follow-up (P = 0.02),
adjusted for age, sex, and obesity (Table 1).
Subjects who drank coffee at least once per day were more likely
to sleep fewer hours per night, both at baseline (8.11 1.17 vs
8.43 1.35 h; P = 0.005 adjusted for age, sex and obesity) and at the
6-year follow-up (7.82 1.35 vs 8.15 1.40 h; P = 0.005 adjusted for

1400

Table 1
General description of the population at the three study points according to the number of hours of night-time sleep.
Characteristics

Baseline
7 h (N = 300) (26.5%)
38.9 12.2
41.7/58.3
45.3
41.0
28.23 12.14
32.7
9.2
53.6
2.50 1.76
21.3
18.1

39.29 14.16
37.8/62.2
49.6
39.3
27.08 15.10
26.3
14.1
54.4
2.68 2.27
25.9
19.3

2221.17 740.46
247.8 12.44
84.06 27.05
101.67 33.5
44.14 7.08
15.53 2.89
41.85 6.06
9.89 2.18
18.44 4.47
5.13 1.93
0.45 0.16

2203.51 692.88
250.32 91.85
83.47 27.54
99.66 33.45
45.14 7.87
15.56 3.03
40.81 5.82
9.74 2.73
18.08 4.11
4.99 1.77
0.44 0.18

P
0.80
0.20a
0.22b
0.78c
<0.0001b
0.01b
0.05c
0.80c
0.10c
0.20e
0.63

0.53c
0.51c
0.44c
0.86c
0.31c
0.71c
0.19c
0.99c
0.76c
0.67c
0.81c

1.7

0.8

0.14
0.78

8.9
26.6
64.5
62.6

9.9
25.9
65.0
51.6

0.002c

12.35 13.03
2.15 0.81

11.67 11.52
2.29 0.65

0.32c
0.65c

7 h (N = 345) (35.8%)

11-year follow-up
P

7 h (N = 376) (55.9%)

8 h (N = 297) (54.1%)

47.90 14.86
34.3/65.1
54.8
27.7
28.33 5.23
38.6
22.5
61.5
2.15 1.80
39.0
41.7
10.5

0.001
0.40a
0.32b
0.16c
0.02b
0.84b
0.04c
0.26c
0.86c
0.38e
0.01c
0.01c

52.75 12.99
31.6/68.4
56.4
21.1
29.58 6.09
41.7
24.2
48.3

50.57 14.01
36.7/66.3
50.3
28.3
28.80 5.55
35.8
22.7
45.4

0.03
0.35a
0.50b
0.09c
0.22b
0.35b
0.81c
0.33c

35.9
11.2

42.1
8.4

0.02c
0.02c

2077.55 823.0
217.06 97.74
78.84 37.72
99.60 48.03
42.04 10.02
15.70 5.28
42.55 10.31

1919.78 659.33
197.73 82.20
72.43 26.86
94.30 43.27
41.39 10.43
15.42 4.13
43.50 10.76

0.23c
0.06c
0.08c
0.91c
0.13c
0.25c
0.07c

2.01 1.46
60.6

2.34 1.48
60.0

0.07b
0.26

1.89 1.26
55.3
13.3

2.07 1.65
56.2
11.1

0.22b
0.55
0.39

18.5
40.8
59.3
46.0
34.1

18.4
30.8
52.2
45.1
32.2

43.12 11.45
41.7/58.3
41.6
35.1
29.15 5.36
31.9
16.9
51.9
2.26 2.38
36.0
33.7
20.5

>8 h (N = 623) (64.4%)

6.1

4.5

9.9
26.8
63.2
50.0
36.0

12.9
20.6
66.7
40.1
38.0

13.25 10.94
2.18 3.49
3.12 4.18
8.47 10.26
1.01 2.42
12.27 17.82
24.71 14.13
97.14 94.54
4.32 1.77

14.26 1.82
2.54 4.97
3.38 4.37
9.81 10.85
1.09 2.65
11.82 13.61
27.80 17.12
102.06 142.2
4.49 1.81

0.54
0.10

0.03c
0.28

0.12

0.89c
0.48

0.98c
0.54c
0.38c
0.12c
0.62c
0.51c
0.26c
0.30c
0.16c

BMI, body mass index; T2D, type 2 diabetes; HOMA-IR, homeostatic model assessment insulin resistance; MUFA, monounsaturated fatty acids; TV, television; IL, interleukin; TNF, tumor necrosis factor; hs-CRP, highsensitivity C-reactive protein; FABP, fatty acid binding protein.
a Adjusted for age.
b Adjusted for age and sex.
c
Adjusted for age, sex and obesity.
d Excluding those subjects with known T2D.
e Adjusted for age, sex, obesity and abnormal glucose regulation (excluding known T2D).

