Effects of aerobic, resistance, and combined exercise training on insulin
resistance markers in overweight or obese children and adolescents: A
systematic review and meta-analysis

Elisa Corrêa Marson, Rodrigo Sudatti Delevatti, Alexandre Konig Gar-

cia Prado, Nathalie Netto, Luiz Fernando Martins Kruel

PII: S0091-7435(16)30342-5
DOI: doi: 10.1016/j.ypmed.2016.10.020
Reference: YPMED 4821

To appear in: Preventive Medicine

Received date: 2 March 2016

Revised date: 18 October 2016
Accepted date: 19 October 2016

Please cite this article as: Marson Elisa Corrêa, Delevatti Rodrigo Sudatti, Prado Alexan-
dre Konig Garcia, Netto Nathalie, Kruel Luiz Fernando Martins, Effects of aerobic,
resistance, and combined exercise training on insulin resistance markers in overweight
or obese children and adolescents: A systematic review and meta-analysis, Preventive
Medicine (2016), doi: 10.1016/j.ypmed.2016.10.020

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT
1

Review Article

Effects of aerobic, resistance, and combined exercise training on insulin

T
resistance markers in overweight or obese children and adolescents: A

IP
systematic review and meta-analysis

R
SC
Elisa Corrêa Marson1
Rodrigo Sudatti Delevatti1,2

NU
Alexandre Konig Garcia Prado1,3
Nathalie Netto1
MA
Luiz Fernando Martins Kruel1

1
Universidade Federal do Rio Grande do Sul. Exercise Research Laboratory.
D

Department of Physical Education. Porto Alegre. RS, Brazil.

TE

2
Faculdade Sogipa de Educação Física. Porto Alegre. RS, Brazil.
3
Faculdade Cenecista de Osório. Osório. RS, Brazil
P
CE

Authors' Address:

Felizardo street, 750. Jardim Botânico. Physical Education School/Swimming

AC

center, room 18 – 90690-200 – Porto Alegre, RS – Brazil.

Phone: 51 3308-5820.

E-mail addresses:

Rodrigo Sudatti Delevatti: rsdrodrigo@hotmail.com;

Elisa Corrêa Marson: elisa_marson@yahoo.com.br;

Alexandre Konig Garcia Prado: konigg@ig.com.br

Nathalie Netto: nettonathalie@gmail.com

Luiz Fernando Martins Kruel: kruel@esef.ufrgs.br.

Corresponding author: Rodrigo SudattiDelevatti.

 ACCEPTED MANUSCRIPT
2

Word Counts: 3624.

Abstract

T
IP
Objective: To assess the associations of aerobic, resistance, and combined
exercise with changes in insulin resistance, fasting glucose, and fasting insulin

R
in children and adolescents who are overweight or obese.

SC
Data Searches: MEDLINE via Pubmed, Cochrane-CENTRAL, SPORTDiscus,
and LILACS.

NU
Study Selection: Randomized clinical trials of at least six weeks of duration
that evaluated the ability of exercise training to lower at least one of the
MA
following outcomes: insulin resistance-HOMA, fasting glucose, and fasting
insulin in children and/or adolescents classified as obese or overweight.
Data Extraction and analysis: Two independent reviewers extracted data and
D

assessed the quality of the included studies. Differences (exercise training

TE

group minus control group) in the outcomes evaluated were analyzed using a
random effects model.
P

Results: Of 1853 articles retrieved, 17 studies were included. The meta-

CE

analysis showed that physical training in general was not associated with a
reduction in fasting glucose levels compared to the control, but it was
associated with reductions in fasting insulin levels (-3.37Uu/ml; CI 95%,-
AC

5.16Uu/ml to -1.57Uu/ml; I², 54%, p = 0.003) and HOMA (-0.61; CI 95%, -1.19
to -0.02; I², 49%, p=0.040). In addition, each modality (aerobic, resistance, and
combined) was compared to the control group. Aerobic exercise was associated
with declines in fasting insulin levels (-4.52Uu/ml; CI 95%, -7.40 to -1.65; I²,
65%, p=0.002) and in HOMA (-1.33; 95% confidence interval, -2.47 to -0.18; I²,
73%, p=0.005).
Conclusions: Exercise training, especially aerobic training, is associated with
the reduction of fasting insulin levels and HOMA in children and adolescents
with obesity and overweight, and may prevent metabolic syndrome and type 2
diabetes.

Key words: Childhood obesity, insulin resistance, exercise.

 ACCEPTED MANUSCRIPT
3

T
IP
INTRODUCTION
Obesity in children and adolescents has been one of the major public

R
health problems of the last century. Estimates show that about 170 million

SC
children worldwide are overweight and this prevalence is higher in low- and
middle-income countries (WHO, 2012). According to the World Health

NU
Organization (WHO), overweight and obesity are defined by an age-related BMI
greater than one and two standard deviations above the median of the WHO’s
MA
growth reference, respectively (WHO, 2016). Many cardiovascular risk factors
are associated with being overweight, such as higher blood pressure,
development of dyslipidemia, and insulin resistance. However, in obese
D

children, there is a greater concern about insulin resistance, which is highly

TE

associated with obesity, metabolic syndrome, and type 2 diabetes in adulthood

(Liang et al., 2015). A recent cohort (Zhang et al., 2015) containing 1.593
P

individuals with follow-up of 19 years, found higher prevalence of metabolic

CE

syndrome in a group of obese people with insulin resistance compared to obese

insulin sensitive individuals (34.9% vs. 24.3%, p = 0.008), suggesting that
insulin resistance increases the association between childhood obesity and
AC

metabolic disorders in adulthood.

