Sleep Medicine 47 (2018) 60e65

Contents lists available at ScienceDirect

Sleep Medicine
journal homepage: www.elsevier.com/locate/sleep

Original Article

Self-reported nocturnal sleep duration and glycosylated hemoglobin

A in the Study of Cardiovascular Risks in Adolescents (ERICA)
Luciane Gaspar Guedes a, *, Gabriela de Azevedo Abreu b, Katia Vergetti Bloch b
a ~o Gesteira/Universidade Federal do Rio de Janeiro, Brazil
Instituto de Puericultura e Pediatria Martaga
b
Instituto de Estudos em Saúde Coletiva/Universidade Federal do Rio de Janeiro, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Objective: At present, epidemiologic studies regarding the relationship between sleep duration and
Received 4 July 2017 glucose metabolism in adolescents are scarce. The objective was to investigate the association between
Received in revised form self-reported nocturnal sleep duration and glycosylated hemoglobin A in 12- to 17-year-old Brazilian
6 March 2018
adolescents.
Accepted 12 March 2018
Available online 6 April 2018
Patients/Methods: A school-based multicenter cross-sectional study was carried out in private and public
schools from 273 municipalities with more than 100,000 inhabitants. The final sample comprised 24,923
adolescents. A self-administered questionnaire was used. Blood tests included glucose, insulin, glyco-
Keywords:
Sleep length
sylated hemoglobin A and serum lipids. Age, sex, skin color, school type, body mass index (BMI), physical
Glucose metabolism activity, and Brazilian regions were studied as possible effect modifiers and/or confounders using linear
Adolescence regression.
Results: A significant positive association was found between more than 12 h of nocturnal sleep and
glycosylated hemoglobin A in two Brazilian regions: Southeast and South, even after adjustment for age,
sex, skin color, and BMI (coefficients of 0.142 and 0.339, respectively). No association was found with
nocturnal sleep duration <7 h.
Conclusion: Notably, a significant positive relationship was found between more than 12 h of nocturnal
sleep duration and glycosylated hemoglobin A in two Brazilian regions. The specific pubertal sleep
curtailment can be a compensatory mechanism for dealing with the insulin resistance during adoles-
cence. Those that escape from this regulatory strategy and sleep longer than the adequate duration, break
down this balance and tend to damage their glucose metabolism. To our knowledge, this is the first large
scale study, of the association between sleep duration and glucose metabolism in adolescents.
© 2018 Elsevier B.V. All rights reserved.

1. Introduction corresponding to approximately 20% of Brazil's total population in

Cardiovascular diseases are the main cause of mortality world-
wide, including in Brazil [1,2]. The development of atherosclerotic
disease is a continuous process of arterial damage, which begins in
childhood with the appearance of fatty streaks in the inner vascular
wall [3e5]. Furthermore, obesity, insulin resistance, dyslipidemia,
hypertension, smoking, sedentary lifestyle, and inadequate sleep
are recognized changeable risk factors [4,6e8].
Data from the last Census of the Brazilian Institute of Geography
and Statistics (IBGE) showed that there were 34 million adolescents,
~o Gesteira/
* Corresponding author. Instituto de Puericultura e Pediatria Martaga
Universidade Federal do Rio de Janeiro, Rua Bruno Lobo, 50 e Cidade Universit
aria,
Rio de Janeiro, RJ, CEP: 21941-912, Brazil. Fax: þ55 2125904640.
E-mail address: luciane.guedes@ippmg.ufrj.br (L.G. Guedes).
2010 [9]. To date, epidemiologic studies regarding the association
between sleep duration and glucose metabolism among adoles-
cents are scarce [10]. Matthews et al. found an association between
shorter sleep (mainly, during the week) and increased homeostatic
model assessment of insulin resistance (HOMA-IR) score, even after
adjusting for age, sex, race, body mass index (BMI), and waist
circumference, in 245 adolescents aged 14e19 years [11]. Javaheri
et al. showed a quadratic U-shaped relationship in 387 teenagers:
HOMA-IR was 20% greater in subjects who slept for 5 or 10.5 h, in
relation to those who slept 7e8 h when adjusted for age, sex, race,
prematurity, and physical activity. However, there was attenuation
of the association with shorter duration, after adjustment for
adiposity [12]. Spiegel et al., in a review article on the shortened
sleep length and risk for insulin resistance and type 2 diabetes,
reported that two-thirds of the decrease in the peripheral glucose