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

Age (years)
Men/women (%)
Educational level (none or until 14 years) (%)
Smoking (yes) (%)
BMI (kg/m2)
Obesity (>30 kg/m2) (%)
T2D (%)
Hypertension (%)
HOMA-IRd
HOMA: 75th percentile (%)
Leisure physical activity (%)
Intense work activity (%)
Intake of macronutrients
Energy (kJ/day)
Carbohydrates (g)
Proteins (g)
Lipids (g)
Carbohydrates (%)
Proteins (%)
Lipids (%)
Fatty acids (% calories)
MUFA (% calories)
n-6 fatty acids (% calories)
n-3 fatty acids (% calories)
Eating habits
Hours watching TV
Eating watching TV (%)
Snacking (%)
5 meals per day (%)
Alcohol intake (%)
Daily
15 times/week
Never
Coffee (>1/day) (%)
Decaffeinated coffee (>1/day)
Cytokines and inammatory markers
Leptin (ng/mL)
IL-6 (pg/mL)
R1-TNF- (ng/mL)
R2-TNF- (ng/mL)
hs-CRP (mg/L)
Adiponectin (mg/L)
FABP4 (ng/mL)
Ferritin (g/L)
Resistin (ng/mL)

6-year follow-up
8 h (N = 845) (76.8%)

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

1401

Table 2
Incidence of obesity and risk of becoming obese at the 6-year follow-up points.
Baseline BMI

6-year follow-up

11-year follow-up

Night-time sleep duration

<30 kg/m2
>30 kg/m2

Adjusted OR (95% CI)

7 h at baseline

8 h at baseline

18.5%
94.5%

12.5%a
86.7%a

1.99 (1.123.55)b
1.85 (0.516.68)e

Night-time sleep duration

Adjusted OR (95% CI)

7 h at any of the
three study points

8 h at any of the
three study points

28.2%
87.3%

15.8%c
88.8%e

2.73 (1.475.04)d
0.66 (0.202.09)e

BMI, body mass index; OR, odds ratio; CI, condence interval.
a Unadjusted P = 0.07.
b
P = 0.01 adjusted for age, sex, leisure physical activity, and weight gain.
c Unadjusted P = 0.004.
d P = 0.004 adjusted for age, sex, leisure physical activity, and weight gain throughout the 11-year follow-up.
e
Not signicant.

age, sex and obesity). However, this difference was not evident at
the 11-year follow-up or in those subjects who drank decaffeinated coffee. At baseline and at the 6-year follow-up the proportion
of subjects who rested 7 h per night was signicantly higher in the
group of subjects who drank coffee at least once per day (Table 1).
The levels of leptin and IL-6 were not signicantly different according to the number of hours of night-time sleep, either at baseline
or at the 6-year follow-up. At the 6-year follow-up no signicant
differences were seen in the levels of hs-CRP, TNF- receptors R1
and R2, adiponectin, FABP4, ferritin or resistin according to the
number of hours of night-time sleep (Table 1).
The BMI was signicantly higher in those subjects who slept 7 h
per night at baseline (P < 0.0001) and at the 6-year follow-up
(P = 0.02), after adjusting for age and sex (Table 1). The unadjusted prevalence of obesity was 27.9% at baseline, 36.2% at the 6-year
follow-up and 39.1% at the 11-year follow-up. The highest prevalence at the 11-year follow-up was 41% in the group of subjects who
slept 7 h per night. After adjusting for age and sex, at baseline the
prevalence of obesity was signicantly higher in those subjects who
slept 7 h per night (P = 0.01) (Table 1). The prevalence of high blood
pressure at the three study points differed signicantly according
to the night-time sleep duration (Table 1).
The prevalence of T2D (known and unknown) was 14% at baseline, 20.4% at the 6-year follow-up and 23.4% at the 11-year followup. At baseline and the 6-year follow-up, the prevalence of T2D was
signicantly higher in those subjects who slept 8 h per night
(Table 1). No signicant differences were seen according to the nighttime sleep duration either at baseline or at the 6-year follow-up (after
excluding subjects with known T2D) in the HOMA-IR value or in
the prevalence of subjects with HOMA-IR value >75th percentile.
3.2. Incidence of obesity and T2D according to the night-time
sleep duration
After adjusting for age, sex, smoking habit, physical activity, and
coffee consumption, the OR of becoming obese was signicantly
higher in subjects who slept 7 h per night, both at the 6-year

