Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

https://doi.org/10.1007/s40519-019-00841-2

ORIGINAL ARTICLE

Association of sleep, screen time and physical activity with overweight

and obesity in Mexico
Spyros Kolovos1   · Aura Cecilia Jimenez‑Moreno2   · Rafael Pinedo‑Villanueva1   · Sophie Cassidy3   ·
Gerardo A. Zavala4 

Received: 24 June 2019 / Accepted: 19 December 2019 / Published online: 31 December 2019
© The Author(s) 2019

Abstract
Purpose  Approximately 70% of adults in Mexico are overweight or obese. Unhealthy lifestyle behaviors are also prevalent.
We examined the association of three lifestyle behaviors with body mass index (BMI) categories in adults from Mexico.
Methods  We used publicly available data from the ENSANUT 2016 survey (n = 6419). BMI was used to categorize partici-
pants. Differences in sleep duration, suffering from symptoms of insomnia, TV watching time, time in front of any screen,
vigorous physical activity (yes vs no), moderate physical activity (> 30 min/day—yes vs. no) and walking (> 60 min/day—yes
vs. no) were compared across BMI groups using adjusted linear and logistic regression analyses.
Results  Thirty-nine percent of participants were overweight and 37% obese. Time in front of TV, in front of any screen,
sleep duration and physical activity were significantly associated with overweight and obesity. Compared to normal weight
participants, participants in the obese II category spend on average 0.60 h/day (95% CI 0.36–0.84, p = 0.001) and participants
in the obese III category 0.54 h/day (95% CI 0.19–0.89, p < 0.001) more in front of any screen; participants in the obese II
category reported 0.55 h/day less sleep (95% CI − 0.67 to − 0.43, p < 0.001); participants in the obese III category were less
likely to engage in vigorous activity (OR = 0.60, 95% CI 0.43–0.84, p ≤ 0.003), or walking (OR = 0.65, 95% CI 0.49–0.88,
p = 0.005).
Conclusion  Screen time, sleeping hours, and physical activity were associated with overweight and obesity. However, these
associations were not consistent across all BMI categories. Assuming established causal connections, overweight individuals
and individuals with obesity would benefit from reduced screen time and engaging in moderate/vigorous physical activity.
Level of evidence  Level III: observational case-control analytic study.

Keywords  Obesity · Physical activity · Sleep · Lifestyle · ENSANUT 2016 · Mexico

Introduction

This article is part of topical collection on Sleep and Eating and Overweight and obesity rates have increased during the last
Weight Disorders. four decades globally [1]. More than half of the world’s pop-
ulation with obesity live in just ten countries, one of which
Electronic supplementary material  The online version of this
is Mexico. Recent epidemiological studies have shown that
article (https​://doi.org/10.1007/s4051​9-019-00841​-2) contains
supplementary material, which is available to authorized users.

3
* Spyros Kolovos MoveLab, Institute of Cellular Medicine, Newcastle
spyros.kolovos@ndorms.ox.ac.uk University, 4th Floor, William Leech Building, Medical
School, Framlington Place, Newcastle upon Tyne NE2 4HH,
1
Nuffield Department of Orthopaedics, Rheumatology UK
and Musculoskeletal Sciences, University of Oxford, Old 4
Department of Health Sciences, York University, Heslington,
Rd, Oxford OX3 7LD, UK
York YO10 5DD, UK
2
Wellcome Trust Centre for Mitochondrial Research,
Institute of Neurosciences, Newcastle University,
M4.026 Cookson Building, Frammlington Place,
Newcastle upon Tyne NE2 4HH, UK

13
Vol.:(0123456789)

170 Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

approximately 70% of adults in Mexico are either overweight and the increase in unhealthy lifestyle behaviors, the aim
or obese [2, 3]. Obesity is a risk factor for a broad range of of this study was to examine the association between
chronic diseases such as cardiovascular disorders and type modifiable lifestyle factors with overweight and obesity
II diabetes, and it has been associated with decreased life in the Mexican adult population. We addressed four spe-
expectancy, high socio-economic costs, and impaired men- cific research questions: (1) Is screen time associated with
tal health [4, 5]. The Global Burden of Diseases, Injuries, overweight and obesity? (2) Is sleep associated with over-
and Risk Factors Study (GBD) identified diet as the leading weight and obesity? (3) Is physical activity associated with
risk category for premature death, and the second highest overweight and obesity? and (4) Do lifestyle factors cluster
for disability-adjusted life-years (DALYs) worldwide [6]. together and is this cluster associated with overweight and
The GBD Study further estimated that dietary risk factors obesity?
were responsible for 11 million (22%) deaths and 255 mil-
lion (16%) DALYs worldwide in 2017 [7].
Obesity is a multifactorial disease, often expressed as
the result of a long-term energy imbalance between energy Methods
intake (diet) and energy expenditure (physical activity/
sedentary periods) [8]. A reduction in activity levels, pro- Data source and study population
longed sedentary periods and poor sleep patterns have all
been linked to an increased risk of obesity [9–11]. Watching The current study is an analysis of data collected with the
television (TV) is also a lifestyle behavior associated with 2016 Mexican National Health and Nutrition Survey (Span-
obesity, due to the lack of movement involved, the snacking ish acronym: ENSANUT) [29]. ENSANUT is organized by
related to TV viewing, or the combination of both [12, 13]. the National Institute of Public Health and the Federal Min-
Sleep has been linked to energy metabolism and body mass istry of Health in Mexico, and it provides information on
index (BMI), with recent studies suggesting that too much the health and nutritional status of the Mexican population.
or too little sleep is detrimental and may be associated with ENSANUT is a probabilistic, multistage, stratified survey,
obesity [14–16]. Furthermore, sleep, physical activity, and representative of the Mexican population at national, state
screen time have been found to be concurrent in some cases, and municipality levels, with sufficient sampling power to
but in others their correlation was moderate to small and differentiate between urban (≥ 2500 inhabitants) and rural
their associations were not statistically significant [17–19]. (< 2500 inhabitants) areas. The sampling frame consisted
Several studies have characterized these lifestyle behav- of 11,000 households 6000 in urban and 5000 in urban
iors in Mexico [20–22]. One study found that the average regions. Three age groups were defined, including children
time spent in front of the TV is 3 h per day country-wide for (5–9 years), adolescents (10–19 years), and adults (more
children between 10 and 18 years of age [21]; however, the than 20 years), and, when available, one member of each age
respective information for adults is missing. Moreover, the group was randomly selected from each household. Sam-
overall average screen time is expected to be higher due to pling weights were used to estimate nationally representa-
the increased use of mobile phones, tablets and computers tive values. A more detailed description of the sampling
[23]. Despite the World Health Organization (WHO) recom- procedures and survey methodology has been described
mendations for physical activity (i.e., a minimum of 150 min elsewhere [30]. The ENSANUT 2016 survey was approved
of moderate or 75 min of vigorous physical activity per week by the Research, Ethics and Biosafety Committees at the
[24]), in the last 12 years, there has been a significant reduc- Mexican National Institute of Public Health, and it allows
tion in physical activity levels in the Mexican population third parties to access and analyze data that will support
[20, 25]. Furthermore, daily sedentary time has increased standards of health and nutrition in the country. Written
by 8%, with an average increase of 18 min during a 9-year informed consent was obtained from all study participants
period (2006–2015) [22]. There is significant concern, more- and trained personnel administered all questionnaires and
over, that these behaviors seem to cluster in overweight and measurements face-to-face [30].
individuals with obesity [26], further increasing their risk We used available data from 6419 adults, 20 years old
of developing cardiovascular diseases, diabetes and other and above, with complete records on BMI, physical activ-
health problems [27, 28]. ity and sleep duration. We used the STROBE checklist and
Although previous studies have examined the preva- guidelines for cross-sectional studies, which provide recom-
lence of overweight and obesity, and other studies have mendations of reporting the study’s details for each section
described lifestyle behaviors potentially linked to over- of the current manuscript [31]. The dataset used for the cur-
weight and obesity, to the best of our knowledge, no stud- rent study is publically available in the Encuesta Nacional de
ies have investigated the associations between all these Salud y Nutrición 201 repository, https​://ensan​ut.insp.mx/
factors in Mexico. Given the obesity epidemic in Mexico ensan​ut201​6/desca​rga_bases​.php.

