You are on page 1of 12

Eating Frequency and Overweight and Obesity in

Children and Adolescents: A Meta-analysis


abstract

AUTHORS: Panagiota Kaisari, MSc, Mary Yannakoulia, PhD,


and Demosthenes B. Panagiotakos, PhD

OBJECTIVES: To determine the effect of eating frequency on body


weight status in children and adolescents.

Department of Nutrition and Dietetics, Harokopio University,


Athens, Greece

METHODS: In this meta-analysis, original observational studies published


to October 2011 were selected through a literature search in the PubMed
database. The reference list of the retrieved articles was also used to
identify relevant articles; researchers were contacted when needed.
Selected studies were published in English, and they reported on the
effect of eating frequency on overweight/obesity in children and
adolescents. Pooled effect sizes were calculated using a random
effects model.
RESULTS: Ten cross-sectional studies and 1 case-control study (21
substudies in total), comprising 18 849 participants (aged 219 years),
were included in the analysis. Their combined effect revealed that the
highest category of eating frequency, as compared with the lowest,
was associated with a benecial effect regarding body weight status
in children and adolescents (odds ratio [OR] = 0.78, log OR = 0.24,
95% condence interval [CI] 0.41 to 0.06). The observed benecial
effect remained signicant in boys (OR = 0.76, log OR = 0.27, 95% CI
0.47 to 0.06), but not in girls (OR = 0.96, log OR = 0.04, 95% CI 0.40
to 0.32) (P for sex differences = 0.14).
CONCLUSIONS: Higher eating frequency was associated with lower
body weight status in children and adolescents, mainly in boys. Clinical
trials are warranted to conrm this inverse association, evaluate its
clinical applicability, and support a public health recommendation;
more studies are also needed to further investigate any sex-related
differences, and most importantly, the biological mechanisms.
Pediatrics 2013;131:958967

958

KEY WORDS
eating frequency, obesity, children
ABBREVIATIONS
CIcondence interval
ORodds ratio
Ms Kaisari searched the literature and extracted data, drafted
the manuscript, and contributed to the interpretation of the
results; Dr Yannakoulia conceptualized the analysis, searched
the literature and extracted data, contributed to the
interpretation of the results, and critically reviewed the
manuscript; Dr Panagiotakos conceptualized the analysis,
performed the statistical analysis, contributed to the
interpretation of the results, and reviewed and revised the
manuscript; and all authors approved the nal manuscript as
submitted.
www.pediatrics.org/cgi/doi/10.1542/peds.2012-3241
doi:10.1542/peds.2012-3241
Accepted for publication Jan 22, 2013
Address correspondence to Mary Yannakoulia, PhD, Department
of Nutrition and Dietetics, Harokopio University, El. Venizelou 70,
Athens 17671, Greece. E-mail: myiannak@hua.gr
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
Copyright 2013 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they have
no nancial relationships relevant to this article to disclose.
FUNDING: No external funding.

KAISARI et al

Downloaded from by guest on December 26, 2016

REVIEW ARTICLE

The rates of childhood obesity have


risen to an alarming level over the past
3 decades.1 Evidence so far attributes
this phenomenon to the interplay between genetic and environmental factors.2 As far as diet is concerned, total
energy, macronutrients, and food group
intake have long been investigated.36
However, the existing dietary etiological
models cannot fully explain the development and maintenance of childhood obesity: research needs to move
beyond specic nutrients and foods and
focus on dietary patterns and eating
behaviors, including eating frequency.7
Most studies in adults suggest that
greater eating frequency is associated
with a healthier weight status.812 Fabry
and colleagues,13 in the 1960s, were the
rst to explore this effect in children.
They performed an interventional
study on 226 children, aged 6 to 16, and
found that children who were provided
3 meals per day had an increased
tendency to deposit fat, compared with
those who were given 7 or 5 meals per
day in smaller portions. Interestingly,
these differences were found in older
(1016 years), but not younger children (boys: 611 years and girls: 610
years), and were more obvious in girls
compared with boys. The authors
speculated that the hormonal changes
during puberty may mediate the effect
of eating frequency on body weight.
Since then, numerous epidemiologic
studies have been conducted to explore the potential relationship between the total number of eating
events or the number of meals and
snacks and obesity risk in children and
adolescents.1416 However, the effect, if
any, still remains unclear, as well as
the potential underlying mechanisms.
Kirk17 proposed that there are 4 possible physiologic advantages associated
with frequent eating that may improve
body-weight control: frequent eating
may help people to regulate appetite and
daily energy compensation, to increase

the carbohydrate-to-fat ratio of the diet,


to shift the temporal distribution of energy intake toward earlier parts of day
and to keep a more physically active
lifestyle.
The research question addressed here
was to evaluate at what extent eating
frequency is associated with body
weight status (overweight/obesity), in
children and adolescents, through a
meta-analysis of the existing ndings of
published original observational studies. As there is no scientic consensus
on the most appropriate denition to
categorize the different eating occasions (ie, meals versus snacks), total
eating frequency was evaluated.

