University of Southern California Dissertations and Theses

Objectively measured physical activity and body fat distribution in overweight Hispanic and African American adolescents

(USC Thesis Other)

Objectively measured physical activity and body fat distribution in overweight Hispanic and African American adolescents

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

OBJECTIVELY MEASURED PHYSICAL ACTIVITY AND BODY
FAT DISTRIBUTION IN OVERWEIGHT HISPANIC AND
AFRICAN AMERICAN ADOLESCENTS

by

Courtney E. Byrd-Williams

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PREVENTIVE MEDICINE)

August 2009

Copyright 2009 Courtney E. Byrd-Williams
ii

ACKNOWLEDGEMENTS

I wish to thank my committee members (Drs. Donna Spruijt-Metz, Jaimie
Davis, Kiros Berhane, and Florence Clark) and my committee chair (Dr. Michael
Goran) for their encouragement, support, and guidance through the course of my
time at USC. Dr. Spruijt-Metz, thank you for your tireless efforts in helping me to
understand accelerometry. Dr. Davis, thank you for your advice and for being a
great sounding board. Dr. Berhane, thank you for your kindness and patience.
Dr. Florence Clark, thank you for your enthusiasm and your insightful questions.
Dr. Goran, thank you for being a strong and supportive advisor. I have learned
so much under your guidance and by your example. Thank you also for helping
me to take the next steps in my career. I am indebted to you for helping me to
secure my fellowship in Texas.
I would also like to thank others who have helped me along the way. Dr.
Mary Ann Pentz, thank you for your mentorship, and Dr. Terry Huang, thank you
for your unflagging encouragement. To Marny Barovich, thanks for answering
my seemingly endless series of questions. To my fellow students, Dr. Emily
Ventura and Claudia Toledo-Corral, thanks for always keeping me smiling. To
Karla Wagner, thank you for trekking along with me; going through this program
with you has greatly eased the pain of this process. To Dr. Kate Coronges, thank
you for reminding me how to be spontaneous (and for introducing me to
Barbara).
iii

I also want to thank my family. To my husband, Thad, you have been
unimaginably supportive. Thank you for your love, sacrifice, and
encouragement. I am a very lucky girl. To my parents, thank you for allowing
me to forge my own path, even when none of us was sure where that path might
lead. Your support has meant the world to me!
I have been fortunate to have so much support through the last five years
that, unfortunately, I cannot thank you each individually. Collectively, I would like
to thank everyone who has helped me in this endeavor. It has been an amazing
journey.
iv

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

ii
LIST OF TABLES

vi
LIST OF FIGURES

viii
ABBREVIATIONS

x
ABSTRACT
.
xi

CHAPTER 1. INTRODUCTION
.

1
Significance
.
1
Background
.
2
Adiposity
.
2
Physical activity and adiposity
.
5
Sex differences in physical activity and adiposity
.
9
Influence of energy intake on the physical activity and
adiposity association
.
.
9
Ethnic differences in physical activity and adiposity
.
10
Specific aims
.
11
Research methods
.
14
Power analysis
.
15
Inclusion criteria
.
22
SANO-LA intervention
.
23
Sexual maturation and anthropometry
.
24
Body composition and abdominal adipose tissue 24
Dietary intake
.
26
Physical activity by accelerometry
.
26
Statistical analysis 32

CHAPTER 2. STUDY 1: CROSS-SECTIONAL EXAMINATION OF
ETHNIC DIFFERENCES IN PHYSICAL ACTIVITY AND ADIPOSITY
AMONG OVERWEIGHT HISPANIC AND AFRICAN AMERICAN
ADOLESCENTS
.

38

Introduction
.
39
Methods
.
41
Results
.
46
Discussion 50

v

CHAPTER 3. STUDY 2: CROSS-SECTIONAL EXAMINATION OF SEX
DIFFERENCES IN PHYSICAL ACTIVITY AND ADIPOSITY AMONG
OVERWEIGHT HISPANIC ADOLESCENTS
.
68
Introduction
.
69
Methods
.
71
Results
.
76
Discussion
.
79

CHAPTER 4. STUDY 3: SHORT-TERM CHANGES IN PHYSICAL
ACTIVITY AND ADIPOSITY IN OVERWEIGHT HISPANIC
ADOLESCENTS
.

91
Introduction
.
92
Methods
.
94
Results
.
100
Discussion
.
103

CHAPTER 5. SUMMARY, FUTURE DIRECTIONS, AND CONCLUSIONS

115
Summary of findings
.
115
Strengths and limitations
.
119
Future Research
.
123
Conclusions
.
127

REFERENCES
.

129
vi

LIST OF TABLES

Table 1.1 Pearson correlation coefficients between body composition and
physical activity variables for total sample of study 1

18
Table 1.2 Pearson correlation coefficients between body composition and
physical activity variables for total sample of study 2

18
Table 1.3. Minimum detectable differences for each independent and
dependent variable combination in Study 1 given the listed
sample size, alpha of 0.05, and power of 0.80 for the total
sample and by ethnicity

20
Table 1.4. Minimum detectable differences for each independent and
dependent variable combination in Study 2 given the listed
sample size, alpha of 0.05, and power of 0.80 for the total
sample and by sex

20

Table 1.5. Minimum detectable differences for each independent and
dependent variable combination in Study 3 given the listed
sample size, alpha of 0.05, and power of 0.80

21
Table 2.1. Participant characteristics by ethnicity

60
Table 2.2. Pearson correlations between physical activity and body
composition variables by ethnicity

61
Table 2.3. Linear regression of total fat on percentage of time spent in
moderate to vigorous physical activity

62
Table 2.4. Linear regression of subcutaneous abdominal adipose tissue
on total physical activity by ethnicity

63
Table 2.5. Linear regression of visceral adipose tissue on total physical
activity

65
Table 2.6. Linear regression of hepatic fat fraction on total physical
activity by ethnicity

66
Table 3.1. Participant characteristics by sex

86
Table 3.2. Pearson correlations between body composition, energy intake,
and physical activity variables by sex
87
vii

Table 3.3. Linear regression of total fat mass on total physical activity

88
Table 3.4. Linear regression of subcutaneous abdominal adipose tissue on
total physical activity

88
Table 3.5. Linear regression of visceral adipose tissue on total physical
activity

89
Table 3.6. Linear regression of hepatic fat fraction on total physical activity

90
Table 4.1. Characteristics of overweight Hispanic participants at pre-test
and post-test

110
Table 4.2. Pearson correlations of change in adiposity variables, change in
physical activity variables, change in energy intake and
demographic variables

112
Table 4.3. Regression analyses of changes in total physical activity and
changes in total fat mass and % body fat

112
Table 4.4. Regression analyses of changes in percentage of time spent in
moderate to vigorous physical activity (MVPA) and changes in
% body fat with and without controlling for total PA
114

viii

LIST OF FIGURES

Figure 1.1. Conceptual model for the role of sex as a moderator in the
relationship between physical activity and adiposity

13
Figure 1.2. Power and sample analysis

16
Figure 1.3. Conceptual model of sex moderating the relationship between
physical activity and fat mass

35
Figure 1.4. Conceptual model of energy intake confounding relationship
between physical activity and fat mass

36
Figure 2.1. Individual coefficient plot of total fat mass regressed on
percent of time spent in moderate to vigorous physical
activity

63
Figure 2.2. Individual coefficient plot of subcutaneous adipose tissue
regressed on total physical activity in a sample of overweight
Hispanic and African American adolescents, by ethnicity

64
Figure 2.3. Individual coefficient plot of visceral adipose tissue regressed
on total physical activity in a sample of overweight Hispanic
and African American adolescents

65
Figure 2.4. Individual coefficient plot of hepatic fat fraction regressed on
total physical activity in a sample of overweight Hispanic and
African American adolescents

67
Figure 3.1. Individual coefficient plot of SAAT regressed on total physical
activity in a sample of overweight Hispanic adolescents

89
Figure 3.2. Individual coefficient plot of VAT regressed on total physical
activity in a sample of overweight Hispanic adolescents

90
Figure 4.1. Mean change in total physical activity and change in percent
time in moderate to vigorous physical activity by intervention
group

110
Figure 4.2. Change in total physical activity (cpm) by individual

111
ix

Figure 4.3. Graph of predicted values of change in total physical activity by
change in fat mass

113
Figure 4.4. Individual coefficient plot of change in total fat regressed on
change in total physical activity in a sample of overweight
Hispanic adolescents
113
Figure 5.1. Conceptual model of diet mediating the relationship between
physical activity and fat mass

122

x

ABBREVIATIONS

AA = African American
BMI = Body Mass Index
DEXA = Dual Energy X-Ray Absorptiometry
FFA = Free Fatty Acids
GCRC = General Clinical Research Center
Hisp = Hispanic
MDD = Minimum Detectable Difference
PA = Physical Activity
PM = Particulate Matter
SAAT = Subcutaneous Abdominal Adipose Tissue
USDHHS = US Department of Health and Human Services
VAT = Visceral Adipose Tissue
HFF = Hepatic Fat Fraction
rmANCOVA = repeated measures Analysis of Covariance

Note on wording and definitions
Throughout this dissertation, the term overweight will be used for youth who have
a body mass index (BMI) > 85
th
percentile of sex- and age-specific CDC growth
charts. The term obese will be used for youth with a BMI > 95
th
percentile of sex-
and age-specific CDC growth charts.
xi

ABSTRACT

Purpose: This dissertation sought to examine the relationship between physical
activity (PA) and adiposity in overweight Hispanic (Hisp) and African American
(AA) adolescents. Three objectives were addressed, each in a separate study,
and the three studies comprise this dissertation. The objectives were 1) to cross-
sectionally explore ethnic differences in the relationship between PA and body fat
distribution among overweight Hisp and AA adolescents; 2) to cross-sectionally
examine sex differences in the relationship between PA and body fat in
overweight Hisp adolescents; 3) to examine whether short-term increases in
physical activity were associated with improvements in body composition in
overweight Hisp adolescents.

Methods: The participants were overweight (BMI > 85
th
percentile) Hisp and AA
adolescents, grades 9 – 12, who were participating in one of three obesity
intervention studies. Accelerometry was used to assess total PA and percent
time spent in MVPA and sedentary. Energy intake was assessed by 3-day diet
records. Fat and lean mass were assessed by DEXA, and fat stores, including
SAAT, VAT, and HFF, were assessed by MRI.

Results: In study 1, ethnic-specific relationships between PA and adiposity were
observed. Higher PA of 100 cpm, or roughly 30%, was associated with 22%
lower HFF in AA, but not Hisp. Higher PA was associated with 9% lower SAAT
xii

in Hisp, but not AA. In study 2, which only included overweight Hisp adolescents,
higher total PA was associated with 9.2% lower SAAT. No sex differences in the
association between PA and adiposity were observed. In study 3, which also
only included overweight Hisp adolescents, a 30% increase in total PA was
associated with a decrease of 1.4 kg in total fat mass and 1% body fat in
overweight Hisp adolescents after accounting for the influence of energy intake.

Conclusions: Total PA may be sufficient to improve body composition in
overweight Hisp and AA adolescents. Ethnic-specific associations between PA
and adipose depots suggest that physical activity interventions may improve
body composition differently in Hisp and AA adolescents.

1

CHAPTER 1. INTRODUCTION
Significance
The most recent national data indicate that 34% of adolescents, ages 12-
19 years, are overweight and 18% are obese
100
. These rates have risen radically
in the last 30 years. Thirty years ago only 5% of adolescents were obese
120
.
The current overweight and obese prevalence rates in Mexican American and
African American adolescents are even higher than the national rates; 38-39%
are overweight and 21-23% are obese
100
.
These high rates of overweight and obesity among adolescents are an
important public health concern. Being overweight in adolescence can contribute
to the development of serious risk factors for chronic diseases such as type 2
diabetes, cardiovascular disease, and some cancers
88, 101, 130
. Additionally,
obese youth are more likely to become obese adults
129
, and in adulthood, obesity
is again tied to a myriad of chronic diseases
62, 64, 108, 116
.
Hispanic and African American youth are disproportionately affected by
the obesity epidemic, and they are at higher risk for obesity-related comorbidities,
such as insulin resistance
51, 130
. In addition to having more total fat mass,
Hispanic youth are more likely to accumulate fat in the liver and skeletal muscle,
which also contributes to metabolic dysfunction
114
. Therefore, it is important to
understand the etiology of body fat distribution, particularly in these susceptible
populations.
2

The cause of obesity is clearly multifaceted, and the dearth of physical
activity
61
and the abundance of sedentary behaviors
38, 61, 71
in our youth are
several risk factors that are associated with an accumulation of fat mass. A
recent, nationally representative study indicates that physical activity declines
dramatically across age groups between childhood and adolescence, which
results in only 8% of adolescents meeting the recommended 60 mins/day of at
least moderate physical activity
149
. Physical activity decreases with age
127
, as
children become adolescents
70, 97
, and as adolescents become young adults
42, 74
.
To successfully develop future behavioral interventions aimed at preventing and
treating pediatric obesity, it is imperative that the relationship between physical
activity and adiposity is fully understood, especially in vulnerable populations.
This dissertation will help to inform future interventions by examining the
relationship between physical activity and body fat, including total body fat,
abdominal fat depots, and ectopic liver fat, in overweight African American and
Hispanic adolescents.
Background
Adiposity
Adipose tissue is loose connective tissue composed of adipocytes (fat
cells), which store excess energy primarily in the form of triglycerides
128
. As
discussed previously, being overweight or obese, or having an abundance of
stored fat, is linked to many deleterious health outcomes in adults and
adolescents. Adipose cells are not inert as previously thought, but they are
3

metabolically active cells
27
. Adipocytokines, which are cell-to-cell signaling
proteins, are secreted by adipocytes and include adiponectin, leptin, tumor
necrosis factor alpha (TNF-α), and interleukin-6 (IL-6). These adipocytokines act
as mediators by influencing the metabolism and insulin sensitivity
114
of distant
tissues and may help to explain the relationship between adiposity and insulin
resistance, cardiovascular disease, and type 2 diabetes
27
.
The location of body fat deposition is an important consideration for
metabolic health, because the level of metabolic functioning of the adipose cells
varies by fat tissue depot. Visceral adipose tissue is located in the space
contained between the two layers of visceral peritoneum, forming the omentum
and mesenterium in the peritoneal cavity. Visceral adipose tissue is associated
with a greater risk for type 2 diabetes, cardiovascular disease, hypertension, and
certain cancers
6, 12
, and it may contribute to metabolic dysfunction with its
increased rates of lipolysis and increased free fatty acids (FFA) delivered to the
liver via the portal vein
6
. SAAT is another compartment of adipose tissue that is
located in the area underlying the skin and overlying the parietal peritoneum.
Abnormalities in subcutaneous adipose tissue lipolysis could be an important
cause of peripheral insulin resistance
78
, which is a precursor to type 2 diabetes.
In some individuals, the positive energy balance created by our
obesogenic environment sometimes exceeds the storage capacity of adipose
tissue, and as a result, excess lipids are stored as triglyceride in liver and skeletal
muscle
114
. The accumulation of intracellular lipid in the liver and skeletal
4

muscle, also referred to as ectopic fat, may be caused by impaired whole body
fat oxidation or by an inability of adipocytes to properly proliferate and
differentiate. Regardless of the cause, ectopic storage of fat in the liver is
associated with multiple metabolic abnormalities, including hypertension,
dyslipidemia, and insulin resistance
15
.
In adults, it is widely accepted that disparate adipose stores (e.g.,
subcutaneous abdominal fat, visceral fat, and ectopic hepatic fat) have different
metabolic and health consequences
6, 8, 9
. In children and adolescents, though,
the relationship between adipose depots and health risk is less clear than it is in
adults. In obese adolescents, a significant decrease in visceral adiposity is
associated with a decrease in nonalcoholic fatty liver disease
29
and metabolic
syndrome
32
, suggesting the relationship may be similar to that in adults. On the
other hand, in obese Hispanic children an accumulation of both total and visceral
adiposity increases the risk of type 2 diabetes
26
, and in Black and White
adolescents, visceral adiposity does not explain any more of the variance in the
risk for cardiovascular disease than total adiposity. It has been suggested that
the deleterious effects of visceral adiposity may become apparent later in life
58
,
though the debate is ongoing.
In obese adolescents, ectopic liver fat has been identified as an increasing
problem, especially in Hispanic adolescents. Liska et al.
80
reported undetectable
levels of hepatic fat in obese African American adolescents, while obese
Hispanic adolescents had a mean of 13% intrahepatic fat content, which is
5

extremely elevated above the reference value of 5.5% used by clinicians to
diagnose steatosis. In obese adolescents, increasing levels of hepatic fat have
been associated with glucose dysregulation, metabolic syndrome, and higher
levels of proinflammatory markers, such as IL-6
19
. This dissertation will examine
the effects of physical activity on total body fat and fat depots, including visceral
fat, subcutaneous abdominal fat, and ectopic liver fat in a sample of overweight
Hispanic and African American adolescents.
Physical activity and adiposity
Many cross-sectional and longitudinal studies have found that obese
youth are less active than their non-obese counterparts
36, 121, 150
. Cross-sectional
studies in adolescents have found a significant negative association between
physical activity and adiposity measures
125, 134
. Additionally, longitudinal studies
in children and adolescents have found that increased physical activity is
associated with decreased adiposity over time
94
. However, there have been
mixed reports, and there is not an absolute consensus
136
, especially on the
magnitude of the relationship between physical activity and adiposity
96
. One
response to the apparent discrepancies in the literature about whether physical
activity contributes to obesity is that physical activity levels are related to
adiposity in older children and adults, but not in younger children
126
.
It has also been suggested that the lack of consensus is due to the use of
imprecise measures of activity, such as self-report measures, which weaken the
observed relationships between activity and adiposity
96
. In addition to the use of
6

self-report measures of activity, imprecise measures of adiposity, such as body
mass index (BMI), may be contributing to the confusion. Longitudinal
studies
138, 139
using BMI to measure adiposity that have failed to find a
relationship with physical activity support the notion that more precise measures
are needed.
One of the strengths of this dissertation is the use of objective measures
of physical activity and precise measures of fat mass. Few studies using
objective measures of physical activity, such as accelerometry, and more precise
measures of adiposity, such as DEXA to measure total body fat mass, have been
conducted
17, 42, 98
. Even fewer studies have examined the effect of physical
activity on specific fat depots in youth, using precise measurements, such as
multi-slice MRI, to measure visceral and subcutaneous abdominal adipose tissue
(VAT and SAAT, respectively)
95, 125
. To our knowledge, no studies have
examined the relationship between objectively measured physical activity and
hepatic fat. Additionally, no studies have examined the associations between
physical activity measured by accelerometry and adiposity measured by DEXA
and MRI in susceptible pediatric populations. The studies in this dissertation
contribute to the field by using objective and precise measures to examine
physical activity and adiposity relationships in overweight Hispanic and African
American adolescents.
Moderate to vigorous physical activity and adiposity. One way that
physical activity can be reported is as the proportion of time spent at different
7

physical activity intensities. A common physical activity intensity used in the
exercise physiology literature is time spent engaged in moderate to vigorous
physical activity (MVPA). MVPA includes activities such as walking briskly,
bicycling, or running. Whether physical activity intensity, such as MVPA,
influences changes in adiposity differently than total physical activity is
unclear
3, 14, 42, 53, 138, 147, 161-163
.
Ekelund et al.
42
report that change in overall activity, but not change in any
of the subcomponents of activity, such as MVPA, is inversely associated with
changes in fat mass. Another study using accelerometry to measure physical
activity reports a significant relationship between objectively measured MVPA
and body fat, but the authors did not control for total physical activity
138
.
Therefore it remains unclear whether the reductions in body fat were associated
with MVPA specifically, or if MVPA was simply a proxy measure for total physical
activity. A cross-sectional study in a large cohort of 12-year olds found that
MVPA was negatively associated with fat mass after adjustment for total physical
activity
98
, suggesting that the physiological effects of MVPA may be importantly
related to adiposity. To identify the effect of MVPA independent of total activity,
the research in this dissertation examines the effect of time spent in MVPA on fat
mass in overweight Hispanic and African American adolescents, while
statistically adjusting for total physical activity.
Sedentary behavior and adiposity. Physical activity can also be reported
as the proportion of time spent in sedentary behaviors, which include activities
8

such as watching television and using the computer. Recent national data
indicate that adolescents spend approximately seven to eight hours per day
sedentary
85
. It is currently being debated in the literature whether the large
amounts of time spent sedentary are independently contributing the obesity in
youth. Some advocate that youth can engage in both sedentary behaviors, e.g.,
using the computer for homework, and MVPA, e.g., playing on the soccer team
after school
7
, thus not necessarily increasing their obesity risk by being
sedentary.
Sedentary behaviors have been associated with increases in
obesity
11, 61, 67
, but this relationship is not always detected
39, 66
. Findings from a
meta-analysis argue that the effects of sedentary behaviors on adiposity are not
clinically meaningful
7, 83
. While time in sedentary behaviors has been associated
with decreased total physical activity, this relationship is small, suggesting that
sedentary activities are largely uncorrelated with physical activity
83
. On the
other hand, in adult studies, sedentary behaviors are associated with waist
circumference, independent of time spent in MVPA
60
. To date, no studies have
examined the effects of time spent in sedentary behavior measured via
accelerometry on precise measures of adiposity in ethnic minority adolescents.
This dissertation contributes to the literature by reporting the effect of sedentary
behaviors on total body fat, subcutaneous abdominal, visceral, and hepatic fat,
independent of total physical activity in overweight Hispanic and African
American adolescents.
9