13
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179 171

Variables Clustering of lifestyle behaviors

Anthropometric measurements We investigated the clustering of the lifestyle behaviors

related to energy expenditure. To this aim, participants with
To calculate BMI, weight and height of participants were no vigorous activity, poor sleep duration (< 7 or > 8 h/day)
measured directly by trained personnel [32]. Participants [41], and prolonged screen time (> 3 h/day) [26] were char-
were classified according to their BMI following the estab- acterized as having an “unhealthy phenotype”. We used the
lished WHO classification: underweight (BMI < 18.49), nor- cut-off of > 3 h/day to operationalize prolonged screen time
mal weight (BMI = 18.5–24.9), overweight (BMI = 25–29.9), because it has been used before in the literature [26] and
obesity I (BMI 30–34.9), obesity II (BMI 35–39.9) and it is above the previously reported average screen time for
obesity III (BMI > 40) [33]. As an alternative indicator of Mexico [21]. Thus, we created two clusters, one including
obesity, waist circumference was measured at the nearest participants meeting all three of the aforementioned criteria,
0.1 cm at the minimum circumference between the bottom and the other including the rest of the participants.
of the ribs and the top of the iliac using a flexible fiberglass
anthropometric tape [34]. Covariates

A number of variables were extracted from the ENSANUT

Screen time 2016 data set and included as covariates in the analyses: sex
(women, men), age of participants (years), area of residence
Screen time was measured using a self-reported question- (rural, urban), previous diagnosis of diabetes (yes, no), and
naire, in which participants reported h/day spent on average socioeconomic status (SES). Previous diagnosis of diabetes
during the last 30 days including weekdays and weekends (1) was self-reported, and it was included as covariate because
watching TV, and (2) in front of any screen (including TV, previous studies have shown a very strong link to obesity and
mobile phones, tablets and computers) [35]. The question- lifestyle behaviors [42–44]. SES was defined by an index,
naire has previously shown acceptable validity and repro- constructed by combining eight variables that assessed the
ducibility in a Mexican population [35]. household properties and available services including con-
struction materials of the floor, ceiling, and walls; sleeping
rooms; water accessibility; vehicle ownership; household
Sleep goods (refrigerator, washing machine, microwave, stove,
boiler); and electrical goods (TV, radio, telephone, and com-
Sleep was assessed by means of: (1) sleep duration (partici- puter). SES index was divided into tertiles into the ENSA-
pants were asked the average hours of sleep per day), and (2) NUT 2016 data set and used as a proxy for low, medium,
suffering from symptoms of insomnia or not. To operational- and high SES.
ize symptoms of insomnia participants were asked (1) if they
had difficulties to fall asleep more than three times a week Statistical analysis
and (2) for how long they have been having this problem.
If they had difficulties to fall asleep more than three times a We extracted information for participants, selecting only
week for the last three weeks, they were categorized as hav- those entries with complete data for the predefined variables
ing symptoms of insomnia [36, 37]. for the study. Their clinical and demographic characteristics
were detailed using descriptive statistics. Linear regression
analyses were performed on dependent continuous variables
Physical activity and logistic regression on categorical variables. Dependent
variables were each measure of screen time, sleep duration,
The International Physical Activity Questionnaire (IPAQ) and physical activity, with BMI categories as explanatory
short form was used to measure physical activity. IPAQ variables (i.e., underweight, normal weight, overweight,
estimates for each participant were provided by the ENSA- obese I, obese II, and obese III as dummy variables). Cate-
NUT 2016 and were established using the IPAQ data pro- gorical dependent variables were symptoms of insomnia and
cessing rules (i.e., truncating all physical activity estimates having an unhealthy phenotype. Continuous dependent vari-
exceeding ‘3 h’ or ‘180 min’) [38]. The questionnaire has ables were sleep duration, TV time, and total screen time.
been previously validated in Mexican adults and has shown The three physical activity variables were continuous (i.e.
modest validity [39, 40]. Self-reported minutes per day were min/day) but they did not meet the assumption of linearity of
included for: (1) vigorous physical activity, (2) moderate residuals when included in the linear regression model and
physical activity, and (3) walking. were, therefore, dichotomized using the median as cut-off.

13

172 Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

In particular, they were defined as ‘some vigorous physi- and low physical activity; poor sleep duration and low physi-
cal activity’ if reported ≥ 1 min/day; ‘high moderate physi- cal activity; poor sleep duration and high screen time).
cal activity’ if reported > 30 min/day; and ‘high walking
time’ if reported > 60 min/day. The choice of independent
and dependent variables included in the regression models
was based on existing evidence from studies examining the Results
associations between BMI and lifestyle factors obesity [10].
As a second step, the variables sex, age, area of resi- Population characteristics
dence, SES, and history of diabetes were included as covar-
iates in the adjusted models. These variables were tested From the 8824 adults eligible in the ENSANUT 2016 data
as effect modifiers. Coefficients (b) and odds ratios (OR) set, 6419 participants with complete information for the var-
were reported for the linear and logistic regression analyses, iables of interest were included in the analyses. The 2405
respectively, together with their 95% confidence intervals excluded participants did not have complete data for at least
(95% CI). Because of the increased risk of type I error due one of the variables we included in our analysis; 838 had
to multiple statistical tests, we adjusted the p value using the no BMI or waist circumference data, 1053 no screen time,
Bonferroni correction. A p value of < 0.0072 was considered physical activity or sleep duration data, and 514 no soci-
statistically significant for all statistical tests. The R “survey” odemographic data. There were no significant differences
package (version 3.5.2) was used to account for the complex in demographic and clinical characteristics between the
survey design of ENSANUT, and the sampling weights were included and excluded participants.
used in all the analyses. The mean age of the included participants was 43 years
We performed four sensitivity analyses to test the robust- (SD = 14) and 67% were women (Table 1). Half of the par-
ness of our findings. In the first sensitivity analysis, we used ticipants were living in urban areas and half in rural set-
BMI but as a continuous variable, and in the second, we tings. From the sample, 1% was classified as underweight,
used waist circumference as a continuous variable instead 24% normal weight, 39% overweight, 25% as having obe-
of BMI as a proxy for overweight and obesity. To ensure sity I, 8% as having obesity II and 4% as having obesity
a linear association in the latter, underweight participants III (Table 1). The mean waist circumference was 95.9 cm
were excluded from the analysis. These were performed on (SD = 17.2), and 10% of the participants had history of dia-
the basis of previous evidence that the association between betes. The detailed description of clinical and demographic
obesity and other factors or diseases is sometimes influenced characteristics separately for each sex can be found in the
by the choice of anthropometric measure used to detect obe- supplementary material (Table S1).
sity [45]. The third sensitivity analysis, excluded participants The mean total screen time for the whole sample was
in the underweight group, because of the very small sam- 2.36 (SD = 2.45) h/day and for TV 1.58 (SD = 1.49) h/day.
ple size of this group, and merged all three obesity sever- The average sleep duration was 7.46 (SD = 1.43) h/day and
ity groups, because again obesity groups had relatively low 19% of the participants suffered from symptoms of insom-
sample sizes. Finally, we analyzed all the combinations of nia. Finally, the average time for vigorous activity was 39
two clustered ‘unhealthy’ behaviors (i.e., high screen time (SD = 66) min/day, for moderate activity 84 (SD = 74) min/
day, and for walking 55 (SD = 58) min/day.