follow-up (OR, 1.99; 95% CI, 1.123.55) and at the 11-year followup (OR, 2.73; 95% CI, 1.475.04). The majority of obese subjects at
baseline remained obese at the six-year and 11-year follow-up points,
independently of the number of hours of night-time sleep (Table 2).
The incidence of T2D at the 6-year follow-up in subjects without
T2D at baseline was higher in those who slept 7 h per night (OR,
1.96; 95% CI, 1.103.50). However, this association was not independent of obesity, weight gain or abnormal glucose regulation at
baseline; when these variables were excluded from the logistic regression model, the signicance of the association between nighttime sleep duration and incidence of T2D disappeared (OR, 1.51;
95% CI, 0.892.84). At the 11-year follow-up, however, there was
no association between night-time sleep duration and the incidence of T2D (Table 3).
In those subjects who slept 10 h at baseline and through the
follow-up, the incidence of obesity and T2D was no higher than in
those subjects who slept between 7 and 9 h (data not shown).

3.3. Changes in night-time sleep duration

Subjects who increased their night-time sleep duration by 2 h
at 6-year follow-up were signicantly older (45.9 13.0 vs 39.3 13.5
years; P = 0.001) and had higher BMI at baseline (29.2 5.1 vs
27.1 5.1 kg/m2; P = 0.009) and at 6-year follow-up (30.3 5 vs
28.1 5.2 kg/m2; P = 0.006) than those who did not increase their
night-time sleep duration.
The prevalence of obesity was higher in those subjects who increased their night-time sleep duration, but the risk of becoming
obese was not higher (OR, 1.4; 95% CI, 0.37.5; P = 0.6) after adjusting for age and sex.
At 6-year follow-up, the percentage of subjects who reduced their
night-time sleep duration of 2 h was 10.6%. These subjects were
older (39.0 13.2 years vs 46.2 13.8; P 0.001) and had higher BMI
both at baseline (27.3 5.0 vs 28.6 4.9 kg/m2; P = 0.014) and at
6-year follow-up (28.4 5.2 vs 29.8 5.0 kg/m2; P = 0.012) than those
subjects who did not reduce their night-time sleep duration.

Table 3
Incidence and risk of becoming diabetic at the 6-year and the 11-year follow-up points.
6-year follow-up

11-year follow-up

Night-time sleep duration

Baseline (excluding T2D)

7 h at baseline

8 h at baseline

14.0%

10.2%a

Adjusted OR (CI 95%)

1.96 (1.103.50)b

Night-time sleep duration

7 h at any of the
three study points

8 h at any of the
three study points

11.8 %

9.8%a

T2D, type 2 diabetes; OR, odds ratio; CI, condence interval.

a Unadjusted P: non-signicant.
b
P = 0.02 adjusted for age, sex, physical activity, smoking habit, weight gain, and abnormal glucose regulation at baseline.
c Adjusted P: non-signicant.

Adjusted OR (95% CI)

1.28 (0.602.69)c

1402

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

Fig. 2. Incidence of obesity at the 11-year follow-up (excluding subjects with body
mass index >30 kg/m2 at baseline). (A) Night-time sleep duration 7 h at all three
study points. (B) Night-time sleep duration 7 h at baseline and at one of the two
follow-up points. (C) Night-time sleep duration 8 h at baseline and at one of the
two follow-up points. (D) Night-time sleep duration 8 h at all three study points.

4. Discussion

Fig. 1. Correlation plot between night-time sleep duration at baseline and the 6-year
follow-up.

In these subjects who reduced their night-time sleep duration,

the risk of becoming obese was not signicantly different after adjusting for age and sex (OR, 1.3; 95% CI, 0.34.8; P = 0.6) and excluding
obese subjects at baseline.