https://doi.org/10.1016/j.sleep.2018.03.013
1389-9457/© 2018 Elsevier B.V. All rights reserved.
 L.G. Guedes et al. / Sleep Medicine 47 (2018) 60e65
utilization during sleep correspond to diminished brain metabolism
related to the predominance of slow waves [13]. Like Javaheri et al.
[12], Koren et al., in a cross-sectional study with 62 obese adoles-
cents, found the same quadratic relationship between sleep
duration and serial glucose in an oral glucose tolerance test, as well
as with glycosylated hemoglobin A (HbA1c), regardless of obesity,
pubertal stage, sex, and sleep apnea measures. They also found
associations between decreased slow wave sleep and reduced in-
sulin secretion [14]. Likewise, the survey in the area of Iwaki-machi,
Japan, with subjects older than 20 years of age, using the definitions
of pre-diabetes and diabetes mellitus (DM) of the American Dia-
betes Association (ADA) as cut-off points, found the highest prev-
alence of high fasting glucose levels and of HbA1c in those with
short or long sleep periods. Logistic regression showed an adjusted
odds ratio of 4.96 (95% confidence interval (CI) ¼ 1.03e23.96) for
the association between sleep of <6 h, and a high HbA1c level and
an adjusted odds ratio of 4.96 (95% CI ¼ 1.70e14.50) when the sleep
length was 9 h [15].
The HbA1c measurement has many advantages, such as its
stability, easy interpretation, fasting-independent blood test, and
its direct relationship with cardiovascular consequences from
uncontrolled diabetes. The disadvantages are few, related to some
specific groups: anemic patients, blood recipients, and those with
the human immunodeficiency virus [16,17].
Considering the incidence and the prevalence of type 2 diabetes
among children and adolescents around the world [18], including
in Brazil [19,20], and the consequences for cardiovascular health,
this research aims to investigate the association between self-
reported nocturnal sleep duration and HbA1c in 12- to 17-year-
old Brazilian adolescents, from public and private schools. It es-
tablishes the hypothesis that nocturnal sleep length is associated
with the HbA1c levels, exploring the possibility of a quadratic
relationship between the exposure and the outcome variables, as
found in other studies [12,14,15].
2. Materials and methods
This research is part of a multicenter school-based countrywide
cross-sectional study funded by the Brazilian Ministry of Health,
the Study of Cardiovascular Risks in Adolescents (Portuguese
acronym ‘ERICA’).
Brazil is a continental country, sub-divided into five geograph-
ical macro-regions (North, Northeast, South, Southeast and Mid-
west), with several socioeconomic and cultural differences among
them. Therefore, the relationship between nocturnal sleep duration
and HbA1c was assessed nationally and by its five macro-regions.
2.1. Study sample
The ERICA sample was composed of adolescents aged
12e17 years, enrolled in the morning or afternoon shifts of private
and public schools, from 273 municipalities with more than
100,000 inhabitants, on July 1, 2009. The overall sample size was
estimated in 75,060 subjects. In each of the 3753 selected classes,
every student was invited to participate, but only the morning
students were included in the present research because of the need
for fasting blood dosages. The ERICA sample is considered a com-
plex sample, because it uses stratification, clustering, and unequal
probabilities in its selection stages. The study sampling in detail has
already been published [21].
The data collection started in the first half of 2013 and finished
in November 2014. At the end of the collection 36,956 students had
blood exams. HbA1c, glucose, and insulin were tested in 36,628
subjects. Exclusion criteria were applied to those adolescents
(reported or treated diabetes or any other health problem or
61
medicines used in the previous day before the blood testing). There
was a 9.9% loss of data, related to 2739 subjects with incoherent
answers for sleep hours.
2.2. Instruments for data collection
The ERICA Questionnaire of the Adolescent (EQA) was self-
administered using a personal digital assistant (PDA). The main
areas covered were socioeconomic status, adolescent work, smok-
ing, alcohol consumption, physical activity assessment, health and
medical history, sleeping hours, eating behavior, oral health,
common mental disorder, and reproductive health.
The ERICA questionnaire of the responsible parents/care giver
provided information on mothers' educational achievement, family
history of cardiovascular and metabolic diseases, and circum-
stances related to the student health (birth weight, breastfeeding,
diseases, and sleep duration). A printed form was sent to the
parents or caregivers through the students.
2.3. Study variables
Age (in years) was initially studied as a continuous variable in
the models. Age was also categorized as 17 years of age or younger
than 17 years of age, as a proxy for insulin resistance: the oldest
subjects being considered the least-resistant. Other variables were
sex (female or male), skin color (white or non-white), and school
type (private or public).
Sleep duration (h) was based on specific questions about sleep
listed in the EQA: “On a typical weekday, at what time do you
usually sleep?”, “On a typical weekday, at what time do you usually
wake up?” “On weekends, at what time do you usually sleep?” and
“On weekends, at what time do you usually wake up?” For the
evaluation of the hours/minutes of nighttime sleep by this method,
the weighted average number of hours/minutes of sleep informed
on weekdays and on weekends was used. It was studied as a
continuous variable, but additionally it was divided into three
groups of sleep duration: <7 h, 7e12 h, and >12 h of nocturnal
sleep, following the National Sleep Foundation's recommendations
[22], adapted for the age range of our study. Maximum limits of
±4 h above or below those considered as “may be appropriate” for
the age range of this research were accepted, as recommended by
the National Sleep Foundation [22].
Physical activity (min/week) was assessed through a question-
naire included in the EQA, as previously validated [23], containing a
list of several activities considered moderate to vigorous. The
subjects chose the activities exercised in the previous week, using
the PDA, describing the number of days and the daily duration
of each activity. This instrument allows the calculation of the
weekly physical activity time, in minutes. Those who performed
300 min/week or more were considered as active [24].
For BMI, the weight and height were measured using a digital
scale (model P150m, 200 kg of capacity and 50 g of precision), and a
portable stadiometer (with millimeter resolution and height up to
213 cm), respectively. The adolescent nutritional status was
obtained by the index BMI for age and sex, expressed in z-scores,
considering adapted World Health Organization reference values
[25,26], and analyzed in three categories: malnourished/low weight,
normal/overweight, obese.
All of the anthropometric procedures described above have
been detailed elsewhere [26].
Tanner's pubertal stage was determined by the EQA answers,
through self-evaluation boards. Despite its known relationship to
insulin resistance [27e29], pubertal stage was not used in the
model because of the subjective aspect of the self-evaluation.
Therefore, age was studied as proxy for insulin resistance.
62 L.G. Guedes et al. / Sleep Medicine 47 (2018) 60e65