Table 1  Demographic and clinical characteristics of participants

Underweight Normal weight Overweight Obese I Obese II Obese III Total

Participants, n (%) 65 (1%) 1519 (24%) 2479 (39%) 1579 (25%) 546 (8%) 231 (4%) 6419 (100%)
Women, n (%) 42 (65%) 903 (59%) 1596 (64%) 1146 (73%) 435 (80%) 183 (86%) 4305 (67%)
Age, mean (SD) 35 (15) 40 (15) 43 (13) 45 (13) 44 (12) 43 (12) 43 (14)
Rural area, n (%) 28 (43%) 788 (52%) 1279 (52%) 751 (48%) 258 (47%) 94 (44%) 3198 (50%)
Waist circumference, (cm) mean (SD) 69 (6) 83 (13) 94 (13) 102 (62) 112 (11) 125 (19) 96 (16)
Socioeconomic status
 Low, n (%) 26 (40%) 616 (40%) 846 (34%) 504 (32%) 148 (28%) 67 (31%) 2207 (34%)
 Middle, n (%) 15 (23%) 506 (32%) 816 (33%) 540 (34%) 246 (45%) 76 (35%) 2199 (34%)
 High, n (%) 24 (37%) 437 (28%) 794 (32%) 540 (34%) 145 (27%) 73 (33%) 2013 (31%)
Previous diagnosis of diabetes, n (%) 4 (6%) 86 (6%) 288 (12%) 183 (12%) 72 (13%) 29 (14) 662 (10%)

SD standard deviation

13
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179 173

Screen time screen time, with the strongest association found for SES
(b = 0.96, 95% CI 0.88–1.04, p < 0.001).
In the adjusted linear regression model, obese III cate-
gory (b = 0.36, 95% CI 0.16–0.56, p < 0.001) was asso- Sleep
ciated with higher TV time compared to normal weight
category (Table 2). In contrast, no association was found Compared to normal weight, overweight (b =− 0.18, 95%
for underweight, overweight, obese I and obese II catego- CI − 0.25 to − 0.09, p < 0.001), obese I (b =− 0.21, 95%
ries. When total screen time was considered, overweight CI − 0.31 to − 0.11, p = 0.003) and obese II categories
(b = 0.33, 95% CI 0.17–1.49, p < 0.001), obese I (b = 0.23, (b =− 0.55, 95% CI − 0.67 to − 0.43, p < 0.001) were associ-
95% CI 0.05–0.41, p = 0.006), obese II (b = 0.60, 95% CI ated with fewer sleeping hours per day, as shown in adjusted
0.36–0.84, p = 0.001) and obese III categories (b = 0.54, regression model (Table 3). No statistical differences were
95% CI 0.19–0.89, p < 0.001) were associated with higher found for the underweight and obese III categories as com-
screen time per day as compared to normal weight cat- pared with the normal weight category. There were no dif-
egory (Table  2). The covariates sex, age, area of resi- ferences in the odds of reporting symptoms of insomnia
dence, and SES were significantly associated with TV and between normal weight participants and participants in
the other BMI categories (Table 3). The covariates area of

Table 2  Adjusted linear model TV time (h/day) Total screen t­imea (h/day)

for screen time
b 95% CI p b 95% CI p

Underweight 0.22 − 0.13–0.57 0.213 0.62 0.03–1.21 0.040

Normal weight (reference category) – – – – – –
Overweight − 0.01 − 0.09–0.07 0.818 0.33 0.17–0.49 < 0.001
Obesity I 0.10 0.00–0.20 0.052 0.23 0.05–0.41 0.006
Obesity II 0.06 − 0.08–0.20 0.401 0.60 0.36–0.84 0.001
Obesity III 0.36 0.16–0.56 < 0.001 0.54 0.19–0.89 < 0.001
Area (urban) 0.13 0.03–0.23 < 0.001 0.81 0.65–0.97 < 0.003
Age (years) − 0.01 − 0.01 to − 0.01 < 0.001 − 0.07 − 0.07 to − 0.07 < 0.001
Sex (man) 0.13 0.05–0.21 < 0.001 0.46 0.34–0.58 < 0.001
History of diabetes (yes) − 0.03 − 0.17–0.11 0.710 − 0.10 − 0.34–0.14 0.372
Socioeconomic status (tertiles) 0.27 0.23–0.31 < 0.01 0.96 0.88–1.04 < 0.001
Constant 1.05 0.85–1.25 < 0.01 2.05 1.72–2.38 < 0.001

95% CI 95% confidence interval, b beta coefficient, p p value

a
 Total screen time encompasses time in front of TV, mobile phones, tablets and computers

Table 3  Adjusted linear and Sleep duration (h/day) Symptoms of insomnia

logistic models for sleep
b 95% CI p OR 95% CI p

Underweight 0.18 − 0.13–0.49 0.25 1.39 0.84–2.30 0.931

Normal weight (reference category) – – – 1.00 – –
Overweight − 0.17 − 0.25 to − 0.09 0.001 1.01 0.88–1.17 0.392
Obesity I − 0.21 − 0.31 to -0.11 0.003 1.08 0.92–1.27 0.842
Obesity II − 0.55 − 0.67 to − 0.43 < 0.001 1.17 0.94–1.46 0.151
Obesity III − 0.17 − 0.37–0.03 0.091 1.25 0.92–1.70 0.852
Area (urban) − 0.29 − 0.37 to − 0.21 < 0.001 1.35 1.20–1.52 0.007
Age (years) − 0.01 − 0.06–0.04 0.423 1.01 1.00–1.01 0.062
Sex (man) − 0.34 − 0.40 to − 0.28 < 0.001 0.62 0.54–0.70 0.018
History of diabetes (yes) 0.12 0.00–0.24 0.038 1.31 1.10–1.55 0.032
Socioeconomic status (tertiles) − 0.14 − 0.18 to − 0.10 < 0.001 1.13 1.05–1.22 0.001
Constant 8.67 8.49–8.85 < 0.001 NA NA NA