METHODS
Search Strategy
Original research, observational studies
published until October 2011, examining
the association between eating frequency and overweight/obesity status in
children and adolescents, were selected
through a literature search in the
PubMed database. Specic key words
were used for this search: eating frequency, meal frequency, meals, and
eating episodes, in combination with
the term overweight or obesity. In
addition, the reference list of the retrieved and eligible articles was used
to identify relevant articles that were
not extracted through the searching
procedure. The initial search resulted
in 1069 entries in PubMed on eating frequency and obesity/overweight,
155 entries on meal frequency and
overweight/obesity, 414 entries on meals
and overweight/obesity, and 69 entries
on eating episodes and overweight/
obesity.
Study Selection
The relevance of studies was assessed
by using a hierarchical approach based
on title, abstract, and full manuscript.
Studies eligible for inclusion were those

PEDIATRICS Volume 131, Number 5, May 2013

published in English. All of the selected


studies were conducted in children
and/or adolescents (219 years old),
and provided results in a form that
could be used for the present analysis
(point estimates of odds ratio [OR] with
95% condence intervals [CIs] or SEs
available or derivable). The selected
studies were cross-sectional, apart from
1 case-control. In the present metaanalysis, the Preferred Reporting Items
for Systematic Reviews and Metaanalyses guidelines were followed.18
Data Extraction
The following study characteristics were
extracted from the original reports by
using a standardized data extraction
form and included in the meta-analysis:
name of the study (if there was a specic
one), lead author, year of publication,
country of origin, sample size, mean age
and sex of participants, effect size
measurements (OR), and variables that
entered into the multivariable model as
potential confounding factors. This information was also extracted: 95% CIs,
the number of daily meals/eating episodes used as reference category, the
method used to evaluate eating frequency, the denition of a meal/eating
episode, and the denition criteria for
overweight/obesity in children and
adolescents. In some cases, additional
information or data were requested
from the authors; 4 authors were contacted,1922 but only 3 of them responded to our request.19,21,22
One investigator (P.K.) collected the
relevant articles; 2 authors reviewed
the literature (P.K. and M.Y.) independently and disagreements were
solved by consensus and by the opinion
of the third author (D.B.P.). Quality of
studies was not assessed because the
total number of studies was limited.
Statistical Analysis
Results of the individual studies are presented as ORs and their corresponding

959

Downloaded from by guest on December 26, 2016

95% CIs. Data were collected from the


most adjusted model. To calculate the
combined effect of the selected studies,
a random effects meta-analysis was
applied. Each study was represented by
a dummy variable, and the use of
a random effects model was possible
because SEs of the point estimates
within studies were provided by the
investigators. With regard to the analysis for overweight/obesity, each data
point consisted of the number of meals/
eating episodes consumed daily and the
OR of being overweight/obese. The
combined effect size measures derived
from the meta-analysis are presented
as ORs, log ORs, and their 95% CIs.
Heterogeneity that was attributed to
studies rather than to chance was
assessed by using Cochrans Q and I2
(I2 ranges between 0% and 100%, with
lower values representing less heterogeneity) and evaluated by using the
x2 test. To assess the presence of
publication bias, the funnels plot was
created, and tested with the Egger
linear regression method of asymmetry (test of the intercept). All statistical calculations were performed in
NCSS/PASS 2005 software (Number
Cruncher Statistical Systems Co,
Kaysville, UT).

RESULTS
Study Selection
Figure 1 shows results from the literature search and study-selection process. Overall, 11 studies were nally
included in this meta-analysis, after
screening 1707 titles and abstracts and
reviewing 30 full articles. A manual
search of references cited in these
articles did not yield new eligible articles. Among these 11 studies, some
studies presented their results according to participants sex,14,15,2123
ethnic origin,14 and age group.22 Moreover, the report of Toschke et al16 included separate results of OR for
overweight and for obesity. Therefore,
960

FIGURE 1
Flow diagram for selection of studies.

21 substudies of the initial 11 original


studies were included in the metaanalysis.
Study Characteristics
Study characteristics are presented in
Tables 1 and 2. The overall working
sample consisted of 18 849 subjects,
aged 2 to 19 years. Sample sizes of the
21 substudies varied between 265 and
4370 participants. With regard to the
geographical distribution of the studies, 8 were conducted in Mediterranean populations (Greece, Portugal,
Italy, and Turkey), 2 in North American
North European populations (United
States and Germany), and 1 in a Latin
American population (Brazil). No racerelated differences of the participants
were reported, with the exception of
the Bogalusa Heart Study; in this case,
the study population consisted of 1562
children, aged 10 years, with a biracial
background (65% European American,
35% African American). In most studies, participants were categorized as
overweight/obese according to International Obesity Task Force BMI cutoff
points.24 However, other criteria were

also used: the reference standards of


the Centers for Disease Control and Prevention,25 and the Greek growth charts
developed by the First Department of
Pediatrics of Athens University Medical
School.26
As it was revealed from the studies
methodology section, the evaluation of
eating frequency was mainly based on
self-reported questionnaires completed by children (aged 1019 years)
themselves,15,23,27,28 or by their caregivers/parents.16,19,21,2931 However, in
2 studies,14,22 information from the
24-hour dietary recalls was used. There
are also differences between studies in
the denition used for a meal or an
eating episode. For example, in the study
of Turkkahraman et al,29 the total number of daily meals consumed referred to
the number of meals considered as
formal eating events, and, thus, snacks
were excluded from analysis, whereas in
the study of Kontogianni et al,22 any eating or drinking occasion separated by at
least 15 minutes from other events during the day was dened as an eating
episode, accounting for the total daily
number of eating episodes.