Sex differences in physical activity and adiposity
In some of the previous work from our research group, Byrd-Williams et
al.
18
reported that accelerometry-measured activity levels do not vary by weight
status in Hispanic girls; specifically, Hispanic girls who were overweight were not
less active than their normal weight counterparts. Similar sex differences have
also been found in studies of physical activity and adiposity in both youth and
adults. Using doubly labeled water to measure total energy expenditure, physical
activity levels (total energy expenditure/resting energy expenditure) were
associated with increased body fat mass in boys, but not in girls
2, 123
. A meta-
analysis conducted in adults that used doubly labeled water to measure energy
expenditure also revealed a significant inverse relationship between activity-
related energy expenditure and percent body fat in males, but not in females
160
.
While a sex difference was not found in an accelerometry study using less
sophisticated methods to assess adiposity, such as skinfolds
121
, studies
employing more precise measures of adiposity, such as DEXA, have found that
increased physical activity is more strongly related to improvements in body
composition in boys than in girls
17, 40, 98
.
Influence of energy intake on the physical activity and adiposity association
In the studies examining sex differences in activity and adiposity that are
discussed above, the associations between physical activity and adiposity were
less clear in girls than boys, indicating that in the girls something other than a
lack of physical activity was contributing to their overweight status. Excessive
10

energy intake contributes to weight gain
13, 111
, and some have suggested that
energy intake, not low physical activity or low total energy expenditure, is the
influential factor in the development of overweight and obesity in youth
37, 41, 126,
141, 142
.
Many studies examining the relationship between activity and adiposity fail
to account for the influence of energy intake
17, 121, 152
. Recent longitudinal studies
in adolescents have acknowledged the potential limitations of not accounting for
the influence of energy intake when examining the relationship between
objectively measured physical activity and adiposity
42, 138
. Not accounting for
energy intake could mask the magnitude of the relationships between physical
activity and adiposity. Stevens et al
138
did not measure energy intake, because
of financial and logistical considerations, and acknowledge that adjusting for
energy intake may have strengthened their findings.
Though Saelens et al.
125
adjusted for energy intake when examining
activity and adiposity, they did not examine the effect of adjusting for energy
intake. This dissertation addresses an important deficit in the literature by
systematically investigating the importance of the role of energy intake in the
relationship between physical activity and adiposity.
Ethnic differences in physical activity and adiposity
The levels of physical activity and obesity appear to be similar in Hispanic
and African American adolescents. In a nationally representative sample, there
were no statistically significant differences in total activity or time spent sedentary
11

between Mexican American and Non-Hispanic black adolescents, ages 12-15
years and 16-19 years
85, 149
. Similarly, in a nationally representative sample, the
rates of overweight and obesity are not statistically different in Non-Hispanic
Black and Mexican American adolescents, ages 12-19 years. Other studies
comparing African Americans and Hispanics have found no significant
differences in physical activity levels
106
, prevalence of obesity
35
, waist
circumference
102
, or percentage body fat
106
.
Although there seem to be consistent similarities in physical activity and
total adiposity among African American and Hispanic adolescents, evidence from
adolescents and adults shows that lipids may accumulate into fat depots
differently in the two ethnic groups. African-Americans have lower volumes of
visceral and hepatic fat than Hispanics
20, 80
independent of total adiposity. It is
currently unknown whether physical activity influences total body fat and adipose
depots similarly in overweight African-American and Hispanic adolescents. This
dissertation will contribute to the literature by comparing the physical activity
levels of overweight Hispanic and African American adolescents and by
investigating if the relationships between physical activity and fat mass
accumulation are different between ethnicities. The specific aims of the three
studies that comprise this dissertation are outlined next.
Specific Aims
The overall aim of this dissertation is to examine how physical activity,
measured via accelerometry, is related to body fat distribution, specifically total
12

fat mass, VAT, SAAT, and hepatic fat, in overweight adolescents and how
potential moderators (sex and ethnicity) and a potential confounder (energy
intake) may affect this relationship. In previous research with elementary school-
age children using BMI to define weight categories, we found that overweight
children are less active than their normal weight peers, with the exception that
there were no differences in physical activity between overweight Hispanic girls
and their normal weight peers
18
.
This dissertation builds on our previous research by examining the
relationship between physical activity and adiposity in Hispanic and African
American adolescents using more precise measures of adiposity and accounting
for the influence of energy intake. Collectively, the three studies involved in this
dissertation explore how ethnicity, sex, and energy intake may influence the
relationship between physical activity and adiposity. This dissertation is
presented in the form of three empirical studies. Specific aims and hypotheses
corresponding to the three studies are discussed below.
Study 1: Cross-sectional examination of ethnic differences in physical activity and
adiposity among overweight Hispanic and African American adolescents
• Aim 1: To explore how physical activity differs by ethnicity (Hispanic and
African American) in overweight adolescents.
• Aim 2: To examine whether the relationship between physical activity (i.e.,
total physical activity, percent time spent in MVPA, and time sedentary) and
13

various adipose depots, including total fat mass, subcutaneous abdominal fat,
visceral fat, and hepatic fat, varied by ethnicity.
o Hypothesis: Engaging in more total physical activity will be associated
with less total fat, hepatic fat, visceral, and subcutaneous abdominal fat
in both Hispanic and African American adolescents.
Study 2: Cross-sectional examination of sex differences in physical activity and
adiposity among overweight Hispanic adolescents
• Aim 1: To examine potential sex differences in the association between
physical activity and adipose depots, including in overweight Hispanic
adolescents.
o Hypothesis: Physical activity will be significantly associated with
adiposity in boys, but not in girls (Figure 1.1).

Figure 1.1. Conceptual model for the role of sex as a moderator in
the relationship between physical activity and adiposity.

14

• Aim 2: To examine energy intake as a potential confounder of the cross-
sectional relationship between physical activity and total body fat in
overweight adolescents.
o Hypothesis: The relationship between physical activity and total fat will
be strengthened by the inclusion of energy intake in boys and girls.
Study 3: Short-term changes in physical activity and adiposity in overweight
Hispanic adolescents
• Aim 1: To examine whether short-term changes in physical activity are
associated with changes in adiposity.
o Hypothesis: An increase in objectively measured physical activity will be
associated with an improvement in body composition, resulting in lower
total fat mass, subcutaneous abdominal adipose tissue, visceral adipose
tissue, and hepatic fat fraction.
• Aim 2: To examine changes in energy intake as a potential confounder in the
relationship between changes in physical activity and changes in total fat
mass in overweight Hispanic adolescents.
o Hypothesis: Including changes in energy intake as a covariate will
strengthen the relationship between changes in physical activity and fat
mass.
Research methods
The general focus of this dissertation is to explore how physical activity
may be related to adiposity in overweight adolescents and how potential
15

moderators (sex and ethnicity) and a potential confounder (energy intake)
operate in this relationship. Collectively, the three studies involved in this
dissertation will examine how ethnicity, sex, and energy intake may influence the
relationship between physical activity and adiposity.
The abbreviated methodologies can be found in the individual studies in
the subsequent chapters. More detailed methodologies are described here.
Each of the studies in this dissertation are secondary analyses from recent
research projects conducted at the University of Southern California, including
SANO-LA (Strength and Nutrition Outcomes for Latino Adolescents; Goran,
Principal Investigator [PI]), STAND (Strength Training and Nutritional
Development in Los Angeles; Goran, PI), and ACT-LA (Adolescent Circuit
Training for Los Angeles; Davis, PI). Because the PIs of the projects collaborate
closely, many of the measurement protocols used in the three studies are
identical. The methods that are identical across studies will be explained in detail
here, and the methods that are study specific will be explained in the subsequent
and relevant chapters.
Power and sample size analysis
As mentioned above, the data for this dissertation was collected for three
different research projects. As a result, this power analysis did not inform the
recruitment of subjects, but was conducted to help ensure that a sufficient
sample was available for analysis. A study by Saelens et al.
125
, which examined
the association between Actigraph-measured physical activity and MRI-
16

measured VAT, was used to provide data needed to estimate the needed sample
size. Prior data, provided by Saelens et al., indicate that the standard deviation
of total physical activity is 160.2 and the standard deviation of the regression
errors will be 544.897. If the true slope of the line obtained by regressing VAT on
physical activity is 1.62, then 37 subjects are needed to be able to reject the null
hypothesis that this slope equals zero with probability, or power, of 0.80 (Figure
1.2). The type I error probability associated with this test of this null hypothesis is
0.05.

Figure 1.2. Power and sample analysis
The above power analysis, which uses data from a previously published
article
124
, is helpful in evaluating the number of subjects needed to detect an
effect in the relationship between total physical activity and VAT. This power
analysis, however, does not reflect the sample size needed for each of the
analyses that are conducted as part of this dissertation that have different
independent and dependent variables. In addition to total physical activity, the
independent variables of interest include percent of time spent in MVPA and
sedentary, and in additional to VAT, the dependent variables include SAAT, HFF,
17

and total fat mass. Power calculations similar to the one described above were
not possible for each combination of the independent and dependent variables
(e.g., total physical activity and SAAT, % time in MVPA and SAAT, % time
sedentary and SAAT, total physical activity and HFF, % time in MVPA and HFF,
etc.), because previously published data that used similar measures and analysis
techniques could not be identified for each combination. Instead, the minimum
detectable differences were calculated to determine how the sample sizes of the
data collected for the projects used in this dissertation may affect the ability to
detect the hypothesized effects.
In linear regression, which is the statistical method used in this
dissertation, the minimal detectable difference represents the smallest coefficient
estimate that would be statistically significant given the specified alpha and
power. Alpha, or the type I error probability, is defined as the probability that the
null hypothesis will be falsely rejected; in other words, it is the probability that it
will be concluded that an effect exists when a true effect does not exist. Power is
the probability of correctly rejecting the null hypothesis that the regression line
has a slope of zero, or said another way, it is the probability that the effect was
correctly identified.
Given an alpha of 0.05 and a power of 0.80, the minimum detectable
differences for each independent and dependent variable combination for each
paper were calculated using the sample sizes, standard deviations of the
independent and dependent variables, and correlation coefficients between the
18

independent and dependent variables from the data used in this dissertation.
The standard deviations for the physical activity and body composition variables
are available in Table 2.1, 3.1, and 4.1. The correlation coefficients for the three
studies are available as follows: 1) study 1 coefficients are in Table 1.1 (total
sample) and Table 2.2 (by ethnicity); 2) study 2 coefficients are in Table 1.2 (total
sample) and Table 3.2 (by sex); 3) study 3 coefficients are in Table 4.2.

19

The calculated absolute values of the minimum detectable differences are
shown in Table 1.1 – 1.3. In chapter 5 of this dissertation, the minimum
detectable differences are discussed and compared to the observed coefficient
estimates. Power analyses were conducted with PS: Power and Sample Size
Calculations, V 3.0.2
34
.

20

21
22

Inclusion criteria
The inclusion and exclusion criteria for the SANO-LA and STAND-LA
research projects were identical, except for the racial/ethnicity criteria.
Participants in the SANO-LA project were of Hispanic ethnicity, and participants
in the STAND-LA project were of African American descent. Hispanic or African
American descent was defined by parental self-report that stated both maternal
and paternal grandparents were of Hispanic/ African American descent.
Participants were recruited from local area high schools, health care
centers, community centers, word of mouth, and newspaper ads.
Participants for all three projects satisfied the following criteria for inclusion:
grades 9
th
thru 12
th
, 14-18 yrs, and BMI ≥ 85
th
(SANO-LA, STAND-LA) or > 95
th

(ACT-LA) age- and sex-specific percentile (CDC, 2000). Participants in all three
projects were excluded based on the following criteria: 1) using medication or
diagnosed with any syndrome or disease that could influence dietary intake,
exercise ability, body composition and fat distribution, or insulin action and
secretion (e.g., prednisone) ; 2) previously diagnosed with any major illness since
birth (e.g., severe intrauterine growth retardation, chronic birth asphyxia, cancer);
3) diagnosed with any criteria for diabetes including polyuria, polydipsia with/or
without unexplained weight loss, fasting plasma glucose > 126 mg/dl, or a 2-hour
plasma glucose >200 mg/dL during an oral glucose tolerance test; or 4)
participated in structured exercise (including strength training), nutrition, or
weight loss program in the past six months.
23

The participants in the ACT-LA research project, which included only
Hispanic girls, were also excluded if they were taking oral contraception, reported
smoking, or had previously been pregnant.
SANO-LA Intervention
Study 3, which is in chapter 4, examines the effect of the SANO-LA
intervention on physical activity levels in Hispanic adolescents, and as a result,
the intervention groups are described in detail here. Study 1 and 2 also include
participants from the SANO-LA project, but those are cross-sectional studies
conducted at baseline so the intervention treatment groups are not relevant to
study 1 and 2. Participants in SANO-LA were randomized to one of three
intervention groups: nutrition education only, nutrition education and strength
training, or the control group. The 16-week intervention is briefly described here.
Nutrition education only group. Participants randomly assigned to the
nutrition education only group attended one 90-minute culturally tailored dietary
intervention class per week. The modified-carbohydrate nutrition education class
specifically aimed to decrease added sugar intake primarily by reducing intake of
sugar sweetened beverages, candy, and syrup and to increase dietary fiber
intake through an increase in fruits and vegetables and modification of breads
and cereals. Participants were not encouraged to increase or change habitual
physical activity.
Strength-training + nutrition education group. In addition to the nutrition
education class described above, the participants assigned to the nutrition
24

education and strength-training group also attended two non-consecutive, 60-
minute strength-training sessions per week. As the participants’ strength
improved, the number of sets, repetitions, and the resistance used were
increased. As in the nutrition education only group, participants were not
encouraged to increase or change habitual physical activity.
Control group. Participants randomized to the control group received no
intervention during the 16 weeks between pre- and post-intervention data
collection. Participants were encouraged not to change dietary intake or habitual
physical activity throughout the 16 weeks. After post-testing, participants were
offered an abbreviated intervention for one month, consisting of biweekly nutrition
and modified strength-training classes developed from the curriculum used in the
intervention.
Sexual maturation and anthropometry
A licensed pediatric health care provider assessed sexual maturation
using Tanner stages as indicators of the development of breasts in girls, genitals
in boys, and pubic hair in each sex
143
. Weight and height were measured to the
nearest 0.1 kg and 0.1 cm, respectively, using a beam medical scale and wall-
mounted stadiometer. BMI percentiles for age and sex were determined using
EpiInfo 2000, Version 1.1 (CDC, Atlanta, GA).
Body composition and abdominal adipose tissue
Body composition. Whole body fat, soft lean tissue (subtracting bone
mineral content), and percentage body fat were measured by dual energy x-ray
25

absorptiometry (DEXA) using a Hologic QDR 4500W (Hologic, Bedford, MA).
This procedure is a non-invasive procedure with a short scan time (<10 minutes)
with minimal radiation exposure (3.6 μSv) that provides accurate and precise
measurements of whole-body fat, soft lean and bone masses
87
. Percentage fat

mass (% fat mass) determined by DEXA is calculated as [fat mass/(fat

mass +
bone-free lean tissue mass + bone mineral content) x

100]
144
.
Abdominal adipose tissue. Hepatic fat fractions (HFF) and subcutaneous
abdominal adipose tissue (SAAT) and visceral adipose tissue (VAT) volumes
were obtained by magnetic resonance imaging (MRI), using a Siemens
Magnetom 1.5T Symphony Maestro Class Syngo 2004A (Siemens AG,
Erlangen, Germany) with a Numaris/4 software. Patients were positioned
supine, and 19 axial images of the abdomen with a thickness of 10 mm were
taken with field of view (FOV) of 420 mm and FOV phase of 75%. The base
resolution of the images was 256 with a phase resolution of 80% and a
bandwidth of 380 Hz/Px. The total acquisition time was 24 s per total abdominal
scan. Three abdominal scans were performed consecutively. After image
acquisition, visceral and subcutaneous abdominal tissue was segmented using
image analysis software (SliceOmatic Tomovision, Montreal Canada) at Image
Reading Center (New York City, New York). Visceral and subcutaneous
abdominal fat volume was calculated from these images. Also, after image
acquisition, the percentage of fat observed in the liver, or HFF, was calculated
using a modification of the Dixon 3-point technique
46
.
26

Dietary intake
Participants completed 3-day (2 weekdays and 1 weekend day) diet
records at home. In youth, 3-day diet records have been shown to be superior to
24-hour recalls and 5-day food frequencies
25
. Research study staff, who were
trained and supervised by a Registered Dietitian, gave participants a short, 10-
minute lesson on how to complete the diet records. Measuring cups and rulers
were given to the participants to aid in accurate estimate of portion sizes. When
the records were returned, research staff clarified all dietary records to ensure
that appropriate data were collected (e.g., amount of food, brand name of food, if
applicable etc.). Nutrition data were analyzed using the Nutrition Data System
for Research (NDS-R version 5.0_35), a software program developed by the
University of Minnesota. Only participants who returned at least 2 days worth of
diet records were included in the analyses.
Physical activity by accelerometry
This section will first include a general review and discussion of
accelerometry as a measure of physical activity. It will discuss important
considerations that must be made when conducting physical activity research
with accelerometry. The section will then include a discussion of the specific
accelerometry methodology employed in the current studies.
Important considerations when measuring physical activity by accelerometry
The following is a review and discussion of important points on
accelerometry, a relatively newer method of objectively measuring physical
27

activity
50
. Physical activity measured by self-report among youth has low to
moderate validity
77
and is subject to recall bias, especially among overweight
children and adolescents
89
. Additionally, research has shown that there is a
significant amount of over-estimation of activity reported in self-report
measures
10
. More objective measures include accelerometers, pedometers,
heart rate monitoring, and double-labeled water for assessment of free-living
physical activity-related energy expenditure
24, 50
. Currently, the accelerometer is
the objective method of choice for measuring body movement in free-living
individuals
24, 115
. Accelerometers are easy to use and unobtrusive. They
measure the intensity and total amount of physical activity directly, via
acceleration, and they are able to quantify the movement of the body part to
which the accelerometer is attached
24
. Additionally, research conducted with
sealed pedometers and accelerometers suggests that there is little to no
reactivity, or a change in normal activity patterns when participants know they are
being monitored, over the entire monitoring period
24, 86, 157
.
Choice of accelerometer. In the current studies, Actigraph accelerometers
(models GT1M and 7164, Actigraph, LLC., Pensacola, FL) were used to assess
physical activity. The Actigraph accelerometer is a uniaxial physical activity
monitor that measures acceleration, or the change in velocity with respect to time
(SI unit; m/s
2
), on a vertical plane. The Actigraph measures and records
acceleration using piezoelectric technology, which changes applied mechanical
stress into an electrical potential, and when the monitors detect acceleration the
28

piezosensor emits a voltage signal proportional to the intensity of the
acceleration
43
. The Actigraph is one of the few accelerometers that has been
proven to correlate adequately with energy expenditure measured by doubly
labeled water
110
, and it is a reliable instrument, with an intraclass correlation of
0.99
44
.
Activity counts. The output of the raw acceleration data is in ‘counts’ over
a user-defined time interval, or epoch. In the Actigraph monitors, one ‘count’ is
equal to 16.6 milli G’s per second at 0.75 Hertz, which is part of the filter
frequency to help ensure that the acceleration captured is likely generated by
human movement and not from movement of a car or lawn mower. Thirty
measurements per second are taken and the counts are summed to the user-
defined epoch. Activity counts are the summation of the absolute values of the
sampled change in acceleration, ∆A/∆t, measured during the cycle period.
Activity counts represent a quantitative measure of bodily movement over time
1
.
Cut-points. In addition to reporting raw activity count data, many
researchers use regression equations to convert counts to more physiologically
relevant units, such as activity-related energy expenditure or intensity level.
Intensity levels are defined by METs, or metabolic equivalent, defined as the ratio
of a person’s working metabolic rate relative to their resting metabolic rate.
Researchers commonly apply cut-points, or thresholds to designate engaging in
a specific activity intensity, such as MVPA.
29

There is no agreement on which cut-points to use in a given population,
and there is a broad range of thresholds used to categorize physical activity
intensities. The range of cut-points used that categorize MVPA varies from > 500
cpm
45
to > 3200 cpm
113
. It is the differences in the calibration studies conducted
to determine the cut-points for physical activity intensities that contribute to the
broad range of cut-points. These differences include different sample sizes, age
range of participants, activities (treadmill walking/ running or free-living),
definitions of MVPA (> 3 or 4 METS), and criterion method used (VO
2
or energy
expenditure)
47, 56
.
When applying activity intensity cut-points, the prediction equations used
to derive the cut-points introduce residual error into the measurement
47
. In some
cases, the best measure of activity may be the raw activity counts as opposed to
activity intensity based on a cut-point. Activity counts are not influenced by
researchers using different definitions of MVPA (3 vs. 4 METS) or different
prediction equations, nor are they exposed to the errors of estimated regression
equations
47
. Using activity counts may be especially important when tracking the
physical activity of individuals over time, because the error introduced by the
regression equations may obscure important trends of physical activity within the
same individual over time
47
.
Activity limitations of the accelerometer. Accelerometers cannot measure
activity that is not ambulatory in nature, such as bicycling and resistance training,
or activity that is water-based, such as swimming. Another limitation is the
30

inability to wear accelerometer during contact sports
43
, because of risk of injury to
wearer and others. These limitations may reduce the accuracy of the
measurement when counts are converted into energy expenditure, because
accelerometers are unable to detect energy expenditure resulting from lifting
weight, from walking uphill, or from activities completed while the accelerometer
is not worn. Despite these activity limitations, accelerometry is a preferred
method of measuring physical activity
86, 115
.
Length of measurement period. The number of days of monitoring that
are needed to accurately assess habitual physical activity is contestable.
Different studies have examined the number of days of monitoring needed to
obtain a reliability of 0.80. Measurement periods of three to five days, seven
days, and eight to nine days have been recommended, depending on the age of
the participants
151, 55
. It has also been recommended that studies of adolescents
include both weekend and weekdays in the monitoring time, because
adolescents engage in significantly lower levels of MVPA on weekends
compared to weekdays
151
.
The mechanical calibration of the individual accelerometer is another thing
that could affect the reliability of the measurement, though research has shown
that the calibration of Actigraph monitors used to collect data in free living
conditions in adolescents has little effect on inter-instrument variability, probably
due to more dominant sources of variation
92
.