For biochemical dosages [28,30], fasting blood samples were 3. Results

collected for analyses of HbA1c, glucose and insulin. The samples
were processed and plasma and serum were separated within 2 h
after collection, and kept at 4e10  C while being transported to the
study's single laboratory.
HbA1c (%) analysis was performed using ion exchange chro-
matography. The method is registered by the National Sanitary
Surveillance Agency (ANVISA) and certified by the National
Glycohemoglobin Standardization Program (NGSP), with inter-
assay coefficient of 1.55%. In the regression models, HbA1c was
used as a continuous variable.
Glucose (mg/dL) was analyzed using the hexoquinase method,
by the ADVIA 2400 (Siemens Equipment), with a coefficient of
variation of 1.54%. In the analysis, glucose was used as a continuous
variable.
Insulin (mU/L) was analyzed by chemiluminescence using the
Modular Analytics (Roche), with a coefficient of variation of 3.58%.
In the analysis, insulin was used as a continuous variable.
HOMA-IR was calculated as ((glucose (mmol/L)  insulin
(mU/L))/22.5).
Brazil is a continental country, sub-divided in five geographical
regions (North, Northeast, South, Southeast and Midwest), with
several socioeconomic and cultural differences among them, and
thus Brazilian region was recorded as a study variable.
All of the study's planning and management details have been
published elsewhere [26].
2.4. Statistical analysis
The sample characteristics were described using means and
medians and proportions, for the continuous and categorical vari-
ables, respectively, as measures of central tendency. For measures
of spread standard deviation (SD), the 25th and 75th percentiles,
and the 95% confidence interval (CI) were used.
The relationship between nocturnal sleep duration (the in-
dependent variable) and HbA1c (the dependent variable) was
studied through linear regression, exploring the role of age,
sex, skin color, school type (as a socioeconomic factor), physical
activity, BMI, and Brazilian region as effect modifiers and/or
confounders. The possibility of a quadratic relationship between
the exposure and the outcome variables was explored, using
the intermediate sleep group (from 7 to 12 h) as the reference
group.
The possible interactions were investigated, considering their
magnitude as well as the statistical significance of the test for
heterogeneity assessed using interaction terms. After ruling out the
possibility of effect modification, the variables that resulted in
changes of at least 10% in the b coefficient were also kept in the
final models as confounders, comparing the coefficients with and
without those variables.
The STATA software (version 14) and its SURVEY routine were
used to deal with the complex sample design.
The final sample comprised 24,923 adolescents (Fig. 1), who
slept from 3 to 16 h per night.
The comparison of the clinical characteristics between the study
sample and the group excluded from the analysis due to incoherent
answers for sleep hours showed that the two groups were very
similar, except for the sleep duration, as expected (data not
presented).
The mean age of the studied group was 14.7 years (SD ¼ 1.6) and,
according to the sample design, 50% of both sexes were found.
The other characteristics of the study sample are presented in
Table 1.
The sample was composed of 39% white adolescents. About a
quarter presented with excess weight, and the majority studied in
public schools. The median time spent weekly in physical activity
was 290 min and for night sleep it was about 8 h. The median values
for HbA1c, HOMA-IR, insulin, and glucose were normal. Notably,
the South Region showed the highest percentages of white ado-
lescents, subjects with excess weight, and individuals studying in
public schools, but without important differences regarding phys-
ical activity, sleep duration, or glucose metabolism markers when
compared to Brazil as a whole or to the other regions.
When the study sample was analyzed by sex, boys worked out
almost 2.5-times more than girls. The other characteristics were
similar between the two groups (data not shown).
In Table 2, the HbA1c medians and their 25th and 75th
percentiles by sleep duration groups are presented by Brazilian
macro-regions and for Brazil as a whole.
Again, the South Region showed the highest HbA1c median
value (5.8) for the adolescents who slept >12 h per night.
During the modeling steps, age, sex, and skin color were not
effect modifiers in the linear regression between self-reported
nocturnal sleep duration and HbA1c. Physical activity, school
type, and BMI did not present interactions of significant magnitude.
The only variable with relevant and statistically significant inter-
action was Brazilian regions. Therefore, separated models by this
factor were studied. Bedtime was also explored as a possible
explanation for the verified associations. However, relevant inter-
action and/or confounding were not found.
The linear regression parameters of the final models by Brazilian
macro-regions, adjusted for the sample design are shown in
Table 3. A statistically significant positive association was observed
between self-reported nocturnal sleep duration of >12 h and
HbA1c in two regions, Southeast and South, with a larger coeffi-
cient for the latter. Although reduced, the results were statistically
significant even after adjustment for the confounders: age