95% CI 95% confidence interval, b beta coefficient, OR odds ratio, p p value

13

174 Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

residence and SES included in the models were significantly of these 1% were in the underweight, 21% in the normal
associated with sleep duration and symptoms of insomnia. weight, 38% in the overweight, 27% in the obese I, 8% in
the obese II and 5% in the obese III category. The adjusted
Physical activity logistic regression model showed that participants in
the obese III category (OR = 1.79, 95% CI 1.18–2.72,
Results from the adjusted logistic regression models indi- p = 0.007) were more likely to have an unhealthy pheno-
cated that participants in the obese II (OR = 0.72, 95% CI type as compared with normal weight participants. No
0.58–0.90, p = 0.004), and obese III (OR = 0.60, 95% CI statistically significant differences were found for under-
0.43–0.84, p = 0.003) categories were less likely to engage weight, overweight, obese I, and obese II categories com-
in vigorous physical activity compared to normal weight pared to the normal weight category. The covariates area
participants (Table 4). Participants in the obese I category of residence, age and SES were significantly associated
(OR = 0.84, 95% CI 0.73–0.96, p = 0.006) were less likely with unhealthy phenotype (Table 5).
to engage in higher moderate physical activity compared
to normal weight participants, whereas no differences were
found for any of the other obesity categories. Participants
in the obese III category (OR = 0.65, 95% CI 0.49–0.88, Table 5  Adjusted logistic models for unhealthy phenotype
p = 0.005) were less likely to engage in higher walking time Unhealthy phenotype
as compared to normal weight participants, whereas no dif- (n = 635)
ferences were found for overweight and obesity I and II cat- OR 95% CI p
egories. All the covariates were related to vigorous physical
activity; sex and area of residence were related to moder- Underweight 0.92 0.39–2.20 0.854
ate physical activity; sex, history of diabetes, and area of Normal weight (reference category) 1.00 – –
residence were related to walking time (Table 4). From the Overweight 1.13 0.90–1.42 0.281
included covariates, being a man showed the strongest asso- Obesity I 1.19 0.93–1.53 0.162
ciation with physical activity (vigorous physical activity: Obesity II 1.29 0.93–1.80 0.130
OR = 4.92, 95% CI 4.42–5.49, p  ≤ 0.001; moderate physical Obesity III 1.79 1.18–2.72 0.007
activity: OR = 1.76, 95% CI 1.59–1.95, p  ≤ 0.001; walking: Area (urban) 1.74 1.44–2.10 < 0.001
OR = 1.63, 95% CI 1.48–1.81, p ≤ 0.001). Age (years) 0.98 0.98–0.99 < 0.001
Sex (man) 0.78 0.64–0.94 0.009
Clustering of lifestyle factors History of diabetes (yes) 1.15 0.85–1.54 0.360
Socioeconomic status (tertiles) 1.87 1.66–2.10 < 0.001
Approximately, 10% of the participants (n = 635) had an The unhealthy phenotype is defined as poor sleep duration, prolonged
“unhealthy phenotype” (combined reported poor sleep screen time, and no vigorous activity
duration, prolonged screen time, and no vigorous activity), 95% CI 95% confidence interval, OR odds ratio, p p value

Table 4  Adjusted logistic models for physical activity

Vigorous physical activity Moderate physical activity Walking (> 60 min/day)
(≥ 1 min/day) (> 30 min/day)
OR 95% CI p OR 95% CI p OR 95% CI p

Underweight 0.65 0.39–1.11 0.122 1.39 0.85–2.28 0.194 0.89 0.55–1.44 0.626
Normal weight (reference category) 1.00 – – 1.00 – – 1.00 – –
Overweight 0.98 0.86–1.12 0.791 0.90 0.79–1.02 0.063 1.01 0.90–1.14 0.846
Obesity I 0.85 0.76–0.97 0.009 0.84 0.73–0.96 0.006 0.93 0.81–1.06 0.281
Obesity II 0.72 0.58–0.90 0.004 1.00 0.82–1.20 0.844 0.89 0.74–1.08 0.247
Obesity III 0.60 0.43–0.84 0.003 1.08 0.82–1.43 0.571 0.65 0.49–0.88 0.005
Area (urban) 0.84 0.76–0.94 0.003 1.17 1.06–1.29 0.004 0.94 0.85–1.04 0.254
Age (years) 0.98 0.98–0.98 < 0.001 1.00 1.00–1.00 0.951 1.00 1.00–1.01 0.023
Sex (man) 4.92 4.42–5.49 < 0.001 1.76 1.59–1.95 < 0.001 1.63 1.48–1.81 < 0.001
History of diabetes (yes) 0.71 0.58–0.87 0.001 1.18 1.00–1.39 0.048 0.76 0.71–0.81 < 0.001
Socioeconomic status (tertiles) 0.92 0.86–0.99 0.017 0.96 0.90–1.02 0.225 0.72 0.61–0.85 < 0.001

95% CI 95% confidence interval, b beta coefficient, OR odds ratio, p p value

13
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179 175

Sensitivity analysis and total, including mobile phones, tablets, and comput-
ers). Sleep duration was associated with overweight, obe-
The analysis using BMI as continuous explanatory vari- sity I and obesity II but not with obesity III. Furthermore,
able showed results similar to the main analyses. The dif- the analysis of physical activity indicated that participants
ferent adjusted regression models showed that higher BMI with obesity were less likely to engage in vigorous physical
was associated with higher total screen time (b =0.12, 95% activity; whilst, participants in the obese I category were
CI 0.07–0.17, p < 0.001) and TV time (b =0.14, 95% CI less likely to be involved in moderate physical activity and
0.06–0.23, p < 0.001), and lower vigorous physical activity participants in the obese III category were also less likely
(OR = 0.96, 95% CI 0.95–0.97, p = 0.001) (Table 6). Larger to walk for more than 1 h a day, in all cases compared to
waist circumference was associated with higher total screen normal weight participants. Combined reported poor sleep
time (b =0.54 95% CI 0.39–0.68, p < 0.001) and TV time duration, prolonged screen time, and no vigorous activity
(b =0.63 95% CI 0.38–0.87, p < 0.001), poorer sleep duration were observed for small number of participants (10%), and
(b = − 0.47 95% CI − 0.74 to − 0.21, p < 0.001), and vigorous participants in the obese III category were more likely to
(OR = 0.99, 95% CI 0.98–0.99, p < 0.001) physical activity report this clustering.
(Table 6). Participants in the obese II and obese III categories
The results from the third sensitivity analysis, exclud- reported more than half an hour extra screen time per day
ing underweight participants and merging the three obesity compared to normal weight participants. The respective
severity levels, were in accordance with the results from association for TV time was not so strong, and for partici-
the main analyses. The three tables with the detailed results pants in the overweight and obese II categories, it was not
from the adjusted regression analyses for screen time, sleep, statistically significant. The mean TV time found in our
and physical activity can be found in the supplementary study was lower compared to the mean TV time previously
material (Tables S2, S3, and S4). The sensitivity analysis of reported in Mexican children [21]. Previous research has
the combination of clusters of ‘unhealthy’ behaviors is also shown that overall screen time is a better proxy for sedentary
in accordance with the main analysis of ‘unhealthy pheno- behavior than just TV time [46]. Our results are in line with
type’. In all three regressions, participants in the obesity II previous research indicating an association between overall
and III categories were more likely to be in the cluster of screen time and obesity, demonstrating that reducing screen
two ‘unhealthy’ behaviors compared to participants in the time could contribute to decreasing the prevalence of over-
normal weight category (Table S5). weight and obesity [47, 48]. Thus, the reduction of screen
time should be considered for weight management programs
and it should also be modified for a non-sedentary activity
Discussion to impact effectively on weight [13].
Participants in the overweight, obese I and obese II cat-
This study investigated the association of different lifestyle egories reported less sleeping time per day compared with
factors with overweight and obesity in the Mexican popula- normal weight participants. In particular, participants in the
tion. The modifiable lifestyle factor most strongly associated obese I category reported sleeping approximately half an
with overweight and obesity was screen time (both for TV hour less per day than those of normal weight. The small