KAISARI et al

Downloaded from by guest on December 26, 2016

Portugal,
2008

Brazil,
2003

Italy,
2006

Turkey,
2006

Neutzling et al27

Barba et al19
(ARCA project)

Turkkahraman
et al29

Germany,
2005

Toschke et al16

Mota et al15

US, 2003

Country,
Year

Nicklas et al
(The Bogalusa
Heart Study)

14

Reference/ Name
of the Study

461

2070

PEDIATRICS Volume 131, Number 5, May 2013

Downloaded from by guest on December 26, 2016


2465

3668

508

513 EA/279 AA

Girls

425

2300

497 EA/273 AA

Boys

Sample, n

617

611

1519

1317

57

10

Age

Cross-sectional

Cross-sectional

Case-control

Cross-sectional

Cross-sectional

Cross-sectional

Type of Study

Parental questionnaire

Parental questionnaire
(Yes or No)

Standardized and
pre-coded questionnaire
self-reported
(Number of daily meals)

Self-reported questionnaire
(possible answers:1/2/3/4/5/6)

Age
Frequency and amount of
physical activity
Hours of sleep
Hours of watching television,
playing video games, and using
a computer per day
Dietary habits
Number of daily meals
Sex, maturity
Smoking
Dieting for weight-loss reasons
Socioeconomic status
Total breastfeeding duration
Birth wt
Wt, height, BMI, and
Schooling of Parents
Body wt, height, waist
circumference and BP
Parental clinical history,
demographics and lifestyle
Demographic variables
Birth wt

Breakfast skipping

Parental education
Parental obesity
Watching television or
playing video games
Breastfeeding
Physical activity
Smoking during pregnancy
Eating snacks while watching television
Physical activity

Year participating in the Study

Total eating episodes per day

Parental questionnaire (possible


answers:1/2/3/4/5/.5
and dont know)

Dietary intake, Age

Parameters Evaluated

One 24-h dietary recall

Evaluation of Eating
Frequency

TABLE 1 Basic Characteristics of the Studies Included in the Meta-analysis (Chronological Order)

Sex- and age-specic cutoff points


proposed by IOTF

Not specied

Sex- and age-specic cut off points


proposed by IOTF
Note: Children who were overweight
and those who were obese were
combined to reect the obese group
Age- and sex specic CDC
reference standards:
Overweight:
BMI $85th percentile

Age- and sex specic


CDC reference standards: At risk for
overweight: BMI $85th and ,95th
percentile; overweight: BMI $95th
percentile
Note: Children at risk for overweight
and overweight were combined
to reect the overweight group
Sex- and age-specic cutoff points
proposed by IOTF

Overweight/Obesity
Denition

REVIEW ARTICLE

961

962

KAISARI et al

Downloaded from by guest on December 26, 2016

Greece,
2010

Greece,
2011

Kontogianni
et al22

Cassimos
et al31

344
334

544

987

Girls

335

633

346
265

581

1021

Boys

Sample, n

1112

312
1318

1012

610

1217

Age

Cross-sectional

Cross-sectional

Cross-sectional

Cross-sectional

Cross-sectional

Type of Study

Questionnaire
completed by parents

One 24-h dietary recall


Total eating episodes per day
(eating episode:
any eating/drinking occasion
separated by at least 15 min
from other events)

Interviewer-administered
questionnaire (number of
eating occasions)

Parental questionnaire (mean


number of meals consumed)

Semi-quantitative Food Frequency


Questionnaire

Evaluation of Eating
Frequency
Body wt, height
Dietary habits
Physical activities
Sedentary activities
(watching television, working
on a computer, playing video games)
Body wt, height, upper arm,
waist circumference
Triceps skinfold
Birth wt
Duration of breast-feeding
Dietary habits
Parents wt, height
Sociodemographic variables
Birth order
Paternal education
Dietary patterns
Energy intake
Physical activity
Sedentary activities (watching
television, working on a
computer, playing video games)
Sociodemographic variables
Parental education
Dietary intake
Eating behaviors
Adherence to the Mediterranean diet
guidelines
Energy intake/ BMR: for the
assessment of low energy reporting
Physical activity
Sedentary activities (watching
television, working on a computer,
playing video games)
Body wt, height
Blood pressure
Parents wt, height
Socioeconomic status
Parents educational status
Dietary habits
Time sleeping and time watching
television of the children

Parameters Evaluated

AA, African American; BMR, basal metabolic rate; BP, blood pressure; CDC, Centers for Disease Control and Prevention; EA, European American; IOTF, International Obesity Task Force; wt, weight.