Based on this association, the understanding of the pathophysiological
mechanisms that lead children with overweight or obesity to important metabolic
disorders is very important. Physiological alterations associated with obesity,
such as the increased availability of pro-inflammatory cytokines and free fatty
acids, reduce the tissue sensitivity to insulin, increasing the need for insulin
secretion for maintenance of glycemic homeostasis (Ten and McLaren, 2004).
At first, this compensatory mechanism is efficient, but increased levels of fasting
insulin may already indicate insulin resistance, even with normal glycemic
levels. Without effective therapeutic intervention, pancreatic secretion of insulin
too will fail, increasing the fasting glucose, leading to pre-diabetes and type 2
diabetes (ADA, 2016).
 ACCEPTED MANUSCRIPT
4

One of the main factors related to obesity and being overweight, along
with insulin resistance, is a sedentary lifestyle in childhood and adolescence

T
(Rennie et al., 2005). A study by Rey-López et al., (2008) highlighted the time

IP
that children spent watching television as a determining factor in weight gain, for
which physical exercise was the indicated therapeutic approach. Systematic

R
reviews with meta-analyses showed the beneficial effects of exercise on lipid

SC
profile (Kelley and Kelley, 2007), blood pressure (Garcia-Hermoso et al., 2013),
glucose, and insulin levels (Garcia-Hermoso et al., 2014) for children and

NU
adolescents who are overweight or obese. However, evidence of the effect of
exercise on glucose metabolism (Garcia-Hermoso et al., 2014) focused only on
MA
aerobic exercise, although resistance training is also able to show different
benefits for children and adolescents (Faigenbaum and Myer, 2010) and higher
levels of muscular fitness are inversely associated with cardiometabolic
D

outcomes, such as insulin resistance and systemic inflammation in children and

TE

adolescents (Artero et al., 2012). Besides that, only isolated insulinemic and
glycemic responses have been investigated, without consistent evidences of the
P

effects of different modalities of exercise training (aerobic, resistance, and a

CE

combination of these) on insulin resistance markers by homeostatic model

assessment (HOMA), in addition to isolated glycemic and insulinemic
responses. In this context, the aim of the present meta-analysis was to assess
AC

the association of exercise interventions (aerobic, resistance, and combined

exercise) with changes in insulin resistance, fasting glucose, and fasting insulin
in children and adolescents who are overweight or obese, by means of a
systematic review with meta-analysis of randomized clinical trials.

METHODS

This systematic review and meta-analysis is reported in accordance with

the Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) statement (Liberati et al., 2009).

Search strategy and studies selection

 ACCEPTED MANUSCRIPT
5

The following databases were used without date limits: Medline by

Pubmed, Cochrane CENTRAL, SportDiscus, and LILACS. In addition, a manual

T
search of the references from the studies found in the chosen databases was

IP
carried out. We only included articles that were already published in journals.
Theses and dissertations were excluded.

R
There was no language restriction in the search criteria. Boolean

SC
operators ―OR‖ and ―AND‖ were used. In PubMed, the search was performed
using the MeSH terms and their synonyms. Related terms to MeSH were used

NU
in the other databases. The complete search strategy used for the PubMed
database is shown in the Electronic Supplementary Material (Table 1).
MA
Two independent reviewers (E.C.M and N.N) read the titles and abstracts
of all papers found, and independently read the full article if the paper met the
eligibility criteria. Disagreements were solved by consensus and discussion with
D

a third reviewer (R.S.D).

TE

Eligibility criteria
P

Criteria for inclusion into the present study were: children and
CE

adolescents (until 19 years old) of both sexes, overweight or obese, and not
engaged in structured exercise for at least three months. Clinical trials should
have included at least six-week of intervention period. There were no
AC

restrictions on the exercise modality, intensity, volume, and frequency. For

inclusion, studies showed absolute values pre- and post-intervention period or
the difference between average and dispersion values of at least one of the
following outcomes: fasting glucose, fasting insulin, and insulin resistance-
HOMA.

Data extraction
Data extraction was performed by two independent reviewers (E.C.M and
N.N). Disagreements between the two reviewers about the conditions of the
studies were solved by a consensus meeting with the third reviewer (R.S.D).
For data extraction, a standardized form was used, composed of the following
items: author, population, intervention data, and outcomes. For each item, the
 ACCEPTED MANUSCRIPT
6

following data were extracted: 1) Population (mean age, comorbidities, and

dietary co-intervention); 2) Intervention (modality, intervention period, session,

T
and weekly duration, weekly frequency, intensity, number of set and repetitions,

IP
and rest interval between sets); Results (mean values and standard deviation of
the pre- and post-intervention, and the difference between the means).

R
SC
Study quality assessment (risk of bias)
Study quality assessment included adequate sequence generation,

NU
allocation concealment, blinding of outcomes assessors, use of intention-to-
treat analysis, and description of losses and exclusions. Studies without clear
MA
descriptions of an adequate sequence generation or how the allocation list was
concealed were considered not to have fulfilled these criteria. Quality
assessment was independently performed by two unblinded reviewers (E.C.M.
D

and N.N) and disagreements were resolved by consensus or by a third reviewer

TE

(R.S.D).
The risk of bias was evaluated in the following form: high risk - when the
P

methodological criteria, such as adequate sequence generation, were not

CE

reported or were not performed; low risk - when the methodological criteria were
performed appropriately; unclear risk - when there was no adequate description
of the criteria, making it difficult to evaluate it as high or low risk.
AC

Data analysis
For consideration in the statistical analysis, only studies which results
could be compared in three or more studies were included. Those results that
did not meet this standard were only described in the results. Some studies in
which the exercise group was associated with nutritional counseling, we
considered as control group the one that had nutritional counseling, because in
this cases, it provides a better control for physical exercise than that of the
group without any intervention.
The results were presented as mean difference weighted with confidence
intervals (CI) of 95%. Calculations were performed using a method of random
effects, adopting α = 0.05. Statistical heterogeneity of the intervention effects
 ACCEPTED MANUSCRIPT
7

among the studies was assessed by using Cochran’s Q test and the
inconsistency by I2 test. Values above 25% and 50% were considered

T
indicators of being moderate and high heterogeneity, respectively. All analysis

IP
were performed using Review Manager, version 5.3. A primary analysis was
performed with all studies (Exercise training (overview)) and sub-group

R
analyses were carried out considering the training modalities (aerobic,

SC
resistance, and combined) used in the studies and average age of individuals
(≤12 and >12 years old).