Table 6  Sensitivity analysis of the association between lifestyle behaviors with BMI as continuous score and waist circumference
BMI (explanatory variable) Waist circumference (cm) (explanatory vari-
able)
­ Ra
b or O 95% CI p b or ­ORa 95% CI p

TV (h/day) 0.14 0.06–0.23 < 0.001 0.63 0.38–0.87 < 0.001

Screen (h/day) 0.12 0.07–0.17 < 0.001 0.54 0.39–0.68 < 0.001
Sleep duration (h/day) − 0.31 − 0.39 to − 0.24 0.821 − 0.47 − 0.74 to − 0.21 0.001
Symptoms of insomnia (yes) 1.00 0.99–1.00 0.231 1.00 1.00–1.00 0.081
Vigorous physical activity (higher) 0.96 0.95–0.97 0.002 0.99 0.98–0.99 < 0.001
Moderate physical activity (higher) 1.00 0.99–1.00 0.339 0.99 0.99–1.00 0.231
Walking (higher) 0.98 0.97–0.99 0.022 0.99 0.99–0.99 0.038

All the analyses were adjusted for age, sex, urban/rural, diabetes diagnose and socioeconomic status. Underweight participants were excluded
from these analyses
a
 We reported beta coefficients for continuous outcomes and odd ratios (OR) for dichotomous outcomes

13

176 Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

sample size for the obese III category and the respective may be independent of physical activity [10, 47]. It is pos-
limited statistical power to detect differences between the sible that the public is more informed about the positive
groups may explain the absence of association between impact of physical activity and less informed about the nega-
sleeping hours and obese III category. Symptoms of insom- tive impact of sedentary behavior, which may be reflected by
nia were not related to overweight or obesity in our study. the results from the screen time analysis. Thus, the findings
A number of causal pathways have linked short sleep dura- from this study could be used as evidence to update public
tion with obesity, as demonstrated by experimental stud- health guidelines.
ies [49, 50]. For instance, sleep deprivation may stimulate Participants in the obese III category had 1.79 times
appetite and increase caloric intake [51–53] due to a dys- higher odds to have an “unhealthy phenotype” compared to
regulated production of hormones related to appetite [54, normal weight participants. About 10% of the participants
55]. However, findings regarding the association between were included in the “unhealthy phenotype” group (i.e.,
sleep disturbances and obesity are contradictory, since some reported simultaneously poor sleep duration, high screen
studies showed a relationship but others did not confirm the time, and no vigorous activity). In line with our findings, a
association [56–58]. These results may be explained by the previous study found that 10% of participants with comorbid
difficulty in defining problematic sleep patterns and associ- type 2 diabetes and cardiovascular disease (CVD) reported
ated limitations of self-reported outcomes as compared to poor sleep duration, high TV time, and low physical activ-
objective measures. The ideal sleep duration varies amongst ity clustered together [26]. This suggests that even though
people and it remains subjective, especially for insomnia the clustering of negative lifestyle factors is not very preva-
as it is a self-reported group of symptoms related to sleep lent, it does occur in some individuals, for whom the risk
behaviors [56]. The ideal method to corroborate these find- for various health conditions may be increased. These indi-
ings should use a more objective approach, like the use of viduals should be targeted as priority on health intervention
activity monitors (e.g., accelerometers) or sleep tracking programs.
technology. From the covariates included in the analyses, sex, the area
This study identified that participants with obesity from of residence and SES were consistently associated with the
all three categories were less likely to engage in vigorous different lifestyle factors. For instance, men reported more
physical activity as compared to normal weight participants. screen time and higher odds of engaging in vigorous physi-
However, these associations did not hold for moderate physi- cal activity compared to women. Overall, it appears that
cal activity, where the only statistically significant difference these sociodemographic factors influence the association
was found for participants in the obese I category compared between BMI and the modifiable behaviors. Reasons behind
to normal weight participants. Similarly, only obese III par- these differences could inform health prevention plans in
ticipants were less likely to engage in at least an hour of low- and middle-income countries. These cultural and sex
walking time per day. A previous study using data from the differences should be addressed when designing intervention
ENSANUT 2006 and 2012 in Mexico found only a small dif- programs to improve lifestyle behaviors and these should be
ference between BMI categories and reported moderate-to- designed population specific [61].
vigorous physical activity minutes per week [20]. For exam- Results from the sensitivity analyses were, in general,
ple, overweight participants engaged in four more minutes in accordance with the results from the main analyses. The
of physical activity per week when compared to the normal two analyses including BMI and WC as continuous outcome
weight group based on data from the ENSANUT 2006, and variable yielded similar results. The only difference was that
3 min less as indicated by ENSANUT 2012. In both surveys, WC was associated with sleep duration but BMI was not.
participants in the obese group engaged in lower physical This may be an indication that the association between obe-
activity (i.e., 17–25 min less per week) compared to the sity and lifestyle factors is influenced by the measure used
normal weight group. Therefore, it seems that the associa- to operationalize obesity, which has been reported before in
tion of physical activity and overweight and obesity in the similar studies [5].
Mexican population is not conclusive and more research is
required. Future studies looking into this association should Strengths and limitations
use objective devices, such as accelerometers, to measure
physical activity [59, 60]. This study was undertaken based on a large-scale survey rep-
In this study, screen time was associated with overweight resentative of the Mexican population, thereby signaling that
and obesity, whereas the association between physical its statistically significant findings are widely applicable in the
activity and obesity appears more complex. This may be wide Mexican context. We conducted sensitivity analyses to
explained by the growing evidence that sedentary behavior, examine if the operationalization of overweight and obesity
for which screen time can be considered a proxy, is a distinct influenced our findings and demonstrated that the results were
risk factor for various conditions, including obesity, and it similar across all analyses. Furthermore, obesity was measured

13
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179 177

by trained personnel using objective measures and was not also prevalent. However, no studies have investigated the
self-reported by the participants [62]. associations between obesity and unhealthy lifestyle behav-
However, our analyses faced limitations. The variables iors in Mexico.
included as outcomes were reported by the participants
themselves and, therefore, are subject to recall bias [63],
which may have resulted for instance in an underestimation What does this study add?
of the percentage of participants with unhealthy phenotype
and an overestimation of the reported physical activity [64]. Screen time, sleeping hours, and physical activity were
Moreover, the study has a cross-sectional design that does found to be associated with obesity; although, these asso-
not allow to draw any conclusions on the direction of causal- ciations were not consistent across all obesity levels.
ity between BMI categories and lifestyle factors [65]. The
WHO recommends a minimum of 150 min of moderate
physical activity or 75 min of vigorous physical activity per Compliance with ethical standards 
week, but we could not use this cut-off in our study due to
the very high percentage of participants reporting values Conflict of interest  The authors declare that they have no conflict of
interest.
exceeding those guidelines [24]. It has been reported that
some participants over-report their physical activity levels Ethical approval and informed consent  Ethical approval and informed
in these types of surveys [66], and for this reason we used consent were not required for this study.
the median of each physical activity variable as the cut-off
in corresponding analyses. Finally, the criteria for diagnosis Open Access  This article is licensed under a Creative Commons Attri-
of insomnia based on the Diagnostic and Statistical Manual bution 4.0 International License, which permits use, sharing, adapta-
of Mental Disorders 5th edition (DSM-5) regard the past tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
3 months, whereas the question in the ENSANUT survey provide a link to the Creative Commons licence, and indicate if changes
regarded the past 3 weeks. The question about duration of were made. The images or other third party material in this article are
symptoms can be amended in future surveys to resemble included in the article’s Creative Commons licence, unless indicated
more the DSM criteria for each disorder. otherwise in a credit line to the material. If material is not included in
the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
Conclusion need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/.
This is the first study to examine different modifiable life-
style factors associated with BMI categories in Mexico,
a highly relevant undertaking considering the high rates References
of overweight and obesity in the country’s adult popula-
tion. Screen time, sleeping hours, and physical activity 1. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono
C, Mullany EC, Biryukov S, Abbafati C, Abera SF, Abraham
were found to be associated with obesity. However, these JP, Abu-Rmeileh NME, Achoki T, AlBuhairan FS, Alemu ZA,
associations were not consistent across all obesity levels. Alfonso R, Ali MK, Ali R, Guzman NA, Ammar W, Anwari P,
Unhealthy lifestyle factors seem to cluster together for a few Banerjee A, Barquera S, Basu S, Bennett DA, Bhutta Z, Blore J,
individuals, and the risk of this clustering may be increased Cabral N, Nonato IC, Chang J-C, Chowdhury R, Courville KJ,
Criqui MH, Cundiff DK, Dabhadkar KC, Dandona L, Davis A,
for obese individuals. Given the high prevalence of obesity Dayama A, Dharmaratne SD, Ding EL, Durrani AM, Esteghamati
and unhealthy behaviors in Mexico and the limited effective- A, Farzadfar F, Fay DFJ, Feigin VL, Flaxman A, Forouzanfar
ness of current strategies to tackle obesity, multi-disciplinary MH, Goto A, Green MA, Gupta R, Hafezi-Nejad N, Hankey GJ,
lifestyle interventions should be explored to tackle the obe- Harewood HC, Havmoeller R, Hay S, Hernandez L, Husseini A,
Idrisov BT, Ikeda N, Islami F, Jahangir E, Jassal SK, Jee SH,
sity epidemic in the country. Assuming established causal Jeffreys M, Jonas JB, Kabagambe EK, Khalifa SEAH, Kengne
connections, individuals with obesity would benefit from AP, Khader YS, Khang Y-H, Kim D, Kimokoti RW, Kinge JM,
improved sleep, reduced screen time and engaging in vigor- Kokubo Y, Kosen S, Kwan G, Lai T, Leinsalu M, Li Y, Liang X,
ous physical activity. Liu S, Logroscino G, Lotufo PA, Lu Y, Ma J, Mainoo NK, Men-
sah GA, Merriman TR, Mokdad AH, Moschandreas J, Naghavi
M, Naheed A, Nand D, Narayan KMV, Nelson EL, Neuhouser
ML, Nisar MI, Ohkubo T, Oti SO, Pedroza A, Prabhakaran D,
What is already known on this subject? Roy N, Sampson U, Seo H, Sepanlou SG, Shibuya K, Shiri R,
Shiue I, Singh GM, Singh JA, Skirbekk V, Stapelberg NJC, Stu-
rua L, Sykes BL, Tobias M, Tran BX, Trasande L, Toyoshima H,
Approximately, 70% of adults in Mexico are overweight or van de Vijver S, Vasankari TJ, Veerman JL, Velasquez-Melendez
obese. Unhealthy lifestyle behaviors, such as low physical G, Vlassov VV, Vollset SE, Vos T, Wang C, Wang X, Weider-
activity, poor sleep duration, and prolonged screen time, are pass E, Werdecker A, Wright JL, Yang YC, Yatsuya H, Yoon