Greece,
2008

Portugal,
2008

Ferreira &
MarquezVidal21

Lagiou &
Parava28

Greece,
2007

Country,
Year

Kosti et al
(The Vyronas
study)

23

Reference/ Name
of the Study

TABLE 1 Continued

Sex- and age-specic cut


off points proposed by IOTF

Sex- and age-specic cutoff


points proposed by IOTF

Greek growth charts developed


by the First Department of Pediatrics
of Athens University Medical School

Sex- and age-specic cutoff


points proposed by IOTF

Sex- and age-specic cut off


points proposed by IOTF

Overweight/Obesity
Denition

REVIEW ARTICLE

TABLE 2 ORs and 95% CIs Between Number of Meals and Overweight/Obesity Status, as Reported in the Studies Included in the Meta-analysis
ID

Reference
14

1
2
3
4
5

Nicklas et al 2003 ( boys)


Nicklas et al 2003 (EA girls)14
Nicklas et al 2003 ( boys)14
Nicklas et al 2003 ( girls)14
Toschke et al 200516

Toschke et al 200516

7
Mota et al 2008 (girls)15
8

Mota et al 2008 (boys)15

Neutzling et al 200327

OR

95% CI

Comments

0.97
1.23
0.70
0.56
#3 meals: 1
4 meals: 0.73
$5 meals: 0.56
#3 meals: 1
4 meals: 0.73
$5 meals: 0.51
#3 meals: 1
4 meals: 1.97
$5 meals: 1.33
#3 meals: 1
4 meals: 2.75
$5 meals: 1.44
.3 meals: 0.54

0.631.50
0.801.89
0.381.33
0.330.95
1
0.560.96
0.420.75
1
0.441.21
0.290.89
1
1.003.96
0.752.36
1
1.295.83
0.862.97
0.291.00

497
513
273
279
4370

n+1meals compared with n meals (n.0) for overweight status


Notes: Meals refer to regular meals, snacks not included; OR
adjusted for: total calorie intake, age, and study year
In comparison with 3 or fewer meals for overweight

In comparison with 3 or fewer meals for obesity


Note: OR adjusted for: parental education, childs sex parental
obesity, breastfeeding
#3 meals: Reference category for obesity
461

425
508

10

Barba et al 200619

#3 meals: 1
4 meals: 0.78
$5 meals: 0.48

1
0.581.04
0.360.64

3668

11

Turkkahraman et al 200629

Kosti et al 2007 (boys)23


Kosti et al 2007 (girls)23
Ferreira & Marquez-Vidal 2008 (boys)21

15

Ferreira & Marquez-Vidal 2008 (girls)21

16

Lagiou & Parava 200828

1.563.98
0.340.84
0.212.62
0.360.86
0.461.30
1
0.461.64
0.572.13
1
0.632.83
0.894.09
0.480.76

2465

12
13
14

2 meals: 2.49
3 meals: 0.54
$ 4 meals: 0.59
$ 3 meals: 0.55
$ 3 meals: 0.77
#3 meals: 1
4 meals: 0.87
$5 meals: 1.01
#3 meals: 1
4 meals: 1.34
$5 meals:1.90
0.61

17

Kontogianni et al 2010
(Children: 312, boys)22

18

Kontogianni et al 2010
(Children: 312, girls)22

19

Kontogianni et al 2010
(Adolescents:1318, boys)22

20

Kontogianni et al 2010
(Adolescents:1318, girls)22

21

Cassimos et al 201131

#3 meals: 1
4 meals: 1.30
$5 meals: 0.66
#3 meals: 1
4 meals: 1.02
$5 meals: 1.20
#3 meals: 1
4 meals: 0.90
$5 meals: 0.82
#3 meals: 1
4 meals: 0.47
$5 meals: 0.51
#3 meals: 1.74

1
0.463.68
0.251.75
1
0.362.90
0.453.20
1
0.421.95
0.371.79
1
0.191.15
0.221.21
1.032.94

1021
987
581

Note: OR adjusted for: mothers and fathers BMI, birth wt,


breastfeeding and self-perception of overweight before age 10
#3 meals: reference category for overweight and obesity
Notes: Meals include snacks; data collected through E-mail; OR
adjusted for: age, sex birth wt, physical activity, parental
overweight, and education level
OR for obesity
Note: Meals refer to regular meals, snacks not included
In comparison with less than 3 meals for overweight and obesity
Note: Meals include snacks
#3 meals: Reference category for overweight and obesity

544

633
346

n+1meals compared with n meals (n.0) for overweight status


Note: OR adjusted for: total energy intake
#3 meals: Reference category for overweight and obesity
Note: Meals refer to the total number of eating episodes

344

265

334

335

In comparison with more than 3 meals for overweight and obesity


Notes: Multivariate model

AA, African American; EA, European American; ID, identication number; wt, weight; n, number.