3.4. Sensitivity analyses

Throughout the 11-year follow-up there were signicant correlations between night-time sleep duration at the three study points,
between baseline and the 6-year follow-up (r = 0.31; P < 0.001) (Fig. 1)
and between baseline and the 11-year follow-up (r = 0.25; P < 0.001).
The concordance was higher than that due to chance (P < 0.0001
between the three study points), but the Kappa index between baseline and the 6-year follow-up (Kappa = 0.25) and between baseline
and the 11-year follow-up (Kappa = 0.25) was low. Accordingly,
we dened a new variable using hours of night-time sleep at the
three study points, establishing four categories of subjects who slept:
(A) 7 h per night at the three points of the study.
(B) 7 h per night at baseline and at one of the two follow-up
points.
(C) 8 h per night at baseline and at one of the two follow-up
points.
(D) 8 h per night at all three study points.
After selecting subjects with BMI <30 kg/m2 at baseline, the incidence of obesity at the 11-year follow-up increased signicantly
according to the reduction in hours of night-time sleep throughout the whole follow-up of 11 years (P = 0.02) (Fig. 2).
Selecting criterion D as the reference criterion and after adjusting for age, sex and weight gain throughout the 11-year
follow-up, the OR (95% CI) of becoming obese in group C was 1.14
(0.552.39) (P = 0.71); in group B it was 2.28 (0.895.79) (P = 0.08);
and in group A it was 3.91 (1.569.78) (P = 0.004) (P for trend = 0.001).

The incidence of obesity throughout the 11-year follow-up increased in subjects with fewer hours of night-time sleep, and the
difference in the incidence of T2D according to the hours of nighttime sleep was not independent of obesity or of the carbohydrate
metabolism phenotype.
Several studies in recent years have shown the association
between short sleep duration and the risk of obesity [2022].
Cappuccio et al.s meta-analysis presented 19 cross-sectional studies
in children and 26 in adults [21]. The results showed a consistent
pattern of increased odds of being a short sleeper if you are obese,
both in childhood and in adulthood. However, the diculties of observational studies to establish the directionality of the association
are well known. Another meta-analysis by Magee and Hale included 13 longitudinal studies in adult persons, nding inconsistent
results concerning the relation between sleep duration and weight
gain [14]. The differences in design, the age at inclusion, the duration of follow-up, uncontrolled confounding variables or the
procedures used to measure night-time sleep duration may be some
of the reasons for these differences.
In the Nurses Health Study, 68,183 women who reported habitual sleep duration in 1986 were followed up for 16 years. The
results show that in middle-aged women aged 65 years, a habitual sleep duration of <7 h predicts increased future weight gain,
independent of baseline weight [23].
Our study was undertaken in a representative sample of men
and women from southern Spain. The results show that nonobese subjects who sleep 7 h per night had a higher risk of
becoming obese throughout the 11-year follow-up period.
The mechanisms whereby short sleep duration may increase the
risk of obesity are not clear. A restriction in the number of hours
of night-time sleep can increase daily tiredness and reduce physical activity [24]. Other studies found that physical activity increases
after brief periods of short sleep duration [25], though others found
no change in physical activity [11,23,26]. In our study, subjects who
slept fewer hours did less physical activity in their leisure time and
their normal daily work was considered more intense.
Several studies have evaluated the relationship between diet and
sleep duration. Some have shown that short sleep duration over short
periods of time increases food intake [25,27]. Others, however, failed
to nd this association [24,28]. Some studies have shown that subjects who sleep less consume more fats [24,25,29], fewer vegetables
[30] and in general their diet is of a poorer quality [11].