2.5. Ethical issues

The main study was approved by the Research Ethics Commit-

tees of the Instituto de Estudos em Saúde Coletiva of the
Universidade Federal do Rio de Janeiro (IESC/UFRJ) and of the in-
stitutions that participated in the study. The research complies with
the ethical principles contained in the Resolution 466/2012 of the
National Health Council (CNS, 2012). Only those individuals who
returned the Written Term of Consent signed by their parents or
guardians for blood collection and who signed the Written State-
ment of Consent participated in the study. Fig. 1. Flow diagram.
 L.G. Guedes et al. / Sleep Medicine 47 (2018) 60e65 63

Table 1
Characteristics of the adolescents in the study samplea.

Characteristic Brazil North Northeast Midwest Southeast South

% 95% CI % 95% CI % 95% CI % 95% CI % 95% CI % 95% CI

Skin color
White 39.3 37.2 41.5 21.6 19.6 23.7 27.7 23.7 32.2 33.6 30.5 36.8 40.3 36.8 44.0 67.3 63.5 71.0
Non-white 60.7 58.5 62.8 78.4 76.3 80.4 72.3 67.8 76.3 66.4 63.2 69.5 59.7 56.0 63.2 32.7 29.0 36.5
Weight
Very low weight 0.3 0.2 0.5 0.7 0.3 1.7 0.4 0.3 0.6 0.2 0.1 0.5 0.4 0.2 0.7 0.0 0.0 0.1
Low weight 2.0 1.7 2.4 2.3 1.8 2.9 2.3 1.8 3.0 2.2 1.7 2.8 2.0 1.5 2.7 1.2 0.7 1.8
Adequate weight 71.8 70.1 73.4 74.3 72.0 76.5 70.4 67.6 73.0 72.8 71.0 74.6 72.8 69.9 75.6 67.5 64.0 70.7
Overweight 17.0 15.8 18.2 16.1 14.5 17.9 17.3 15.6 19.1 16.7 15.0 18.6 16.3 14.4 18.5 20.0 17.8 22.4
Obese 8.9 8.1 9.7 6.5 5.6 7.6 9.6 8.1 11.4 8.0 6.9 9.3 8.5 7.3 9.8 11.4 9.0 14.2
School
Public 77.4 71.8 82.2 88.1 81.7 92.5 66.2 54.8 76.0 73.9 66.0 80.5 78.6 68.5 86.1 87.0 77.3 92.9
Private 22.6 17.8 28.2 11.9 7.5 18.3 33.8 24.0 45.2 26.1 19.5 34.0 21.4 13.9 31.5 13.0 7.1 22.7

Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p

Physical activity 290.0 80.0 690.0 300.0 60.0 750.0 260.0 60.0 645.0 345.0 90.0 770.0 275.0 80.0 670.0 360.0 120.0 770.0
(min/week)
Sleep duration (h) 7.9 7.1 8.9 8.0 7.0 9.0 8.0 7.1 8.9 7.9 7.1 8.6 7.9 7.1 8.9 8.0 7.3 8.9
HbA1c (%) 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.3 5.2 5.6
HOMA-IR 1.7 1.2 2.5 1.5 1.0 2.2 1.6 1.1 2.4 1.5 1.0 2.2 1.8 1.2 2.5 2.0 1.4 2.8
Insulin (mU/L) 8.2 5.7 11.6 7.3 5.1 10.2 7.7 5.3 11.0 7.4 5.2 10.6 8.4 5.8 11.7 9.5 6.7 13.0
Glucose (mg/dL) 86.0 82.0 91.0 85.0 80.0 89.0 87.0 82.0 91.0 84.0 80.0 89.0 86.0 81.0 90.0 88.0 83.0 93.0

CI, confidence interval; HbA1c, glycosylated hemoglobin A; HOMA-IR, homeostatic model assessment of insulin resistance; Med, median; 25p, 25th percentile; 75p, 75th
percentile.
a
Brazilian Study of Cardiovascular Risks in Adolescents.

Table 2
Glycosylated hemoglobin A by sleep duration groups in Brazil and Brazilian macro-regions.

Sleep Duration Brazil North Northeast Midwest Southeast South

Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p Med 25p 75p

<7 h 5.4 5.1 5.6 5.5 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.3 5.2 5.5
7e12 h 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.2 5.6 5.4 5.1 5.6 5.3 5.1 5.6
>12 h 5.6 5.4 5.8 5.4 5.2 5.6 5.3 5.2 5.6 5.4 5.2 5.6 5.6 5.4 5.7 5.8 5.8 5.8

Med, median; 25p, 25th percentile; 75p, 75th percentile.

Brazilian Study of Cardiovascular Risks in Adolescents.

Table 3
Linear regression between self-reported nocturnal sleep duration and glycosylated hemoglobin A by Brazilian macro-regionsa.

Region Sleep duration (h)b Unadjusted coefficient 95% CI p Adjusted coefficientc 95% CI p

North <7 0.009 0.036e0.019 0.524 0.005 0.032e0.022 0.725

>12 0.022 0.123e0.078 0.659 0.025 0.123e0.072 0.607
Northeast <7 0.012 0.019e0.044 0.431 0.008 0.025e0.042 0.624
>12 0.002 0.127e0.123 0.975 0.026 0.171e0.120 0.731
Midwest <7 0.013 0.021e0.047 0.450 0.014 0.019e0.048 0.398
>12 0.166 0.507e0.174 0.336 0.176 0.529e0.176 0.325
Southeast <7 0.018 0.061e0.024 0.395 0.001 0.044e0.043 0.975
>12 0.148 0.059e0.237 0.001 0.142 0.043e0.241 0.005
South <7 0.001 0.065e0.068 0.966 0.016 0.056e0.088 0.663
>12 0.412 0.356e0.467 <0.001 0.339 0.227e0.452 <0.001
a
Brazilian Study of Cardiovascular Risks in Adolescents.
b
Reference category: 7e12 h of sleep.
c
Adjusted by age (12e16 years; 17 years), sex, skin color (white; non-white) and body mass index (malnourished/underweight; normal/overweight; obese).

(<17 years; 17 years), sex, skin color (white; non-white) and BMI.
There was no statistically significant association between sleep
<7 h and the outcome studied.
4. Discussion
Sleep influences adolescent health in several ways. It has
recently been the central focus of many studies [10e12,14,31e44].
However, epidemiologic studies about the relationship between
sleep duration and glucose metabolism are scarce. Such studies
point out the link between sleep length and increased insulin
resistance [11,12,35,39,40,44,45]. Despite its wide utilization,
HOMA-IR [11,12,39,40,45] has been questioned for representing
only the insulin hepatic action [46].
HbA1c was chosen as the glucose metabolism marker in our
work; a few studies have used it for the specific purpose of studying
the association between self-reported sleep duration and glucose
metabolism [14,15,47]. Berentzen et al. did not find any association
between those aspects, although the sample comprised 1481
adolescents of a restricted age range, from 11 to 12 years, at the
beginning of the puberty [47]. A quadratic relationship was found
in the other two researches [14,15]; however, Nakajima et al.
64 L.G. Guedes et al. / Sleep Medicine 47 (2018) 60e65
studied 1062 adult subjects [15], and, Koren and colleagues studied
62 obese adolescents, aged 8e17 years [14].
Our results did not show the U-shaped curve expected in the
relationship studied. A statistically significant positive association
was found only for sleep durations longer than 12 h and HbA1c, in
the South and Southeast Brazilian regions, even after adjustment
for potential confounding factors.
Considering our findings, it is meaningful that the National
Sleep Foundation does not recommend more than 12 h of sleep for
children and adolescents from 6 to 13 years, and, not more than
11 h for those from 14 to 17 years of age [22]. These recommen-
dations supported by studies such as Javaheri and colleagues [12],
Koren et al. [14], and Duggan et al. [48], pointing out not only the
importance of sleep deprivation but also the possible undesirable
consequences of excessive hours of sleep.
Still, sleep is a process composed of sequential, alternate and cy-
clic phases that occur in the brain: N1, N2 and N3 stages of non-rapid
eye movement (non-REM) sleep and REM sleep [49]. Some studies
[14,35,44,45,50] show negative association between N1 sleep stage
percentage and insulin secretion and/or insulin sensitivity, and pos-
itive association between N3 sleep stage percentage and the same
conditions [13], independent on the total sleep duration.
Some hypotheses are suggested to explain our findings. First, the
specific pubertal sleep curtailment, in fact, is a compensatory
mechanism of the human body for facing the insulin resistance that
occurs during this time. Some studies corroborate this hypothesis
[51,52]. Crowley et al. showed decreased salivary melatonin during
puberty in 69 subjects aged 9.6e17.8 years [51]. Maslowsky and
Ozer, in a longitudinal research with 15,701 adolescents, verified a
progressive decrease in the sleep length across puberty, with initial
recovery after 18 years, reaching the pre-pubertal levels at around
23 years [52]. Those that escape from this functional regulatory
strategy and sleep longer than what is adequate, break down this
balance and tend to increase the HbA1c. This is probably a result of
an increased percentage of a more superficial sleep, with smaller
amounts of slowewave activity, both negatively related to insulin
secretion and insulin sensitivity, with glucose metabolism impair-
ment, as previously described [14,35,44,45,50].
The second hypothesis concerns Brazilian regional diversity,
regarding the different adolescent lifestyles. Other ERICA researches
have shown that adolescents from the South (one of the most
developed regions in Brazil) have greater alcohol consumption [53],
with a relative frequency of 27.5%, preference for beverages with
higher alcohol levels, and higher smoking prevalence compared to
the other regions [54]. In addition, they have more screen time
(use of computer, TV, and electronic games more than 2 h a day by
77.9% of those teenagers) [55], which could be related to lower sleep
duration and quality [56]. Differences related to eating habits among
the adolescents were also found in all regions [57,58]. Considering
our findings, the most relevant results of the last two researches
were the highest prevalence of soft drinks intake [58] and the lowest
water consumption [57] among those from the South. Macronu-
trient percentages were not different among the regions [58]. All
of the mentioned factors might have contributed for the macro-
regions-related differences found in Table 3.
Third, the lack of a negative association between <7 h of
nocturnal sleep and HbA1c levels might have been caused by
compensating naps during the day. Unfortunately, information
about this sleep aspect was not collected.
Finally, residual confounding also might have occurred because
of the broad categories of the sleep groups in our research.
They were adapted and based on the National Sleep Foundation
recommendations [22].
Our study has the subjective sleep measure as its main limita-
tion, related to the memory and also to the interpretation of the
mere act of going to bed as the sleep onset, and overestimating it
[10,59]. Some available more detailed questionnaires could be used,
although they are inappropriate as our questionnaire is already
extensive. Sleep diaries are also not practical for our population
research with adolescents. Despite probably being more accurate,
those strategies are still subjective with respect to sleep duration
measurement. Unfortunately, the objective techniques, such as the
polysomnography and the actigraphy, are also unfeasible for
population-based studies. Another limitation is the lack of evalu-
ation of the sleep quality influence on our findings. Again, as
explained previously, this assessment was not operationally
possible. However, as reported by Koren et al. in their review, the
relationship between obstructive sleep apnea and insulin resis-
tance in children is less clear, without conclusive results [10]. The
third limitation refers to our cross-sectional study design. It does
not allow us to establish directly the temporality between cause
and effect. Hence, clinical trials and prospective longitudinal
studies should be carried out.
Nevertheless, our research presents the strength of being a
population-based study, with a sample as large as 24,923 adoles-
cents, from all the Brazilian regions, with national and regional
representation.
5. Conclusions
A positive association was found between self-reported
nocturnal sleep length of >12 h in adolescents and HbA1c from
two Brazilian regions. Sleep should be included in the adolescent
health agenda and discussed at different levels of national health
policies. It should be also systematically assessed by health
professionals during adolescents' routine consultations.
Acknowledgements
The authors would like to thank the adolescents who partici-
pated in this research, and CAPES (Coordenaça ~o de Aperfeiçoa-
mento de Pessoal de Nível Superior) for L.G.G.’s scholarship
(BEX 3610/15-2). This article is part of the first author's doctorate
thesis. The ERICA Study was funded by the Ministry of Health of
Brazil (MS), Public Call e MCT/FINEP/MS/SCTIE/DECIT e CT-Saúde
and FNS e Síndrome Metabo lica e 01/2008 (Grant No. FINEP:
01090421, CNPq: 565037/2010-2 and 405009/2012-7) (the sponsor
had no involvement in the study design; collection, analysis, and
interpretation of data; writing of the report; or the decision to
submit the paper for publication).
Conflicts of interest
The authors have no conflicts of interest to declare regarding
this study.
The ICMJE Uniform Disclosure Form for Potential Conflicts of
Interest associated with this article can be viewed by clicking on the
following link: https://doi.org/10.1016/j.sleep.2018.03.013.
References
[1] Mansur AP, Favarato D. Mortality due to cardiovascular diseases in Brazil and
in the metropolitan region of Sa ~o Paulo: a 2011 update. Arq Bras Cardiol
2012;99(2):755e61.
[2] Soares GP, Klein CH, Silva NA, et al. Evolution of cardiovascular diseases
mortality in the counties of the State of Rio de Janeiro from 1979 to 2010. Arq
Bras Cardiol 2015;104(5):356e65.
[3] Berenson GS, Srinivasan SR, Bao W, et al. Association between multiple
cardiovascular risk factors and atherosclerosis in children and young adults.
The Bogalusa Heart Study. N Engl J Med 1998;338(23):1650e6.
[4] Raghuveer G. Lifetime cardiovascular risk of childhood obesity. Am J Clin Nutr
2010;91(5). 1514Se9S.
 L.G. Guedes et al. / Sleep Medicine 47 (2018) 60e65
[5] Raitakari OT, Juonala M, Ka € ho
€ nen M, et al. Cardiovascular risk factors in
childhood and carotid artery intima-media thickness in adulthood: the
cardiovascular risk in Young Finns Study. JAMA 2003;290(17):2277e83.
[6] Suriano K, Curran J, Byrne SM, et al. Fatness, fitness, and increased cardio-
vascular risk in young children. J Pediatr 2010;157(4):552e8.
[7] Kong AP, Wing YK, Choi KC, et al. Associations of sleep duration with obesity
and serum lipid profile in children and adolescents. Sleep Med 2011;12(7):
659e65.
[8] Martikainen S, Pesonen AK, Feldt K, et al. Poor sleep and cardiovascular
function in children. Hypertension 2011;58(1):16e21.
[9] Instituto Brasileiro de Geografia e Estatística (Brasil). Sinopse do censo
demogra fico 2010. 2011. Available from: http://www.censo2010.ibge.gov.br/
sinopse/index.php?dados¼12&uf¼00:IBGE.
[10] Koren D, O'Sullivan KL, Mokhlesi B. Metabolic and glycemic sequelae of sleep
disturbances in children and adults. Curr Diabetes Rep 2015;15(1):562.
[11] Matthews KA, Dahl RE, Owens JF, et al. Sleep duration and insulin resistance in
healthy black and white adolescents. Sleep 2012;35(10):1353e8.
[12] Javaheri S, Storfer-Isser A, Rosen CL, et al. Association of short and long sleep
durations with insulin sensitivity in adolescents. J Pediatr 2011;158(4):617e23.
[13] Spiegel K, Knutson K, Leproult R, et al. Sleep loss: a novel risk factor for insulin
resistance and Type 2 diabetes. J Appl Physiol (1985) 2005;99(5):2008e19.
[14] Koren D, Levitt Katz LE, Brar PC, et al. Sleep architecture and glucose and
insulin homeostasis in obese adolescents. Diabetes Care 2011;34(11):2442e7.
[15] Nakajima H, Kaneita Y, Yokoyama E, et al. Association between sleep duration
and hemoglobin A1c level. Sleep Med 2008;9(7):745e52.
[16] Juarez DT, Demaris KM, Goo R, et al. Significance of HbA1c and its measure-
ment in the diagnosis of diabetes mellitus: US experience. Diabetes Metab
Syndr Obes 2014;7:487e94.
[17] Aldasouqi SA, Gossain VV. Hemoglobin A1c: past, present and future. Ann
Saudi Med 2008;28(6):411e9.
[18] Fazeli Farsani S, van der Aa MP, van der Vorst MM, et al. Global trends in the
incidence and prevalence of type 2 diabetes in children and adolescents: a
systematic review and evaluation of methodological approaches. Diabetologia
2013;56(7):1471e88.
[19] Alvarez MM, Vieira ACRS R, Veiga GV. Prevalence of metabolic syndrome and
of its specific components among adolescents from Niteroi City, Rio de Janeiro
State, Brazil. Arquivos Brasileiros Endocrinol Metabol 2011;55(2):164e70.
[20] Seki M, Matsuo T, Carrilho AJ. Prevalence of metabolic syndrome and associated
risk factors in Brazilian schoolchildren. Publ Health Nutr 2009;12(7):947e52.
[21] Vasconcellos MT, Silva PL, Szklo M, et al. Sampling design for the study of
cardiovascular risks in adolescents (ERICA). Cad Saúde Pública 2015;31(5):
921e30.
[22] Hirshkowitz M, Whiton K, Albert SM, et al. National Sleep Foundation's sleep
time duration recommendations: methodology and results summary. Sleep
Health 2015;1:40e3.
[23] de Farias JC, Lopes AaS, Mota J, et al. Validity and reproducibility of a physical
activity questionnaire for adolescents: adapting the Self-Administered Phys-
ical Activity Checklist. Rev Bras Epidemiol 2012;15(1):198e210.
[24] Biddle S, Sallis JF, Cavill NA. Young and active? Young people and health
enhancing physical activity. Evidence and implication. London: Health Edu-
cation Authority; 1998.
[25] de Onis M, Onyango AW, Borghi E, et al. Development of a World Health
Organization growth reference for school-aged children and adolescentes.
Bull World Health Organ 2007;85(9):660e7.
[26] Bloch KV, Szklo M, Kuschnir MC, et al. The Study of cardiovascular risk in
adolescents e ERICA: rationale, design and sample characteristics of a national
survey examining cardiovascular risk factor profile in Brazilian adolescents.
BMC Publ Health 2015;15:94.
[27] Moran A, Jacobs DR, Steinberger J, et al. Insulin resistance during puberty:
results from clamp studies in 357 children. Diabetes 1999;48(10):2039e44.
[28] Goran MI, Gower BA. Longitudinal study on pubertal insulin resistance. Dia-
betes 2001;50(11):2444e50.
[29] Kelly LA, Lane CJ, Weigensberg MJ, et al. Pubertal changes of insulin sensi-
tivity, acute insulin response, and b-cell function in overweight Latino youth.
J Pediatr 2011;158(3):442e6.
[30] Sociedade Brasileira de Cardiologia. I Diretriz de prevença ~o da aterosclerose
na inf^ancia e na adolesce ^ncia. Arq Bras Cardiol 2005;85(Supl VI):1e36.
[31] Moore M, Meltzer LJ. The sleepy adolescent: causes and consequences of
sleepiness in teens. Paediatr Respir Rev 2008;9(2):114e20. quiz 20e21.
65
[32] Leproult R, Van Cauter E. Role of sleep and sleep loss in hormonal release and
metabolism. Endocr Dev 2010;17:11e21.
[33] Narang I, Manlhiot C, Davies-Shaw J, et al. Sleep disturbance and cardiovas-
cular risk in adolescents. CMAJ 2012;184:E913e20.
[34] Tatone-Tokuda F, Dubois L, Ramsay T, et al. Sex differences in the association
between sleep duration, diet and body mass index: a birth cohort study.
J Sleep Res 2012;21(4):448e60.
[35] Armitage R, Lee J, Bertram H, et al. A preliminary study of slow-wave EEG
activity and insulin sensitivity in adolescents. Sleep Med 2013;14(3):257e60.
[36] Countryman AJ, Saab PG, Llabre MM, et al. Cardiometabolic risk in adoles-
cents: associations with physical activity, fitness, and sleep. Ann Behav Med
2013;45(1):121e31.
[37] Maximova K, Kuhle S, Davidson Z, et al. Cardiovascular risk-factor profiles of
normal and overweight children and adolescents: insights from the Canadian
health measures survey. Can J Cardiol 2013;29(8):976e82.
[38] Samson R, Blunden S, Banks S. The characteristics of sleep and sleep loss in
adolescence: a review. Int Rev Soc Sci Humanit 2013;4(2):90e107.
[39] Androutsos O, Moschonis G, Mavrogianni C, et al. Identification of lifestyle
patterns, including sleep deprivation, associated with insulin resistance in
children: the Healthy Growth Study. Eur J Clin Nutr 2014;68(3):344e9.
[40] Cespedes EM, Rifas-Shiman SL, Redline S, et al. Longitudinal associations of
sleep curtailment with metabolic risk in mid-childhood. Obesity (Silver
Spring) 2014;22(12):2586e92.
[41] Iglayreger HB, Peterson MD, Liu D, et al. Sleep duration predicts car-
diometabolic risk in obese adolescents. J Pediatr 2014;164(5). 1085e1090.e1.
[42] Korcarz CE, Stein JH, Peppard PE, et al. Combined effects of sleep disordered
breathing and metabolic syndrome on endothelial function: the Wisconsin
sleep cohort study. Sleep 2014;37(10):1707e13.
[43] Reither EN, Krueger PM, Hale L, et al. Ethnic variation in the association
between sleep and body mass among US adolescents. Int J Obes (Lond)
2014;38(7):944e9.
[44] Zhu Y, Li AM, Au CT, et al. Association between sleep architecture and glucose
tolerance in children and adolescents. J Diabetes 2015;7(1):10e5.
[45] Klingenberg L, Chaput JP, Holmb€ ack U, et al. Acute sleep restriction reduces
insulin sensitivity in adolescent boys. Sleep 2013;36(7):1085e90.
[46] Hoffman RP. Indices of insulin action calculated from fasting glucose and in-
sulin reflect hepatic, not peripheral, insulin sensitivity in African-American
and Caucasian adolescents. Pediatr Diabetes 2008;9(3 Pt 2):57e61.
[47] Berentzen NE, Smit HA, Bekkers MB, et al. Time in bed, sleep quality and
associations with cardiometabolic markers in children: the Prevention and
Incidence of Asthma and Mite Allergy birth cohort study. J Sleep Res
2014;23(1):3e12.
[48] Duggan KA, Reynolds CA, Kern ML, et al. Childhood sleep duration and lifelong
mortality risk. Health Psychol 2014;33(10):1195e203.
[49] National Institutes of Health. Your guide to healthy sleep. Bethesda: NIH.
Publication; 2011.
[50] Tasali E, Leproult R, Ehrmann DA, et al. Slow-wave sleep and the risk of type 2
diabetes in humans. Proc Natl Acad Sci U S A 2008;105(3):1044e9.
[51] Crowley SJ, Acebo C, Carskadon MA. Human puberty: salivary melatonin
profiles in constant conditions. Dev Psychobiol 2012;54(4):468e73.
[52] Maslowsky J, Ozer EJ. Developmental trends in sleep duration in adolescence
and young adulthood: evidence from a national United States sample.
J Adolesc Health 2014;54(6):691e7.
[53] Coutinho ES, França-Santos D, Magliano EaS, et al. ERICA: patterns of alcohol
consumption in Brazilian adolescents. Rev Saude Publica 2016;50(Suppl. 1).
[54] Figueiredo VC, Szklo AS, Costa LC, et al. ERICA: smoking prevalence in Bra-
zilian adolescents. Rev Saude Publica 2016;50(Suppl. 1).
[55] Oliveira JS, Barufaldi LA, Abreu GeA, et al. ERICA: use of screens and con-
sumption of meals and snacks by Brazilian adolescents. Rev Saude Publica
2016;50(Suppl. 1).
[56] Falbe J, Davison KK, Franckle RL, et al. Sleep duration, restfulness, and screens
in the sleep environment. Pediatrics 2015;135(2):e367e75.
[57] Barufaldi LA, Abreu GA, Oliveira JS, et al. ERICA: prevalence of healthy eating
habits among Brazilian adolescents. Rev Saude Publica 2016;50(Suppl. 1).
[58] Souza AM, Barufaldi LA, Abreu GA, et al. ERICA: intake of macro and micro-
nutrients of Brazilian adolescents. Rev Saude Publica 2016;50(Suppl. 1).
[59] Guedes LG, Abreu GA, Rodrigues DF, et al. Comparison between self-reported
sleep duration and actigraphy among adolescents: gender differences. Rev
Bras Epidemiol 2016 ApreJun;19(2):339e47.