RESULTS
NU
MA
Studies description
The initial search resulted in 1853 studies, 78 of which were excluded as
duplicates, and 1721 articles of which the title and abstract were read. Of the 54
D

articles analyzed in full, 18 studies met the eligibility criteria. The research and
TE

studies selection is represented in Figure 1. The included studies contain 961

individuals in total. The characteristics of these studies are summarized in
P

Tables 2 and 3.
CE

―Insert figure 1 here‖

―Insert table 2 and 3 here‖
AC

Assessing risk of bias

Among the included studies, 39% presented adequate sequence
generation (7 of 18), 33% reported allocation concealment (6 of 18), 28% had
blinded assessment of outcomes (5 of 18), 61% described losses to follow-up
and exclusions (11 of 18), and 72% used the intention-to-treat principle for
statistical analyses (13 of 18) (Figure 2 – Supplementary material).

Interventions effects on glucose levels, insulinemic, and on insulin

resisitance (HOMA)

Exercise training (overview)

 ACCEPTED MANUSCRIPT
8

In general, physical training (Figure 3) was not associated with changes

in fasting plasma glucose (Mean difference (MD):0.30mg/dl; CI 95%: -0.69mg/dl

T
to 0.10mg/dl; I² = 81%, p for heterogeneity< 0.001) as compared with the

IP
control. However, physical training (Figure 4) was associated with a reduction in
fasting insulin of-3.37Uu/ml (CI 95%: -5.16Uu/ml to -1.57Uu/ml; I² = 54%, p for

R
heterogeneity = 0.003) and in HOMA of -0.61 (CI 95%: -1.19to -0.02; I² = 49%,

SC
p for heterogeneity = 0.040) as compared with the control. The association
between physical training and changes in insulin resistance (HOMA) is

NU
summarized in Figure 5. MA
Aerobic exercise training
Aerobic exercise training (Figure 3) was not associated with
improvements in fasting plasma glucose (MD:-0.99mg/dl; CI 95%: -2.00mg/dl to
D

0.03mg/dl; I² = 89%, p for heterogeneity< 0.001; 8 studies, 189 individuals) as

TE

compared with the control, but was associated with a reduction in fasting insulin
(Figure 4) levels of-4.52Uu/ml (CI 95%: -7.40Uu/ml to -1.69Uu/ml; I² = 65%, p
P

for heterogeneity = 0.002; 10 studies, 244 individuals) and in HOMA of -1.33 (CI
CE

95%: -2.47to -0.18; I² = 73%, p for heterogeneity = 0.005; 5 studies, 134

individuals) as compared with the control. The association between aerobic
AC

training and changes in HOMA is summarized in figure 5.

Resistance exercise training

Resistance training (Figure 3) did not show an association with
improvements in fasting plasma glucose (MD:0.88mg/dl; CI 95%: -2.51 mg/dl to
0.74 mg/dl; I² = 71%, p for heterogeneity = 0.02; 4 studies, 58 individuals) or in
fasting insulin (Figure 4) (MD:-2.98Uu/ml; CI 95%: -8.77Uu/ml to 2.81Uu/ml; I² =
52%, p for heterogeneity = 0.10; 4 studies, 58 individuals) as compared with the
control. The restricted number of studies with resistance training and HOMA
assessments for this outcome made statistical analysis unfeasible.

Combined exercise training

 ACCEPTED MANUSCRIPT
9

Similar to what is observed with aerobic training and resistance training,

combined training did not show an association with improvements in fasting

T
plasma glucose (Figure 3) (MD:0.33mg/dl; CI 95%: -0.98 mg/dl to 0.31 mg/dl; I²

IP
= 93%, p for heterogeneity < 0.001; 3 studies, 58 individuals) compared to
control. Regarding to fasting insulin, the combined exercise training was not

R
associated with a reduction in fasting insulin (Figure4) (-2.48Uu/ml; CI 95%: -

SC
5.50Uu/ml to 0.54Uu/ml; I² = 52%, p for heterogeneity = 0.10; 4 studies, 75
individuals). As with the assessment of resistance training, the restricted

NU
number of studies with combined training and HOMA assessment made
statistical analysis for this outcome unfeasible.
MA
―Insert figure 3, 4, and 5‖
D

Children versus adolescents

TE

By the fact that biological maturation has an important influence in some

biological outcomes, we analyzed the exercise effects only in children (≤12
P

years old) comparing to the effects in adolescents (>12 years old). In children,
CE

only a maintenance in fasting glucose was found (MD: 0.12 mg/dl; CI 95%: -
1.40 mg/dlto1.64 mg/dl; I² = 71%, p for heterogeneity = 0.03; 3studies,
47individuals). In fasting insulin, a reduction was found (MD: -4.73 Uu/ml; CI
AC

95%: -8.42Uu/ml to-1.04Uu/mll; I² = 0%, p for heterogeneity = 0.71; 3studies, 45

individuals). In a similar manner, the adolescents (>12 years old) experienced
only a maintenance of fasting glucose (MD:-0.25 mg/dl; CI 95%: -0.79 mg/dl to
0.29 mg/dl; I² = 85%, p for heterogeneity = 0.001; 9 studies, 131 individuals)
and a reduction in fasting insulin (MD:-3.40Uu/ml; CI 95%: -4.65Uu/ml to-
2.14Uu/ml l; I² = 54%, p for heterogeneity = 0.008; 10 studies, 169 individuals).