Eating Frequency and Overweight/


Obesity Status in Children and
Adolescents
Figure 2 summarizes the effect of eating frequency on overweight/obesity
status in children and adolescents.
Overall, increased eating frequency, as
compared with the lowest category,

was associated with 0.78-times lower


odds (ie, log OR = 0.24, 95% CI 0.41 to
0.06), which means 22% lower likelihood of being overweight/obese. A
signicant heterogeneity of the effect
sizes of the number of meals on
overweight/obesity was observed (Q =
41.63, P = .003, I2 = 51.9%). Moreover, an

PEDIATRICS Volume 131, Number 5, May 2013

asymmetric plot of the ORs versus SE


{lnOR} was then constructed (not presented here), indicating potential presence of publication bias, as smaller
studies showing no benecial effects
on overweight/obesity were missing.
However, the Egger method of asymmetry clearly suggested no publication
963

Downloaded from by guest on December 26, 2016

analysis by age group, as there were


only 2 studies in adolescents (Table 1).
Finally, the impact of the selected
studies on the combined effect was
tested, by removing each study and
recalculating the combined effect and
the corresponding log-transformed
95% CI of the remaining studies. The
combined effect sizes under each scenario of the studys removal remained
unaltered (the 95% CIs varied within
0.45 to 0.01).

DISCUSSION
Summary and Interpretation of
Findings

FIGURE 2
Eating frequency and overweight/obesity status in children and adolescents. Forest plot of studies that
evaluated the effect of eating frequency on the overweight/obesity status in children and adolescents
(squares and diamonds represent effect size; extended lines show 95% CIs). Increased eating frequency,
as compared with the reference category, was associated with a benecial effect regarding the body
weight status. Numbers in brackets are the study identication numbers, presented in Table 2.

bias (b0=1.29, 95% CI 2.1 to 4.6,


Pone-tailed = .21).
Sensitivity Analysis
When stratied by sex the observed
benecial effect remained signicant in
boys (n = 3408, OR = 0.76, log OR =
0.27, 95% CI 0.47 to 0.06), but not in
girls (n = 3462, OR = 0.96, log OR =
0.04, 95% CI 0.40 to 0.32) (Fig 3) (P for
sex differences = 0.14). The a-posterior
statistical power was 95% for the
subanalysis in boys and 19% for the
subanalysis in girls. No signicant
heterogeneity of the effect measures
regarding overweight/obesity in boys
was observed (Q = 7.78, P = .35, I2 =
10.0%). However, heterogeneity of the
effect sizes of the number of meals on
overweight/obesity was observed in
girls (Q = 17.46, P = .008, I2 = 65.6%).

964

Then, a subgroup analysis was performed according to the tool used to


measure eating frequency. Based on the
studies that evaluated eating frequency by using precoded, self-reported
questionnaires, the number of meals/
eating episodes was still inversely associated with overweight/obesity (OR =
0.77, log OR = 0.26, 95% CI 0.49
to 0.04); whereas, in the 2 studies that
used 24-hour recalls, the number of
meals/eating episodes was not associated with the outcome (OR = 0.84, log
OR = 0.17, 95% CI 0.45 to 0.11). Heterogeneity of the effect sizes was observed in the subanalysis that
evaluated data only from the questionnaires (Q = 30.71, P = .006, I2 =
54.4%), but not in the subanalysis that
evaluated information from recalls
(Q = 9.71, P = .08, I2 = 48.5%). No attempt
was made to perform a subgroup

To the best of our knowledge, this is


the rst work that has systematically
assessed the association between eating frequency and overweight/obesity in
children and adolescents. The presented
meta-analysis of 21 substudies, with
an overall incorporated population of
18 849 subjects, revealed an inverse
association between eating frequency
(ie, the total number of meals/eating
episodes consumed on a daily basis)
and overweight/obesity status in children and adolescents. Specically, children and adolescents who had a higher
number of eating episodes per day had
22% lower probabilities of being overweight or obese compared with those
who had fewer episodes. Interestingly,
the inverse association was evident only
in boys and not in girls, suggesting that
there are sex related differences in dietary patterns and behaviors and their
effect on overweight/obesity. However,
at this point it should be noted that the
presence of publication bias, and the
signicant heterogeneity observed in
the results of the selected studies, indicates the need for further investigations
in this important scientic eld.
Several pathways have been proposed
to explain the association between
higher eating frequency and lower body
weight in children. The obvious effect
of thermogenesis that follows food

KAISARI et al

Downloaded from by guest on December 26, 2016

REVIEW ARTICLE

studies included in the meta-analysis


(ie, only observational studies), no
causal relationships can be established.
Moreover, the characteristics of the selected studies did not allow the evaluation of the exact effect of age in the
observed inverse relationship between
eating frequency and overweight/
obesity status. Signicant heterogeneity was present in the pooled analysis,
which introduces a warning about the
generalization of the results.