31

Accelerometry methodology employed in the current studies
To assess physical activity, subjects were instructed to wear the GT1M
and 7164 Actigraph accelerometers (Actigraph, LLC., Pensacola, FL) for seven
days during all waking hours, except during water-based activities or when
sleeping
158
. Participants were shown how to wear the device above the iliac
crest of the right hip on an elastic belt
113
. The accelerometers were set to
monitor activity in 15-second sampling intervals, or epochs. Accelerometry data
downloaded from the Actigraph devices were reduced using an adapted version
of the SAS code used to reduce the accelerometry data in the 2003-2004
National Health and Nutrition Examination Survey (NHANES) available at
http://riskfactor.cancer.gov/tools/nhanes_pam .
Data were screened for invalid data, such as count values equal to
32,767, which indicates voltage signal saturation within the piezosensor
43
. The
amount of time the participant wore the device was determined by subtracting
nonwear time from 24h. Nonwear time was defined by an interval of at least 60
consecutive minutes of zero activity intensity counts, with allowance for 1-2 min
of counts between 0 and 100.
Because two different models of Actigraph accelerometer were employed
in the current studies, a mathematical adjustment was applied to allow for
comparison between the two monitor models. Research has shown that
compared to the GT1M, the Model 7164 classifies less time as sedentary and
more time as light-intensity activity
23
. On average, counts per minute captured
32

by the GT1M are 9% lower than the Model 7164. Based on the research of
Corder et al.
23
, a correction factor of 0.91 was employed (Eq. 1.1) to enable
accurate comparison between the models used in this dissertation.
Model 7164 = GT1M / 0.91
Equation 1.1. Correction factor to compare Actigraph models
26

Statistical analyses
Multivariate linear regression and regression diagnostics
Multivariate linear regression. Multivariate linear regression will be the
primary statistical procedure used in this dissertation. There are generally two
different goals associated with conducting a multivariate linear regression: 1) to
predict the dependent variable with a set of independent variables, or 2) to
quantify the relationship of one or more independent variables to a dependent
variable
76
. The first goal focuses on finding a model that fits the observed data
and predicts future events as well as possible, i.e., attempting to predict
individuals who will develop type 2 diabetes using demographic and clinical
variables. The second goal aims to approximate a relationship between a
predictor and dependent variable by producing an estimate of a regression
coefficient in the model
76
. The regression analyses conducted in this dissertation
pertain to the second goal; specifically, the analyses conducted in this
dissertation seek to quantify the magnitude of the relationship between physical
activity and adiposity. Individual coefficient plots will be to visually depict the
relationship between physical activity and adiposity. Individual coefficient plots,
also known as partial regression plots, display information about a single
33

regression coefficient. Vellemen et al.
155
state that one of the useful properties of
the partial plots include the fact that it is easy to see the influence of the
individual data values on the estimation of the coefficient, and for this reason,
individual coefficient plots will be used throughout this dissertation to illustrate the
relationship between physical activity and adiposity.
Linear regression diagnostics. Diagnostic analyses were conducted to
ensure that regression assumptions were met. The assumptions of linear
regression include 1) linearity of the relationship between dependent and
independent variables; 2) independence of the errors; 3) homoscedasticity, or
constant variance, of the errors; 4) normality of the error distribution. To ensure
that the linearity and homoscedasticity assumptions were not violated, plots of
the residuals by the predicted values were examined. To ensure that the
normality of errors assumption was not violated, normal probability plots of the
residuals were examined.
Diagnostic analyses were also conducted to ensure that predictors in the
models were not collinear. Multicollinearity, a statistical phenomenon in which
two or more predictor variables are highly correlated, can affect the calculations
of the individual predictors, and it was assessed by examining tolerance and
variance inflation factor (VIF) values. Tolerance is a measure of collinearity, and
the tolerance of a variable is 1-R
2
of the predictor. A small tolerance value
indicates that the variable under consideration is almost a perfect linear
combination of predictors already in the equation and that it should not be added
34

to the regression equation. In general, a tolerance value less than 0.1 should be
further investigated. VIF measures the impact of collinearity among the variables
in a regression model. VIF is 1/tolerance, and in general, a value greater than 10
indicates possible multicollinearity.
Evaluating moderators and confounders in linear regression
The following section contains a description of the conceptual methods
that will be used to evaluate the potential confounders and moderators in the
dissertation studies. These descriptions are included here, because they are
outside the scope of the study specific methods that are provided in the individual
study chapters.
Evaluating sex and ethnicity as moderators. This dissertation examines
ethnicity and sex as potential moderators, also known as effect modifiers, in the
relationship between physical activity and adiposity. In this dissertation, a
moderator is a qualitative variable (e.g., sex) that affects the direction and/or
strength of the relation between a predictor variable (e.g., physical activity) and a
dependent variable (e.g., total fat mass)
4
. In other words, sex would be classified
as a moderator if there was a significant relationship between physical activity
and adiposity in boys, but not in girls.
The model diagrammed in Figure 1.3 is a conceptual model of
moderation, in which sex is acting as the moderating variable. The relationship
between the predictor variable (e.g., physical activity) and the outcome variable
(e.g., fat mass) is the main effect (path a). The moderator variable (e.g., sex) is
35

shown to affect the relationship between physical activity and fat mass (path b).
Using this example, sex is a moderator if the relationship between physical
activity and fat mass is different in boys than it is in girls. In this dissertation, both
sex and ethnicity will be examined as moderators. As such, a similar moderator
model could be illustrated using ethnicity as the moderator variable, in which the
relationship between physical activity and fat mass would be different in
Hispanics and African Americans.

Figure 1.3. Conceptual model of sex moderating the relationship
between physical activity and fat mass

Evaluating energy intake as a confounder. This dissertation examines
energy intake as a potential confounder in the relationship between physical
activity and adiposity. When investigating the association between a predictor
(e.g., physical activity) and an outcome (e.g., fat mass), a confounding variable is
a third variable that may be distorting the true relationship between the predictor
and the outcome
33
. In this dissertation, it is hypothesized that not accounting for
the influence of energy intake attenuates the relationship between physical
activity and adiposity.
36

Figure 1.4. Conceptual model of energy intake confounding
relationship between physical activity and fat mass

As seen in Figure 1.4, the confounding variable is associated with the
exposure of interest (path a) and is also a potential cause of the outcome of
interest (path b). In addition, the predictor variable is also a potential cause of
the outcome (path c).
To evaluate whether energy intake is a confounder, the change-in-
estimate criteria
90
will be used. In the change-in-estimate approach, first the
model containing the outcome of interest (e.g., fat mass), the predictor of interest
(e.g., physical activity), and covariates will be run to estimate the parameter
coefficient, or B, for the predictor of interest. This is the reduced model. Then,
the confounder variable will be added to the model as an additional covariate.
This is the expanded model. The parameter estimates of the predictor of interest
(e.g., physical activity) from the reduced and expanded models will be compared.
If including the potential confounder variable changes the parameter estimate of
the predictor of interest by 10%, then the variable will be considered a
confounder
90
.
To summarize the research methods section, the relationship between
physical activity and body fat stores (i.e., total fat mass, SAAT, VAT, and HFF) in
37

overweight Hispanic and African American adolescents will be examined using
objective measures of physical activity, precise measures of body fat stores, and
multivariate regression. The following three chapters include the three empirical
studies that comprise the body of this dissertation.
38

CHAPTER 2. STUDY 1: CROSS-SECTIONAL EXAMINATION OF ETHNIC
DIFFERENCES IN PHYSICAL ACTIVITY AND ADIPOSITY AMONG
OVERWEIGHT HISPANIC AND AFRICAN AMERICAN ADOLESCENTS

Abstract
Purpose: Objectives of this study were to 1) test for ethnic differences in
objectively measured physical activity in overweight Hispanic and African
American adolescents and 2) examine whether the relationship between
physical activity and various adipose depots, including total fat mass,
subcutaneous abdominal fat, visceral fat, and hepatic fat, varied by ethnicity.
Methods: Participants were 48 Hispanic (Hisp) and 29 African American (AA)
overweight adolescents (BMI > 85
th
percentile for age and gender). Total fat and
lean mass were assessed by DEXA. Hepatic fat fraction (HFF), visceral (VAT)
and subcutaneous abdominal adipose tissue (SAAT) volumes were assessed by
multiple-slice MRI. PA was assessed by 7-day accelerometry. Total PA is
expressed as counts per minute (cpm). Linear regression was used to examine
whether PA was associated with adiposity measures (i.e., total fat, VAT, SAAT,
and HFF) after controlling for sex, age, sexual maturation, ethnicity, caloric
intake, and relevant adiposity covariates. To evaluate ethnic differences in the
relationship between PA and adiposity, ethnicity*PA interactions were included in
the models. Variables were log transformed as needed to meet regression
assumptions.
Results: Overweight Hisp adolescents spent significantly more time in MVPA
than their AA counterparts (mean + SD, Hisp 2.6 + 2.0, AA 1.5 + 1.8m, p<0.05).
39

No significant differences between Hisp and AA adolescents were detected for
total PA or % of time spent sedentary. In both Hisp and AA, higher levels of PA
were associated with higher values of log VAT after adjusting for SAAT (B =
0.0014, p<0.05). Stratification by ethnicity revealed higher PA (100 cpm) was
associated with 12% lower SAAT among Hisp (B = -0.00096, p<0.01), but not AA
adolescents. Conversely, higher PA was associated with 22% lower log HFF
among overweight AA, but not Hisp, adolescents regardless of whether the
model was adjusted for VAT (AA B = -0.0025, p = 0.04) or total fat mass (AA B =
-0.0025, p = 0.04).
Conclusions: Higher levels of PA were associated with adiposity differently in
this relatively sedentary sample of overweight AA and Hisp adolescents.
Approximately 30% higher PA (100 cpm) was related to 22% lower HFF among
AA adolescents and 8% lower SAAT among Hisp adolescents. These findings
suggest that PA may confer different levels of health benefits in overweight
Hispanic and African American adolescents.
Introduction
Recent data from the nationally-representative National Health and
Nutrition Examination Survey (NHANES) indicate that 34% of adolescents are
overweight, and 18% are obese
100
. It is also reported in NHANES that the
overweight and obese prevalence rates in Mexican American and African
American adolescents are even higher than the national rates; 38 - 39% are
overweight and 21 - 23% are obese, respectively
100
. Other large-scale studies
40

have also reported that African American and Hispanic are more obese than their
white counterparts
35, 102
.
Though there are similarities in total adiposity among African American
and Hispanic adolescents, evidence has shown that African Americans and
Hispanics may accumulate fat differently. Specifically, African-Americans have
lower volumes of visceral adipose tissue
20
, have a greater tendency to expand
subcutaneous adipose tissue stores, and accumulate less ectopic fat in the liver
than Hispanics
80
. Numerous studies have shown that increased physical activity
is associated with lower total fat mass and lower visceral adipose
tissue
72, 95, 96, 125
, but only a few studies have examined the relationship between
physical activity and hepatic fat
107, 135, 137
, with none to our knowledge in youth.
Moreover, the studies that have examined physical activity and liver fat
generally use self-report measures to assess physical activity
107, 137
. To our
knowledge, no studies have used objectively measured physical activity to
assess the relationship between physical activity and hepatic fat accumulation,
nor has any study examined ethnic differences in this relationship. It is currently
unknown whether physical activity is associated with total body fat and adipose
depots similarly or differently in African-American and Hispanic adolescents.
Therefore, the aims of the current cross-sectional study were two-fold: 1)
test for ethnic differences in objectively measured physical activity in overweight
Hispanic and African American adolescents, and 2) examine whether the
relationship between physical activity and various adipose depots, including total
41

fat mass, subcutaneous abdominal fat, visceral fat, and hepatic fat, varied by
ethnicity.
Methods
Participants
Study participants were Hispanic (n = 48) and African American (n = 29)
adolescents (ages 13 – 18 yrs) who participated in a randomized nutrition
education and strength training intervention trial aimed at preventing type 2
diabetes. This analysis used baseline data, and participants were included if
they had complete baseline data on demographic, physical activity and dietary
variables, and body composition variables. Due to a delay in finalizing the MRI
methodology, only 41 Hispanics and 17 African American participants had SAAT
and VAT data, and only 31 Hispanic and 17 African American participants had
HFF data. Three boys were excluded from the analysis when total fat mass was
included in the model, because they exceed the weight limit for the DEXA table.
Participants excluded from the analysis were not different on weight, BMI, fat
mass, lean mass, percent female, percent Hispanic, or pubertal stage than those
included in the analysis (all p>0.25). Though the difference in mean age was
less than a year, participants included in the analyses were significantly younger
than those not included (mean + SD, 15.3 + 1.2 years and 16.0 + 1.2 years,
respectively). Study methods have previously been reported elsewhere
28
,
therefore only a brief overview of the methods will be described here. The
42

protocol for the study was approved by the Institutional Review Board at the
University of Southern California.
Procedures
Screening visit. Participants arrived at the General Clinical Research
Center after an overnight fast. A licensed pediatric health care provider
conducted a medical history exam and determined sexual maturation via Tanner
staging
143
. To screen for diabetes, an oral glucose tolerance test was conducted.
Participants who met the following criteria were invited back for further testing: 1)
age- and gender-specific BMI ≥ 85
th
percentile; 2) African American or Hispanic
ethnicity; 3) grades 9
th
thru 12
th
; 4) not currently taking medication or diagnosed
with any syndrome or disease that influences fat distribution or insulin action; 5)
not diagnosed with diabetes at screening or any major illness (e.g., cancer) since
birth; 6) reported not participating in a structured exercise, nutrition, or weight
loss program in the past six months.
Anthropometric and Body Composition. Weight and height were
measured in triplicate using a beam medical scale and wall-mounted
stadiometer, respectively, and then averaged. BMI percentiles for age and
gender were determined using EpiInfo 2000, Version 1.1 (CDC, Atlanta, GA).
Whole body fat, lean tissue, and percent body fat were measured by dual energy
x-ray absorptiometry (DEXA) using a Hologic QDR 4500W (Hologic, Bedford,
MA).
43

Abdominal adipose tissue. Hepatic fat fractions (HFF), and subcutaneous
abdominal adipose tissue (SAAT), and visceral adipose tissue (VAT) volumes
were obtained by magnetic resonance imaging, using a Siemens Magnetom 1.5T
Symphony Maestro Class Syngo 2004A (Siemens AG, Erlangen, Germany) with
a Numaris/4. Patients were positioned supine, and 19 axial images of the
abdomen with a thickness of 10 mm each were taken. After image acquisition,
HFF was calculated using a modification of the Dixon 3-point technique
46
, and
visceral and subcutaneous abdominal tissue was segmented and calculated
using image analysis software (SliceOmatic Tomovision, Montreal Canada) at
Image Reading Center (New York City, New York). These procedures have
been described in more detail elsewhere (Fisher et al, unpublished data).
Physical activity and dietary assessment. To assess physical activity,
subjects were instructed to wear Actigraph accelerometers (GT1M or 7164,
Actigraph, LLC., Pensacola, FL) for seven days, except during water-based
activities or when sleeping
113, 158
. Accelerometers were set to monitor activity in
15-second epochs, which were collapsed to 60-second epochs during analysis.
Data were reduced using an adapted version of the SAS code used for the 2003-
2004 National Health and Nutrition Examination Survey available at
http://riskfactor.cancer.gov/tools/nhanes_pam. A correction factor was applied to
allow for comparison between the two Actigraph monitor models
23
.
The amount of time the participant wore the device was determined by
subtracting nonwear time from 24h. Nonwear time was defined by an interval >
44

60 consecutive minutes of 0 activity counts, with allowance for 1-2 mins of counts
between 0 and 100. Days with less than 8h of wear data were not considered
acceptable, and participants with > 4 days of acceptable accelerometry data
were included. There is no clear consensus on the length of acceptable
monitoring periods
151
, and the monitoring period of the current study was similar
in duration to monitoring periods in other accelerometry studies
17, 68, 70
.
Participants with acceptable data wore the accelerometers for a mean + SD of
12.9 + 1.4 hours/day for 6.2 + 1.9 days, which resulted in a mean monitoring
period of 81.3 + 29.8 hours.
Data from all acceptable days were averaged, and included the following
variables: number of wear days, average number of minutes worn, total physical
activity represented by average counts per minute (cpm) on wear periods from all
valid days, minutes and percent of wear time spent engaged in sedentary
behavior and moderate to vigorous physical activity (MVPA). The intensity cut-
points applied to categorize MVPA were age-dependent thresholds based on the
Freedson pediatric equation
47, 48
, and sedentary behavior was defined as less
than 100 cpm
85
.
To assess dietary intake, participants completed 3-day diet records at
home after being trained by study staff, who were supervised by a Registered
Dietician. Total energy intake in kilocalories was the primary dietary variable of
interest. Nutrition data were analyzed using the Nutrition Data System for
Research (NDS-R version 5.0_35) developed by the University of Minnesota.
45

Statistical analysis. First, means were calculated and Student’s t-tests
were conducted to evaluate ethnic differences in demographic, physical activity,
and adiposity variables. Second, Pearson’s correlations were conducted to
describe the bivariate relationships between the adiposity and physical activity
variables. Next, multiple regression analyses were conducted to assess whether
physical activity (i.e., total PA, % time in MVPA, and % time spent sedentary)
was associated with adiposity (i.e., total fat mass, SAAT, VAT, and HFF). To test
whether the physical activity and adiposity relationships were different in
Hispanics and African Americans, an ethnicity*physical activity interaction term
was included in each model. If the interaction term was significant (p<0.05), then
the regression analyses were conducted in ethnicity-specific samples.
Sex, age, and Tanner stage were the a priori covariates included in all
statistical models. Total physical activity (cpm) was also included as a covariate
when physical activity intensity (i.e., % time in MVPA and sedentary) was the
predictor of interest. Additional body fat covariates included VAT when SAAT
was the dependent variable, SAAT or total fat mass when VAT was the
dependent variable, total fat mass or VAT when HFF was the dependent
variable, and soft lean and bone mineral content when fat mass was the
dependent variable. Throughout the regression analyses, total physical activity is
discussed in increments of 100 cpm.
Residual diagnostic analyses were completed to ensure that the
assumptions of regression were not violated. Dependent variables were log
46

transformed as needed to meet regression assumptions. Coefficient estimates
(B), standard errors (SE), and p-values will be reported. Results were interpreted
as the percent difference (i.e., (exp(B)-1)*100) when the dependent variables
were log transformed. Last, energy intake will be included in the models to
assess the influence of energy intake as a potential confounder of the
relationship between physical activity and adiposity. If the inclusion of the dietary
variable changes the coefficient estimate of the physical activity variable by >
10%, then the variable was considered a confounder
54
. Individual coefficient
plots, also known as partial regression plots, are presented to illustrate the
relationship between the predictor of interest, e.g., physical activity, and the
outcome, e.g., total fat mass. Analyses were conducted using SPSS for windows
(V16, SPSS Inc. Chicago, IL, USA) and SAS (v9.1, SAS Institute, Cary, NC). P <
0.05 denotes statistical significance.
Results
Participant characteristics by ethnicity are shown in Table 2.1. Participants
were 15.3 + 1.1 years old (mean + SD), 64% female, and 62% Hispanic.
Compared to Hispanic adolescents, African American adolescents had 7%
higher mean soft lean tissue (p = 0.02), 9% higher bone mineral content (p =
0.002), 70% higher SAAT (p < 0.001), and Hispanics had 120% more HFF (p =
0.003). Total physical activity did not differ between ethnic groups (p = 0.60), but
Hispanic adolescents spent significantly more time in MVPA than African
American adolescents (p = 0.01). Caloric intake did not significantly differ by
47

ethnicity (p = 0.60), but Hispanic adolescents reported consuming a significantly
higher percentage of calories from carbohydrate (p = 0.02) and African American
adolescents consumed a significantly higher percent of calories from fat (p =
0.001).
Correlations
To describe bivariate associations, Pearson’s correlations were conducted
between physical activity variables (i.e., total PA, % of time sedentary and % of
time spent in MVPA) and body fat distribution variables (i.e., total fat mass, soft
lean tissue, bone mineral tissue, SAAT, VAT, and HFF; Table 2.2). Higher total
physical activity was associated with lower total fat mass (r = -0.32, p = 0.03) and
SAAT (r = -0.46, p = 0.001) in Hispanic, but not African American, adolescents.
Additionally, higher total physical activity was positively associated with soft lean
tissue in both Hispanic (r = 0.30, p < 0.05) and African American adolescents (r =
0.61, p = 0.0002). Higher % of time spent in MVPA was also correlated with
lower fat mass (r = -0.46, p = 0.004) and lower SAAT (r = -0.60, p < 0.0001) in
Hispanics, and MVPA was correlated with higher soft lean tissue in African
American adolescents (r = 0.53, p = 0.0002). Percentage of time spent
sedentary was not correlated with any body fat variables in either ethnicity (all p >
0.10). Bone mineral content and HFF were not correlated with any physical
activity variables in either ethnicity (all p > 0.12).

48

Regression of total fat mass on physical activity
Because the ethnicity*MVPA interaction was not significant (p > 0.40), the
analysis was not stratified by ethnicity. In the total sample, regression analyses
revealed that spending 1% more time spent in MVPA was marginally associated
with 7.4% less total fat after adjusting for sex, age, ethnicity, maturation, soft lean
tissue, bone mineral content, and total physical activity (Table 2.3; Figure 2.1;
B = -0.057, p = 0.051). Adjusting for energy intake did not change the estimate
coefficient for percentage of time in MVPA by more than 10%. Neither total
physical activity nor percentage of time spent sedentary were significant
independent predictors of log total fat mass (all p > 0.17). When an
ethnicity*physical activity interaction term was included in the models, results
indicated that the relationship between physical activity variables (i.e., total
physical activity and % time sedentary) and total fat mass was not significantly
different by ethnicity (all p > 0.4).
Regression of SAAT on physical activity
Regression analyses revealed that there was a significant ethnicity*total
physical activity interaction (p = 0.01), and subsequent analyses were stratified
by ethnic group. Higher total physical activity (100 cpm) was significantly
associated with 9.2% lower SAAT in overweight Hispanic adolescents (Table 2.4,
Model 1; Figure 2.2; B = -0.096, p = 0.03) after adjusting for the standard
covariates and VAT, but total physical activity was not significantly associated
with log SAAT in African American adolescents (Table 2.4, Model 2.2; B = 0.10,
49

p = 0.08). MVPA was not associated with log SAAT (p > 0.10), and this
relationship was similar in Hispanic and African American adolescents
(ethnicity*MVPA interaction term p = 0.22). Percentage of time sedentary was
not associated with log SAAT (p = 0.90), and when an ethnicity*percent time
sedentary interaction term was included in the model, it was not significant (p =
0.28). Energy intake was not identified as a confounder.
Regression of VAT on physical activity
Because the ethnicity*physical activity interaction was not significant (p >
0.20), the analysis was not stratified by ethnicity. In the complete sample, higher
total physical activity (per 100 cpm) was significantly associated with 15% higher
VAT after adjusting for total fat mass and the a priori covariates (B = 0.14, p =
0.03). Similar results were found when adjusting for SAAT, instead of total fat
mass (B = 0.16, p = 0.01). Table 2.5 presents the regression of log VAT on total
physical activity adjusting for a priori covariates (Figure 2.3). Energy intake was
not identified as a confounder. Neither percentage of time spent in MVPA nor
percentage of time spent sedentary were associated with log VAT (all p > 0.50).
When an ethnicity*physical activity interaction term was included in the other
models, results indicated that the relationships between the physical activity
variables (i.e., % time in MVPA and % time sedentary) and VAT were not
significantly different by ethnicity (all p > 0.20).