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

In our study, diet was evaluated at baseline and at the 6-year

follow-up. Despite a decrease in daily energy intake, as was expected over the years, neither the total daily energy nor the
macronutrients varied according to the number of hours of nighttime sleep, either at baseline or at the 6-year follow-up. There were
no differences in the habits related to food intake, hours watching
TV or alcohol intake according to the night-time sleep duration. Subjects who had coffee at least once per day slept 7 h per night. The
effect of caffeinated drinks on sleep duration is well known [31],
and it was observed at two of the three study points.
Experimental studies in humans have found that short periods
of sleep duration result in important changes in the secretion of cortisol, growth hormone [32], and levels of leptin and ghrelin [33],
both hormones related to the regulation of hunger and satiety.
Changes in the levels of leptin and ghrelin have also been found in
cross-sectional studies [34].
The activation of the inammation cascade due to restriction in
sleep duration has been proposed as an intermediary mechanism
in the association with obesity [35]. Controlled studies of sleep restriction have found increased levels of TNF- [36] and IL-6 [36] and
decreased levels of leptin [37] and adiponectin [38]. However, in
our study, none of the adipokines or inammatory markers studied
at the 6-year follow-up was associated with sleep duration.
In parallel with the association with obesity, we also assessed
the association between night-time sleep duration and the prevalence and incidence of T2D, hypertension and insulin resistance,
which are all related with obesity. Neither the prevalence and incidence of hypertension nor the prevalence of insulin resistance
measured by the HOMA-IR was associated with night-time sleep
duration.
The results of the association between the prevalence and incidence of T2D and the night-time sleep duration are paradoxical.
In at least two of the study points (baseline and the 6-year followup) the prevalence of T2D was lower in those subjects with fewer
hours of night-time sleep, but after adjusting for obesity, the incidence of T2D at the 6-year follow-up was signicantly higher in those
subjects who slept fewer hours. Studies of the association between
sleep duration and the risk of T2D show inconsistent results. Several
cross-sectional or prospective studies suggest a U-shaped association between sleep duration and T2D [5,6,8,39,40], though others
only found the association in women [41]. Cappuccio et al., in a metaanalysis of 10 prospective studies involving 107,756 subjects
followed-up for 4.432 years and 3586 incident cases of T2D, concluded that the quantity and quality of sleep consistently and
signicantly predict the risk of the development of T2D [42].
Our study has some limitations. The hours of night-time sleep
were measured using questionnaires in which the subjects were
asked about hours of night-time resting rather than hours of sleep
per 24 h period. At baseline, we noticed that subjects had some difculty in establishing the difference between the quantity and the
quality of the sleep, and also in separating hours of night-time sleep
from hours of day-time sleep (for example the afternoon nap), but
there was no problem when they were asked about the hours of
night-time sleep. In addition, certain common causes of obesity and
restriction of sleep duration were not included in our study; for
example, psychiatric comorbidities such as depression [43]. Other
chronic diseases, physical disability, or use of hypnotics could also
be confounders.
There are several denitions of short and long sleepers in literature. The denitions of normal sleep duration ranged greatly across
studies varying from 6 h to 9 h [44,45]. Some studies have included in the design mid-range sleepers as a reference category (68
or 78 h), whereas others have not included mid-range sleepers in
the comparisons. Also, the proportion of natural short and long
sleepers in a population is unknown. These subjects represent a small
but signicant proportion of the population for whom too much or