Other studies
Three studies were described only (Foschini et al., 2010; Inoue et al.,
2014; Tsang et al., 2009) and where subsequently considered inadequate for
statistical analysis as they present single characteristics of interventions, not
 ACCEPTED MANUSCRIPT
10

found elsewhere in the study for this review. Two of these (Foschini et al., 2010;
Inoue et al., 2014) showed a reduction in insulinemic levels and insulin

T
resistance. The study by Foschini et al., (2010) found a positive result in a

IP
combined training group periodization of wave form during 14 weeks, which did
not happened with a linear periodization. Inoue et al., (2014) showed reduction

R
in insulinemic levels, glucose, and insulin resistance in two combined training

SC
models of one year duration (wave periodization and linear), which did not occur
with aerobic training alone. Tsang et al., (2009) did not find any changes in the

NU
outcomes after 27 weeks of Kung Fu or Taichi.
MA
DISCUSSION
The characteristics of the exercise training interventions were different
between studies. In terms of an aerobic training modality, the majority of the
D

studies used a treadmill, cycle ergometer and elliptical trainer, with others
TE

studies using sports (Wong et al., 2008; Tsang et al., 2009), dancing games
(Murphy et al., 2009) and jump rope (Lee et al., 2010a). In these studies, the
P

intensity was prescribed by percentages of the maximal heart rate, reserve

CE

heart rate, maximal uptake oxygen, and by first ventilatory threshold. In terms of
resistance training, the majority of the studies used weight machines, with one
study using elastic bands and own body weight (Farpour et al., 2009). In these
AC

studies, the intensity was prescribed by 1RM percentage and maximal

repetitions. In general, weekly frequency was from two to five sessions and
sessions lasted between 30 and 60 minutes.
The present study showed that physical training, in general, is associated
with insulin resistance and fasting insulin reduction in children and adolescents
who are overweight or obese. In addition, a sub-analysis assessing the isolated
effects of the modalities of training (aerobic, resistance, and combined) showed
that only aerobic training is associated with insulin resistance markers in the
studied population. Therefore, aerobic training can be indicated as a therapeutic
approach for children and adolescents with insulin resistance—which is very
common in cases of overweight and obesity—and seems to maintain proper
glucose levels due to excessive insulin production (Chiarelli et al., 2008).
 ACCEPTED MANUSCRIPT
11

This excessive insulin production was found in 17 (94%) of 18 studies

included in the present review. Only in the study by Kelly et al., (2004) are the

T
fasting insulin levels below the normal cut point of 15–20μU/ml (SBC, 2005). In

IP
children and adolescents, these high insulin levels are still able to maintain the
glycemia at suitable levels. This may explain the lack of association between

R
physical training and glycemic reduction. For this reason, a focus on glycemic

SC
outcomes does not seem to be advisable in this population, which shows
increases in insulin levels preceded by increases in glycemic levels. Despite

NU
this important discussion about the sensibility of the outcomes and metabolic
status of the children and adolescents, the primary aim of this study was to
MA
analyze how the investigated outcomes were affected by exercise training
(overview) and by different training modalities.
Our findings related to fasting plasma glucose differ from the results of
D

Garcia-Hermoso et al., (2014), who in a recent meta-analysis found

TE

associations between aerobic training and a reduction in fasting plasma glucose

in children and adolescents with obesity. The difference in the results of these
P

studies might be due to methodological differences: the study by Garcia-

CE

Hermoso et al., (2014) did not consider nutritional intervention or children and
adolescents who are overweight in their analysis, and therefore has less
heterogeneity when compared to the present study.
AC

Regarding insulinemic levels, resistance and combined training were not

associated with a reduction in fasting insulin. This information has extremely
practical implications, because it indicates the importance of aerobic exercise
training in children and adolescents with greater insulin production. To our
knowledge, this is the first study to show the effects of different training
modalities on this outcome. Despite only aerobic training presenting an
association with fasting insulin reduction, caution in interpretation of the findings
is needed, because the number of studies looking at resistance and combined
training with this population is smaller when compared to aerobic training, with
great methodological heterogeneity.
However, there are some specific physiological and methodological
characteristics of aerobic training that may explain its association with a
 ACCEPTED MANUSCRIPT
12

reduction in fasting insulin levels, including duration of the sessions and great
energy expenditure. It is more usual to design an aerobic training regime with a

T
focus on metabolic changes compared to resistance training, usually focused

IP
on musculoskeletal aspects, in children and adolescents. This is probably why
there are relatively few studies using resistance training in the assessment of

R
fasting insulin levels; this is evidenced by the present study, which found only

SC
four studies, with a total of 58 individuals. In addition to this methodological
characteristic, lower insulin production results from lower resistance to this

NU
hormone, as a consequence, among other factors, of the anti-inflammatory
effect of exercise, which is very evident with aerobic training (Hayashino et al.,
MA
2014; Delevatti et al., 2015). These benefits might be obtained even without a
reduction in body fat (Goran et al., 2003), suggesting that these metabolic
improvements are not totally dependent on morphological changes.
D

Insulin resistance reduction, assessed by HOMA levels, is not broadly

TE

studied in children and adolescents. The limited number of studies allowed us to

perform a comparative analysis of physical training (overview) and aerobic
P

training, both showing an association with a reduction in HOMA levels.