50

Regression of HFF on physical activity
Based on a significant ethnicity by total physical activity interaction (p =
0.04), the sample was stratified by ethnicity. Higher total physical activity (per
100 cpm) was significantly associated with 22% lower HFF in African American
adolescents when adjusting for total fat mass (B = -0.25, p = 0.04) or when
adjusting for VAT (Table 2.6, Model 2.1; Figure 4; B = -0.25, p = 0.04). There
was no relationship between total physical activity and log HFF in Hispanic
adolescents when adjusting for either total fat mass or VAT (both p > 0.30).
Percentage of time spent sedentary and in MVPA were not associated with log
HFF after adjusting for either VAT (all p > 0.2) or total fat mass (all p >0.3).
Neither the ethnicity*MVPA nor ethnicity*sedentary interaction terms were
significant (all p > 0.40), indicating that there was no significant ethnic difference
in the relationship between log HFF and percentage of time spent in MVPA or
sedentary. Energy intake was not identified as a confounder.
Discussion
The primary purpose of this cross-sectional, observational study was to
examine whether physical activity is associated with total fat mass and fat stores,
i.e., SAAT, VAT, and HFF, differently among overweight Hispanic and African
American adolescents. In African American adolescents only, we found higher
levels of physical activity were associated with lower levels of ectopic hepatic fat,
and in Hispanic adolescents only, more time spent in total physical activity was
51

associated with less SAAT. In the total sample, we found that spending more
time in MVPA was marginally associated with having less total fat mass.
In the total sample of overweight Hispanic and African American
adolescents, the percentage of time spent in MVPA was marginally and
negatively associated with total fat mass after adjusting for total physical activity,
lean tissue mass, and a priori covariates. The decision of which accelerometry
cut-points should be used to categorize physical activity into intensities (e.g.,
MVPA) is somewhat subjective, especially in populations in which validation
studies have not been conducted, and this is particularly true in the current
sample of overweight adolescents. Because the mean age of the adolescents is
15 years, the cut-points developed in pediatric populations are plainly applicable.
It could also be argued, based on their mean weight of 95 kg and median Tanner
stage of 5, that the biomechanics of these adolescents are similar to that of an
adult, justifying the use of the adult-derived cut-points to categorize MVPA.
When this analysis was replicated using adult, instead of pediatric cut-points,
there was no significant relationship between MVPA and total fat mass (b = -
0.03, p = 0.16). This, coupled with the fact that total physical activity was not
related to fat mass, suggests that the marginal relationship observed between
MVPA and total fat mass may be an artifact of the pediatric cut-points used. To
avoid this doubt, future studies could conduct individual cut-points calibrations for
each participant to more accurately categorize MVPA.
52

An ethnic-specific relationship was observed between total physical
activity and SAAT. Hispanic adolescents who roughly 30% more active had 9%
lower SAAT, but this relationship was not observed in African American
adolescents. It is unclear whether SAAT, independent of visceral fat, contributes
to metabolic dysregulation. Studies removing SAAT via liposuction and
investigating subsequent changes in adult participants metabolic profiles have
had mixed results
49, 75
,
133
. Some investigators distinguish between deep and
superficial subcutaneous adipose tissue. Superficial SAAT is located in the
space immediately underlying the skin and overlying the abdominal wall fascia.
Deep SAAT, is located in the space immediately underlying the abdominal wall
fascia and overlying the parietal peritoneum
93
, and is more strongly related to
metabolic variables than superficial SAAT
73
. The current study did not
distinguish between deep and superficial SAAT. Future studies should examine
whether there is an ethnic-specific effect of physical activity on superficial and
deep SAAT among Hispanic and African American adolescents.
An unexpected positive association between physical activity and VAT
was observed among adolescents of both ethnicities, which could be attributable
to the effect of an unmeasured third variable that is correlated with both physical
activity and visceral fat. Possibilities include stress/cortisol
91, 104
, smoking
95
,
alcohol
95
, or air pollution
140
. Of these potential third variables, the most likely to
explain a positive relationship between physical activity and adiposity is air
pollution. A recent study has shown that exposure to air pollution, specifically
53

PM 2.5, increased VAT in obese mice
140
. The participants in the current study
were recruited from central Los Angeles, which has PM 2.5 concentrations higher
than both the state and national averages
79
. From 2004 – 2007, the American
Lung Association designated the Los Angeles area as the city most polluted by
year-round particle pollution, specifically PM 2.5
145
. It is plausible that
participants who were more physically active were more likely to be exposed to
air pollution, and as a result of that exposure be more likely to differentially
deposit visceral fat. Future studies could assess air quality to help examine the
independent effects of air pollution and physical activity on VAT.
An ethnic-specific relationship between physical activity and HFF was also
observed. Among overweight African American adolescents, but not their
Hispanic counterparts, those who had 30% higher levels of total physical activity
had 22% less hepatic fat. Few cross-sectional studies have examined the
relationship between physical activity and hepatic fat. The studies that do exist
are generally conducted in adults and do not include African American or
Hispanic participants. In Tiikkainen el al.
146
, obese Finnish women with previous
gestation diabetes who had lower amounts of liver fat engaged in more MVPA
than similar women with higher amounts of liver fat. Similar evidence by Hsieh et
al.
63
demonstrated that adult Japanese men who engaged in regular physical
activity (> 3 days/wk) had less fatty liver than their sedentary counterparts.
Perseghin et al
107
reported that white, healthy adults who reported higher levels
of habitual physical activity had lower volumes of hepatic fat. To our knowledge,
54

the current study is the first to examine the relationship between physical activity
and hepatic fat among African American or Hispanic adolescents, and it is the
first to compare and report ethnic differences.
The accumulation of ectopic hepatic fat may be due to increases in non-
esterified fatty acids, or free fatty acids (FFA), from endogenous and exogenous
sources. The exogenous source of FFA includes dietary fat, and the
endogenous sources include FFA from visceral adipose stores and de novo
lipogenesis, the metabolic pathway that converts excess dietary carbohydrates
into fat
31
. Possible causes of the ethnic difference in the relationship between
physical activity and hepatic fat may be ethnic-specific behaviors, such as dietary
intake. In the current study, African American adolescents reported consuming
significantly more calories from fat than Hispanic adolescents. Evidence from
rodent models suggests that exercise may be beneficial for reducing ectopic
hepatic fat deposition in the presence is a high-fat diet
21
, which would suggest a
greater response to exercise among African American adolescents in part due to
their consumption of a high fat diet.
The ethnic difference in the relationship between physical activity and
hepatic fat could also be due to ethnic differences in the underlying physiology in
storage of excess fat. One possibility is that African Americans uptake more
circulating FFA into skeletal muscle as a response to physical activity or exercise
than Hispanics, though to our knowledge this has not been tested. Few studies
have been conducted that investigate ethnic-specific physiological responses to
55

exercise. Chitwood et al.
22
studied the effect of individual bouts of exercise on
FFA in blacks and whites, and found no significant difference in serum FFA.
Hasson et al. (unpublished data, 2008) evaluated the metabolic response of a
single exercise bout on FFA, but the ethnic comparison was also between black
and white adults. To our knowledge, no studies have investigated ethnic
differences in metabolic responses to exercise in African American and Hispanic
adults or adolescents.
Though the total levels of physical activity did not differ between African
American and Hispanic adolescents, Hispanic adolescents spent significantly
more of their time in MVPA. However, the magnitude of this difference was
small, equating to a difference of less than 10 minutes spent in moderate activity.
The 2008 physical activity guidelines from the U.S. Department of Health &
Human Services (USDHHS)
154
recommend that adolescents engage in at least
60 minutes of at least moderate physical activity per day. Less than 3% of the
adolescents in the current study meet those recommendations, compared to less
than 8% of adolescents from a nationally representative sample
149
. The lack of
activity observed in this sample, while disheartening, is not surprising given the
overweight status of the sample. Results from this and other studies suggest
that increasing the physical activity levels could improve their body composition.
It should be acknowledged that the observed ethnic differences in the
relationship between physical activity and adiposity may not be due to differing
underlying physiological responses to physical activity. Those conclusions
56

cannot be made given the cross-sectional design of the current study. It is
possible that physical activity is lower among those with more adipose tissue
because of physiological mechanisms that may encourage inactivity. It could
also be that the ethnic differences in the association are due to an ethnic-specific
behavioral reaction to the measurement. In other words, if Hispanic, but not
African American, adolescents with higher SAAT volumes were more likely to be
less active during the measurement period as a reaction to the accelerometer,
then that could explain the association found between physical activity and SAAT
in Hispanics, but not African Americans. However, this is unlikely, because
research conducted with sealed pedometers and accelerometers suggests that
there is little reactivity, or a change in normal activity patterns when participants
know they are being monitored
24, 86, 157
.
Reactivity is especially an unlikely explanation for the ethnic difference
found in the association between physical activity and hepatic fat accumulation,
which found a significant relationship among African American adolescents, but
not Hispanic. It is unlikely that adolescents who had higher volumes of hepatic
fat were aware of their higher HFF and reacted by being less physically active.
Given the current study design, however, these questions cannot be answered.
Randomized controlled physical activity trials are needed to fully examine
whether increases in physical activity reduces HFF.
One of the strengths of current study is the use of objective measures of
physical activity and precise measures of total, visceral, hepatic, and
57

subcutaneous abdominal adiposity in adolescents. Few studies using objective
measures of physical activity, such as accelerometry, and more precise
measures of adiposity, such as DEXA to measure total body fat mass, have been
conducted
17, 42, 98
, and even fewer studies have used precise methods to
measure specific adipose depots, such as MRI to measure visceral and
subcutaneous abdominal adipose tissue
125
. To date, no studies have examined
the associations between physical activity measured by accelerometry and HFF,
especially in adolescents.
The small sample size and cross-sectional nature of this analysis are a
limitation of the study. The uniqueness of the sample used in this study should
also be noted. The participants in this sample were relatively sedentary,
overweight, urban, ethnic minority adolescents. As a result of these potential
limitations, the findings presented in this manuscript should be interpreted
cautiously. Larger, longitudinal studies would help to inform whether these
results could be replicated. Because of the sample size, the moderating effects
of sex within each of the ethnicities could not be examined in the current study.
Hispanic adolescent males have significantly higher HFF than their Hispanic
female and African American male and female counterparts (Fisher et al,
unpublished data). Future studies should examine whether physical activity
influences HFF differently in male and female Hispanic adolescents.
The results from this study may have several important public health
implications. First, these results suggest that ethnic-specific expectations of
58

intervention results may be helpful to reduce attrition. Frequently, investigators
culturally-tailor physical activity interventions to improve adherence and
compliance
117-119
. These findings suggest that, in addition to culturally tailoring
intervention content to make content more relevant to the participants, it may
also be important to explain that intervention results may be ethnic-specific
based on differing biological responses to physical activity.
Another important public health message that results from this work is that
an increase in total physical activity may be able to reduce fat stores of
overweight adolescents. Importantly, it may not necessarily be an increase in
moderate to vigorous physical activity or exercise that affects body composition;
an increase in total physical activity may improve metabolically important fat
stores, such as hepatic fat, specifically in sedentary, overweight African
American adolescents.
In conclusion, the current study suggests that Hispanic adolescents who
engage in more MVPA have less SAAT and African American adolescents who
engage in more total physical activity have less ectopic hepatic fat. The
relationship between physical activity and adipose stores, especially hepatic fat,
has important implications for metabolic health, and results suggest that the
ethnic specific associations between physical activity and adiposity may result in
reductions of different adipose stores in response to physical activity
interventions among African American and Hispanic youth.

59

Acknowledgements:
This work was supported by the National Institutes of Cancer (NCI),
NCI Centers for Transdisciplinary Research on Energetics and Cancer (TREC,
U54 CA 116848), NCI training grant (T32 CA 09492), and the National Institute of
Child Health and Human Development (RO1 HD/HL 33064). We would also like
to thank all of the staff at USC and USC-GCRC. Most importantly, we would like
to thank all of the participants and their families for making this study possible.

60

61

62

63

Figure 2.1. Individual coefficient plot of total fat mass regressed on percent of
time spent in moderate to vigorous physical activity (n = 76). Adjusted for sex,
age, ethnicity, Tanner stage, soft lean tissue, bone mineral content, total physical
activity. MVPA, moderate to vigorous physical activity. Total fat values were log
transformed. Dotted lines are 95% confidence intervals.

64
Figure 2.2. Individual coefficient plot of SAAT regressed on total physical activity in a sample of overweight Hispanic
and African American adolescents, by ethnicity. Adjusted for sex, age, Tanner stage, and VAT. SAAT, subcutaneous
abdominal adipose tissue; VAT, visceral adipose tissue; cpm, counts per minute. SAAT values were log transformed.
Dotted lines are 95% confidence intervals.

65

Figure 2.3. Individual coefficient plot of VAT regressed on total physical activity
in a sample of overweight Hispanic and African American adolescents (n = 59).
Adjusted for sex, age, ethnicity, Tanner stage, and SAAT. VAT, visceral adipose
tissue; cpm, counts per minute; SAAT, subcutaneous abdominal adipose tissue.
VAT values were log transformed. Dotted lines are 95% confidence intervals.

66

67
Figure 2.4. Individual coefficient plot of HFF regressed on total physical activity in a sample of overweight Hispanic and
African American adolescents (n = 59). Adjusted for sex, age, ethnicity, Tanner stage, and VAT. HFF, hepatic fat
fraction; VAT, visceral adipose tissue; cpm, counts per minute. HFF values were log transformed. Dotted lines are 95%
confidence intervals.

68

CHAPTER 3. STUDY 2: CROSS-SECTIONAL EXAMINATION OF SEX
DIFFERENCES IN PHYSICAL ACTIVITY AND ADIPOSITY AMONG
OVERWEIGHT HISPANIC ADOLESCENTS

Abstract
Purpose: The principal objective of this study was to examine whether there
was a sex difference in the relationship between physical activity and body fat
distribution in overweight Hispanic adolescents. It was hypothesized that the
relationship between physical activity and body fat distribution would be stronger
in overweight Hispanic boys, as compared to girls.
Methods: Participants were 60 Hispanic (30% males) overweight adolescents
(BMI > 85
th
percentile for age and gender). Total fat and lean mass were
assessed by DEXA. Hepatic fat fraction (HFF), visceral (VAT) and subcutaneous
abdominal adipose tissue (SAAT) volumes were assessed by multiple-slice MRI.
Physical activity was assessed by 7-day accelerometry. Total physical activity is
expressed as counts per minute (cpm). Linear regression was used to examine
whether physical activity was associated with adiposity measures (i.e., total fat,
VAT, SAAT, and HFF) after controlling for sex, age, sexual maturation, ethnicity,
and relevant adiposity covariates. To evaluate sex differences in the relationship
between physical activity and adiposity, sex*PA interactions were included in the
models. Variables were log transformed as needed to meet regression
assumptions. The change-in-estimate method was used to determine whether
energy intake confounded the activity and adiposity relationship.
69

Results: Overweight Hispanic boys were significantly more active than their girl
counterparts. Boys engaged in 47% more total physical activity, spent over twice
as much time in MVPA, and spent 2.5% less of their time sedentary than girls (all
p < 0.05). In overweight Hispanic boys and girls, higher levels of physical activity
by 100 cpm were associated with 9.2% lower SAAT (B = -0.00096, p = 0.03) and
19% higher VAT volumes (B = 0.0019, p = 0.001). Physical activity was not
associated with total fat mass or HFF (all p > 0.05). No sex differences were
detected in the activity and adiposity associations (all sex*PA p > 0.05). Energy
intake was not identified as a confounder in the physical activity and adiposity
associations.
Conclusions: Among overweight Hispanic adolescents, increasing total physical
activity may help to reduce subcutaneous abdominal fat. Sex differences in the
relationship between physical activity and adiposity were not observed between
overweight Hispanic boys and girls.
Introduction
Hispanics are the largest and fastest growing ethnic minority in the United
States
153
. Approximately 38% of Hispanic adolescents are overweight or obese,
and these rates are significantly higher than those in their non-Hispanic white
counterparts
100
. Overweight Hispanic adolescents are at increased risk for
developing type 2 diabetes, due to both their overweight status
30
and their
Hispanic ethnicity
51
. Given the growth rate of this population and the incidence
rates of obesity and related co-morbidities among the youth of this population, it
70

is a public health imperative to identify factors that affect fat mass among
Hispanic adolescents.
Lack of physical activity
61
and an abundance of sedentary
behaviors
38, 61, 71
have been identified as risk factors that are associated with
obesity. There are mixed reports, though, on how physical activity is associated
with fat mass in girls. A meta-analysis using doubly labeled water to measure
activity-related energy expenditure concluded that a higher level of physical
activity was related to a lower percent body fat in males, but not females
160
.
Some studies using accelerometry to assess physical activity have not detected
a sex difference in the association between physical activity and obesity
36, 121
,
while others have found that higher physical activity is more strongly related to
improvements in body composition in boys than it is in girls
17, 40, 98
. Using
accelerometry, we have previously shown a sex interaction in the relationship
between physical activity and obesity in Hispanic youth. Specifically, overweight
Hispanic boys were less physically activity than non-overweight boys, but
overweight Hispanic girls were not less active than normal weight Hispanic
girls
18
.
Few previous studies examining sex differences in the relationship
between physical activity and adiposity have examined specific fat depots, such
as visceral fat or subcutaneous abdominal adipose tissue, especially in a
pediatric population. To our knowledge, no studies have examined sex
differences in the relationship between physical activity and hepatic fat
71

deposition. As a result, the aims of the current study are two-fold: 1) to examine
whether there is a sex difference in the relationship between physical activity and
body fat distribution and 2) to determine whether accounting for energy intake will
help to strengthen the relationship between physical activity and total fat mass,
especially in girls.
Methods
Participants
Study participants were 60 Hispanic adolescents (ages 13 – 18 yrs) who
were recruited to participate in one of two projects: Strength and Nutritional
Outcomes for Hispanic Adolescents (SANO-LA) or Adolescent Circuit Training
for Los Angeles (ACT-LA). These projects were physical activity and/or nutrition
education randomized controlled trials that aimed to reduce obesity and risk
factors for type 2 diabetes. Adolescents included in the current cross-sectional
analysis had complete baseline data on demographic, physical activity, and body
composition variables. Only 41 participants had SAAT and VAT data and only 31
had HFF data due to a delay in finalizing the MRI methodology. Three boys were
excluded from the analysis when total fat mass was the dependent variable,
because they exceed the weight limit for the DEXA table. There was no
significant difference between those included and excluded from the analysis on
weight, BMI, fat mass, lean mass, age, pubertal stage, or percentage female (all
p>0.10). Study methods have previously been reported elsewhere
28
, therefore
only a brief overview of the methods will be described here. The protocol for
72

both studies was approved by the Institutional Review Board at the University of
Southern California.
Procedures
Screening visit. After an overnight fast, participants arrived at the General
Clinical Research Center. A medical exam was performed by a licensed
pediatric health care provider, and sexual maturation was determined
143
. An
oral glucose tolerance test was conducted to screen for diabetes. Participants in
both studies met the following inclusion criteria: 1) Hispanic ethnicity; 2) grades
9
th
thru 12
th
; 3) not currently taking medication or diagnosed with any syndrome
or disease that influences fat distribution or insulin action; 4) not diagnosed with
diabetes at screening or any major illness (e.g., cancer) since birth; 5) reported
not participating in a structured exercise, nutrition, or weight loss program in the
past six months; 6) age- and gender-specific BMI ≥ 85
th
percentile (SANO-LA)
and > 95
th
percentile (ACT-LA).
Anthropometric and Body Composition. Weight and height were
measured in triplicate using a beam medical scale and wall-mounted
stadiometer, respectively, and then averaged. BMI percentiles for age and
gender were determined using EpiInfo 2000, Version 1.1 (CDC, Atlanta, GA).
Whole body fat, lean tissue, and percent body fat were measured by dual energy
x-ray absorptiometry (DEXA) using a Hologic QDR 4500W (Hologic, Bedford,
MA).
73

Abdominal adipose tissue. Hepatic fat fractions (HFF), and subcutaneous
abdominal adipose tissue (SAAT) and visceral adipose tissue (VAT) volumes
were obtained by magnetic resonance imaging, using a Siemens Magnetom 1.5T
Symphony Maestro Class Syngo 2004A (Siemens AG, Erlangen, Germany) with
a Numaris/4. Patients were positioned supine, and 19 axial images of the
abdomen with a thickness of 10 mm each were taken. After image acquisition,
HFF was calculated using a modification of the Dixon 3-point technique
46
, and
visceral and subcutaneous abdominal tissue was segmented and calculated
using image analysis software (SliceOmatic Tomovision, Montreal Canada) at
Image Reading Center (New York City, New York).
Physical activity and dietary assessment. Physical activity was assessed
with accelerometers. Subjects were instructed to wear Actigraph accelerometers
(GT1M or 7164, Actigraph, LLC., Pensacola, FL) for seven days, except during
water-based activities or when sleeping
113, 158
. Data were reduced using an
adapted version of the SAS code used for the 2003-2004 National Health and
Nutrition Examination Survey available at
http://riskfactor.cancer.gov/tools/nhanes_pam. A correction factor was applied to
allow for comparison between the two Actigraph monitor models
23
.
Days with less than 8h of wear data were not considered acceptable, and
participants with > 4 days of acceptable accelerometry data were included in the
analysis. Participants with “valid” data wore the accelerometers for a mean + SD
of 13.1 + 1.4 hours/day for 6.3 + 2.1 days, which resulted in a mean monitoring
74

period of 83.8 + 31.8 hours. Data from all acceptable days were averaged, and
included the following variables: number of wear days, average number of
minutes worn, total physical activity represented by average counts per minute
(cpm) on wear periods from all valid days, minutes and percent of wear time
spent engaged in sedentary behavior and moderate to vigorous physical activity
(MVPA). The intensity cut-points applied to categorize MVPA were age-
dependent thresholds based on the Freedson pediatric equation
47, 48
, and
sedentary behavior was defined as less than 100 cpm
85
.
Diet was assessed by 3-day diet records. Participants were asked to
complete the 3-day diet records at home after being trained by study staff, who
were supervised by a Registered Dietician. The dietary variable of interest was
total energy intake and was calculated by averaging the total kilocalories from
each diet record. Nutrition data were analyzed using the Nutrition Data System
for Research (NDS-R version 5.0_35) developed by the University of Minnesota.
Statistical analysis. First, means were calculated and Student’s t-tests
were conducted to evaluate sex differences in demographic, physical activity,
and adiposity variables. Second, Pearson’s correlations were conducted to
describe the associations between the adiposity, physical activity, and dietary
variables. Next, multiple regression analyses were conducted to assess whether
physical activity (i.e., total PA, % time in MVPA, and % time spent sedentary) is
associated with body fat stores (i.e., total fat mass, SAAT, VAT, and HFF). To
test for a sex difference in the physical activity and adiposity associations, a
75

sex*physical activity interaction term was included in the models. If the
interaction term was significant, then the relationship between physical activity
and fat was estimated for boys and girls separately.
Sex, age, and Tanner stage were a priori covariates included in all linear
regression models. Total physical activity (cpm) was also included as a covariate
when physical activity intensity (i.e.,% time in MVPA and sedentary) was the
predictor of interest. Additional body composition covariates included VAT when
SAAT was the dependent variable, SAAT or total fat mass when VAT was the
dependent variable, total fat mass or VAT when HFF was the dependent
variable, and soft lean and bone mineral content when fat mass was the
dependent variable. Throughout the regression analyses, total physical activity is
discussed in increments of 100 cpm.
Residual diagnostic analyses were conducted to ensure that the
assumptions of linear regression were not violated. Dependent variables were
log transformed as needed to meet regression assumptions. Coefficient
estimates (B), standard errors (SE), and p-values are reported. Results were
interpreted as the percent difference (i.e., (exp(B)-1)*100) when the dependent
variables were log transformed. Individual coefficient plots were created when
the physical activity coefficient estimate was significant to illustrate the
relationship between the physical activity and adiposity variables after accounting
for all covariates. Finally, energy intake was included in the models to assess its
influence as a potential confounder of the relationship between physical activity
76

and adiposity. If the inclusion of energy intake changed the coefficient estimate
of the physical activity variable by > 10%, then the variable was considered a
confounder
54
. Individual coefficient plots, also known as partial regression plots,
are presented to illustrate the relationship between the predictor of interest, i.e.,
physical activity, and the outcome, i.e., adiposity variable. Analyses were
conducted using SPSS for windows (V16, SPSS Inc. Chicago, IL, USA) and SAS
(v9.1, SAS Institute, Cary, NC). P < 0.05 denotes statistical significance.
Results
Participant characteristics by sex are shown in Table 3.1. The participants
had a mean + SD age of 15.2 + 1.0 years and Tanner stage of 4.7 + 0.5. Boys
were taller (p = 0.005), had more soft lean tissue (p = 0.0001), more bone
mineral content (p = 0.005), and more hepatic fat (p = 0.04) than girls. Boys
were also significantly more active (cpm; p = 0.0001), spent more time in MVPA
(p = 0.04), and less time sedentary (p = 0.001) than girls. Additionally, boys
reported consuming more energy (p = 0.003).
Correlations
Pearson correlations were conducted separately for boys and girls (Table
3.2). In boys, total physical activity was negatively correlated to total fat mass (r
= -0.60, p = 0.01) and SAAT (r = -0.66, p = 0.003). Additionally in boys,
percentage of time spent in MVPA was also significantly negatively correlated
with total fat mass (r = -0.74, p = 0.001) and SAAT (r = -0.77, p = 0.002). In girls,
energy intake is negatively correlated with total fat mass (r = -0.34, p = 0.0.3) and
77