1403

too little sleep does not seem to be a health problem. These could
make the interpretation of the results dicult.
One of the strengths of our study is the evaluation of the nighttime sleep duration at the three study points, which showed a
positive correlation between the night-time sleep duration throughout the study. In addition, this was a prospective study in which
not only night-time sleep duration but also obesity, T2D (using OGTT)
and other confounding variables were measured at the three study
points.
In summary, the results of this prospective study undertaken in
a cohort of subjects in southern Spain support the hypothesis of an
association between short sleep duration and the incidence of
obesity. This association could be partly explained by the lower level
of physical activity undertaken by the subjects with fewer hours of
night-time sleep. On the other hand, the association between nighttime sleep duration and the incidence of T2D does not seem to be
independent of that found with the obesity.
Funding
This study was undertaken with nance from the Fondo de
Investigacin Sanitaria of the Instituto de Salud Carlos III (PI041883,
PI051307, and PI06/90564), Servicio Andaluz de Salud de la Junta
de Andaluca (0124/2005 and 0258/2007) and with nance from
the Spanish Biomedical Research Centre in Diabetes and Associated Metabolic Disorders (CIBERDEM) funds. CIBERDEM is an
initiative of the Instituto de Salud Carlos III.
Conict of interest
The ICMJE Uniform Disclosure Form for Potential Conicts of Interest associated with this article can be viewed by clicking on the
following link: http://dx.doi.org/10.1016/j.sleep.2014.06.014.
References
[1] Bixler E. Sleep and society: an epidemiological perspective. Sleep Med
2009;10(Suppl. 1):S36.
[2] Bin YS, Marshall NS, Glozier N. Secular trends in adult sleep duration:
a systematic review. Sleep Med Rev 2012;16:22330.
[3] Bonnet M, Arand D. We are chronically sleep deprived. Sleep 1995;18:90811.
[4] Liu Y, Croft JB, Wheaton AG, Perry GS, Chapman DP, Strine TW, et al. Association
between perceived insucient sleep, frequent mental distress, obesity and
chronic diseases among US adults, 2009 behavioral risk factor surveillance
system. BMC Public Health 2013;13:84.
[5] Gottlieb DJ, Punjabi NM, Newman AB, Resnick HE, Redline S, Baldwin CM, et al.
Association of sleep time with diabetes mellitus and impaired glucose tolerance.
Arch Intern Med 2005;165:8637.
[6] Chaput JP, Desprs JP, Bouchard C, Tremblay A. Association of sleep duration
with type 2 diabetes and impaired glucose tolerance. Diabetologia
2007;50:2298304.
[7] Lou P, Chen P, Zhang L, Yu J, Zhang N, Wu H, et al. Relation of sleep quality
and sleep duration to type 2 diabetes: a population-based cross-sectional survey.
BMJ Open 2012;2(4). pii: e000956.
[8] Ayas NT, White DP, Al-Delaimy WK, Manson JE, Stampfer MJ, Speizer FE, et al.
A prospective study of self-reported sleep duration and incident diabetes in
women. Diabetes Care 2003;26:3804.
[9] Gangwisch JE, Steven B, Heymseld SB, Boden-Albala B, Buijs RM, Kreier F, et al.
Short sleep duration as a risk factor for hypertension: analyses of the rst
national health and nutrition survey. Hypertension 2006;47:8339.
[10] Bo S, Ciccone G, Durazzo M, Ghinamo L, Villois P, Canil S, et al. Contributors
to the obesity and hyperglycemia epidemics. A prospective study in a
population-based cohort. Int J Obes (Lond) 2011;35:14429.
[11] Haghighatdoost F, Karimi G, Esmaillzadeh A, Azadbakht L. Sleep deprivation
is associated with lower diet quality indices and higher rate of general and
central obesity among young female students in Iran. Nutrition 2012;28:1146
50.
[12] Morselli L, Leproult R, Balbo M, Spiegel K. Role of sleep duration in the regulation
of glucose metabolism and appetite. Best Pract Res Clin Endocrinol Metab
2010;24:687702.
[13] Engeda J, Mezuk B, Ratliff S, Ning Y. Association between duration and quality
of sleep and the risk of pre-diabetes: evidence from NHANES. Diabet Med
2013;30:67680.
[14] Magee L, Hale L. Longitudinal associations between sleep duration and
subsequent weight gain: a systematic review. Sleep Med Rev 2012;16:23141.

1404

C. Gutirrez-Repiso et al./Sleep Medicine 15 (2014) 13981404

[15] Soriguer F, Rojo-Martinez G, Almaraz MC, Esteva I, Ruiz de Adana MS, Morcillo
S, et al. Incidence of type 2 diabetes in southern Spain (Pizarra Study). Eur J
Clin Invest 2008;38:12633.
[16] Alberti KG, Zimmet PZ. Denition, diagnosis and classication of diabetes
mellitus and its complications. Part 1: diagnosis and classication of diabetes
mellitus provisional report of a WHO consultation. Diabet Med 1998;15:53953.
[17] World Health Organization Expert Committee. Physical status: the use and
interpretation of anthropometry. Report of a WHO Expert Committee. World
Health Organ Tech Rep Ser 1995;854:1452.
[18] Soriguer F, Almaraz MC, Garca-Almeida JM, Cardona I, Linares F, Morcillo S,
et al. Intake and home use of olive oil or mixed oils in relation to healthy
lifestyles in a Mediterranean population. Findings from the prospective Pizarra
study. Br J Nutr 2010;103:11422.
[19] Mataix J, Garca L, Maas M, Martnez-Victoria E, Llopis J. Tabla de composicin
de alimentos. Granada: Universidad de Granada; 2003.
[20] Patel SR, Hu FB. Short sleep duration and weight gain: a systematic review.
Obesity (Silver Spring) 2008;16:64353.
[21] Cappuccio FP, Taggart FM, Kandala NB. Meta-analysis of short sleep duration
and obesity in children and adults. Sleep 2008;31:61926.
[22] Gangwisch JE, Malaspina D, Boden-Albala B, Heymseld SB. Inadequate sleep
as a risk factor for obesity: analyses of the NHANES I. Sleep 2005;28:128996.
[23] Patel SR, Malhotra A, White DP, Gottlieb DJ, Hu FB. Association between reduced
sleep and weight gain in women. Am J Epidemiol 2006;164:94754.
[24] Schmid SM, Hallschmid M, Jauch-Chara K, Wilms B, Benedict C, Lehnert H, et al.
Short-term sleep loss decreases physical activity under free-living conditions
but does not increase food intake under time-deprived laboratory conditions
in healthy men. Am J Clin Nutr 2009;90:147682.
[25] Brondel L, Romer MA, Nougues PM, Touyarou P, Davenne D. Acute partial sleep
deprivation increases food intake in healthy men. Am J Clin Nutr 2010;91:1550
9.
[26] Benece E, Garnier D, Ndiaye G. Nutritional status, growth and sleep habits
among Senegalese adolescent girls. Eur J Clin Nutr 2004;58:292301.
[27] Nedeltcheva AV, Kilkus JM, Imperial J, Kasza K, Schoeller DA, Penev PD. Sleep
curtailment is accompanied by increased intake of calories from snacks. Am J
Clin Nutr 2008;89:12633.
[28] Weiss A, Xu F, Storfer-Isser A, Thomas A, Ievers-Landis CE, Redline S. The
association of sleep duration with adolescents fat and carbohydrate
consumption. Sleep 2010;33:12019.
[29] Shi Z, McEvoy M, Luu J, Attia J. Dietary fat and sleep duration in Chinese men
and women. Int J Obes 2008;32:183540.
[30] Ohida T, Kamal AM, Uchiyama M. The inuence of lifestyle and health status
factors on sleep loss among the Japanese general population. Sleep
2001;24:3338.