CE

However, there are not enough data to perform a comparative analysis between
resistance and combined training. It is interesting to highlight that one of the
selected articles using resistance training (Davis et al., 2009b) did not find
AC

improvements in this outcome, whereas a study with combined training (Davis

et al., 2011) found a reduction in HOMA levels.
Evidence pointing to the effects of physical training on insulin resistance
reduction is extremely important, as this outcome is associated with
cardiometabolic risk in obese children (Ten and McLaren, 2004), a central factor
of metabolic syndrome which, despite having a genetic component, is related to
the excessive consumption of carbohydrates, fats, and lack of exercise (Ten
and McLaren, 2004). Continued actions against childhood obesity, particularly
for insulin resistance, are necessary, otherwise compensatory insulin secretion
will fail, resulting in an increased number of children and adolescents who are
overweight with type 2 diabetes, and a consequent increase in the
cardiovascular mortality risk (Stamler et al., 1993; Ten and McLaren, 2004).
 ACCEPTED MANUSCRIPT
13

Faced by this worrying scene, our findings are promising, because the
prescription of aerobic training seems to be an effective alternative, reducing

T
insulin levels to maintain glycemic homeostasis. Aiming to prevent problems in

IP
adulthood and to reduce health care costs, the practice of exercise training,
especially aerobic training, should be encouraged in schools and in other

R
locations for children and adolescents.

SC
As physiological alterations associated with puberty may influence
glycemic metabolism, a sub-analysis for age was also performed. The findings

NU
were similar in the two analyses, the fasting insulin being reduced by means of
exercise training in children (≤12 years old) and adolescents (> 12 years old),
MA
without alterations in fasting glucose in both analyses. We were not able to
analyze HOMA results due to the small number of studies that analyzed only
children or adolescents.
D

This study presents some limitations. The meta-analysis had high

TE

heterogeneity, especially in the analysis of fasting glucose. The methodological

quality analysis highlighted that most of the studies presented methodological
P

error in at least one of the analyzed items, which may have contributed to the
CE

studies’ heterogeneity. Otherwise, the study strongly contributes to pediatric

exercise sciences, showing important associations between exercise training
and insulin resistance, fasting insulin, and fasting glucose, with a detailed
AC

description of the training protocols used in literature and an important sub-

analysis of the isolated effects of training modalities.

CONCLUSIONS
In general, exercise training interventions for at least eight weeks are
associated with reductions in fasting insulin and HOMA levels. More specifically,
aerobic exercise training is associated with reductions in fasting insulin and
HOMA levels, whereas combined exercise training and resistance exercise
training are not associated with the evaluated metabolic outcomes. Exercise
training interventions are not associated with fasting glucose. Based on these
results, it seems likely that physical training could play an important role in
preventing or delaying metabolic syndrome and type 2 diabetes. We emphasize
 ACCEPTED MANUSCRIPT
14

the need for future intervention studies investigating the effects of resistance
and combined training on insulin resistance markers in overweight or obese

T
children and adolescents. It is important that future intervention studies have

IP
higher methodological quality, with adequate random sequence generation,
allocation concealment, and blinding of outcomes assessment, aiming to

R
increase the internal validity and thus diminish the between-studies

SC
heterogeneity of future meta-analyses.
CONFLICT OF INTEREST STATEMENT

NU
The authors declare no conflict of interest associated with this manuscript.
MA

REFERENCES
D

American Diabetes Association. Standards of Medical Care in Diabetes.

Diabetes Care. 2016; 39(Sup.1): S1-S109.
TE

Artero, E.E., Lee, D.C., Lavie, C.J., España-Romero, V., Sui, X., Church, T.S.,
Blair, S.N. Effects of muscular strength on cardiovascular risk factors and
P

prognosis. J Cardiopulm Rehabil Prev. 2012; Nov-Dec; 32(6):351-8. doi:

10.1097/HCR.0b013e3182642688..
CE

Ackel-D’elia, C., Carnier, J., Bueno, C.R., Campus R.M., Sanches, P.L. Effects
of different physical exercises on leptin concentration in obese adolescents. Int
AC

J Sports Med. 2014;35(2):164-71. doi: 10.1055/s-0033-1345128.

Balducci, S.,Zanuso, S., Nicolucci, A., Fernando, F., Cavallo, S., Cardelli, P.
Anti-inflammatory effect of exercise training in subjects with type 2 diabetes and
the metabolic syndrome is dependent on exercise modalities and independent
of weight loss. Nutr Metab Cardiovasc Dis. 2010; 20(8): 608-617, 2010.

Chiarelli, F; Marcovvechio, M.L. Insulin resistance and obesity in childhood.

European Journal of Endocrinology (2008) 159 S67–S74.

Davis, J.,Gyllenhammer, L.E., Vanni, A.A., et al. Startup circuit training program
reduces metabolic risk in Latino adolescents. Med Sci Sports Exerc.2011Nov;
43 (11):2195-203. doi: 10.1249/MSS.0b013e31821f5d4e.

Davis,J., Kelly, L.A., Lane, C.J., Ventura, E.E., Byrd-Willians, C.E., Alexandar,
K.A. Randomized control trial to improve adiposity and insulin resistance in
overweight Latino adolescents. Obesity (Silver Spring). 2009 Aug;17(8):1542-8.
doi: 10.1038/oby.2009.19.
 ACCEPTED MANUSCRIPT
15

Davis, J., Tung, A., Chak, S.S., Ventura, E.E., Alexander K.E., Lane, C.J.
Aerobic and strength training reduces adiposity in overweight Latina
adolescents. Med Sci Sports Exerc. 2009 Jul; 41(7):1494-503. doi:

T
10.1249/MSS.0b013e31819b6aea.

IP
Delevatti R.S.,Kanitz A.C.,Alberton C.L.,et al.Glucose control can be similarly
improved after aquatic or dry-land aerobic training in patients with type 2

R
diabetes: A randomized clinical trial. Sports Med Australia. 2015; 1440-
2440(15)00208-X. doi: 10.1016/j.jsams.2015.10.008.