FIGURE 3
Eating frequency and overweight/obesity in children and adolescents, by sex. Forest plot of studies that
evaluated the effect of eating frequency on the overweight/obesity status in children and adolescents
(separate data for boys and girls; squares and diamonds represent effect size; extended lines show 95%
CIs). Increased frequency of meal consumption, as compared with the reference category, was associated with a signicant benecial effect regarding body weight status in boys, but not in girls.
Numbers in brackets are the study identication numbers, presented in Table 2.

consumption on total energy expenditure


was not supported by laboratory experiments that evaluated differences in
thermogenesis rates between nibbling
and gorging regimens.32 In contrast, the
benecial effect of increased eating frequency on insulin metabolism may be
a potential biological mechanism, but
there are no data in young people. Relevant evidence in adults indicates that
increased eating frequency attenuates a
series of postprandial metabolic and endocrine responses to dietary intake.33,34
The mediating or confounding role of
physical activity or the active lifestyle
has also been proposed in adults,35,36
and it may also apply to children.
Regarding the sex-related differences in
the association between eating frequency and overweight/obesity in children and adolescents, the available
scientic data are still limited and conicting. The presented results are in

accordance with a previous crosssectional study in Greek adolescents,


showingthatthedailyfrequencyofeating
episodes was associated with obesity
indices in boys, but not in girls.23 At this
point it should be underlined that the
subanalysis in girls was underpowered,
and a substantial heterogeneity of the
effect sizes of the selected studies was
evident, leading to the conclusion of serious inconsistencies between studies.
Moreover, other studies that attempted
to identify possible differences between
boys and girls did not report signicant
results.15,21,22 Additional research is
needed to clarify this issue and, most
importantly, to propose potential mechanisms that may be responsible for
a sex-dependent effect.
Limitations
Certain limitations of this work warrant
consideration. Because of the type of

PEDIATRICS Volume 131, Number 5, May 2013

Methodological differences between


studies may also raise concerns;
the most important difference is
the denition used for overweight/
obesity in children and adolescents
and the way that eating frequency was
assessed. In most studies, participants were categorized as overweight and obese according to
International Obesity Task Force criteria for BMI.15,16,2123,2931 However,
other criteria were also used. There
are differences between reference
standards and the corresponding
BMI cutoff points, resulting in differences in overweight/obesity classication of the children.37,38 There is
also variability in the method of assessment of eating frequency (selfreported questionnaires, by children 1019 years old or by their
caregivers, or 24-hour dietary recalls) and in the respective denition
given to the eating frequency. This
variability is a common limitation in
reviews and meta-analyses evaluating eating habits and behaviors.39,40
Specically, in relation to eating
frequency, there is no general agreement on the most appropriate definition to categorize the eating
occasions into meals or snacks, as
well as on the best assessment
method.41 Several suggestions have
been made to dene eating events
according to the time of day, the
amount and/or energy of food, the
presence/absence of companions,
965

Downloaded from by guest on December 26, 2016

the quality of food, or combinations of


these factors. Moreover, in some
studies, participants decide by
themselves whether they eat a meal or
a snack, and this was the case in
a number of investigations included
in the present meta-analysis. Interestingly, the only subgroup of
studies in which we found a negative
association between eating frequency
and overweight/obesity was the 1
study that used self-reported questionnaires; however, there were only 2
studies that used recalls for the evaluation of this information. Confounding factors included in the analysis
also varied between studies (Table 2).
Finally, some researchers presented
OR values for overweight,14,16,28,42
whereas others for obesity,15,16,29,30
and others for both overweight and
obesity.19,2123,27,31

The results of the present meta-analysis


suggest an inverse association between eating frequency and overweight/
obesity status in children and adolescents. Sex-related differences emerged
when this association was assessed
separately in boys and girls, as the
effect of eating frequency was evident
only in boys. Despite the aforementioned considerations, these results
are of considerable importance, as
changes in the number of eating episodes consumed may be an interventional goal for prevention or early
treatment of overweight and obesity.
Schools may play an important role
in improving childrens eating preferences and patterns, and in this

perspective, integrating smaller, more


frequent meals in the school environment may be an effective way to battle
childhood obesity. Clinical trials are
warranted to conrm this inverse association, evaluate the clinical practice
applicability, and support a public
health recommendation. Future studies should also evaluate the size and
types of meals and snacks in terms of
volume, energy content, macronutrient
composition, or glycemic index and
take this evaluation into account when
investigating associations between
eating/meal patterns and obesity. Finally, in relation to the sex-related
differences observed, additional research is needed to explore these differences and, most importantly, the
underlying biological and/or social
mechanisms.

in children and adolescents? Nutr Metab


Cardiovasc Dis. 2006;16(4):294301
Berg C, Lappas G, Wolk A, et al. Eating
patterns and portion size associated with
obesity in a Swedish population. Appetite.
2009;52(1):2126
Edelstein SL, Barrett-Connor EL, Wingard
DL, Cohn BA. Increased meal frequency
associated with decreased cholesterol
concentrations; Rancho Bernardo, CA,
1984-1987. Am J Clin Nutr. 1992;55(3):664
669
Fabry P, Hejl Z, Fodor J, Braun T, Zvolankova
K. The frequency of meals. Its relation to
overweight, hypercholesterolaemia, and
decreased glucose-tolerance. Lancet. 1964;
2(7360):614615
Kant AK, Schatzkin A, Graubard BI, BallardBarbash R. Frequency of eating occasions
and weight change in the NHANES I Epidemiologic Follow-up Study. Int J Obes Relat
Metab Disord. 1995;19(7):468474
Metzner HL, Lamphiear DE, Wheeler NC,
Larkin FA. The relationship between frequency of eating and adiposity in adult men
and women in the Tecumseh Community
Health Study. Am J Clin Nutr. 1977;30(5):
712715