SAAT (r = -0.47, p = 0.03) and marginally correlated with VAT (r = -0.36, p<0.10).
HFF was not significantly correlated with either energy intake or any physical
activity measures in boys or girls.
Regression of total fat mass on physical activity
In the models in which log total fat mass was the dependent variable, the
sex*physical activity interaction terms were not significant, which indicates that
there were no significant sex differences in the relationship between physical
activity (e.g., total physical activity, time spent in MVPA or sedentary) and total
fat mass (for all sex*physical activity interaction terms p > 0.30). As a result, all
subsequent analyses in which total fat mass was the dependent variable were
conducted on the sample containing both females and males. After adjusting for
the a priori covariates, log total fat mass was not associated with total physical
activity (Table 3.3; B = -0.02, p = 0.59), percent of time spent in MVPA (B = -
0.06, p = 0.15), or percent of time sedentary (B = -0.37, p = 0.56). Energy intake
was not identified as a confounder in any models of physical activity and log total
fat mass.
Regression of SAAT on physical activity
In models in which log SAAT was the dependent variable, the
sex*physical activity interaction terms were not significant (p > 0.48), and
subsequent analyses were conducted on the total sample. Higher levels of total
physical activity (per 100 cpm) were significantly associated with a 9.2%
decrease in SAAT (Figure 3.1; Table 3.4; B = -0.096, p = 0.03) after adjusting for
78

a priori covariates. Neither percent of time spent in MVPA (B = -0.09, p = 0.10)
nor percent of time sedentary (B = 0.01, p = 0.32) were associated with log
SAAT. Energy intake was not identified as a confounder in any models
containing physical activity and log SAAT.
Regression of VAT on physical activity
In models in which log VAT was the dependent variable, the sex*physical
activity interaction terms were not significant (p > 0.44), and subsequent
analyses were conducted on the total sample. Higher total physical activity (per
100 cpm) was associated with a 19% higher VAT volumes (Figure 3.2; Table 3.5,
B = 0.19, p = 0.01) after adjusting for a priori covariates. Energy intake was not
identified as a confounder in any of the models regressing VAT on physical
activity. Neither MVPA nor sedentary behavior was significantly associated with
log VAT (all p > 0.69).
Regression of HFF on physical activity
In models in which log HFF was the dependent variable, the sex*physical
activity interaction terms were not significant (p > 0.44), and subsequent
analyses were conducted on the total sample. Total physical activity was not
associated with HFF in the total sample (Table 3.6; B = -0.00066, p = 0.63). In
addition, neither MVPA nor sedentary behavior was associated with HFF in boys
or girls (all p > 0.40). Energy intake was not identified as a confounder in any of
the models regressing HFF on physical activity.

79

Discussion
The primary objective of this study was to examine sex differences in the
relationship between physical activity and measures of body fat distribution,
including total fat mass, SAAT, VAT, and HFF. The secondary objective was to
examine the effect of accounting for energy intake on the relationship between
physical activity and body composition. The principal findings of this study were
that total physical activity was negatively associated with SAAT and positively
associated with VAT in the total sample of overweight Hispanic adolescents after
adjusting for Tanner stage, age, and the complementary body composition
variable (VAT and SAAT, respectively). We did not observe stronger
associations between physical activity and adiposity in boys than in girls, as
hypothesized based on previous studies. Physical activity was not associated
with total fat mass or HFF. Accounting for energy intake did not influence any of
the activity and fat relationships.
Interestingly, neither total physical activity nor time spent in MVPA was
significantly related to total fat mass in the current study despite the use of
objective measurements of physical activity (i.e., accelerometry) and precise
measurements of fat mass (i.e., DEXA). Other studies using similar objective
and precise measures of physical activity and adiposity have reported a
significant and inverse relationship. In a cross-sectional study of normal and
overweight Hispanic youth, Butte et al.
17
observed that higher levels of physical
activity were significantly, though weakly, associated with lower percent body fat
80

with a partial correlation of -0.13 after adjusting for age and gender. In another
large, cross-sectional study of 12 year old children (unspecified ethnic
composition), MVPA was negatively associated with fat mass
98
. The difference in
findings between the current study and the cited studies may be the overweight
status of the participants. The current study, which only included overweight
adolescents, had a restricted range of adiposity in comparison to the cited
studies, which included both normal weight and overweight youth. The variations
in fat mass in overweight youth may be more heavily influenced by factors other
than physical activity, such as genetics or caloric intake. Future studies
examining the relationship between physical activity and total fat mass should
test whether the relationship is different in normal and overweight adolescents.
Few studies have investigated the relationship between physical activity
and specific fat depots, such as SAAT and VAT. In the current study, overweight
Hispanic adolescents who were more physically active by 100 cpm had 8.6%
less SAAT, after adjusting for total fat mass. Saelens et al.
125
reported that
among largely prepubescent 8 year olds who were at risk for obesity (>75
BMI%ile and at least one overweight parent) physical activity was not associated
with SAAT after accounting for total fat mass. Similarly, a study of adolescent
girls did not detect a relationship between physical activity and SAAT either
cross-sectionally or longitudinally
95
. The differing participant characteristics (e.g.,
weight status, age, and ethnicity) make it difficult to compare the above studies.
81

An unanticipated positive association between total physical activity and
VAT was observed. This differs from previous studies, which either found a
negative relationship between physical activity and VAT
125
or failed to find a
relationship
95
. A possible explanation for the positive relationship between
physical activity and VAT is that it is a spurious finding. It is possible that another
factor, such as stress
91, 104
, smoking
95
, or air pollution
140
, is confounding the
relationship between activity and VAT. A recent study of mice with diet-induced
obesity found that exposure to air pollution caused preferential visceral fat
deposition
140
. The overweight adolescents in this study were recruited from Los
Angeles, CA, where air pollution levels are among the highest in the
country
79
.Overweight adolescents who were more physically active may have
been more likely to spend time outdoors, and thus, be more likely to be exposed
to air pollution. Future studies should take into account possible effects of
factors such as stress, smoking, and air pollution when examining predictors of
VAT.
A notable finding from this study is that, while boys are significantly more
active than girls, the majority of neither boys nor girls met the current physical
activity guidelines. The 2008 physical activity guidelines recommend that
adolescents engage in at least 60 mins of MVPA/day. Eight percent of
adolescents from a nationally representative sample met the 2008
recommendations
149
, but only 5.5% of boys (1/18) and 0% of girls (0/42) from the
this study were engaged in MVPA for more than 60 mins/day. This dearth of
82

physical activity was not surprising given the overweight status of the sample.
Even though the overweight Hispanic boys from the current study did not meet
the national recommendations for being active, they were markedly more active
than the overweight Hispanic girls. The boys were 47% more active as
measured by total physical activity, spent over twice as much time in MVPA, and
spent 2.5% less of their time sedentary than girls. It is well established that girls
are less physically active than boys. As a result, numerous intervention studies
have aimed to increase physical activity, specifically in girls, with the intent of
improving body composition
5, 69, 99, 103, 105
, though not all have been successful
69,
99, 105
. Future intervention studies should continue to attempt to identify ways to
increase physical activity and improve adiposity in adolescent girls.
Several studies, including a meta-analysis of studies using doubly labeled
water to measure energy expenditure, have noted a stronger relationship
between physical activity and obesity in boys than in girls
40, 98, 160
. In contrast,
the current study did not detect sex differences in the relationships between
physical activity and adiposity. Several other studies in youth have failed to
observe a sex difference in physical activity and adiposity
36, 121
. Ekelund et al.
36

compared 18 (8 males,10 females) overweight and matched, normal-weight
adolescents, reported that obese adolescents were less physically active than
their normal-weight counterparts, and did not detect a significant sex difference.
Another study conducted by Treuth et al.
148
found that physical activity is
associated to fat mass in girls, and not in boys. These inconsistent findings
83

warrant further investigations into the sex-specific relationships between physical
activity and adiposity.
Accounting for energy intake did not help to elucidate the relationship
between physical activity and any of the adiposity measures. This may be
partially explained by the negative correlation between energy intake and three
measures of adiposity (i.e., total fat, SAAT, and VAT) that was observed in girls.
This negative association suggests that the 3-day diet records may not have
assessed the habitual energy intake of the girls. Girls with more fat mass, SAAT,
and VAT reported consuming fewer calories. It is possible that heavier girls
within this overweight sample were restricting or underreporting their dietary
intake, as has been seen in previously
156
. The self-report nature inherent in the
measurement of free-living dietary intake makes assessment difficult, especially
in overweight populations. Methods, such as digital recording of food, are being
developed to make dietary measurement more valid
84
, but these methods are still
subject to underreporting. Future studies, especially those conducted in
overweight populations, would benefit from objective measurements of dietary
intake.
One of the strengths of the current study is the use of objectively
measured physical activity, such as accelerometry; a meta-analysis has shown
that the magnitude of the relationship between activity and adiposity is
strengthened when physical activity is measured objectively. An additional
strength of the study is the precise measures of body fat distribution, such as
84

DEXA to measure total and lean tissue mass and MRI to measure SAAT and
VAT. Despite these strengths, there are also limitations that should be noted.
These include the small sample size and cross-sectional nature of the analysis,
which precludes any causality conclusions.
An additional limitation, and a potential explanation for why a sex
difference was not observed, may be the lackluster participant compliance in
wearing the accelerometer for 8 h/d for 4d. The lack of compliance reduced the
sample of boys to less than 20. Almost 40% (14/36) of the male participants did
not wear the accelerometer for a minimum of 4d for 8h/d. It does not appear that
the non-compliance introduced a systematic bias in boys, as boys who were
compliant were not different from the boys who were not in age or any body fat
variables (all p > 0.25). The compliance issue was also not different between
boys and girls. The percentage of boys who were non-compliant was not
significantly different from the percentage of girls who were non-compliant (39%
and 37%, respectively; p = 0.83). Futures studies should consider increasing the
incentive offered to participants for providing complete accelerometry data.
Sirard et al. showed that 95% compliance rates, defined as wearing the
accelerometer for 10 hours for at least four day, are possible among urban
adolescents by offering an additional $5 gift card for every day the accelerometer
was worn for at least 10 hours
131
.
In conclusion, in this sample of overweight Hispanic adolescents, higher
physical activity by 100 cpm, or about 30%, was associated with 8.6% less SAAT
85

and 19% more VAT. No sex difference was detected in the relationship between
physical activity and adiposity measures, and energy intake did not influence the
relationship. More research is warranted to investigate the relationship between
activity and adiposity in overweight Hispanic adolescents, who are particularly
vulnerable to the deleterious co-morbidities associated with adiposity.
Acknowledgements:
This work was supported by the National Institutes of Cancer (NCI),
NCI Centers for Transdisciplinary Research on Energetics and Cancer (TREC,
U54 CA 116848), NCI training grant (T32 CA 09492), and the National Institute of
Child Health and Human Development (RO1 HD/HL 33064). We would also like
to thank all of the staff at USC and USC-GCRC. Most importantly, we would like
to thank all of the participants and their families for making this study possible.

86

87

88

89

Figure 3.1. Individual coefficient plot of SAAT regressed on total physical activity
in a sample of overweight Hispanic adolescents (n=41). Adjusted for sex, age,
Tanner stage, and VAT. SAAT, subcutaneous abdominal adipose tissue; VAT,
visceral adipose tissue; cpm, counts per minute. SAAT values were log
transformed. Dotted lines are 95% confidence intervals.

90

Figure 3.2. Individual coefficient plot of VAT regressed on total physical activity
in a sample of overweight Hispanic adolescents (n=41). Adjusted for sex, age,
Tanner stage, and VAT. VAT, visceral adipose tissue; SAAT, subcutaneous
abdominal adipose tissue; cpm, counts per minute. VAT values were log
transformed. Dotted lines are 95% confidence intervals.

91

CHAPTER 4. STUDY 3: SHORT-TERM CHANGES IN PHYSICAL ACTIVITY
AND ADIPOSITY IN OVERWEIGHT HISPANIC ADOLESCENTS

Abstract
Purpose: Objectives of this study were to examine 1) whether changes in
physical activity (PA) differ by treatment group during a randomized intervention,
2) whether changes in PA (total PA [counts/minute, cpm] and time spent
sedentary and in moderate to vigorous PA [MVPA]) are associated with changes
in adiposity and 3) whether energy intake confounds the relationship between
changes in PA and changes in adiposity.
Methods: Analysis included 38 overweight (BMI > 85
th
%ile) Hispanic
adolescents with complete pre- and post-test data on relevant variables after
participating in a 16-week intervention. Participants were randomized to one of
three groups: nutrition education only, nutrition and strength training, or control.
Body fat distribution by DEXA and MRI, 7-day physical activity by accelerometry,
and dietary intake by 3-day diet records were assessed pre- and post-
intervention.
Results: There were no intervention effects on total PA or MVPA (all p>0.1); in
subsequent analyses the treatment groups were combined and analyses
controlled for group. Within individuals, the mean increase of PA (n = 19) and
mean decrease of PA (n = 19) was approximately 105 cpm. A 100 cpm increase
in total PA was associated with a decrease of 1.3 kg fat mass and 0.8% body fat
after adjusting for pre-test adiposity, PA, age, sex, and treatment (p < 0.05).
92

Controlling for energy intake strengthened the relationships between total PA and
fat mass and percent body fat by up to 25%. Changes in MVPA were not related
to changes in adiposity after controlling for total PA (p>0.05).
Conclusion: Increasing total PA by 28% (100 cpm) was associated with a
decrease of 1.4 kg of fat mass and 1% body fat over 16 weeks in overweight
Hispanic adolescents independent of intervention group assignment.
Introduction
National data from 2003-2006 indicate that 34% of adolescents in the
United States are at risk for overweight and 17% are overweight
100
. The
prevalence rates among Mexican American adolescents are even higher than the
national average, with 39% at risk for overweight and of those 21% are
overweight
100
. These high rates of overweight among adolescents are an
important public health concern, because being overweight in adolescence can
contribute to the development of chronic diseases, such as type 2 diabetes,
cardiovascular disease, and obesity-related cancers
88, 101
.
Undoubtedly, there are many factors contributing to the alarming obesity
rate, and one of these modifiable risk factors is a lack of physical activity
74, 112
.
Physical activity declines dramatically across age groups
149
; it decreases as
children become adolescents
52, 70, 97
and as adolescents become young adults
42
.
Only 8% of adolescents and 5% of adults meet the CDC physical activity
recommendations (60 and 30 mins/day of at least moderate-intensity activity for
adolescents and adults, respectively
149
.
93

While many studies in children and adolescents have found that increased
physical activity is associated with decreased adiposity
94
, there have been mixed
reports
136
. These discrepancies in results may be due to subjective methods
used to assess physical activity and adiposity, such as self-report
questionnaires
115
and body mass index, respectively. The use of objective and
precise measures of physical activity and adiposity, such as accelerometry and
DEXA
50, 115, 132, 159
, can reduce measurement error, thereby increasing the ability
to detect effects and elucidate the relationship between physical activity and
adiposity
122
. Another way to increase the ability to detect an effect is to control
for potential confounders, such as energy intake. Several studies have stated
the importance of exploring the influence of energy intake when examining the
relationship between activity and adiposity
42, 98, 138
.
Recent cross-sectional studies have shown that objectively measured
physical activity is associated with adiposity in adolescents
16, 81
, but to date, no
study has examined the short-term effects of changes in objectively measured
physical activity on changes in adiposity, particularly in Hispanic youth. The
current study examines 16-week changes in physical activity among overweight
Hispanic adolescents enrolled in a nutrition and strength training type 2 diabetes
prevention intervention and whether changes in habitual physical activity are
related to changes in adiposity. Although the intervention did not specifically
target daily physical activity (only dietary behaviors were targeted outside of the
94

intervention), previous research has suggested that health behaviors during
adolescence, including diet and physical activity, may cluster
82
.
Therefore, the specific objectives of the current research were to 1)
investigate whether changes in physical activity differ by treatment group during
a 16-week intervention among Hispanic adolescents, 2) investigate whether
changes in objectively measured physical activity (i.e., total physical activity,
percent of time spent sedentary, and percent time in moderate to vigorous
physical activity [MVPA]) are associated with changes in adiposity measures
(i.e., total and percentage fat mass, subcutaneous abdominal [SAAT] and
visceral fat [VAT]) and 3) examine the influence of energy intake on the
relationship between changes in physical activity and changes in adiposity.
Methods
Participants
Study participants consisted of a sub-group of 54 adolescents with
complete data who participated in a randomized nutrition and strength training
type 2 diabetes prevention intervention. Except for the accelerometry
methodology that will be described in detail below, a complete description of the
study methods have been reported elsewhere
28
, so only a brief overview of the
methods will be given here. Participants were 38 adolescents (19 girls, 19 boys)
who had complete data for all relevant measures at pre- and post-intervention
(10 in the control group, 20 in the nutrition only group, 8 in the nutrition + strength
training group); only 30 participants had SAAT and VAT data due to a delay in
95

finalizing the MRI methodology. Participants included in the analyses were not
significantly different than those excluded (sex, age, Tanner stage, BMI%ile,
weight, total fat or lean tissue mass, all p > 0.10). Informed written parental
consent and child assent were obtained prior to testing. The Institutional Review
Board of the University of Southern California approved the study.
Procedures
Screening visit. Participants arrived at the General Clinical Research
Center after an overnight fast. A licensed pediatric health care provider
conducted a medical history exam and determined sexual maturation
143
. To
screen for diabetes, an oral glucose tolerance test was conducted. Participants
who met the following criteria were invited back for further testing: 1) age- and
gender-specific BMI ≥ 85
th
percentile; 2) Hispanic ethnicity and grades 9
th
thru
12
th
;

3) not currently taking medication or diagnosed with any syndrome or
disease that influences fat distribution or insulin action; 4) not diagnosed with
diabetes at screening or any major illness (e.g., cancer) since birth; 5) reported
not participating in a structured exercise, nutrition, or weight loss program in the
past six months.
Intervention. Participants were randomized to one of three intervention
groups: nutrition education only, nutrition education and strength training, or
control group. Participants randomized to the control group (C; n = 10 in current
analysis) received no intervention during the 16-week period. After the end of
the study, control participants were offered an abbreviated intervention for one
96

month, consisting of biweekly nutrition and strength training classes. Participants
randomly assigned to the nutrition education only group (NUT; n = 20 in current
analysis) attended one 90-minute dietary intervention class per week that
specifically aimed to modify carbohydrate intake by decreasing added sugar and
increasing dietary fiber intake. In addition to the nutrition education class
described above, participants assigned to the nutrition education and strength-
training group (NUT + ST; n = 8 in current analysis) also attended two non-
consecutive, 60-minute strength training sessions per week.
Anthropometry and Body Composition. Weight and height were measured
in triplicate using a beam medical scale and wall-mounted stadiometer,
respectively, and then averaged. BMI percentiles for age and gender were
determined using EpiInfo 2000, Version 1.1 (CDC, Atlanta, GA). Whole body fat,
lean tissue, and percent body fat were measured by dual energy x-ray
absorptiometry (DEXA) using a Hologic QDR 4500W (Hologic, Bedford, MA).
Abdominal adipose tissue. Subcutaneous abdominal adipose tissue
(SAAT) and visceral adipose tissue (VAT) volumes were obtained by magnetic
resonance imaging, using a Siemens Magnetom 1.5T Symphony Maestro Class
Syngo 2004A (Siemens AG, Erlangen, Germany) with a Numaris/4. Patients
were positioned supine, and 19 axial images of the abdomen with a thickness of
10 mm each were taken. After image acquisition, visceral and subcutaneous
abdominal tissue was segmented and calculated using image analysis software
97