SC
Faigenbaum, A.D., Myer, G.D. Resistance training among young athletes:
safety, efficacy and injury prevention effects. Br J Sports Med. 2010 Jan; 44

NU
(1):56-63. doi: 10.1136/bjsm.2009.068098.

Farpour-lambert,N.J., Aggoun, Y., Marchand, L.M., Martin, X.E., Herrmann,

F.R., Beghetti, M. Physical activity reduces systemic blood pressure and
MA
improves early markers of atherosclerosis in pre-pubertal obese children. J Am
Coll Cardiol. 2009 Dec 15;54(25):2396-406. doi: 10.1016/j.jacc.2009.08.030.

Foschini, D., Araújo, R.C., Bacurau, R.F., De Piano, A., De Almeida S.S.,
D

Carnier, J. Treatment of obese adolescents: the influence of periodization

models and ACE genotype. Obesity (Silver Spring). 2010 Apr;18(4):766-72. doi:
TE

10.1038/oby.2009.247.

Garcia-hermoso, A., Saavendra, J.M., Escalante, I. Effects of exercise on

P

resting blood pressure in obese children: a meta-analysis of randomized

CE

controlled trials. Obes Rev. 2013 Nov; 14(11):919-28.

Garcia-hermoso, A., Saavendra, J.M., Escalante, I., Sanchez-Lopez, M.,

Martinez-Viscaino, V. Aerobic exercise reduces insulin resistance markers in
AC

obese youth: a meta-analysis of randomized controlled trials. Eur J Endocrinol.

2014 Oct;171(4):R163-71.

Goran, M., Ball, G., Cruz, M. Obesity and risk of type 2 diabetes and
cardiovascular disease in children and adolescents. J Clin Endocrinol Metab.
2003 Apr; 88(4):1417-27.

Hayashino, Y., Jackson, J.L., Hirata, T.,Fukumori, N., Nakamura, F., Fukuhara,
S. et al. Effects of exercise on C-reactive protein, inflammatory cytokine and
adipokine in patients with type 2 diabetes: A meta-analysis of randomized
controlled trials. Metabolism. 2014; 63 (3): 431-440.

Inoue, D.S., De mello, M.T., Foschini, D., Lira, F.S., De Piano Ganen, A., Da
Silveira Campos, R.M. Linear and Undulating Periodized Strength Plus Aerobic
Training Promote Similar Benefits and Lead to Improvement of Insulin
Resistance on Obese Adolescents. Journal of Diabetes and Its Complications
(2014), doi:10.1016/j.jdiacomp.2014.11.002
 ACCEPTED MANUSCRIPT
16

Karacabey, K. The effect of exercise on leptin, insulin, cortisol and lipid profiles
in obese children. J Int Med Res. 2009 Sep-Oct;37(5):1472-8.

T
Kelley, G., Kelley, K. Aerobic exercise and lipids and lipoproteins in children and
adolescents: a meta-analysis of randomized controlled trials. Atherosclerosis.

IP
2007 Apr; 191(2): 447–453.

R
Kelly, A.S., Wetzsteon, R.J., Kaiser, D.R., Steinberger, J., Bank, A.J., Dengel,
D.R. Inflammation, insulin, and endothelial function in overweight children and

SC
adolescents: the role of exercise. J Pediatr. 2004 Dec; 145 (6):731-6.

Lee, S., Bacha, F., Hannon, T., Kuk, J.L., Boesch, C., Arslanian, S. Effects of

NU
aerobic versus resistance exercise without caloric restriction on abdominal fat,
intrahepatic lipid, and insulin sensitivity in obese adolescent boys: a
randomized, controlled trial. Diabetes. 2012 Nov;61(11):2787-95. doi:
10.2337/db12-0214.
MA
Lee, S., Deldin, A.R., White, D., Kim, Y., Libman, I., Rivera-Vega., M. Aerobic
exercise but not resistance exercise reduces intrahepatic lipid content and
visceral fat and improves insulin sensitivity in obese adolescent girls: a
D

randomized controlled trial. Am J Physiol Endocrinol Metab. 2013 Nov

15;305(10):E1222-9. doi: 10.1152/ajpendo.00285.2013.
TE

Lee, K.J., Shin, Y.A., Lee, K.Y., Jun, T.W., Song, W. Aerobic exercise training-
induced decrease in plasma visfatin and insulin resistance in obese female
P

adolescents. Int J Sport Nutr Exerc Metab. 2010 Aug;20(4):275-81.

CE

Leite,N., Carvalho, H.M., Padez, C., Lopes, W.A., Milano, G.E., Radominski,
R.B. Age and menarcheal status do not influence metabolic response to aerobic
training in overweight girls. Diabetol Metab Syndr. 2013 Feb 25;5(1):7. doi:
AC

10.1186/1758-5996-5-7.

Liang, Y. et al. Childhood obesity affects adult metabolic syndrome and

diabetes. Endocrine. 2015; 60, 87–92.

Liberati A., Altman DG.,Tetzlaff J., et al. The PRISMA statement for reporting
systematic reviews and meta-analyses of studies that evaluate health care
interventions: explanation and elaboration. PLoS Medicine. 2009; 6 (1000100).
(doi:10.1371/journal.pmed.1000100)

Meyer, A.A., Kundt, G., Lenschow, U., Schuff-Werner, P., Kienast, W.

Improvement of early vascular changes and cardiovascular risk factors in obese
children after a six-month exercise program. J Am Coll Cardiol. 2006 Nov
7;48(9):1865-70.