13

178 Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179

J, Yoon S-J, Zhao Y, Zhou M, Zhu S, Lopez AD, Murray CJL, 16. Norton MC, Eleuteri S, Cerolini S, Ballesio A, Conte SC, Falaschi
Gakidou E (2014) Global, regional, and national prevalence of P, Lucidi F (2018) Is poor sleep associated with obesity in older
overweight and obesity in children and adults during 1980–2013: adults? A narrative review of the literature. Eat Weight Disord
a systematic analysis for the Global Burden of Disease Study EWD 23(1):23–38. https​://doi.org/10.1007/s4051​9-017-0453-2
2013. Lancet 384(9945):766–781. https​://doi.org/10.1016/S0140​ 17. Aguilar MM, Vergara FA, Velásquez EJA, García-Hermoso A
-6736(14)60460​-8 (2015) Physical activity, screen time and sleep patterns in Chilean
2. Gutierrez J, Rivera-Dommarco J, Shamah-Levy T, Villalpando- girls. Anales de Pediatría (English Edition) 83(5):304–310. https​
Hernandez S, Franco A, Cuevas-Nasu L, Romero-Martínez M, ://doi.org/10.1016/j.anped​e.2014.12.004
Hernández-Ávila M (2012) Encuesta Nacional de Salud y Nutri- 18. Tandon PS, Sasser T, Gonzalez ES, Whitlock KB, Christakis DA,
ción 2012. ResultadosNacionales. Instituto Nacional de Salud Stein MA (2019) Physical activity, screen time, and sleep in chil-
Pública, México dren with ADHD. J Phys Act Health 16(6):416–422. https​://doi.
3. Rtveladze K, Marsh T, Barquera S, Sanchez Romero LM, Levy org/10.1123/jpah.2018-0215
D, Melendez G, Webber L, Kilpi F, McPherson K, Brown M 19. Laurson KR, Lee JA, Gentile DA, Walsh DA, Eisenmann JC
(2014) Obesity prevalence in Mexico: impact on health and (2014) Concurrent associations between physical activity, screen
economic burden. Public health Nutr 17(1):233–239. https​:// time, and sleep duration with childhood obesity. ISRN Obes
doi.org/10.1017/s1368​98001​30000​86 2014:204540. https​://doi.org/10.1155/2014/20454​0
4. Dandona P, Aljada A, Chaudhuri A, Mohanty P, Garg R (2005) 20. Medina C, Janssen I, Campos I, Barquera S (2013) Physical inac-
Metabolic syndrome a comprehensive perspective based on tivity prevalence and trends among Mexican adults: results from
interactions between obesity, diabetes, and inflammation. Cir- the National Health and Nutrition Survey (ENSANUT) 2006 and
culation 111(11):1448–1454 2012. BMC Public Health 13(1):1063
5. Zavala GA, Kolovos S, Chiarotto A, Bosmans JE, Campos- 21. Janssen I, Medina C, Pedroza A, Barquera S (2013) Screen time in
Ponce M, Rosado JL, Garcia OP (2018) Association between Mexican children: findings from the 2012 National Health and Nutri-
obesity and depressive symptoms in Mexican population. Soc tion Survey (ENSANUT 2012). Salud Publica Mex 55(5):484–491
Psychiatry Psychiatr Epidemiol 53(6):639–646. https​: //doi. 22. Medina C, Tolentino-Mayo L, Lopez-Ridaura R, Barquera S
org/10.1007/s0012​7-018-1517-y (2017) Evidence of increasing sedentarism in Mexico city during
6. GBD 2017 Risk Factor Collaborators (2018) Global, regional, the last decade: sitting time prevalence, trends, and associations
and national comparative risk assessment of 84 behavioural, with obesity and diabetes. PLoS One 12(12):e0188518. https​://
environmental and occupational, and metabolic risks or clusters doi.org/10.1371/journ​al.pone.01885​18
of risks for 195 countries and territories, 1990–2017: a sys- 23. Mark AE, Boyce WF, Janssen I (2006) Television viewing, com-
tematic analysis for the Global Burden of Disease Study 2017. puter use and total screen time in Canadian youth. Paediatr Child
Lancet 392 (10159):1923–1994. https​://doi.org/10.1016/s0140​ Health 11(9):595–599
-6736(18)32225​-6 24. Stegenga H, Haines A, Jones K, Wilding J (2014) Identification,
7. Afshin A, Sur PJ, Fay KA, Cornaby L, Ferrara G, Salama JS, assessment, and management of overweight and obesity: summary
Mullany EC, Abate KH, Abbafati C, Abebe Z (2019) Health of updated NICE guidance. BMJ 349:g6608
effects of dietary risks in 195 countries, 1990–2017: a system- 25. Ahmadi N, Chung SA, Gibbs A, Shapiro CM (2008) The Berlin
atic analysis for the Global Burden of Disease Study 2017. Lan- questionnaire for sleep apnea in a sleep clinic population: rela-
cet 393(10184):1958–1972 tionship to polysomnographic measurement of respiratory distur-
8. Hall KD, Sacks G, Chandramohan D, Chow CC, Wang YC, Gort- bance. Sleep Breath 12(1):39–45
maker SL, Swinburn BA (2011) Quantification of the effect of 26. Cassidy S, Chau JY, Catt M, Bauman A, Trenell MI (2016) Cross-
energy imbalance on bodyweight. Lancet 378(9793):826–837 sectional study of diet, physical activity, television viewing and
9. McGuire MT, Wing RR, Klem ML, Lang W, Hill JO (1999) What sleep duration in 233 110 adults from the UK Biobank; the behav-
predicts weight regain in a group of successful weight losers? J ioural phenotype of cardiovascular disease and type 2 diabetes.
Consult Clin Psychol 67(2):177–185 BMJ Open 6(3):e010038
10. Cassidy S, Chau JY, Catt M, Bauman A, Trenell MI (2017) Low 27. Santos A-C, Ebrahim S, Barros H (2007) Alcohol intake, smok-
physical activity, high television viewing and poor sleep duration ing, sleeping hours, physical activity and the metabolic syndrome.
cluster in overweight and obese adults; a cross-sectional study of Prev Med 44(4):328–334
398,984 participants from the UK Biobank. Int J Behav Nutr Phys 28. Gopinath B, Hardy LL, Baur LA, Burlutsky G, Mitchell P (2012)
Act 14(1):57. https​://doi.org/10.1186/s1296​6-017-0514-y Physical activity and sedentary behaviors and health-related qual-
11. Beccuti G, Pannain S (2011) Sleep and obesity. Curr Opin ity of life in adolescents. Pediatrics 130:e167
Clin Nutr Metab Care 14(4):402–412. https​://doi.org/10.1097/ 29. Romero-Martinez M, Shamah-Levy T, Franco-Nunez A, Vil-
MCO.0b013​e3283​47910​9 lalpando S, Cuevas-Nasu L, Gutierrez JP, Rivera-Dommarco JA
12. Ekelund U, Brage S, Froberg K, Harro M, Anderssen SA, Sardinha (2013) National Health and Nutrition Survey 2012: design and
LB, Riddoch C, Andersen LB (2006) TV viewing and physical coverage. Salud Publica Mex 55(Suppl 2):S332–S340
activity are independently associated with metabolic risk in chil- 30. Romero-Martinez M, Shamah-Levy T, Cuevas-Nasu L, Men-
dren: the European Youth Heart Study. PLoS Med 3(12):e488 dez Gomez-Humaran I, Gaona-Pineda EB, Gómez-Acosta LM,
13. Duggan MP, Taveras EM, Gerber MW, Horan CM, Oreskovic Rivera-Dommarco JÁ, Hernandez-Avila M (2017) Methodo-
NM (2019) Presence of small screens in bedrooms is associated logical design of the National Health and Nutrition Survey 2016.
with shorter sleep duration and later bedtimes in children with Salud Publica Mex 59(3):299–305
obesity. Acad Pediatr 19(5):515–519. https​://doi.org/10.1016/j. 31. Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Van-
acap.2018.11.004 denbroucke JP, Initiative S (2007) The Strengthening the Report-
14. Taheri S, Lin L, Austin D, Young T, Mignot E (2004) Short sleep ing of Observational Studies in Epidemiology (STROBE) state-
duration is associated with reduced leptin, elevated ghrelin, and ment: guidelines for reporting observational studies. Prev Med
increased body mass index. PLoS Med 1(3):e62 45(4):247–251
15. Taheri S (2006) The link between short sleep duration and obesity: 32. Barquera S, Campos-Nonato I, Hernandez-Barrera L, Pedroza A,
we should recommend more sleep to prevent obesity. Arch Dis Rivera-Dommarco JA (2013) Prevalence of obesity in Mexican
Child 91(11):881–884 adults 2000–2012. Salud publica de Mexico 55:S151–S160