(SliceOmatic Tomovision, Montreal Canada) at Image Reading Center (New
York City, New York).
Energy intake and physical activity. To assess energy intake, participants
completed 3-day diet records at home after being trained by study staff, who
were supervised by a Registered Dietician. Staff clarified records when they
were collected. Nutrition data were analyzed using the Nutrition Data System for
Research (NDS-R version 5.0_35) developed by the University of Minnesota.
To assess physical activity (PA), subjects were instructed to wear Actigraph
accelerometers (GT1M or 7164, Actigraph, LLC., Pensacola, FL) for seven days,
except during water-based activities or when sleeping
113, 158
. Accelerometers
were set to monitor activity in 15-second epochs, which were collapsed to 60-
second epochs during analysis. Data were reduced using an adapted version of
the SAS code used for the 2003-2004 National Health and Nutrition Examination
Survey available at http://riskfactor.cancer.gov/tools/nhanes_pam. A correction
factor was applied to allow for comparison between the two Actigraph monitor
models
23
.
The amount of time the participant wore the device was determined by
subtracting nonwear time from 24h. Nonwear time was defined by an interval >
60 consecutive minutes of 0 activity counts, with allowance for 1-2 mins of counts
between 0 and 100. Days with less than 6h of wear data were not considered
acceptable, and participants with > 2 days of acceptable accelerometry data at
pre- and post-testing were included. There is no clear consensus on the length
98

of acceptable monitoring periods
151
, and the monitoring period of the current
study was similar in duration to monitoring periods in other accelerometry
studies
17, 68, 70
. At pre-test, participants with “valid” data wore the accelerometers
for a mean + SD of 12.6 + 1.3 hours/day for 6.2 + 2.3 days, which resulted in a
mean monitoring period of 78.1 hours. At post-test, the participants with “valid”
data wore the accelerometers for 12.4 + 1.4 hours/day for 5.6 + 2.7 days, which
resulted in a mean monitoring period of 69.7 hours. Statistical analyses were
repeated in a subsample (n = 36) of participants with > 3 days of accelerometry
data, and similar results were obtained.
Data from all acceptable days were averaged, and included the following
variables: number of wear days, average number of minutes worn, total physical
activity represented by average counts per minute (cpm) on wear periods from all
valid days, percent of wear time spent in MVPA. The intensity cut-points applied
to categorize MVPA were those used for adults and older adolescents in
NHANES ( > 2020 cpm
149
), because the current sample of adolescents had an
average weight of 94 kg and median Tanner stage of 5, suggesting their
biomechanics may be closer to those of adults than children. To ensure that
results were not an artifact of the MVPA cut-point used, the analyses were
replicated using the age-dependent MVPA cut-points based on the Freedson
pediatric equation
47, 48
.
Statistical Analysis. The effect of the intervention on physical activity was
tested using Analysis of Covariance (ANCOVA) on the post-pre change score for
99

total physical activity and MVPA, adjusting for age, sex, and pre-test physical
activity values. Paired sample t-tests were conducted to determine if there were
mean differences between pre- and post-test in the anthropometric, adiposity,
PA, or dietary measures. To describe the bivariate relationships between the
variables, Pearson correlations were conducted.
Multiple regression analyses were conducted to assess whether changes
in physical activity (e.g., total physical activity and MVPA) were associated with
changes in adiposity (e.g., DEXA data) after controlling for covariates. The
following standard covariates were included in all models: sex, age, pre-test PA,
pre-test adiposity, and intervention group. Additional covariates included pre-
and post-test DEXA lean mass when change in fat mass was the dependent
variable. To ensure that the relationships between changes in physical activity
and adiposity were not biased by differing measurement times at pre- and post-
test, the total hours of measurement at pre- and post-test were also included as
covariates.
Residual diagnostic analyses were completed to ensure that the
assumptions of regression were not violated. Further diagnostic analyses,
specifically tolerance and variance inflation factor, were also evaluated to identify
collinear predictors. For the regression models, coefficient estimates (B),
standard errors, and p-values are reported. Throughout the regression analyses,
total physical activity is discussed in increments of 100 cpm to ease the
interpretation of the coefficient estimates. The value of the coefficient estimate is
100

the amount of change in the dependent variable that is associated with a unit
change in the physical activity independent variable, e.g., MVPA (1%) or total
physical activity (100 cpm).
To assess the influence of energy intake on the relationship between
change in physical activity and change in adiposity, energy intake was included
in the regression models. If including energy intake changed the coefficient
estimate of the physical activity variable by > 10%, then it was considered a
confounder
54, 65
. Individual coefficient plots, also known as partial regression
plots, are presented to illustrate the relationship between the predictor of interest,
i.e., physical activity, and the outcome, i.e., adiposity variable. Analyses were
conducted using SPSS for windows (V16, SPSS Inc. Chicago, IL, USA) and SAS
(v9.1, SAS Institute, Cary, NC). P < 0.05 denotes statistical significance.
Results
Table 4.1 shows the participant characteristics at pre- and post-test. The
participants consisted of 19 boys and 19 girls; 92% were Tanner stage 4 or 5.
The participants were an average of 15 years, 95 kg, and 97
th
BMI %ile for age
and sex at pre-test. There were statistically significant decreases between pre-
and post-test in SAAT (p = 0.03) and VAT (p = 0.005).
There were no significant differences among the intervention groups in
mean change of total physical activity after 16 weeks [Figure 4.1 (left); mean +
SE; C = -5.6 + 42.3cpm, NUT = 23.9 + 29.8cpm , NUT+ST = 46.1 + 47.3cpm;
p>0.10] or percent time spent in MVPA [Figure 4.1 (right); C = -0.1 + 1.1%, NUT
101

= 1.5 + 0.7%, NUT+ST = 0.5 + 1.1%; all p>0.10]. As a result, participants were
grouped together and intervention group was included as a covariate in
subsequent regression models.
Although there were no overall intervention effects on PA, there were
noteworthy individual changes in total physical activity (maximum decrease was
-317.8 cpm vs. maximum increase of 339.3 cpm) and percent of time spent in
MVPA (maximum decrease was -8.9% vs. maximum increase 6.6%). Figure 4.2
shows the individual values of change (post-pre) in total physical activity for each
participant; 22 participants increased and 16 decreased their total PA. Similarly,
23 participants increased (mean + SD, 2.8 + 2.1%) and 15 decreased (-2.2 + 2.5
%) their percent of time spent in MVPA. Table 4.2 shows the Pearson
correlations of change in adiposity variables, change in physical activity
variables, change in energy intake and demographic variables. Change in total
physical activity and % time in MVPA were not correlated to change in adiposity.
Age was the only variable significantly correlated with change in adiposity; older
participants had greater reductions in fat mass and percent body fat (p < 0.05).
Changes in total physical activity and adiposity
Regression analyses revealed that an increase in total physical activity
(cpm) was significantly associated with a decrease in total fat mass after
controlling for the standard covariates (Table 4.3; B = -1.3, p = 0.02). Including
energy intake in the model increased the coefficient estimate 8% (Table 4.3; B =
102

-1.4, p = 0.02), thus it cannot be concluded that energy intake confounded the
relationship (> 10% change in estimate).
An increase in total physical activity (cpm) was significantly associated
with a decrease in percent body fat (Table 4.3; B = -0.8, p = 0.03), and energy
intake confounded the relationship between total physical activity and percent
body fat, increasing the coefficient estimate by 25% (Table 4.3; B = -1.0, p =
0.01). Figure 4.3 is a scatter plot of the change in total physical activity by the
predicted values of change in total fat mass after controlling for the standard
covariates, and Figure 4.4 is an individual coefficient plot of change in total fat
regressed on change in total physical activity after removing the influence of the
standard covariates. Both show a negative relationship between change in
physical activity and predicted change in fat mass. Changes in total physical
activity were not significantly associated with changes in SAAT or VAT,
regardless of whether energy intake was included in the models or not (all p >
0.05).
Changes in percent time spent in components of physical activity and adiposity
After controlling for standard covariates, regression analyses revealed that
an increase in percent time spent in MVPA was marginally associated with a
decrease in percent body fat (B = -0.26, p = 0.10). When change in energy
intake was included in the model, this relationship became significant (Table 4.4;
B = -0.33, p = 0.04). The relationships between change in percent time in MVPA
and percent body fat were no longer significant after accounting for change in
103

total PA, regardless of whether change in energy intake was in the model (B = -
0.29, p = 0.43) or not (Table 4.4; B = 0.17, p = 0.64).
Increases in percent of time spent in MVPA were associated with
decreases in fat mass both before accounting for change in energy intake (B = -
0.49, p = 0.04) and after (B = -0.51, p = 0.03), but not after controlling for change
in total physical activity regardless of whether energy intake was excluded (B = -
0.02, p = 0.96) or included in the model (B = -0.15, p = 0.75). Increases in MVPA
were significantly associated with decreases in percent body fat (B = -0.37, p =
0.09) and fat mass (B = -0.81, p = 0.009) after accounting for energy intake, but
not after adjusting for total physical activity (all p > 0.60). Parallel results were
obtained when using MVPA cut-points generated by the Freedson pediatric
equation
47, 48
. Changes in percent time spent in MVPA were not significantly
associated with SAAT or VAT (all p > 0.05).
Including the total hours of measurement at pre- and post-test as
covariates in the models did not impact the results and were not included in the
final models. Changes in sedentary behavior were not significantly associated
with changes in total fat mass, percent body fat, SAAT, or VAT (all p > 0.05).
Discussion
A major objective of this study was to examine how changes in physical
activity over a 16-week period are associated with changes in adiposity in
overweight Hispanic adolescents. The primary findings are that a short-term
increase in objectively measured total physical activity is significantly associated
104

with a decrease in both total fat mass and percent body fat. Specifically, an
increase of 28% of total physical activity, or 100 cpm, was associated with a
decrease of 1.4 kg fat mass and 1% body fat after controlling for intervention
group assignment, energy intake, and a priori covariates. To translate the
accelerometry unit of counts per minute (cpm) into more physiologically relevant
terms, prediction equations based on previous observations in normal weight
adolescents
42
were used to estimate that an increase of 100 cpm is broadly
similar to an increase of 250 kcal of energy expenditure.
A secondary objective was to examine energy intake as a confounder of
the relationship between changes in physical activity and adiposity. Recent
studies in adolescents state that adjusting for energy intake may strengthen the
observed relationships between objectively measured physical activity and
adiposity
42, 138
. The current study supports these assertions, finding that
accounting for changes in energy intake strengthened the relationship between
changes in physical activity and adiposity up to 25%. Change in energy intake
was identified as a confounder in the relationship between total physical activity
and percent body fat, because adjusting for change in energy intake increased
the coefficient estimate by 25%, which was greater than the 10% change in the
coefficient needed to identify it as a confounder. Change in energy intake was
not identified as a confounder in the relationship between total physical activity
and total fat mass, because adjusting for change in energy intake only change
the coefficient estimate by 8%. The 10% change in coefficient to define a
105

confounder is a rather arbitrary rule of thumb and is not based on a statistical
test
65
. As such, future studies attempting to quantify the magnitude of the
relationship between changes in physical activity and adiposity should consider
whether to account for energy intake.
Change in percent of time spent in MVPA was not associated with
changes in percent body fat or with total fat mass, independent of total physical
activity. The correlation of change in total physical activity and change in MVPA
was very high (r = 0.93, p<0.001), and it may be that change in total activity was
driving the relationship between MVPA and adiposity. These findings suggest
that if overweight Hispanic adolescents increase their total physical activity, then
they may improve their body composition. Because increases in MVPA may be
difficult for and not well tolerated by this population, it is important for future
randomized trials to determine whether increases in total physical, independent
of MVPA, can reduce adiposity in overweight Hispanic adolescents.
There is no clear consensus on whether the intensity level of physical
activity influences changes in adiposity, especially in overweight adolescents.
Some longitudinal studies that report a relationship between objectively
measured MVPA and body fat have accounted for total physical activity
42
, while
others have not
138
. Therefore, it remains unclear whether the reductions in body
fat observed in these studies were associated with MVPA independent of total
physical activity. In a large cohort of 12-year olds, time spent in MVPA was
negatively associated with fat mass after adjusting for total physical activity
98
,
106

suggesting that the physiological effects of MVPA may be importantly related to
adiposity. Future longitudinal studies examining the relationship between MVPA
and adiposity should control for total physical activity to determine whether
decreases in adiposity are related specifically to an increase in MVPA or whether
the effects are due simply to an increase in total physical activity. This has
important public health implications given that the barriers to increasing MVPA
may be different than the barriers to increasing total daily physical activity,
especially in overweight youth
164
.
It is also possible that a significant relationship between changes in
percent of time spent in MVPA and adiposity was not detected because of the
cut-points used to designate time spent in MVPA. There is no agreement on
which cut-points should be used in different populations, and using cut-points
derived from different prediction equations can yield markedly different results
56
.
The prediction equations, also, introduce residual error into the measurement,
and thus counts per minute may be a more valid measure of activity than
intensity
47
. It is unlikely, though, that the MVPA cut-point used in the current
study contributed to the failure to detect a relationship between changes in
MVPA and adiposity independent of total activity, because when cut-points from
a different prediction equation were used to designate MVPA
47, 48
, the results
were replicated.
In a prospective four year study with older adolescents, Ekelund et al.
42

also reported that change in total activity, not MVPA, is related to change in
107

adiposity, though they only reported these results in normal weight participants.
The authors postulate that a failure to detect an association in overweight
adolescents may be due to the lack of change in activity in this group. In the
current study, we also reported a lack of overall change, but there was a nicely
distributed variability in change in physical activity at the individual level. One
reason for the apparently discrepant findings may be that the distribution of
change in activity was skewed in Ekelund et al., making it difficult to identify a
relationship. It is difficult to compare results from these studies, though, due to
differences in the populations studied, the length of the follow-up period, and the
different covariates, such as dietary intake, included in the analyses.
Some longitudinal studies using body mass index to measure adiposity
have failed to find a relationship with physical activity
138, 139
. One strength of the
current study was the ability to examine the relationship with objective measures
of physical activity and adiposity, e.g., activity by accelerometry
122
and total fat
mass and percent body fat measured by DEXA
50
. Another strength of the current
study is the homogeneity of the demographic characteristics (e.g., ethnicity, age,
Tanner stage, overweight status) of the sample of the participants. This
homogeneity helps to reduce the effects of maturation bias, though this also
restricts the generalizability of the findings to those who are similar to the study
participants.
Our findings are supported by cross-sectional studies that have found a
significant association between physical activity and adiposity
125, 134
. Coupled
108

with cross-sectional and longitudinal studies, the current analysis lends support
to a causal inverse association between changes in physical activity and
adiposity, but caution should be used when interpreting these findings as
definitive evidence of causation.
A potential limitation to this study is the fact that accelerometry is not a
perfect measure of physical activity, because it is worn on the hip and does not
accurately detect bicycling or upper body movement, such as weight lifting.
Similarly, the accelerometer cannot be worn during water-based activities, such
as swimming. A recent large cohort study reported that swimming is not a
common activity among Hispanic adolescents
16
, so this limitation may not be of
concern in the current sample. The definition of valid accelerometry data used in
the current study, 2 days and 6 hours, is another potential limitation, though the
average measurement period was approximately 70 hours. The self-report
nature of the dietary data may also be a limitation, especially given the
overweight status of the sample and the likelihood of overweight participants to
underreport their dietary intake
109
. Another possible limitation of the study is the
fact that dietary intake was not examined as either an effect modifier or a
mediator of the relationship between physical activity and adiposity. The authors
acknowledge the probability that energy intake acts as a mediator and/or effect
modifier, but that investigation is outside the scope of the current study.
To our knowledge, this is the first study to examine the relationship
between short-term changes in objectively measured activity by accelerometry
109

and adiposity and to examine the influence of energy intake on this relationship
in overweight Hispanic adolescents. In summary, an increase of 28% of total
physical, or roughly 250 kcal, was associated with a modest, yet significant,
decrease in 1.4 kg of fat mass and 1% percent body fat in overweight Hispanic
adolescents over 16-weeks, while there was no association between change in
time spent in MVPA and adiposity. Additionally, controlling for energy intake may
help to further elucidate the relationship between physical activity and adiposity.
Acknowledgements:
This work was supported by the National Institutes of Cancer (NCI),
NCI Centers for Transdisciplinary Research on Energetics and Cancer (TREC,
U54 CA 116848), NCI training grant (T32 CA 09492), and the National Institute of
Child Health and Human Development (RO1 HD/HL 33064). We would also like
to thank all of the staff at USC and USC-GCRC. Most importantly, we would like
to thank all of the participants and their families for making this study possible.

110

Figure 4.1. Mean change in total physical activity (left) and change in percent
time in moderate to vigorous physical activity (right) by intervention group (n =
38; control group n = 10, nutrition group n = 20, nutrition + strength training n =
8). There were no statistically significant differences in mean change by group.
MVPA, moderate to vigorous physical activity. Bars shown are error bars.

111

Figure 4.2. Change in total physical activity (counts per minute) by individual
(n=38).

112

113

Figure 4.3. Graph of predicted values of change in total physical activity by
change in fat mass. Predicted values are adjusted for sex, age, intervention
group, lean tissue mass, and pre-test physical activity (n = 38).

Figure 4.4. Individual coefficient plot of change in total fat regressed on change
in total physical activity in a sample of overweight Hispanic adolescents (n = 38).
Adjusted for sex, age, intervention group, lean tissue mass, and pre-test physical
activity. Dotted lines are 95% confidence intervals.

114

115

CHAPTER 5. SUMMARY, CONCLUSIONS, AND FUTURE DIRECTIONS
Summary of Findings
The general purpose of this dissertation was to examine the relationship
between objectively measured physical activity and body fat deposition variables,
including total fat mass, SAAT, VAT, and HFF. This dissertation also sought to
examine potential moderators of this relationship, such as sex and ethnicity as
well as potential confounders, such as energy intake. Three studies compose
this dissertation. The objectives of these studies were 1) to cross-sectionally
explore ethnic differences in the relationship between physical activity and fat
(i.e., total fat mass, SAAT, VAT, and HFF) in overweight Hispanic and African
American adolescents (study 1); 2) to cross-sectionally examine sex differences
in the relationship between physical activity and fat in overweight Hispanic
adolescents (study 2); 3) to examine the association between changes in
physical activity and changes in fat in overweight Hisp adolescents (study 3).
The primary findings from study 1 indicated that the relationships between
total physical activity and HFF and SAAT are ethnic-specific. Higher total
physical activity, by approximately 30%, was associated with 22% lower HFF in
overweight African American, but not Hispanic, adolescents. In contrast, roughly
30% higher physical activity was associated with 9% lower SAAT in overweight
Hispanic, but not African American, adolescents. The findings suggest that
responses to physical activity may differ in Hispanic and African American youth.
116

The key findings from study 2, which focused on Hispanic youth only,
indicated that higher total physical activity was associated with 9% lower SAAT in
overweight Hispanic boys and girls. The results for study 2 also suggest that the
relationship between physical activity and adiposity is similar in overweight
Hispanic boys and girls, as no sex differences were detected. Accounting for
energy intake did not influence the activity and adiposity association.
The principal findings from study 3, which also focused on Hispanic youth,
indicated that among overweight Hispanic adolescents increasing physical
activity about 30% over 16 weeks is associated with a reduction of 1.4 kg of fat
mass and 1% body fat. These results were modestly strengthened by
accounting for changes in energy intake (0.1 kg total fat and 0.2 % body fat).
These results suggest that increasing physical activity over 16 weeks may help to
improve the body composition of overweight Hispanic adolescents.
Conclusions common throughout the dissertation
When the findings from the three studies that comprise this dissertation
are considered as a whole, several re-occurring conclusions become apparent.
The first consistent finding that is seen in all three studies is that total physical
activity, and not percent of time spent in MVPA or sedentary, is related to
reductions in fat mass in overweight Hispanic and African American adolescents.
In all three studies, the percent of time spent engaged in MVPA or spent
sedentary was not significantly associated with fat mass after adjusting for total
physical activity.
117

Throughout this dissertation, we attempted to examine the effect of
physical activity intensity on fat mass independent of total physical activity
because of the overweight status of the adolescents in the studies. Total
physical activity can be increased by reducing sedentary behaviors, such as TV
watching; by increasing light intensity activities, such as walking to school; or by
increasing moderate to vigorous intensity activities, such as playing soccer.
Higher intensity activities, including exercise, may not be well tolerated in
overweight participants, who have more barriers to exercise
164
. The findings
from this dissertation highlight the fact that higher total physical activity may be
sufficient to reduce body fat in overweight Hispanic and African American
adolescents.
It should be noted that while the studies in this dissertation suggest that
total physical activity is sufficient to reduce body fat in overweight adolescents,
the conclusions from these studies do little to answer the question of whether
MVPA or total physical activity is more effective in reducing adiposity. The
observational aspect of the study designs in this dissertation could not address
whether MVPA leads to greater reductions in fat mass than total physical activity.
To adequately compare the effects of total physical activity and activity intensity
on fat mass, a randomized control trial is needed. One potential randomized trial
that is informed by these conclusions is discussed in the future research section
of this dissertation.
118

Another common conclusion in this dissertation is that increases in
physical activity may be associated with decreases in body fat stores (i.e., SAAT
and HFF), even when physical activity is not associated with total fat mass. In
studies 1 and 2, total physical activity is negatively associated with SAAT and
HFF, but not associated with total fat mass. A similar conclusion is not reached
in study 3, though that may be due to the reduced sample size that had MRI data
in study 3 (n=30). The power analysis, which is described in chapter 1 of this
dissertation, is based on a study of physical activity and VAT
124
and suggested a
sample size of at least 37 participants to detect a relationship. Future studies
with a larger sample size could help to answer whether changes in total physical
activity could be related to changes in VAT and SAAT.
A third consistent finding in this dissertation was the intriguing discovery of
a positive association between PA and VAT, independent of total fat mass or
SAAT. In the discussion sections of chapters 3 and 4 (studies 1 and 2), air
pollution, specifically particulate matter 2.5 (PM
2.5
), was discussed as a possible
unmeasured third variable that could help to explain this apparently spurious
relationship. Particulate matter (PM) is fine or tiny particles of solid or liquid
suspended in a gas, and PM 2.5 refers to particles less than 2.5 micrometers.
Particles less than 2.5 micrometers may penetrate into gas-exchange regions of
the lung, and very small particles (< 100 nanometers) may pass through the
lungs to affect other organs. As previously cited, a recent study in a mouse
model of diet-induced obesity indicates that exposure to ambient PM 2.5 led to
119

increased visceral and mesenteric fat mass, but did not result in increased
subcutaneous or retroperitoneal fat
140
.
To insure that the positive relationship between physical activity and VAT
was not due to a data error, the data were visually inspected, and no improbable
values were identified. In addition, correlations between VAT and other
measures of body fat suggest that the variable was not corrupted. In Study 1,
VAT is significantly and positively correlated with total fat (r = 0.61, p < 0.0001),
SAAT (r = 0.35, p = 0.01), and HFF (r = 0.38, p = 0.01), and similar correlations
between VAT and other fat measures were seen in Study 2 (all p < 0.05). As
expected, as VAT increases so do the other measures of body fat, which
indicates that the variable is not reverse coded. Future studies examining the
relationship between physical activity and VAT in populations recruited from
highly polluted areas should measure time spent outdoors and exposure to air
pollution.
Strengths and limitations
One of the strengths of this dissertation is the use of accelerometry.
Youth tend to report that they have engaged in more physical activity, especially
MVPA, than is measured by more objective measures. For example, in one
study that measured physical activity with a self-report measure and
accelerometry, two times as much MVPA was reported in the questionnaire than
was detected by the accelerometer
125
. Another strength of the studies is the use
of precise measures of adipose stores, such as DEXA and multi-slice MRI.
120

Several potential limitations of studies presented in this dissertation should
also be noted. One of the limitations of the study designs is the cross-sectional
and observational nature of studies 1 and 2 and the observational nature of study
3. The observational nature limits the causal conclusions that could be made
about whether decreased physical activity leads to increased adiposity or
whether increased adiposity leads to decreased physical activity. To adequately
address the question of whether decreased physical activity precedes increased
adiposity, experimental study designs are required.
The study samples in this dissertation were a homogenous group, i.e., all
sedentary, overweight Hispanic or African American adolescents who are
primarily postpubescent and recruited from greater Los Angeles, CA. The
homogeneity prevents generalizing the findings to other adolescents, especially
those who live in rural areas, are not overweight, or belong to a different racial or
ethnic group. It is possible that the homogeneity of the sample contributed to the
lack of hypothesized findings, especially the lack of a relationship between
physical activity and total fat mass. It is possible that the relationship would have
been detected in a sample including both lean and overweight youth with a
greater variability in physical activity.
In an attempt to determine if a relationship between physical activity and
fat mass is present in a ethnically similar sample with more variation in physical
activity and weight status, preliminary analyses were conducted in an ongoing
study of lean and overweight, prepubescent, Hispanic girls (TRANSITIONS;
121

Spruijt-Metz, PI). The cross-sectional sample included 27 Hispanic girls (50%
Tanner stage 1, 50% Tanner stage 2) with a mean + SD age of 9.7 + 0.9 years
and BMI percentile of 81.5 + 21.9%, and the statistical analysis approach
described in studies 1 and 2 was used. In the sample of prepubescent Hispanic
girls, multivariate regression analyses revealed that higher total PA (per 100
cpm) was marginally associated with lower log SAAT (B = -0.15, p = 0.07) and
was not associated with log total fat mass (B = -0.16, p = 0.13), log VAT (B =
0.01, p = 0.88), or log HFF (B = 0.001, p = 0.94) after adjusting for age, Tanner,
and relevant body fat covariate. These preliminary results support the lack of a
relationship between physical activity and total fat that was reported in
overweight Hispanic adolescents in studies 1 and 2. In addition, the marginal
inverse relationship between physical activity and SAAT in the younger sample
agrees with the significant relationship reported in the older sample (studies 1
and 2).
Another limitation that should be noted is that the sample sizes of the
studies in this dissertation made it difficult to examine the potential mediator
effects of diet. A mediator is a third variable (e.g., diet) that is intermediate in the
causal relationship relating to predictor and outcome variables
4
. In other words,
diet is a mediator if physical activity causally influences diet and then diet
causally influences fat mass (Figure 5.1). While this is a feasible hypothesis and
worthy of investigation, it was outside of the scope of this dissertation.
122

Figure 5.1. Theoretical model of diet mediating the relationship
between physical activity and fat mass.