Murphy, E.C., Carson, L., Neal, W., Baylis, C., Donley, D., Yeater, R. Effects of
an exercise intervention using Dance Dance Revolution on endothelial function
 ACCEPTED MANUSCRIPT
17

and other risk factors in overweight children. Int J Pediatr Obes. 2009;4(4):205-
14. doi: 10.3109/17477160902846187.

T
Rennie, K., Johnson, L.,Jebb, S. Behavioural determinants of obesity. Best
Practice & Research Clinical Endocrinology & Metabolism. 2005; 19 (3): 343–

IP
358.

R
Rey-López, J. P., Vicente-Rodríguez G.,Biosca M., Moreno L.A. Sedentary
behaviour and obesity development in children and adolescents. Nutr Metab

SC
Cardiovasc Dis. 2008 Mar; 18 (3):242-51.

Rosebaum, M., Nonas C., Weil R., et al. School-based intervention acutely

NU
improves insulin sensitivity and decreases inflammatory markers and body
fatness in junior high school students. J Clin Endocrinol Metab. 2007
Feb;92(2):504-8.
MA
Stamler J.,Vaccaro O.,Neaton J.D.,Wentworth D. Diabetes, other risk factors,
and 12-yr cardiovascular mortality for men screened in the Multiple Risk Factor
Intervention Trial. Diabetes Care. 1993 Feb;16(2):434-44.
D

SOCIEDADE BRASILEIRA DE CARDIOLOGIA. I Diretriz de prevenção da

aterosclerose na infância e adolescência. Arq Bras Cardiol. 2007; 85 (supl6): 1-
TE

36

Ten, S., Mclaren, N. Insulin Resistance Syndrome in Children. J Clin Endocrinol

P

Metab. 2004; 89(6):2526 –2539.

CE

Tsang, T., Kohn, M., Chow, C.M., Singh, M.F. A randomized controlled trial of
Kung Fu training for metabolic health in overweight/obese adolescents: the
"martial fitness" study. J Pediatr Endocrinol Metab. 2009 Jul;22(7):595-607.
AC

WHO. Population-based approaches to childhood obesity prevention. Word

Health Organization Press, Geneva, Switzerland; 2012.

Wong, P.C., Chia, M.Y., Tsou, I.Y, Wansaicheong, G.K., Tan, B., Wang, J.C.
Effects of a 12-week exercise training programme on aerobic fitness, body
composition, blood lipids and C-reactive protein in adolescents with obesity.
Ann Acad Med Singapore. 2008 Apr;37(4):286-93.

WORLD HEALTH ORGANIZATION. Report of the commission on ending

childhood obesity Geneva: WHO; 2016.

Zhang H., ZhangT.,Li S., Li Y.,Hussain A., Fernandez C et al. Long-term Impact
of Childhood Adiposity on Adult Metabolic Syndrome Is Modified by Insulin
Resistance: The Bogalusa Heart Study. Sci Rep. 2015; 5: 17885.
 ACCEPTED MANUSCRIPT
18

Figure 1. Flow diagram of articles identified.

T
IP
Eletronic Search
(n=1853)

R
SC
Records after
duplicates removes
(n=78)

Records screened (titles NU

MA
and abstracts)
(n=1775)

Records excluded
D

(n= 1721)
TE

Full-text articles assessed

for eligibility
(n=54)
P
CE

Full-text articles excluded (n=36)

AC

-Not fulfilling inclusion criteria: n=19

- No variable of interest: n= 11

- No comparator group: n=3

- No outcomes pre and post intervention: n=3

Studies included in
review
(n=18)
 ACCEPTED MANUSCRIPT
19

T
R IP
Figure 3. Absolute changes in fasting glucose levels of individual studies of

SC
exercise training (all modalities, aerobic, resistance and combined exercise vs.
no intervention.

NU
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT
20

Figure 4. Absolute changes in fasting insulin levels of individual studies of

exercise training (all modalities), aerobic, resistance and combined exercise vs.
no intervention

T
R IP
SC
NU
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT
21

T
Figure 5. Absolute changes in HOMA of individual studies of exercise training

IP
(all modalities) and aerobic exercise vs. no intervention

R
SC
NU
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT
22

Table2. Characteristics of the studies included in the review.

Studies Number Age (mean) and Dietary Dropouts

T
of study gender Cointervention
participants

IP
1. Ackel et al. (2014) 72 16.4 ± 1.5 (F/M) Yes NR

R
2. Davis et al. (2009a) 21 15.2± 1.1 (F) Yes 9
3. Davis et al. (2009b) 38 15.5 ± 1.0 (F/M) Yes 12

SC
4. Davis et al. (2011) 26 15.8 ± 1.1 (F) No 6
5. Farpour et al. (2009) 44 8.9 ± 1.5 (F/M) No 8
6. Foshini et al.(2010) 32 16.5 ± 1.7 (F/M) Yes N

NU
7. Inoue et al. (2014) 45 16.2 ± 1.3 (F/M) Yes 35
8. Karacabey (2009) 40 11.5 ±0,65 (M) No N
9. Kelly et al. (2004) 20 10.9 ± 0.4 (F/M) No 5
10. Lee et al. (2010a) 18 16.7 ± 0.6 (F) No N
MA
11. Lee et al. (2012) 42 14.8± 1.6 (M) Yes 3
12. Lee et al. (2013) 36 14.8±2.0 (F) Yes 6
13. Leite et al. (2013) 59 12.7 ± 1.7 (F) Yes 26
14. Meyer et al. (2006) 67 13.9 ± 2.2 (F/M) No 19
D

15. Murphy et al. (2009) 35 10.2± 1.6 (F/M) No NR

16.Rosenbaum et al. 53 13.6 ±1.1 (F/M) Yes 18
TE

(2006)
17. Tsang et al (2009) 19 13.1 ± 2.1 (F/M) No 1
18. Wong et al. (2008) 24 14±1.3 (M) No N
P
CE