13
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2021) 26:169–179 179

33. World Health Organization (2000) Obesity: preventing and man- 52. Rechtschaffen A, Bergmann BM (1995) Sleep deprivation
aging the global epidemic. World Health Organization, vol 894 in the rat by the disk-over-water method. Behav Brain Res
34. World Health Organization (2011) Waist circumference and waist- 69(1–2):55–63
hip ratio: report of a WHO expert consultation, Geneva, 8–11 53. Kandeger A, Selvi Y, Tanyer DK (2019) The effects of individual
December 2008 circadian rhythm differences on insomnia, impulsivity, and food
35. Hernandez B, Gortmaker S, Laird N, Colditz G, Parra-Cabrera S, addiction. Eat and Weight Disord EWD 24(1):47–55. https​://doi.
Peterson K (2000) Validity and reproducibility of a questionnaire org/10.1007/s4051​9-018-0518-x
on physical activity and non-activity for school children in Mexico 54. Spiegel K, Tasali E, Penev P, Van Cauter E (2004) Brief com-
city. Salud Publica Mex 42(4):315–323 munication: sleep curtailment in healthy young men is associated
36. Guerrero-Zúñiga S, Gaona-Pineda EB, Cuevas-Nasu L, Torre- with decreased leptin levels, elevated ghrelin levels, and increased
Bouscoulet L, Reyes-Zúñiga M, Shamah-Levy T, Pérez-Padilla hunger and appetite. Ann Intern Med 141(11):846–850
R (2018) Prevalencia de síntomas de sueño y riesgo de apnea 55. Kohatsu ND, Tsai R, Young T, VanGilder R, Burmeister LF,
obstructiva del sueño en México. Salud publica de Mexico Stromquist AM, Merchant JA (2006) Sleep duration and body mass
60(3):347–355 index in a rural population. Arch Intern Med 166(16):1701–1705
37. American Academy of Sleep Medicine (2005) International clas- 56. Crönlein T (2016) Insomnia and obesity. Curr Opin Psychiatry
sification of sleep disorders: diagnostic and coding classification 29(6):409–412
of sleep disorders 697 manual, 2nd edn. American Academy of 57. Zhou Q, Zhang M, Hu D (2019) Dose-response association
Sleep Medicine, Westchester between sleep duration and obesity risk: a systematic review and
38. Medina C, Jauregui A, Campos-Nonato I, Barquera S (2018) Prev- meta-analysis of prospective cohort studies. Sleep Breath. https​
alence and trends of physical activity in children and adolescents: ://doi.org/10.1007/s1132​5-019-01824​-4
results of the Ensanut 2012 and Ensanut MC 2016. Salud Publica 58. Tambalis KD, Panagiotakos DB, Psarra G, Sidossis LS (2018)
Mex 60(3):263–271 Insufficient Sleep duration is associated with dietary habits, screen
39. Committee IR (2005) Guidelines for data processing and analysis time, and obesity in children. J Clin Sleep Med 14(10):1689–
of the International Physical Activity Questionnaire (IPAQ)-short 1696. https​://doi.org/10.5664/jcsm.7374
and long forms. http://www.ipaq.ki.se/scori​ng.pdf 59. Celis-Morales CA, Perez-Bravo F, Ibanez L, Salas C, Bailey ME,
40. Medina C, Barquera S, Janssen I (2013) Validity and reliability of Gill JM (2012) Objective vs. self-reported physical activity and
the International Physical Activity Questionnaire among adults in sedentary time: effects of measurement method on relationships
Mexico. Revista Panamericana de Salud Pública 34(1):21–28 with risk biomarkers. PloS One 7(5):345
41. Beihl DA, Liese AD, Haffner SM (2009) Sleep duration as a risk 60. Wyke S, Bunn C, Andersen E, Silva MN, van Nassau F, McSkim-
factor for incident type 2 diabetes in a multiethnic cohort. Ann ming P, Kolovos S, Gill JMR, Gray CM, Hunt K, Anderson
Epidemiol 19(5):351–357 AS, Bosmans J, Jelsma JGM, Kean S, Lemyre N, Loudon DW,
42. Stein CJ, Colditz GA (2004) The epidemic of obesity. J Clin Endo- Macaulay L, Maxwell DJ, McConnachie A, Mutrie N, Nijhuis-van
crinol Metab 89(6):2522–2525 der Sanden M, Pereira HV, Philpott M, Roberts GC, Rooksby
43. Hu FB, Manson JE, Stampfer MJ, Colditz G, Liu S, Solomon CG, J, Roynesdal OB, Sattar N, Sorensen M, Teixeira PJ, Treweek
Willett WC (2001) Diet, lifestyle, and the risk of type 2 diabetes S, van Achterberg T, van de Glind I, van Mechelen W, van der
mellitus in women. N Engl J Med 345(11):790–797 Ploeg HP (2019) The effect of a programme to improve men’s
44. Eckel RH, Kahn SE, Ferrannini E, Goldfine AB, Nathan DM, sedentary time and physical activity: the European Fans in
Schwartz MW, Smith RJ, Smith SR (2011) Obesity and type 2 Training (EuroFIT) randomised controlled trial. PLoS Med
diabetes: what can be unified and what needs to be individualized? 16(2):e1002736. https​://doi.org/10.1371/journ​al.pmed.10027​36
J Clin Endocrinol Metab 96(6):1654–1663 61. Narcisse MR, Long CR, Felix HC, Howie EK, Purvis RS, McElf-
45. Sebo P, Beer-Borst S, Haller DM, Bovier PA (2008) Reliability ish PA (2019) Adherence to sleep guidelines reduces risk of over-
of doctors’ anthropometric measurements to detect obesity. Prev weight/obesity in addition to 8-5-2-1-0 guidelines among a large
Med 47(4):389–393 sample of adolescents in the United States. Sleep Health. https​://
46. Healy GN, Clark BK, Winkler EA, Gardiner PA, Brown WJ, doi.org/10.1016/j.sleh.2019.03.004
Matthews CE (2011) Measurement of adults’ sedentary time in 62. Pursey K, Burrows TL, Stanwell P, Collins CE (2014) How accu-
population-based studies. Am J Prev Med 41(2):216–227 rate is web-based self-reported height, weight, and body mass
47. Thorp AA, Owen N, Neuhaus M, Dunstan DW (2011) Sedentary index in young adults? J Med Internet Res 16(1):e4
behaviors and subsequent health outcomes in adults: a system- 63. Ezzati M, Martin H, Skjold S, Hoorn SV, Murray CJ (2006)
atic review of longitudinal studies, 1996–2011. Am J Prev Med Trends in national and state-level obesity in the USA after cor-
41(2):207–215 rection for self-report bias: analysis of health surveys. J R Soc
48. Mitchell JA, Rodriguez D, Schmitz KH, Audrain-McGovern J Med 99(5):250–257
(2013) Greater screen time is associated with adolescent obesity: 64. Prince SA, Adamo KB, Hamel ME, Hardt J, Gorber SC, Tremblay
a longitudinal study of the BMI distribution from ages 14 to 18. M (2008) A comparison of direct versus self-report measures for
Obesity 21(3):572–575 assessing physical activity in adults: a systematic review. Int J
49. Capers PL, Fobian AD, Kaiser KA, Borah R, Allison DB (2015) Behav Nutr Phys Act 5(1):56
A systematic review and meta-analysis of randomized controlled 65. Ekelund U, Brage S, Besson H, Sharp S, Wareham NJ (2008)
trials of the impact of sleep duration on adiposity and components Time spent being sedentary and weight gain in healthy adults:
of energy balance. Obes Rev 16(9):771–782 reverse or bidirectional causality? Am J Clin Nutr 88(3):612–617
50. Eleuteri S, Norton MC, Livi F, Grano C, Falaschi P, Violani C, 66. Rzewnicki R, Auweele YV, De Bourdeaudhuij I (2003) Address-
Lucidi F, Lombardo C (2018) Sleep quality as predictor of BMI ing overreporting on the International Physical Activity Question-
in non-depressed caregivers of people with dementia. Eat Weight naire (IPAQ) telephone survey with a population sample. Public
Disord EWD 23(5):553–560. https​://doi.org/10.1007/s4051​ Health Nutr 6(3):299–305
9-018-0560-8
51. Baron KG, Reid KJ, Kern AS, Zee PC (2011) Role of sleep timing Publisher’s Note Springer Nature remains neutral with regard to
in caloric intake and BMI. Obesity 19(7):1374–1381 jurisdictional claims in published maps and institutional affiliations.

13