In the introduction of this dissertation, power calculations were conducted
to determine the minimum detectable differences (MDD), which are the smallest
coefficient estimates that could be considered statistically significant given the
sample size of the studies and the standard deviations of and correlations
between the independent and dependent variables. Tables 1.1 – 1.3 shows the
absolute values for the MDD for each study. In general, the observed coefficient
estimates were close to or less than the minimum detectable differences. For
example in Table 1.1, the MDD for the coefficient estimate of the relationship
between total physical activity and log total fat mass in the total sample (n=77) is
0.1, and the observed coefficient estimate is 0.051 (p=0.18). The MDD for the
coefficient estimate of the relationship between percent of time spent in MVPA
and log total fat mass in the total sample sample (n=77) is 0.06 and the observed
coefficient estimate is 0.057 (p=0.02). The MDD for the coefficient estimate of
the relationship between percent of time spent sedentary and log total fat mass in
the total sample is 0.02, and the observed coefficient estimate is 0.005 (p=0.41).
Similar comparisons are seen when looking within ethnicity. For example,
the MDD for the relationship between total physical activity and SAAT in
123

Hispanics (n=38) is 0.1, and the observed coefficient is 0.96 (p=0.02). The MDD
for the same relationship in African American adolescents (n=18) is 0.2, and the
observed coefficient estimate is 0.10 (p=0.08). Similar comparisons can be
made between the MDD and observed coefficient estimates in study 2 and study
3. Overall, these comparisons show that the sample sizes are not adequate, or
are just barely adequate, to reject the null hypotheses in the three studies that
comprise this dissertation.
Future Research
The most novel findings that are reported in this dissertation are the
ethnic-specific relationships between physical activity and body fat distribution
(study 1) and the significant relationships between total physical activity, as
opposed to physical activity intensity, and body fat in overweight Hispanic and
African American adolescents (study 1, 2, and 3). The notable findings from
these three studies generated additional research questions that will be further
explored in this section.
The results from study 1, indicating that physical activity is associated with
hepatic fat in overweight African American adolescents, warrant further research.
The next step in showing a causal relationship between physical activity and
hepatic fat is investigating whether short-term increases in total physical activity
are associated with decreases in HFF in overweight African American
adolescents. This question was not examined as part of this dissertation, but in
124

the future, data to answer this question will be available for analysis from the
STAND and/or TRANSITIONS studies.
The potential mechanisms that explain the ethnic-specific relationship
between physical activity and HFF observed in African American adolescents
also warrant further research. One potential explanation of the ethnic-specific
association is that there may be ethnic differences in the underlying physiological
response to physical activity in Hispanic and African American adolescents. As
discussed in study 1, one possibility is that African American adolescents uptake
more FFA into skeletal muscle in response to physical activity, which could
reduce the FFA delivered to and subsequently deposited in the liver. Another
possibility is Hispanic adolescents have increased hepatic FFA uptake in
response to physical activity, potentially due in part to the increased FFA
delivered to the liver from visceral fat. To test whether there are differences
between African American and Hispanic adolescents in FFA uptake to liver and
skeletal muscle, positron emission tomography (PET) could be used to measure
FFA uptake in skeletal muscle and liver during a low-intensity physical activity
(e.g., single leg extensions), using a similar methodology as Hannukainen et
al.
59
. PET is a nuclear medicine imaging technique that produces a three-
dimensional image of functional processes in the body, and by injecting and
measuring a FFA radiotracer (e.g., FTHA), the FFA uptake by skeletal and
hepatic tissues can be calculated. To determine if ethnic differences in FFA
uptake by skeletal and hepatic tissues exist, a study could be conducted during
125

which the FFA uptake of Hispanic and African American young adults, who are
fasting at measurement and matched on fitness, habitual physical activity, age,
tanner, and insulin resistance, are measured by PET during a bout of single leg
extensions. The results of this study may help to explain the ethnic-specific
relationship between total physical activity and HFF in African American
adolescents.
As discussed previously in the summary section of this dissertation, in
order to adequately compare the effect of total physical activity and physical
activity intensities, such as MVPA, a randomized, controlled trial is required.
Gutin et al.
57
conducted an elegant randomized exercise intervention designed to
determine the effects of physical training intensity on total fat mass and visceral
fat in overweight white and black adolescents. They reported that physical
training of at least a moderate intensity decreased percent body fat and VAT
more than a lifestyle education program that delivered counseling on diet and
physical activity. They did not, however, discuss whether increases in total
physical activity within the lifestyle education group also resulted in reductions in
fat mass. It is important to determine if increases in total physical activity result
in reductions in fat mass, because increasing total physical activity may be an
easier goal for overweight Hispanic and African American adolescents to attain
than the current USDHHS recommendations for adolescents of 60 min/day of at
least moderate physical activity. It is also important to determine if increases in
126

total physical activity result in ethnic-specific reductions in fat mass, as
suggested by the findings of the current dissertation.
A future randomized, controlled trial could address the two research
questions concurrently: 1) do increases in total physical activity, as compared to
MVPA, result in decreases in body fat in adolescents? 2) are there ethnic-
specific responses to physical activity in overweight Hispanic and African
American adolescents? The trial would consist of four experimental groups and
compare the effect of total physical activity and MVPA on total body fat and
specific adipose depots (e.g., SAAT, VAT, and HFF) in overweight Hispanic and
African American adolescents. The four experimental groups would be 1) a
waitlist control group, 2) a healthy lifestyle intervention class that advocates
increased physical activity throughout the day, 3) a low intensity training +
healthy lifestyle intervention class, and 4) a moderate to vigorous intensity
training group + healthy lifestyle intervention class. The purpose of the four
experimental groups is to stagger the amount of increases in total physical
activity, aiming for no increase in physical activity in the control group, the
smallest increase in physical activity in the healthy lifestyle class group, and the
largest increase in physical activity in the moderate to vigorous intensity training
group. Throughout the 16-week intervention, the physical activity of the
participants in all experimental groups would be periodically measured via
accelerometry to investigate the effects of the intervention groups on daily total
physical activity. Based on the findings reported in this dissertation, it would be
127

expected that adolescents in the treatment groups (e.g., healthy lifestyle, low-
intensity training, and moderate-intensity training groups) would see incremental
reductions in fat mass with the smallest reduction in the healthy lifestyle group
and the greatest reduction in the moderate training group and that those
reductions would be different in African American and Hispanic adolescents.
Specifically, findings from this dissertation suggest that Hispanic adolescents
would reduce more fat mass in the subcutaneous abdominal area as compared
to African American adolescents, and African American adolescents would be
more likely to reduce the amount of fat mass in the liver in response to increases
in physical activity as compared to Hispanic adolescents.
In summary, future research directions that were generated from the
findings of this dissertation include examining whether changes in physical
activity are associated with changes in HFF in overweight African American
adolescents, examining potential ethnic differences in FFA uptake, and a
randomized intervention comparing the effect of total physical activity and
physical activity intensity on body fat stores in overweight Hispanic and African
American adolescents.
Conclusions

To summarize, in a sample of overweight Hispanic and African American
adolescents ethnic-specific relationships between physical activity and body fat
distribution were observed. Higher physical activity of 100 cpm, which is roughly
30% higher physical activity, was associated with 22% lower hepatic fat in
128

African American, but not Hispanic, adolescents. Higher physical activity (by 100
cpm) was associated with 9% lower SAAT in Hispanic, but not African American,
adolescents. In a sample of overweight Hisp adolescents, 30% increases in total
physical activity were associated with decreases of 1.4 kg in total fat mass and
1% body fat in overweight Hisp adolescents after accounting for the influence of
energy intake. These results suggest that total physical activity may be sufficient
to improve body composition in overweight Hispanic and African American
adolescents. Ethnic-specific associations between physical activity and adipose
depots suggest that interventions may improve body composition differently (i.e.,
reducing SAAT versus HFF) in overweight Hispanic and African American
adolescents.

129

REFERENCES

1. ActiLife User Manual Rev F. Actigraph, LLC. Available at:
http://www.theactigraph.com/index.php?option=com_docman&task=cat_vi
ew&gid=53&Itemid=64. Accessed July 27, 2008.

2. Ball EJ, O'Connor J, Abbott R, et al. Total energy expenditure, body
fatness, and physical activity in children aged 6-9 y. Am J Clin Nutr. Oct
2001;74(4):524-528.

3. Ballor DL, McCarthy JP, Wilterdink EJ. Exercise intensity does not affect
the composition of diet- and exercise-induced body mass loss. Am J Clin
Nutr. Feb 1990;51(2):142-146.

4. Baron RM, Kenny DA. The moderator-mediator variable distinction in
social psychological research: conceptual, strategic, and statistical
considerations. J Pers Soc Psychol. Dec 1986;51(6):1173-1182.

5. Bayne-Smith M, Fardy PS, Azzollini A, Magel J, Schmitz KH, Agin D.
Improvements in heart health behaviors and reduction in coronary artery
disease risk factors in urban teenaged girls through a school-based
intervention: the PATH program. Am J Public Health. Sep
2004;94(9):1538-1543.

6. Bergman RN, Kim SP, Catalano KJ, et al. Why visceral fat is bad:
mechanisms of the metabolic syndrome. Obesity (Silver Spring). Feb
2006;14 Suppl 1:16S-19S.

7. Biddle SJ, Gorely T, Marshall SJ, Murdey I, Cameron N. Physical activity
and sedentary behaviours in youth: issues and controversies. J R Soc
Health. Jan 2004;124(1):29-33.

8. Bjorntorp P. Abdominal fat distribution and disease: an overview of
epidemiological data. Ann Med. Feb 1992;24(1):15-18.

9. Bjorntorp P. The regulation of adipose tissue distribution in humans. Int J
Obes Relat Metab Disord. Apr 1996;20(4):291-302.

10. Boon RM, Hamlin MJ, Steel GD, Ross JJ. Validation of the New Zealand
Physical Activity Questionnaire (NZPAQ-LF) and the International Physical
Activity Questionnaire (IPAQ-LF) with Accelerometry. Br J Sports Med.
Nov 3 2008.
130

11. Booth FW, Laye MJ, Lees SJ, Rector RS, Thyfault JP. Reduced physical
activity and risk of chronic disease: the biology behind the consequences.
Eur J Appl Physiol. Mar 2008;102(4):381-390.

12. Boyko EJ, Fujimoto WY, Leonetti DL, Newell-Morris L. Visceral adiposity
and risk of type 2 diabetes: a prospective study among Japanese
Americans. Diabetes Care. Apr 2000;23(4):465-471.

13. Bray GA, Popkin BM. Dietary fat intake does affect obesity! Am J Clin
Nutr. Dec 1998;68(6):1157-1173.

14. Bryner RW, Toffle RC, Ullrich IH, Yeater RA. The effects of exercise
intensity on body composition, weight loss, and dietary composition in
women. J Am Coll Nutr. Feb 1997;16(1):68-73.

15. Bugianesi E, Zannoni C, Vanni E, Marzocchi R, Marchesini G. Non-
alcoholic fatty liver and insulin resistance: a cause-effect relationship? Dig
Liver Dis. Mar 2004;36(3):165-173.

16. Butte NF, Cai G, Cole SA, et al. Metabolic and behavioral predictors of
weight gain in Hispanic children: the Viva la Familia Study. Am J Clin Nutr.
Jun 2007;85(6):1478-1485.

17. Butte NF, Puyau MR, Adolph AL, Vohra FA, Zakeri I. Physical activity in
nonoverweight and overweight Hispanic children and adolescents. Med
Sci Sports Exerc. Aug 2007;39(8):1257-1266.

18. Byrd-Williams C, Kelly LA, Davis JN, Spruijt-Metz D, Goran MI. Influence
of gender, BMI and Hispanic ethnicity on physical activity in children. Int J
Pediatr Obes. 2007;2(3):159-166.

19. Cali AM, De Oliveira AM, Kim H, et al. Glucose dysregulation and hepatic
steatosis in obese adolescents: Is there a link? Hepatology. Jan 23 2009.
20. Carroll JF, Chiapa AL, Rodriquez M, et al. Visceral fat, waist
circumference, and BMI: impact of race/ethnicity. Obesity (Silver Spring).
Mar 2008;16(3):600-607.

21. Charbonneau A, Unson CG, Lavoie JM. High-fat diet-induced hepatic
steatosis reduces glucagon receptor content in rat hepatocytes: potential
interaction with acute exercise. J Physiol. Feb 15 2007;579(Pt 1):255-267.

22. Chitwood LF, Brown SP, Lundy MJ, Dupper MA. Metabolic propensity
toward obesity in black vs white females: responses during rest, exercise
and recovery. Int J Obes Relat Metab Disord. May 1996;20(5):455-462.
131

23. Corder K, Brage S, Ramachandran A, Snehalatha C, Wareham N,
Ekelund U. Comparison of two Actigraph models for assessing free-living
physical activity in Indian adolescents. J Sports Sci. Dec
2007;25(14):1607-1611.

24. Corder K, Ekelund U, Steele RM, Wareham NJ, Brage S. Assessment of
physical activity in youth. J Appl Physiol. Sep 2008;105(3):977-987.

25. Crawford PB, Obarzanek E, Morrison J, Sabry ZI. Comparative advantage
of 3-day food records over 24-hour recall and 5-day food frequency
validated by observation of 9- and 10-year-old girls. J Am Diet Assoc. Jun
1994;94(6):626-630.

26. Cruz ML, Bergman RN, Goran MI. Unique effect of visceral fat on insulin
sensitivity in obese Hispanic children with a family history of type 2
diabetes. Diabetes Care. Sep 2002;25(9):1631-1636.

27. Cruz ML, Shaibi GQ, Weigensberg MJ, Spruijt-Metz D, Ball GD, Goran MI.
Pediatric obesity and insulin resistance: chronic disease risk and
implications for treatment and prevention beyond body weight
modification. Annu Rev Nutr. 2005;25:435-468.

28. Davis JN, Kelly LA, Lane CJ, et al. Randomized control trial of a Nutrition
Education and Strength Training Program to Prevent Obesity Related
Diseases in Overweight Latino Adolescents. Obesity (Silver Spring).Feb
2009:16(2):415-420.

29. de Piano A, Prado WL, Caranti DA, et al. Metabolic and nutritional profile
of obese adolescents with nonalcoholic fatty liver disease. J Pediatr
Gastroenterol Nutr. Apr 2007;44(4):446-452.

30. Dietz WH. Health consequences of obesity in youth: childhood predictors
of adult disease. Pediatrics. Mar 1998;101(3 Pt 2):518-525.

31. Donnelly KL, Smith CI, Schwarzenberg SJ, Jessurun J, Boldt MD, Parks
EJ. Sources of fatty acids stored in liver and secreted via lipoproteins in
patients with nonalcoholic fatty liver disease. J Clin Invest. May
2005;115(5):1343-1351.

32. Druet C, Baltakse V, Chevenne D, et al. Independent effect of visceral
adipose tissue on metabolic syndrome in obese adolescents. Horm Res.
2008;70(1):22-28.

132

33. Dupont WD. Statistical Modeling for Biomedical Researchers. New York:
Cambridge University Press; 2002.

34. Dupont WD, Plummer WD, Jr. Power and sample size calculations for
studies involving linear regression. Control Clin Trials. Dec
1998;19(6):589-601.

35. Dwyer JT, Stone EJ, Yang M, et al. Prevalence of marked overweight and
obesity in a multiethnic pediatric population: findings from the Child and
Adolescent Trial for Cardiovascular Health (CATCH) study. J Am Diet
Assoc. Oct 2000;100(10):1149-1156.

36. Ekelund U, Aman J, Yngve A, Renman C, Westerterp K, Sjostrom M.
Physical activity but not energy expenditure is reduced in obese
adolescents: a case-control study. Am J Clin Nutr. Nov 2002;76(5):935-
941.

37. Ekelund U, Brage S, Franks PW, et al. Physical activity energy
expenditure predicts changes in body composition in middle-aged healthy
whites: effect modification by age. Am J Clin Nutr. May 2005;81(5):964-
969.

38. Ekelund U, Brage S, Froberg K, et al. TV viewing and physical activity are
independently associated with metabolic risk in children: the European
Youth Heart Study. PLoS Med. Dec 2006;3(12):e488.

39. Ekelund U, Griffin SJ, Wareham NJ. Physical activity and metabolic risk in
individuals with a family history of type 2 diabetes. Diabetes Care. Feb
2007;30(2):337-342.

40. Ekelund U, Neovius M, Linne Y, Brage S, Wareham NJ, Rossner S.
Associations between physical activity and fat mass in adolescents: the
Stockholm Weight Development Study. Am J Clin Nutr. Feb
2005;81(2):355-360.

41. Ekelund U, Sardinha LB, Anderssen SA, et al. Associations between
objectively assessed physical activity and indicators of body fatness in 9-
to 10-y-old European children: a population-based study from 4 distinct
regions in Europe (the European Youth Heart Study). Am J Clin Nutr. Sep
2004;80(3):584-590.

42. Ekelund U, Sarnblad S, Brage S, Ryberg J, Wareham NJ, Aman J. Does
physical activity equally predict gain in fat mass among obese and
nonobese young adults? Int J Obes (Lond). Jan 2007;31(1):65-71.
133

43. Esliger D, Copeland J, Tremblay M. Standardizing and Optimizing the Use
of Accelerometer Data fro Free-Living Physical Activity Monitoring. Journal
of Physical Acitivty and Health. 2005;3:366-383.

44. Esliger DW, Tremblay MS. Technical reliability assessment of three
accelerometer models in a mechanical setup. Med Sci Sports Exerc. Dec
2006;38(12):2173-2181.

45. Eston RG, Rowlands AV, Ingledew DK. Validity of heart rate, pedometry,
and accelerometry for predicting the energy cost of children's activities. J
Appl Physiol. Jan 1998;84(1):362-371.

46. Fishbein MH, Gardner KG, Potter CJ, Schmalbrock P, Smith MA.
Introduction of fast MR imaging in the assessment of hepatic steatosis.
Magn Reson Imaging. 1997;15(3):287-293.

47. Freedson P, Pober D, Janz KF. Calibration of accelerometer output for
children. Med Sci Sports Exerc. Nov 2005;37(11 Suppl):S523-530.

48. Freedson P, Sirard J, Debold E, et al. Calibration of the Computer Science
and Applications INC. (CSA) Accelerometer. Med Sci Sports Exer.
1997;29(Suppl):S45.

49. Gonzalez-Ortiz M, Robles-Cervantes JA, Cardenas-Camarena L, Bustos-
Saldana R, Martinez-Abundis E. The effects of surgically removing
subcutaneous fat on the metabolic profile and insulin sensitivity in obese
women after large-volume liposuction treatment. Horm Metab Res. Aug
2002;34(8):446-449.

50. Goran MI. Measurement issues related to studies of childhood obesity:
assessment of body composition, body fat distribution, physical activity,
and food intake. Pediatrics. Mar 1998;101(3 Pt 2):505-518.

51. Goran MI, Bergman RN, Cruz ML, Watanabe R. Insulin resistance and
associated compensatory responses in African-American and Hispanic
children. Diabetes Care. Dec 2002;25(12):2184-2190.

52. Goran MI, Gower BA, Nagy TR, Johnson RK. Developmental changes in
energy expenditure and physical activity in children: evidence for a decline
in physical activity in girls before puberty. Pediatrics. May
1998;101(5):887-891.

134

53. Grediagin A, Cody M, Rupp J, Benardot D, Shern R. Exercise intensity
does not effect body composition change in untrained, moderately overfat
women. J Am Diet Assoc. Jun 1995;95(6):661-665.

54. Greenland S, Mickey RM. Re: "The impact of confounder selection criteria
on effect estimation." Am J Epidemiol. Nov 1989;130(5):1066.

55. Gretebeck RJ, Montoye HJ. Variability of some objective measures of
physical activity. Med Sci Sports Exerc. Oct 1992;24(10):1167-1172.

56. Guinhouya CB, Hubert H, Soubrier S, Vilhelm C, Lemdani M, Durocher A.
Moderate-to-vigorous physical activity among children: discrepancies in
accelerometry-based cut-off points. Obesity (Silver Spring). May
2006;14(5):774-777.

57. Gutin B, Barbeau P, Owens S, et al. Effects of exercise intensity on
cardiovascular fitness, total body composition, and visceral adiposity of
obese adolescents. American Journal of Clinical Nutrition. May
2002;75(5):818-826.

58. Gutin B, Johnson MH, Humphries MC, et al. Relationship of visceral
adiposity to cardiovascular disease risk factors in black and white teens.
Obesity (Silver Spring). Apr 2007;15(4):1029-1035.

59. Hannukainen JC, Nuutila P, Borra R, et al. Increased physical activity
decreases hepatic free fatty acid uptake: a study in human monozygotic
twins. J Physiol. Jan 1 2007;578(Pt 1):347-358.

60. Healy GN, Wijndaele K, Dunstan DW, et al. Objectively measured
sedentary time, physical activity, and metabolic risk: the Australian
Diabetes, Obesity and Lifestyle Study (AusDiab). Diabetes Care. Feb
2008;31(2):369-371.

61. Hills AP, King NA, Armstrong TP. The contribution of physical activity and
sedentary behaviours to the growth and development of children and
adolescents: implications for overweight and obesity. Sports Med.
2007;37(6):533-545.

62. Hjartaker A, Langseth H, Weiderpass E. Obesity and diabetes epidemics:
cancer repercussions. Adv Exp Med Biol. 2008;630:72-93.

63. Hsieh SD, Yoshinaga H, Muto T, Sakurai Y. Regular physical activity and
coronary risk factors in Japanese men. Circulation. Feb 24
1998;97(7):661-665.
135

64. Hsing AW, Sakoda LC, Chua S, Jr. Obesity, metabolic syndrome, and
prostate cancer. Am J Clin Nutr. Sep 2007;86(3):s843-857.