F, female; M, male; NR, notrelated; N. no dropouts

AC
 ACCEPTED MANUSCRIPT
23

Table 3. Characterization of interventions

Study Intervention Modalities and Characteristics of training

T
duration number of

IP
study
participants

R
1. Ackel et 27 weeks Aerobic (24) Aerobic: 3x/wk.60 min in VT1
al. (2014) Combined (24) Combined: 3x/wk. AERO: 60

SC
Leisure (24) min in VT1; RESISTANCE:
progressive training: 20–
15RM, 12–10RM, 6–8RM

NU
Leisure: Leisure activities
2. Davis et 16 weeks Resistance (9) Resistance: 2x/wk.60 min
al. (2009a) Combined (15) (intensity not related)
MA
Control (7) Combined: 2x/wk. AERO: 30
min; RESISTANCE: 30 min
(intensity not related)
D

3. Davis et 16 weeks Resistance and Resistance: 2x/wk.60 min, 10

al. (2009b) nutrition(17) exercises (intensity not
TE

Nutrition (21) related)

4. Davis et 16 weeks Combined (14) Combined: 2x/wk.AERO: 30-
P

al. (2011) Control (12) 45 min; RESISTANCE: 12–

14,10–12,8–10RM
CE

5. Farpour et 13 weeks Combined (22) Combined: AERO: 3x/wk. 30

al. (2009) Control (22) min at 55–60%VO2max;
RESISTANCE: 3x/wk. 20 min,
AC

2–3 sets of 10–15RM.

6. Foschini 14 weeks Combined: Aerobic: 3x/wk.30 min in VT1
et al.(2010) linear Resistance: 3x/wk.30 min
periodization following the
(16) Combined: recommendations of the
wave American College of Sports
periodization Medicine. The group of linear
(16) periodization and wave had
modified its volume and
intensity differently; however,
their final values were
equivalent
7. Inoue et 48 weeks Aerobic (20) Aerobic: 3x/wk.30 min in VT1
al. (2014) Combined: Resistance: 3x/wk.30 min
linear following the
periodization recommendations of the
 ACCEPTED MANUSCRIPT
24

(13) American College of Sports

Combined: Medicine. The group of linear
wave periodization and wave had

T
periodization modified its volume and
(12) intensity differently; however,

IP
their final values were
equivalent

R
8. 12 weeks Aerobic (20) Aerobic: 3x/wk. 60 min, 50–

SC
Karacabey Control (20) 65%HRreserve
(2009)
9. Kelly et al. 8 weeks Aerobic (10) Aerobic: 4x/wk. 30–50 min,
(2004) Control (10) 50–90% VO2peak

NU
10. Lee et al. 12 weeks Aerobic (11) Aerobic: 4x/wk. 30–40 min,
(2010a) Control (7) 40–80% HRmax
MA
11. Lee et al. 12 weeks Aerobic (15) Aerobic: 3x/wk. 40-60 min
(2012) Resistance (16) (main part), 50–75% VO2peak
Control (11) Resistance: 3x/sem. First 4
weeks (1 or 2 sets of 8-12
D

repetitions at 60% of 1RM.

During weeks 4-12 (2 sets of
TE

8-12 repetitions to fatigue.

12. Lee et al. 12 weeks Aerobic (14) Aerobic: 3x/wk. 40-60 min
(2013) Resistance (14) (main part), 50–75% VO2peak
P

Control (8) Resistance: 3x/sem. First 4

CE

weeks (1 or 2 sets of 8-12

repetitions at 60% of 1RM.
During weeks 4-12 (2 sets of
8-12 repetitions to fatigue.
AC

13. Leite et 12 weeks Aerobic (43) Aerobic: 3x/wk. 90 min, 35–

al. (2013) Control (16) 75% HRreserve

14. Meyer et 24 weeks Aerobic (33) Aerobic: 3x/wk. Day 1:

al. (2006) Control (34) sessions comprised 60 min of
water activities (swimming or
water aerobics); day 2: 90 min
of sports; day 3: 60 mins of
walking
15.Murphy 12 weeks Aerobic (23) Aerobic: Aerobic exercise in
et al. (2009) Control (12) a dance video game. The
intensity increased
progressively during training.
Subjects were encouraged to
use the video game, 5x/wk.
 ACCEPTED MANUSCRIPT
25

16. 12 weeks Aerobic (49) Aerobic: 3x/wk. Dance and

Rosenbaum Control (24) dance/noncontact kickboxing
et al. (2006) activities were offered.

T
Participants could choose
between these activities and

IP
regular gym participation
17. Tsang et 27 weeks Kung Fu (11) Kung Fu: The group was

R
al. (2009) Control: Tai Chi guided by a professional Kung

SC
(8) Fu teacher; classes 3x/wk, 40
min.
Control Tai Chi: The group
was guided by a professional

NU
in Tai Chi; classes 3x/wk, 40
min.
MA
18. Wong et 12 weeks Combined (12) Combined: 2x/wk. 45–62
al. (2008) Control (12) min, 50–85% HRmax
x/wk: times a week; VT1: first ventilatory threshold; AERO: aerobic; HRmax:
Maximum heart rate; RM: Maximum repetitions, VO2peak: Peak of oxygen
D

uptake.
PTE
CE
AC
 ACCEPTED MANUSCRIPT
26

HIGHLIGHTS

 Childhood obesity is associated with insulin resistance;

T
 Childhood obesity is associated with high fasting insulin levels;

IP
 Aerobic training is associated with reduction in insulin resistance;

R
 Aerobic training is associated with reduction in fasting insulin.

SC
NU
MA
D
P TE
CE
AC