13. Fbry P, Hejda S, Cern K, Osancov K,


Pechar J. Effect of meal frequency in
schoolchildren. Changes in weight-height
proportion and skinfold thickness. Am J
Clin Nutr. 1966;18(5):358361
14. Nicklas TA, Yang SJ, Baranowski T, Zakeri I,
Berenson G. Eating patterns and obesity in
children. The Bogalusa Heart Study. Am J
Prev Med. 2003;25(1):916
15. Mota J, Fidalgo F, Silva R, et al. Relationships between physical activity, obesity and
meal frequency in adolescents. Ann Hum
Biol. 2008;35(1):110
16. Toschke AM, Kchenhoff H, Koletzko B, von
Kries R. Meal frequency and childhood
obesity. Obes Res. 2005;13(11):19321938
17. Kirk TR. Role of dietary carbohydrate and
frequent eating in body-weight control.
Proc Nutr Soc. 2000;59(3):349358
18. Moher D, Liberati A, Tetzlaff J, Altman DG;
PRISMA Group. Preferred reporting items
for systematic reviews and meta-analyses:
the PRISMA statement. BMJ. 2009;339:
b2535
19. Barba G, Troiano E, Russo P, Siani A; ARCA
Project Study group. Total fat, fat distribution and blood pressure according to eating frequency in children living in southern

CONCLUSIONS AND IMPLICATIONS


FOR POLICY, PRACTICE, AND
RESEARCH

REFERENCES
1. Orsi CM, Hale DE, Lynch JL. Pediatric obesity
epidemiology. Curr Opin Endocrinol Diabetes
Obes. 2011;18(1):1422
2. Skelton JA, Irby MB, Grzywacz JG, Miller G.
Etiologies of obesity in children: nature and
nurture. Pediatr Clin North Am. 2011;58(6):
13331354, ix
3. Guillaume M, Lapidus L, Lambert A. Obesity
and nutrition in children. The Belgian Luxembourg Child Study IV. Eur J Clin Nutr.
1998;52(5):323328
4. Gazzaniga JM, Burns TL. Relationship between diet composition and body fatness,
with adjustment for resting energy expenditure and physical activity, in preadolescent
children. Am J Clin Nutr. 1993;58(1):2128
5. Receveur O, Morou K, Gray-Donald K, Macaulay AC. Consumption of key food items is
associated with excess weight among elementary-school-aged children in a Canadian
rst nations community. J Am Diet Assoc.
2008;108(2):362366
6. Yannakoulia M, Karayiannis D, Terzidou M,
Kokkevi A, Sidossis LS. Nutrition-related
habits of Greek adolescents. Eur J Clin
Nutr. 2004;58(4):580586
7. Rodrguez G, Moreno LA. Is dietary intake
able to explain differences in body fatness

966

8.

9.

10.

11.

12.

KAISARI et al

Downloaded from by guest on December 26, 2016

REVIEW ARTICLE

20.

21.

22.

23.

24.

25.

26.

Italy: the ARCA project. Int J Obes (Lond).


2006;30(7):11661169
Barbiero SM, Pellanda LC, Cesa CC,
Campagnolo P, Beltrami F, Abrantes CC.
Overweight, obesity and other risk factors
for IHD in Brazilian schoolchildren. Public
Health Nutr. 2009;12(5):710715
Ferreira RJ, Marques-Vidal PM. Prevalence
and determinants of obesity in children in
public schools of Sintra, Portugal. Obesity
(Silver Spring). 2008;16(2):497500
Kontogianni MD, Farmaki AE, Vidra N,
Sofrona S, Magkanari F, Yannakoulia M.
Associations between lifestyle patterns and
body mass index in a sample of Greek
children and adolescents. J Am Diet Assoc.
2010;110(2):215221
Kosti RI, Panagiotakos DB, Mihas CC,
et al. Dietary habits, physical activity and
prevalence of overweight/obesity among
adolescents in Greece: the Vyronas
study. Med Sci Monit. 2007;13(10):CR437
CR444
Cole TJ, Bellizzi MC, Flegal KM, Dietz WH.
Establishing a standard denition for child
overweight and obesity worldwide: international survey. BMJ. 2000;320(7244):
12401243
Kuczmarski RJ, Ogden CL, Grummer-Strawn
LM, et al. CDC growth charts: United States.
Adv Data. 2000;(314):127
Hiotis D, Tsiftis G, Hatzisimeon M, et al.
Height and body weight of Greek children
aged 018 years of age (20012002). A
comparison with data from a study undertaken in 19781979 [in Greek]. Ann Clin
Peadiatr. 2003;50(2):136155

27. Neutzling MB, Taddei JA, Gigante DP. Risk


factors of obesity among Brazilian adolescents: a case-control study. Public Health
Nutr. 2003;6(8):743749
28. Lagiou A, Parava M. Correlates of childhood
obesity in Athens, Greece. Public Health
Nutr. 2008;11(9):940945
29. Turkkahraman D, Bircan I, Tosun O, Saka O.
Prevalence and risk factors of obesity in
school children in Antalya, Turkey. Saudi
Med J. 2006;27(7):10281033
30. Toschke AM, Thorsteinsdottir KH, von Kries
R; GME Study Group. Meal frequency,
breakfast consumption and childhood obesity. Int J Pediatr Obes. 2009;4(4):242248
31. Cassimos D, Sidiropoulos H, Batzios S,
Balodima V, Christoforidis A. Sociodemographic and dietary risk factors for excess
weight in a Greek pediatric population living in Kavala, Northern Greece. Nutr Clin
Pract. 2011;26(2):186191
32. Bellisle F, McDevitt R, Prentice AM. Meal
frequency and energy balance. Br J Nutr.
1997;77(suppl 1):S57S70
33. Jenkins DJ, Khan A, Jenkins AL, et al. Effect
of nibbling versus gorging on cardiovascular risk factors: serum uric acid and blood
lipids. Metabolism. 1995;44(4):549555
34. Jenkins DJ, Wolever TM, Vuksan V, et al.
Nibbling versus gorging: metabolic advantages of increased meal frequency. N Engl J
Med. 1989;321(14):929934
35. Yannakoulia M, Melistas L, Solomou E,
Yiannakouris N. Association of eating frequency with body fatness in pre- and
postmenopausal women. Obesity (Silver
Spring). 2007;15(1):100106

PEDIATRICS Volume 131, Number 5, May 2013

36. Duval K, Strychar I, Cyr MJ, Prudhomme D,


Rabasa-Lhoret R, Doucet E. Physical activity is a confounding factor of the relation
between eating frequency and body composition. Am J Clin Nutr. 2008;88(5):1200
1205
37. Vidal E, Carlin E, Driul D, Tomat M, Tenore A.
A comparison study of the prevalence of
overweight and obese Italian preschool
children using different reference standards. Eur J Pediatr. 2006;165(10):696700
38. Savva SC, Kourides Y, Tornaritis M, Epiphaniou-Savva M, Chadjigeorgiou C, Kafatos A.
Obesity in children and adolescents in
Cyprus. Prevalence and predisposing factors. Int J Obes Relat Metab Disord. 2002;26
(8):10361045
39. Hammons AJ, Fiese BH. Is frequency of
shared family meals related to the nutritional health of children and adolescents?
Pediatrics. 2011;127(6). Available at: www.
pediatrics.org/cgi/content/full/127/6/e1565
40. Wang Y, Beydoun MA, Li J, Liu Y, Moreno LA.
Do children and their parents eat a similar
diet? Resemblance in child and parental
dietary intake: systematic review and metaanalysis. J Epidemiol Community Health.
2011;65(2):177189
41. Oltersdorf U, Schlettwein-gsell D, Winkler G.
Assessing eating patternsan emerging
research topic in nutritional sciences: introduction to the symposium. Appetite.
1999;32(1):17
42. Nicklas TA, Morales M, Linares A, et al.
Childrens meal patterns have changed
over a 21-year period: the Bogalusa Heart
Study. J Am Diet Assoc. 2004;104(5):753761

967

Downloaded from by guest on December 26, 2016

Eating Frequency and Overweight and Obesity in Children and Adolescents: A


Meta-analysis
Panagiota Kaisari, Mary Yannakoulia and Demosthenes B. Panagiotakos
Pediatrics 2013;131;958; originally published online April 8, 2013;
DOI: 10.1542/peds.2012-3241
Updated Information &
Services

including high resolution figures, can be found at:


/content/131/5/958.full.html

References

This article cites 41 articles, 9 of which can be accessed free


at:
/content/131/5/958.full.html#ref-list-1

Citations

This article has been cited by 5 HighWire-hosted articles:


/content/131/5/958.full.html#related-urls

Subspecialty Collections

This article, along with others on similar topics, appears in


the following collection(s):
Adolescent Health/Medicine
/cgi/collection/adolescent_health:medicine_sub
Obesity
/cgi/collection/obesity_new_sub

Permissions & Licensing

Information about reproducing this article in parts (figures,


tables) or in its entirety can be found online at:
/site/misc/Permissions.xhtml

Reprints

Information about ordering reprints can be found online:


/site/misc/reprints.xhtml

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright 2013 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from by guest on December 26, 2016

Eating Frequency and Overweight and Obesity in Children and Adolescents: A


Meta-analysis
Panagiota Kaisari, Mary Yannakoulia and Demosthenes B. Panagiotakos
Pediatrics 2013;131;958; originally published online April 8, 2013;
DOI: 10.1542/peds.2012-3241

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
/content/131/5/958.full.html

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


publication, it has been published continuously since 1948. PEDIATRICS is owned,
published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point
Boulevard, Elk Grove Village, Illinois, 60007. Copyright 2013 by the American Academy
of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from by guest on December 26, 2016