[31] Butt MS, Sultan MT. Coffee and its consumption: benets and risks. Crit Rev
Food Sci Nutr 2011;51:36373.
[32] Spiegel K, Leproult R, Colecchia EF, LHermite-Balriaux M, Nie Z, Copinschi G,
et al. Adaptation of the 24-h growth hormone prole to a state of sleep debt.
Am J Physiol Regul Integr Comp Physiol 2000;279:R87483.
[33] 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;141:84650.
[34] 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.
[35] Miller MA, Cappuccio FP. Inammation, sleep, obesity and cardiovascular
disease. Curr Vasc Pharmacol 2007;5:93102.
[36] Vgontzas AN, Zoumakis E, Bixler EO, Lin HM, Follett H, Kales A, et al. Adverse
effects of modest sleep restriction on sleepiness, performance, and inammatory
cytokines. J Clin Endocrinol Metab 2004;89:211926.
[37] Spiegel K, Leproult R, LHermite-Baleriaux M, Copinschi G, Penev PD, Van Cauter
E. Leptin levels are dependent on sleep duration: relationships with
sympathovagal balance, carbohydrate regulation, cortisol, and thyrotropin. J
Clin Endocrinol Metab 2004;89:576271.
[38] Kotani K, Sakane N, Saiga K, Kato M, Ishida K, Kato Y, et al. Serum
adiponectin levels and lifestyle factors in Japanese men. Heart Vessels
2007;22:2916.
[39] Trenell MI, Marshall NS, Rogers NL. Sleep and metabolic control: waking to a
problem? Clin Exp Pharmacol Physiol 2007;34:19.
[40] Yaggi HK, Araujo AB, McKinlay JB. Sleep duration as a risk factor for the
development of type 2 diabetes. Diabetes Care 2006;29:65761.
[41] Tuomilehto H, Peltonen M, Partinen M, Sepp J, Saaristo T, Korpi-Hyvlti E,
et al. Sleep duration is associated with an increased risk for the prevalence of
type 2 diabetes in middle-aged women the FIN-D2D survey. Sleep Med
2008;9:2217.
[42] Cappuccio FP, DElia L, Strazzullo P, Miller MA. Quantity and quality of sleep
and incidence of type 2 diabetes: a systematic review and meta-analysis.
Diabetes Care 2010;33:41420.
[43] Foley DJ, Monjan AA, Izmirlian G, Hays JC, Blazer DG. Incidence and
remission of insomnia among elderly adults in a biracial cohort. Sleep
1999;22:S3738.
[44] Shigeta H, Shigeta M, Nakazawa A, Nakamura N, Yoshikawa T. Lifestyle, obesity,
and insulin resistance. Diabetes Care 2001;24:608.
[45] Vioque J, Torres A, Quiles J. Time spent watching television, sleep duration and
obesity in adults living in Valencia, Spain. Int J Obes Relat Metab Disord
2000;24:16838.