65. Hu FB. Obesity Epidemiology. New York: Oxford University Press; 2008.

66. Hughes AR, Henderson A, Ortiz-Rodriguez V, Artinou ML, Reilly JJ.
Habitual physical activity and sedentary behaviour in a clinical sample of
obese children. Int J Obes (Lond). Oct 2006;30(10):1494-1500.

67. Hussey J, Bell C, Bennett K, O'Dwyer J, Gormley J. Relationship between
the intensity of physical activity, inactivity, cardiorespiratory fitness and
body composition in 7-10-year-old Dublin children. Br J Sports Med. May
2007;41(5):311-316.

68. Jago R, Anderson CB, Baranowski T, Watson K. Adolescent patterns of
physical activity differences by gender, day, and time of day. Am J Prev
Med. Jun 2005;28(5):447-452.

69. Jamner MS, Spruijt-Metz D, Bassin S, Cooper DM. A controlled evaluation
of a school-based intervention to promote physical activity among
sedentary adolescent females: project FAB. J Adolesc Health. Apr
2004;34(4):279-289.

70. Janz KF, Burns TL, Levy SM. Tracking of activity and sedentary behaviors
in childhood: the Iowa Bone Development Study. Am J Prev Med. Oct
2005;29(3):171-178.

71. Jebb S, Moore MS. Contribution of sedentary lifestyle and inactivity to the
etiology of overweight and obesity: current evidence and research issues.
Medicine and Science in Sports and Exercise. 1999;31:S534-S541.

72. Kay SJ, Fiatarone Singh MA. The influence of physical activity on
abdominal fat: a systematic review of the literature. Obes Rev. May
2006;7(2):183-200.

73. Kelley DE, Thaete FL, Troost F, Huwe T, Goodpaster BH. Subdivisions of
subcutaneous abdominal adipose tissue and insulin resistance. Am J
Physiol Endocrinol Metab. May 2000;278(5):E941-948.

74. Kimm SY, Glynn NW, Kriska AM, et al. Decline in physical activity in black
girls and white girls during adolescence. N Engl J Med. Sep 5
2002;347(10):709-715.
136

75. Klein S, Fontana L, Young VL, et al. Absence of an effect of liposuction on
insulin action and risk factors for coronary heart disease. N Engl J Med.
Jun 17 2004;350(25):2549-2557.

76. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression
analysis and other multivariable methods. 3rd ed. Belmont: Brooks/Cole
Publishing; 1998.

77. Kohl HW, 3rd, Fulton J, Caspersen CJ. Assessment of Physical Activity
among Children and Adolescents: A Review and Synthesis. Preventive
Medicine 2000(31).

78. Koutsari C, Jensen MD. Thematic review series: patient-oriented
research. Free fatty acid metabolism in human obesity. J Lipid Res. Aug
2006;47(8):1643-1650.

79. Kunzli N, McConnell R, Bates D, et al. Breathless in Los Angeles: the
exhausting search for clean air. Am J Public Health. Sep 2003;93(9):1494-
1499.

80. Liska D, Dufour S, Zern TL, et al. Interethnic differences in muscle, liver
and abdominal fat partitioning in obese adolescents. PLoS ONE.
2007;2(6):e569.

81. Lohman TG, Ring K, Schmitz KH, et al. Associations of body size and
composition with physical activity in adolescent girls. Med Sci Sports
Exerc. Jun 2006;38(6):1175-1181.

82. Lytle L, Kelder SH, Klepp KI, Perry CP. Covariance of adolescent health
behaviors: the Class of 1989 study. Health Education Research.
1995;10(2):133-146.

83. Marshall SJ, Biddle SJ, Gorely T, Cameron N, Murdey I. Relationships
between media use, body fatness and physical activity in children and
youth: a meta-analysis. Int J Obes Relat Metab Disord. Oct
2004;28(10):1238-1246.

84. Martin CK, Han H, Coulon SM, Allen HR, Champagne CM, Anton SD. A
novel method to remotely measure food intake of free-living individuals in
real time: the remote food photography method. Br J Nutr. Jul 11 2008:1-
11.

137

85. Matthews CE, Chen KY, Freedson PS, et al. Amount of time spent in
sedentary behaviors in the United States, 2003-2004. Am J Epidemiol. Apr
1 2008;167(7):875-881.

86. Mattocks C, Ness A, Leary S, et al. Use of accelerometers in a large field-
based study of children: protocols, design issues, and effects on precision.
J Phys Act Health. 2008;5 Suppl 1:S98-111.

87. Mazess RB, Barden HS, Bisek JP, Hanson J. Dual-energy x-ray
absorptiometry for total-body and regional bone-mineral and soft-tissue
composition. Am J Clin Nutr. Jun 1990;51(6):1106-1112.

88. McMillan DC, Sattar N, Lean M, McArdle CS. Obesity and Cancer. BMJ.
2006;333:1109-1111.

89. McMurray RG, Ward DS, Elder JP, et al. Do overweight girls overreport
physical activity? Am J Health Behav. Sep-Oct 2008;32(5):538-546.

90. Mickey RM, Greenland S. The impact of confounder selection criteria on
effect estimation. Am J Epidemiol. Jan 1989;129(1):125-137.

91. Miller DB, O'Callaghan JP. Neuroendocrine aspects of the response to
stress. Metabolism. Jun 2002;51(6 Suppl 1):5-10.

92. Moeller NC, Korsholm L, Kristensen PL, Andersen LB, Wedderkopp N,
Froberg K. Unit-specific calibration of Actigraph accelerometers in a
mechanical setup - is it worth the effort? The effect on random output
variation caused by technical inter-instrument variability in the laboratory
and in the field. BMC Med Res Methodol. 2008;8:19.

93. Monzon JR, Basile R, Heneghan S, Udupi V, Green A. Lipolysis in
adipocytes isolated from deep and superficial subcutaneous adipose
tissue. Obes Res. Apr 2002;10(4):266-269.

94. Moore LL, Gao D, Bradlee ML, et al. Does early physical activity predict
body fat change throughout childhood? Prev Med. Jul 2003;37(1):10-17.

95. Must A, Bandini LG, Tybor DJ, Janssen I, Ross R, Dietz WH. Behavioral
risk factors in relation to visceral adipose tissue deposition in adolescent
females. Int J Pediatr Obes. 2008;3 Suppl 1:28-36.

96. Must A, Tybor DJ. Physical activity and sedentary behavior: a review of
longitudinal studies of weight and adiposity in youth. Int J Obes (Lond).
Sep 2005;29 Suppl 2:S84-96.
138

97. Nader PR, Bradley RH, Houts RM, McRitchie SL, O'Brien M. Moderate-to-
vigorous physical activity from ages 9 to 15 years. Jama. Jul 16
2008;300(3):295-305.

98. Ness A, Leary SD, Mattocks C, et al. Objectively measured physical
activity and fat mass in a large cohort of children. Public Library of
Science. 2007;4(3):476-484.

99. Neumark-Sztainer D, Story M, Hannan PJ, Rex J. New Moves: a school-
based obesity prevention program for adolescent girls. Prev Med. Jul
2003;37(1):41-51.

100. Ogden CL, Carroll MD, Flegal KM. High body mass index for age among
US children and adolescents, 2003-2006. Jama. May 28
2008;299(20):2401-2405.

101. Ogden CL, Yanovski SZ, Carroll MD, Flegal KM. The epidemiology of
obesity. Gastroenterology. May 2007;132(6):2087-2102.

102. Okosun IS, Tedders SH, Choi S, Dever GE. Abdominal adiposity values
associated with established body mass indexes in white, black and
hispanic Americans. A study from the Third National Health and Nutrition
Examination Survey. Int J Obes Relat Metab Disord. Oct
2000;24(10):1279-1285.

103. Park TG, Hong HR, Lee J, Kang HS. Lifestyle plus exercise intervention
improves metabolic syndrome markers without change in adiponectin in
obese girls. Ann Nutr Metab. 2007;51(3):197-203.

104. Pasquali R, Vicennati V. Activity of the hypothalamic-pituitary-adrenal axis
in different obesity phenotypes. Int J Obes Relat Metab Disord. Jun
2000;24 Suppl 2:S47-49.

105. Pate RR, Ward DS, Saunders RP, Felton G, Dishman RK, Dowda M.
Promotion of physical activity among high-school girls: a randomized
controlled trial. Am J Public Health. Sep 2005;95(9):1582-1587.

106. Perry AC, Rosenblatt EB, Wang X. Physical, behavioral, and body image
characteristics in a tri-racial group of adolescent girls. Obes Res. Oct
2004;12(10):1670-1679.

107. Perseghin G, Lattuada G, De Cobelli F, et al. Habitual physical activity is
associated with intrahepatic fat content in humans. Diabetes Care. Mar
2007;30(3):683-688.
139

108. Pichard C, Plu-Bureau G, Neves ECM, Gompel A. Insulin resistance,
obesity and breast cancer risk. Maturitas. May 20 2008;60(1):19-30.

109. Pikholz C, Swinburn B, Metcalf P. Under-reporting of energy intake in the
1997 National Nutrition Survey. N Z Med J. Sep 24
2004;117(1202):U1079.

110. Plasqui G, Westerterp KR. Physical activity assessment with
accelerometers: an evaluation against doubly labeled water. Obesity
(Silver Spring). Oct 2007;15(10):2371-2379.

111. Prentice A, Jebb S. Energy intake/physical activity interactions in the
homeostasis of body weight regulation. Nutr Rev. Jul 2004;62(7 Pt 2):S98-
104.

112. Proctor MH, Moore LL, Gao D, et al. Television viewing and change in
body fat from preschool to early adolescence: The Framingham Children's
Study. Int J Obes Relat Metab Disord. Jul 2003;27(7):827-833.

113. Puyau MR, Adolph AL, Vohra FA, Butte NF. Validation and calibration of
physical activity monitors in children. Obes Res. Mar 2002;10(3):150-157.

114. Ravussin E, Smith SR. Increased fat intake, impaired fat oxidation, and
failure of fat cell proliferation result in ectopic fat storage, insulin
resistance, and type 2 diabetes mellitus. Ann N Y Acad Sci. Jun
2002;967:363-378.

115. Reilly JJ, Penpraze V, Hislop J, Davies G, Grant S, Paton JY. Objective
measurement of physical activity and sedentary behaviour: review with
new data. Arch Dis Child. Jul 2008;93(7):614-619.

116. Renehan AG, Roberts DL, Dive C. Obesity and cancer: pathophysiological
and biological mechanisms. Arch Physiol Biochem. Feb 2008;114(1):71-
83.

117. Resnicow K, Jackson A, Braithwaite R, et al. Healthy Body/Healthy Spirit:
a church-based nutrition and physical activity intervention. Health Educ
Res. Oct 2002;17(5):562-573.

118. Resnicow K, Taylor R, Baskin M, McCarty F. Results of go girls: a weight
control program for overweight African-American adolescent females.
Obes Res. Oct 2005;13(10):1739-1748.

140

119. Robinson TN, Kraemer HC, Matheson DM, et al. Stanford GEMS phase 2
obesity prevention trial for low-income African-American girls: design and
sample baseline characteristics. Contemp Clin Trials. Jan 2008;29(1):56-
69.

120. Rosenbloom AL, Joe JR, Young RS, Winter WE. Emerging epidemic of
type 2 diabetes in youth. Diabetes Care. Feb 1999;22(2):345-354.

121. Rowlands AV, Eston RG, Ingledew DK. Relationship between activity
levels, aerobic fitness, and body fat in 8- to 10-yr-old children. J Appl
Physiol. Apr 1999;86(4):1428-1435.

122. Rowlands AV, Ingledew DK, Eston RG. The effect of type of physical
activity measure on the relationship between body fatness and habitual
physical activity in children: a meta-analysis. Ann Hum Biol. Sep-Oct
2000;27(5):479-497.

123. Rush EC, Plank LD, Davies PS, Watson P, Wall CR. Body composition
and physical activity in New Zealand Maori, Pacific and European children
aged 5-14 years. Br J Nutr. Dec 2003;90(6):1133-1139.

124. Saelens BE, Sallis JF, Wilfley DE, Patrick K, Cella JA, Buchta R.
Behavioral weight control for overweight adolescents initiated in primary
care. Obes Res. Jan 2002;10(1):22-32.

125. Saelens BE, Seeley RJ, van Schaick K, Donnelly LF, O'Brien KJ. Visceral
abdominal fat is correlated with whole-body fat and physical activity
among 8-y-old children at risk of obesity. Am J Clin Nutr. Jan
2007;85(1):46-53.

126. Salbe AD, Weyer C, Harper I, Lindsay RS, Ravussin E, Tataranni PA.
Relation between physical activity and obesity. Am J Clin Nutr. Jul
2003;78(1):193-194; author reply 194-195.

127. Sallis JF. Age-related decline in physical activity: a synthesis of human
and animal studies. Med Sci Sports Exerc. Sep 2000;32(9):1598-1600.
128. Scanlon V, Sanders T. Essentials of Anatomy and Physiology. 5th ed.
Philadelphia: F.A. Davis Company; 2007.

129. Serdula MK, Ivery D, Coates RJ, Freedman DS, Williamson DF, Byers T.
Do obese children become obese adults? A review of the literature. Prev
Med. Mar 1993;22(2):167-177.

141

130. Singhal A. Endothelial dysfunction: role in obesity-related disorders and
the early origins of CVD. Proc Nutr Soc. Feb 2005;64(1):15-22.

131. Sirard JR, Kubik MY, Fulkerson JA, Arcan C. Objectively measured
physical activity in urban alternative high school students. Med Sci Sports
Exerc. Dec 2008;40(12):2088-2095.

132. Sirard JR, Pate RR. Physical activity assessment in children and
adolescents. Sports Med. 2001;31(6):439-454.

133. Snijder MB, van Dam RM, Visser M, Seidell JC. What aspects of body fat
are particularly hazardous and how do we measure them? Int J Epidemiol.
Feb 2006;35(1):83-92.

134. Snitker S, Le KY, Hager E, Caballero B, Black MM. Association of physical
activity and body composition with insulin sensitivity in a community
sample of adolescents. Arch Pediatr Adolesc Med. Jul 2007;161(7):677-
683.
135. Spassiani NA, Kuk JL. Exercise and the fatty liver. Appl Physiol Nutr
Metab. Aug 2008;33(4):802-807.

136. Spear BA, Barlow SE, Ervin C, et al. Recommendations for treatment of
child and adolescent overweight and obesity. Pediatrics. Dec 2007;120
Suppl 4:S254-288.

137. Stefan N, Kantartzis K, Haring HU. Causes and metabolic consequences
of Fatty liver. Endocr Rev. Dec 2008;29(7):939-960.

138. Stevens J, Murray DM, Baggett CD, et al. Objectively assessed
associations between physical activity and body composition in middle-
school girls: the Trial of Activity for Adolescent Girls. Am J Epidemiol. Dec
1 2007;166(11):1298-1305.

139. Stevens J, Suchindran C, Ring K, et al. Physical activity as a predictor of
body composition in American Indian children. Obes Res. Dec
2004;12(12):1974-1980.

140. Sun Q, Yue P, Deiuliis JA, et al. Ambient air pollution exaggerates adipose
inflammation and insulin resistance in a mouse model of diet-induced
obesity. Circulation. Feb 3 2009;119(4):538-546.

141. Swinburn BA, Jolley D, Kremer PJ, Salbe AD, Ravussin E. Estimating the
effects of energy imbalance on changes in body weight in children. Am J
Clin Nutr. Apr 2006;83(4):859-863.
142

142. Swinburn BA, Sacks G, Lo SK, et al. Estimating the changes in energy
flux that characterize the rise in obesity prevalence. Am J Clin Nutr. Jun
2009;89(6):1723-1728.

143. Tanner JM. Growth and maturation during adolescence. . Nutr Rev.
1981;39:43-55.

144. Taylor RW, Jones IE, Williams SM, Goulding A. Body fat percentages
measured by dual-energy X-ray absorptiometry corresponding to recently
recommended body mass index cutoffs for overweight and obesity in
children and adolescents aged 3-18 y. Am J Clin Nutr. Dec
2002;76(6):1416-1421.

145. The. The American Lung Association. State of the Air. Available from
http://lungaction.org/reports/sota07_cities.html. Last accessed April 15th,
2009.

146. Tiikkainen M, Tamminen M, Hakkinen AM, et al. Liver-fat accumulation
and insulin resistance in obese women with previous gestational diabetes.
Obes Res. Sep 2002;10(9):859-867.

147. Tremblay A, Simoneau JA, Bouchard C. Impact of exercise intensity on
body fatness and skeletal muscle metabolism. Metabolism. Jul
1994;43(7):814-818.

148. Treuth MS, Hou N, Young DR, Maynard LM. Accelerometry-measured
activity or sedentary time and overweight in rural boys and girls. Obes
Res. Sep 2005;13(9):1606-1614.

149. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M.
Physical activity in the United States measured by accelerometer. Med Sci
Sports Exerc. Jan 2008;40(1):181-188.

150. Trost SG, Kerr LM, Ward DS, Pate RR. Physical activity and determinants
of physical activity in obese and non-obese children. Int J Obes Relat
Metab Disord. Jun 2001;25(6):822-829.

151. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. Using objective
physical activity measures with youth: how many days of monitoring are
needed? Med Sci Sports Exerc. Feb 2000;32(2):426-431.

152. Trost SG, Sirard JR, Dowda M, Pfeiffer KA, Pate RR. Physical activity in
overweight and nonoverweight preschool children. Int J Obes Relat Metab
Disord. Jul 2003;27(7):834-839.
143

153. U.S. Census Bureau. Facts and Features. Available at
http://www.census.gov/Press-
Release/www/releases/archives/facts_for_features_special_editions/0103
27.html. Last accessed May 10th, 2009.

154. U.S. Department of Health and Human Services. 2008 Physical activity
guidelines for Americans.
http://www.health.gov/paguidelines/factsheetprof.aspx. Last assessed
Nov. 2008.

155. Velleman P, Welsch R. Efficient Computing of Regression Diagnostiocs.
The American Statistician. 1981;35(4):234-242.

156. Ventura AK, Loken E, Mitchell DC, Smiciklas-Wright H, Birch LL.
Understanding reporting bias in the dietary recall data of 11-year-old girls.
Obesity (Silver Spring). Jun 2006;14(6):1073-1084.

157. Vincent SD, Pangrazi RP. Does Reactivity Exist in Children When
Measuring Activity Levels with Pedometers? Pediatric Exercise Science.
2002;14:56-63.

158. Ward DS, Evenson KR, Vaughn A, Rodgers AB, Troiano RP.
Accelerometer use in physical activity: best practices and research
recommendations. Med Sci Sports Exerc. Nov 2005;37(11 Suppl):S582-
588.

159. Welk GJ, Corbin CB, Dale D. Measurement issues in the assessment of
physical activity in children. Res Q Exerc Sport. Jun 2000;71(2
Suppl):S59-73.

160. Westerterp KR, Goran MI. Relationship between physical activity related
energy expenditure and body composition: a gender difference. Int J Obes
Relat Metab Disord. Mar 1997;21(3):184-188.

161. Westerterp KR, Plasqui G. Physical activity and human energy
expenditure. Curr Opin Clin Nutr Metab Care. Nov 2004;7(6):607-613.

162. Wittmeier KD, Mollard RC, Kriellaars DJ. Physical activity intensity and
risk of overweight and adiposity in children. Obesity (Silver Spring). Feb
2008;16(2):415-420.

163. Yoshioka M, Doucet E, St-Pierre S, et al. Impact of high-intensity exercise
on energy expenditure, lipid oxidation and body fatness. Int J Obes Relat
Metab Disord. Mar 2001;25(3):332-339.
144

164. Zabinski MF, Saelens BE, Stein RI, Hayden-Wade HA, Wilfley DE.
Overweight children's barriers to and support for physical activity. Obes
Res. Feb 2003;11(2):238-246.

Abstract (if available)

Abstract Purpose: This dissertation sought to examine the relationship between physical activity (PA) and adiposity in overweight Hispanic (Hisp) and African American (AA) adolescents. Three objectives were addressed, each in a separate study, and the three studies comprise this dissertation. The objectives were 1) to cross-sectionally explore ethnic differences in the relationship between PA and body fat distribution among overweight Hisp and AA adolescents

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Ectopic fat and adipose tissue inflammation in overweight and obese African Americans and Hispanics

PDF

Effects of sugar and fiber consumption in minority adolescents and self-tracking as a potential dietary intervention tool

PDF

Sociocultural stress, coping and substance use among Hispanic/Latino adolescents

PDF

Motivation and the meanings of health behavior as factors associated with eating behavior in Latino youth

PDF

Psychosocial and behavioral ractors associated with emotional eating in adolescents

PDF

The metabolic syndrome in overweight Latino youth: influence of dietary intake and associated risk for Type 2 diabetes

PDF

Acute stress reduction with interactive guided imagery in overweight Latino adolescents

PDF

Genetic and dietary determinants of nonalcoholic fatty liver disease in Hispanic children

PDF

The effects of mindfulness on adolescent cigarette smoking: Measurement, mechanisms, and theory

PDF

Marginalization in acculturation is related to objectively measured physical activity in Latina adolecents

PDF

Influences of specific environmental domains on childhood obesity and related behaviors

PDF

A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use

PDF

The vicious cycle of inactivity, obesity, and metabolic health consequences in at-risk pediatric populations

PDF

Genetic variants and smoking progression in Chinese adolescents

PDF

Quantity versus quality: how adipose tissue accumulation and immune cell profile associate with risk for type 2 diabetes in minority children and adults

PDF

Prevalence of amblyopia and strabismus in African-American and Hispanic children

PDF

The role of depression symptoms on social information processing and tobacco use among adolescents

PDF

Contextualizing experiences and developmental stages of immigration and cultural stressors in Hispanic/Latinx adolescents

PDF

Affect, digital media use, physical activity, and ADHD in youth

PDF

Psychosocial and cultural factors in the primary prevention of melanoma targeted to multiethnic children

Asset Metadata

Creator Byrd-Williams, Courtney E. (author)

Core Title Objectively measured physical activity and body fat distribution in overweight Hispanic and African American adolescents

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 07/30/2009

Defense Date 06/09/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag ethnic differences,OAI-PMH Harvest,overweight,sex differences

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Goran, Michael I. (committee chair), Berhane, Kiros (committee member), Clark, Florence A. (committee member), Davis, Jaimie (committee member), Spruijt-Metz, Donna (committee member)

Creator Email byrdc@usc.edu,byrdwilliams@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2429

Unique identifier UC1490353

Identifier etd-ByrdWilliams-3148 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-561302 (legacy record id),usctheses-m2429 (legacy record id)

Legacy Identifier etd-ByrdWilliams-3148.pdf

Dmrecord 561302

Document Type Dissertation

Rights Byrd-Williams, Courtney E.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu