University of Southern California Dissertations and Theses

Objectively measured physical activity and related factors in minority youth

(USC Thesis Other)

Objectively measured physical activity and related factors in minority youth

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content OBJECTIVELY MEASURED PHYSICAL ACTIVITY
AND RELATED FACTORS IN MINORITY YOUTH

by
Britni Ryan Belcher

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

May 2011

Copyright 2011 Britni Ryan Belcher
ii

ACKNOWLEDGEMENTS
I would like to thank my committee for their guidance and committed effort
to my learning and achievement of this degree. I am very grateful for my advisor
Dr. Donna Spruijt-Metz for training me as a transdisciplinary researcher in the
field of pediatric obesity and providing me with numerous opportunities to grow
both personally and professionally. Dr. Spruijt-Metz has been a supportive
mentor, who has challenged me and given me independence to further my
career. I admire her indefatigable spirit and ability to make me feel valued as an
individual. I have also been fortunate to have the guidance of Dr. David Berrigan,
who is both extremely knowledgeable and a patient teacher. I am grateful for the
education and guidance of Dr. Chih-Ping Chou throughout my doctoral program
and this dissertation for his statistical knowledge and willingness to answer all my
questions. Also, I am appreciative of Dr. Rob McConnell‘s insights into this work.
I am very thankful to have the guidance of Dr. Jennifer Unger, whose excellent
statistical knowledge and writing skills have helped me grow as a researcher
during this program and to Dr. Sharon Cermak for her ability to ask insightful
questions that increase the depth of this work.
In addition to my committee members, I have been supported by many
friends and colleagues as USC. They have all contributed to my growth and
development as a researcher and are very much valued. Specifically, I would like
to thank Dr. Adar Emken for modeling excellence in scientific enquiry, for
teaching me that balance in life is essential, and for encouraging me to pursue
iii

my ideas on a larger scale. I am also deeply grateful for Dr. Selena Nguyen-
Rodriguez for her admirable statistical, writing, and listening skills, and for
reassuring me that there is a light at the end of the tunnel. I am also thankful for
the support of Ana Romero, project manager and friend, with whom I have
shared countless hours of data collection and who has supported me throughout
my time at USC. I am also grateful for the company and support of the other
students in the lab: Arianna McClain and Ya-Wen Hsu, who made this journey a
truly unique and more enjoyable experience. I would like to thank other students
in my program for their support: Dr. Courtney Byrd-Williams, Melissa Gunning,
and Dr. Emily Ventura.
This acknowledgement would not be complete without thanking my family
and friends. I would like to thank my parents, Norman and Bari, for loving me
unconditionally, for giving me every opportunity to succeed in life, and for always
encouraging me to follow my dreams. I would also like to thank my grandparents,
aunts, uncles, and cousins for supporting me during this process and reminding
me to keep a sense of humor. I am thankful to my best friends Carrie Nelson and
Devon Bratton for listening to me and helping me remember that what I do is
important; and to Jason Schroeder for keeping me smiling and being a breath of
fresh air in my life. I am very thankful to have had this opportunity, and to use my
skills to serve others in this world.
iv

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ......................................................................................ii
LIST OF TABLES .................................................................................................vi
LIST OF FIGURES ............................................................................................. viii
ABBREVIATIONS .................................................................................................ix
ABSTRACT .......................................................................................................... x
CHAPTER 1 INTRODUCTION ............................................................................. 1
Background and Significance ............................................................................ 1
Pediatric Obesity in the United States ............................................................ 1
Physical Activity in Youth .................................................................................. 6
Physical Activity Measurement ...................................................................... 6
Physical Activity and Individual Demographic Factors ................................. 10
Physical Activity and Selected Hormones ....................................................... 18
Overview of the biological basis of physical activity ..................................... 18
Overview of hormonal energy balance regulation ........................................ 20
Physical Activity and Diet ................................................................................ 26
Specific Aims and Hypotheses ........................................................................ 28
Data Sources and Study Samples ................................................................... 32
CHAPTER 2 PHYSICAL ACTIVITY IN US YOUTH: IMPACT OF
RACE/ETHNICITY, AGE, GENDER, & WEIGHT STATUS ................................ 34

CHAPTER 3 THE LONGITUDINAL EFFECTS OF LEPTIN ON
PHYSICAL ACTIVITY IN MINORITY FEMALE CHILDREN ............................... 67

CHAPTER 4 THE INFLUENCE OF MEAL TYPE ON PHYSICAL
ACTIVITY IN MINORITY ADOLESCENTS: THE FOOD
ADOLESCENCE, MOOD, AND EXERCISE 2 (FAME) STUDY ......................... 86

CHAPTER 5 SUMMARY AND CONCLUSIONS .............................................. 113
Summary of Findings .................................................................................... 113
Implications ................................................................................................... 117
Future Research ............................................................................................ 120
Limitations ..................................................................................................... 123
Contribution to the Literature ......................................................................... 124

v

BIBLIOGRAPHY ............................................................................................... 126

vi

LIST OF TABLES

Table 2-1: Characteristics (mean and standard error) of the analyzed
sample for 2003-2006 .................................................................................. 52

Table 2-2: Mean (standard error) counts per minute by race/ethnicity, age
group, and BMI percentile category
1
............................................................ 54

Table 2-3: Mean (standard error) counts per minute by race/ethnicity,
gender, & BMI percentile category
1
............................................................. 56

Table 2-4: Mean (standard error) minutes per day above specified
thresholds
1
for sedentary behavior, moderate, vigorous, & moderate +
vigorous physical activity by race/ethnicity and age group ........................... 58

Table 2-5: Mean (standard error) minutes per day above specified
thresholds
1
for sedentary behavior, moderate, vigorous, & moderate +
vigorous physical activity by race/ethnicity and gender ................................ 60

Table 2-6: Mean (standard error) minutes per day above specified
thresholds
1
for sedentary behavior, moderate, vigorous, & moderate +
vigorous physical activity by race/ethnicity and BMI percentile
category
2
...................................................................................................... 62

Table 2-7: Multivariable linear regression model predicting MVPA
(min/day) ...................................................................................................... 63

Table 2-8: Logistic regression model predicting meeting the 2008 PA
Guidelines
1
by gender .................................................................................. 64

Table 3-1: Baseline descriptive characteristics (mean (SD)) of analytical
sample by Tanner Stage (N=50) .................................................................. 82

Table 3-2: Baseline Generalized Linear Model (N=50) ....................................... 83

Table 3-3: Longitudinal mixed model predicting change in moderate to
vigorous physical activity over 5 visits (N=50) .............................................. 84

Table 4-1: Timing of in-lab measures ............................................................... 106
vii

Table 4-2: Meal compositions ........................................................................... 107

Table 4-3: Baseline sample descriptive statistics (N=44) ................................. 108

Table 4-4: Mixed model to examine the effect of meal type on activity
levels .......................................................................................................... 109

Table 4-5: Mixed model to examine the effect of meal type on insulin and
glucose IAUC ............................................................................................. 110

viii

LIST OF FIGURES

Figure 1-1: Overview of selected biomarkers as they relate to energy
balance ........................................................................................................ 20

Figure 2-1: 3-way age group-BMI-race/ethnic interaction of MVPA in
males ........................................................................................................... 65

Figure 2-2: 3-way age group-BMI-race/ethnic interaction of MVPA in
females ........................................................................................................ 66

Figure 3-1: Model-based estimates of mean MVPA (min/day) by visit ............... 85

Figure 4-1: Mean moderate to vigorous physical activity levels (min) by
meal type ................................................................................................... 111

Figure 4-2: Mean sedentary behavior levels (min) by meal type ...................... 111

Figure 4-3: Mean insulin values (μU/ml) by meal type ...................................... 112

Figure 4-4: Mean glucose values (mg/dl) by meal type .................................... 112

ix

ABBREVIATIONS
AA= African American
BMI= body mass index
CDC= Centers for Disease Control and Prevention
cpm= counts per minute
CTU= Clinical Trials Unit
DEXA= Dual Energy X-Ray Absorptiometry
FSIVGTT= Frequently Sampled Intravenous Glucose Tolerance Test
GCRC= General Clinical Research Center
IAUC= incremental area under the curve
MET= Metabolic Equivalent Task
min/d= minutes per day
MRI= Magnetic Resonance Imaging
MVPA= moderate to vigorous physical activity
NHANES= National Health and Nutrition Examination Survey
SB= sedentary behavior
SI= Insulin Sensitivity
US= United States
x

ABSTRACT
Obesity prevalence is increasing and physical activity levels are declining
in US youth. The overall goal of this dissertation was to examine the effects of
individual demographic, biological, and dietary factors on objectively measured
physical activity levels in youth. The objectives of this dissertation were: 1) to
describe activity levels across race/ethnic, weight status, age, and gender groups
in a large nationally representative sample of youth; 2) to examine cross-
sectional and longitudinal relationships between leptin and physical activity in a
sample of minority female children; and 3) to assess the effects of a high sugar
versus a high fiber meal on activity levels in an in-lab setting in minority
adolescents.
The participants from Study 1 were drawn from a nationally representative
sample of youth ages 6-19 years who participated in the 2003-2004 and 2005-
2006 National Health and Nutrition Examination Surveys (NHANES). The
samples were combined to allow estimates of activity levels across race/ethnicity,
weight status, age, and gender groups (N=3,106). The participants from Study 2
(N=50) were Hispanic and African American females ages 8-11 years who
participated in the Transitions Study, a longitudinal study on the factors
contributing to the observed decline in moderate to vigorous physical activity
(MVPA). The participants from Study 3 were overweight or obese (BMI ≥ 85
th

percentile) Hispanic males and females aged 14 to 17 years who participated in
a cross-over in-lab feeding study designed to assess the acute effects of two
xi

different meals (high sugar vs. high fiber) on activity levels and insulin/glucose.
All three studies use accelerometers to measure activity levels.
In Study 1, the 6- to 11-yr-olds spent more time (88 min/d) in MVPA than
the 12- to 15-yr-olds (33 min/d) and 16- to 19-yr-olds (26 min/d; p = .001 for
both). Females spent fewer minutes per day in MVPA than males (p = .001).
Overall, obese youth spent 16 fewer minutes per day in MVPA than normal-
weight youth. However, non-Hispanic white males spent three to four fewer
minutes per day in vigorous physical activity than Mexican American (MA; p =
.004) and non-Hispanic black (p = .001) males but had lower obesity rates.
Obese 12- to 15-yr-old Mexican Americans recorded similar minutes in MVPA
per day as normal-weight Mexican Americans (p = .050). There was a significant
three-way age–body mass index–race/ethnicity interaction for mean minutes per
day in MVPA (p = .001) such that although there were differences in physical
activity in younger age groups, levels declined in the oldest age group so that
youth ages 16-19 of all race/ethnic groups spent the same amount of time
(between 24 and 29 minutes) in MVPA (p>.050). In Study 2, there were pubertal
differences in leptin levels such that girls in Tanner stage 1 had lower levels of
leptin (p= .004) than girls in Tanner stage 2. Leptin was negatively associated
with MVPA, but not Tanner stage, independent of adiposity in the cross-sectional
model (p = .013). MVPA declined by 12.3% (6.4 min/d) over one year. In the
longitudinal model, baseline leptin predicted the decline in MVPA over one year
(p = .017) independent of central adiposity and pubertal stage. In Study 3, meal
xii

condition predicted change in insulin and glucose IAUC (p< .001 for both) but not
activity levels over the observation period. Insulin IAUC predicted MVPA, but did
not mediate the meal condition-MVPA relationship. There were also differences
in insulin and glucose IAUC at specific time points (p< .001 for all): these values
were elevated in the high sugar condition at 30- and 60- minutes post-meal.
In conclusion, this dissertation lent support for individual demographic and
biological basis for physical activity. In Study 1, females and older youth were the
least active groups. Obese youth were generally less active, but this did not hold
uniformly across race/ethnic groups. In Study 2, the inverse leptin-MVPA
relationship was stable over time in minority females. Pubertal stage was not a
factor in this relationship. In Study 3, meal condition caused change in insulin
and glucose but not activity. The findings from Study 1 support individual
demographic variations in activity. The findings from Studies 2 and 3 support a
biological basis of activity. These findings inform the understanding of the
individual, biological, and dietary factors that are related to the decline in physical
activity in youth.
1

CHAPTER 1 INTRODUCTION
This dissertation investigates how individual demographic, biological, and
dietary factors influence objectively measured physical activity levels in youth.
Study 1 investigates activity levels in a nationally representative sample of youth
across race/ethnic, age, weight, and gender g roups. Study 2 investigates the
cross-sectional and longitudinal influence of a single biomarker on physical
activity in Hispanic females. Study 3 is a laboratory-based study that investigates
the longitudinal influence of specific nutrients on activity, insulin, and glucose
levels over a four-hour observation period.
Background and Significance
Pediatric Obesity in the United States
Pediatric obesity is one of the foremost public health issues in the US
today because of its substantial impact on health outcomes and rising health
care costs. In youth, overweight (BMI ≥ 85
th
percentile) and obesity (BMI ≥ 95
th

percentile) are categorized according to the age- and gender- specific growth
charts published by the Center for Disease Control and Prevention [102].
Overweight and obesity prevalence increased by 2.2% in female and by 4.2% in
male children and adolescents from 1999 to 2004 [134]. There are long-term
consequences of the rising obesity prevalence as overweight youth are more
likely to be overweight as adults [225], and are at an increased risk for chronic
2

conditions such as type 2 diabetes, high cholesterol, and the Metabolic
Syndrome [92, 174, 218].
Minority populations are disproportionately affected by this public health
problem. Recent national surveillance data from 2003-2006 report that 21.1% of
Hispanic and 22.9% of African American youth are obese compared to 16.0% of
non-Hispanic Whites [135]. Hispanics are the fastest growing minority group in
the US, and were projected to comprise 39% of the nation‘s population growth
from 2000 to 2010 while African Americans were expected to double their
population size between 2000 and 2010. It was projected that the contribution to
the nation‘s population growth by Caucasian Whites would decrease from 35%
between 1990 and 2000 to 23% between 2000 and 2010 [28]. Furthermore,
obesity-related health care costs in youth tripled from 1979 to 1999 [63]. The
significant shift in population demographics and increasing disease burden
underscore the public health implications of obesity. With the health care system
under growing strain, understanding the factors contributing to the obesity
epidemic in minority youth is a public health priority.
Cultural and income-related factors play a role in differences in obesity
risk and prevalence between race/ethnic groups. In a review of forty-five cross-
sectional studies, socioeconomic status (SES) was found to be inversely
associated with adiposity in youth [182]. This finding may be partially due to the
income-related disparities in the availability of healthy food and opportunity for
physical activity [36]. Lovasi et al. (2009) recently reviewed research on the
3

environmental influences on obesity in disadvantaged populations [111]. The
authors found evidence that disadvantaged populations, who were primarily of
low-SES and Black or Hispanic race/ethnicity, were living in environments with
limited food stores and exercise facilities. These disparities in access to healthy
behavioral choices promoted obesity in these groups. In a national study, Powell
et al. (2006) found that higher income areas were more likely to have physical
activity facilities compared to lower income areas and areas with higher
proportions of Hispanic and African American residents [156]. Fast food
restaurant availability is also higher in minority communities, particularly in
African American neighborhoods, and may contribute to race/ethnic differences
in obesity prevalence by influencing the availability of healthy foods [155].
However, even after controlling for SES, the racial/ethnic disparities persist, with
African American youth having the highest obesity prevalence [74]. This
suggests that apart from income and cultural differences, there are other factors
driving the race/ethnic disparities in obesity that make minorities an important
population to study.
Physical activity has been identified as a means to prevent and treat
obesity in minority youth [103]. Current physical activity guidelines recommend at
least 60 minutes per day of moderate to vigorous physical activity (MVPA) for
youth with three or more days per week that include muscle- and bone-
strengthening physical activity [214]. However, physical activity and the
proportion of youth adhering to the guidelines decline with age. Using
4

accelerometer-measured estimates of MVPA in a nationally representative
sample of 4,867 participants, only 42% of children ages 6 to 11 years were
meeting the current physical activity recommendations. This declined in to 8% in
the 12 to 15 and 7.6% in the 16 to 19 year olds [206].
There is strong evidence that low levels of activity are related to obesity.
However, other factors also contribute to obesity risk at the population level. Diet
is a major contributing factor; not only in terms of excess caloric intake, but also
in terms of food type. Specific nutrients, such as added sugars, can affect
adiposity and related outcomes [46, 217]. Higher intakes of sugar sweetened
beverages have been related to greater risk for long-term weight gain in youth
[88]. Also, data from 1998 to 2002 of the National Health and Nutrition
Examination Survey (NHANES) indicated that youth who met the criteria for
central adiposity (waist circumference ≥ 85
th
percentile) reported less dairy, grain,
and fruit and vegetable intake [21]. Although diet and physical activity regulate
energy balance, it is unclear to what extent they influence each other. However,
our pilot work [189] and the work of others [93, 198, 200] suggest that diet and
specific nutrient intake may influence activity levels. In a sample of 472 youth
ages 10 to 14 years, MVPA and energy intake were significantly positively
correlated after controlling for age [69]. However, diet was not found to be a
factor in the MVPA-adiposity relationship in a sample of 38 Hispanic adolescents
[32]. The mechanism through which diet works to influence physical activity is not
well understood and most studies simply control for diet rather than investigate it
5

as a mediator or moderator of activity. These relationships should be elucidated
because if diet adversely influences activity levels, interventions targeting
increasing physical activity will also need to take diet into account.
There is some debate about the relative importance of physical activity
compared to diet in the etiology and treatment of obesity. Some researchers
suggest that is easier to close the energy gap that contributes to obesity through
small decreases in caloric intake [153, 220] than it is to achieve the magnitude of
increase in physical activity that will counteract intake to prevent weight gain or
achieve meaningful reductions in adiposity [29]. However, physical activity has
been shown to be a salient factor in metabolic health and obesity prevention
efforts in youth [79]. Research from several populations indicates that heavier
youths tend to consume fewer calories than their normal weight counterparts [69,
192], although it is unclear whether this is due to underreporting. This finding
may also indicate that while excess intake can lead to overweight, intake may
decline with time irrespective of physical activity behaviors. For example, there is
evidence that weight control behaviors increase as weight increases, perhaps
accounting for decreased energy intake [40, 181, 201]. Also, some researchers
have shown that physical activity is a better predictor of adiposity than diet [69,
121] and has beneficial health outcomes such as lower Metabolic Syndrome
scores and increases in insulin sensitivity independent of adiposity [106, 163,
186]. Youth who are physically inactive over their lifespan are more likely to be
obese and have higher all-cause mortality rates in adulthood [114]. Physical
6

inactivity also increases risk for certain cancers such as breast, endometrial,
colon, and liver [33]. Public health efforts continue to investigate the roles of
physical activity and diet in obesity prevention. However, the lack of longitudinal
objective physical activity and diet measures hinders our knowledge of the
contributions of these behaviors to obesity on the population level.
Physical Activity in Youth
Physical Activity Measurement
The findings on the relationship between physical activity and obesity are
affected by the variability in the measures used to assess activity. Valid and
reliable measures are needed for researchers to accurately measure physical
activity; however the general consensus in the field is that measures are study
population- and resource- dependent [190]. Self-report measures such as the
Previous Day Physical Activity Recall (PD-PAR) are frequently used in
epidemiological studies since they are cost effective and have been proven valid
and reliable in youth [37]. However, self-report measures are not highly
correlated with biological markers of obesity, and are open to desirability and
recall biases. It is also necessary to recognize that these measures of physical
activity can be difficult to use with children, who often rely on adults to estimate
their physical activity levels because they have trouble recalling activity [48].
Furthermore, there is a wide array of self-report instruments available that vary in
validity and reliability, many of them measuring different aspects of activity, which
7

makes it difficult to compare results across studies that have used different self-
report measures [208].
Physical activity has been extensively studied in youth using self-report
measures. Corder et al. (2009) reported on the relationship between several self-
report and objective measures in three age groups (4-5, 12-13, and 16-17 years
of age) [39]. The strength of association between self-report and objective
measures varied from .09 to .46, with greater underreporting at higher activity
levels. The authors concluded that while all self-report measures were valid at
categorizing MVPA, the strength of correlation with objective measures
depended on the age group and individual-level error was larger for some
measures than others. Thus, objective measures may be more appropriate to
use with youth populations. That youth tend to misestimate time spent in physical
activity, may account for the inconsistent relationships seen with health outcomes
[185]. With the development of more objective measures of activity, relationships
found between MVPA and health outcomes have become more consistent [98].
Accelerometry is one of the most popular objective measures and an
increasing number of studies are using accelerometry to measure physical
activity. Accelerometry may be more feasible in youth because of the low
participant burden, and because the devices eliminate recall and social
desirability biases. PD-PAR and accelerometry data have a correlation of
approximately 0.40 [55, 147, 224], indicating that there is only moderate
agreement between a self-report and an objective measure of physical activity.
8

Accelerometers provide the researcher with more control over the amount of data
collected, without influencing the participants since the researcher specifies the
sampling frequency, or epoch, and the number of days it will be worn. The
smallest epoch that can currently be recorded on most accelerometers is 1-
second, and the shorter sampling frequency yields more precise estimates of
activity, particularly in youth because they tend to participate in more short bouts
of activity [208]. However, short epoch lengths do not allow for multiple days of
monitoring in some accelerometer models because the high rate of data
collection uses up battery and storage capabilities. This may result in biased
estimates of activity. Although previous studies have used a wide range of days
of monitoring, 4 to 7 days is considered to be a valid estimate of actual daily
activity in youth [209]. There are several brands of accelerometers. The
Actigraph is the most frequently used and thoroughly validated in youth. The
Actigraph 7164 accelerometer is used in all of the studies in this dissertation
proposal. This monitor adequately correlates with energy expenditure measured
via doubly labeled water [154]. A significant strength of this dissertation is that
accelerometers are used to measure activity across three separate study
populations, providing for comparable estimates between study populations.
Accelerometers can be used to estimate time spent in different intensity
levels. The most basic outcome is counts (e.g.: cpm, total counts), which are a
measure of vertical acceleration and provide an estimate of activity intensity.
Counts are considered the least-processed measure gleaned from the data
9

because they do not require user-defined thresholds. User-defined thresholds
are used to generate estimates of time spent in varying intensity levels (e.g.:
minutes spent in MVPA, VPA). Several calibration equations that identify
intensity thresholds have been developed in multiple youth populations [68, 145,
160, 161, 205, 213]. Accelerometry data is most often reported as time spent in
MVPA and sedentary behaviors.
A growing body of research supports the idea that MVPA and sedentary
behavior are distinct behaviors. Physical activity is conceptualized as bodily
movement produced by the contraction of skeletal muscles and results in
increases in energy expenditure [207]. Sedentary behavior represents a range of
behaviors that do not produce energy expenditure above 1.5 times the resting
rate [140]. As Reilly et al. (2008) note in their review, sedentary behavior ―is not
simply the absence of physical activity, but involves purposeful engagement in
activities that involve minimal movement and low energy expenditure‖ [166] (p.
614). Earlier research tended to view sedentary behavior as lack of physical
activity, or physical inactivity. However, sedentary behavior does not simply
‗displace‘ physical activity [140]. In the first description of time spent in sedentary
behavior in the US population using accelerometry, Matthews et al. (2008)
reported that youth spent approximately six to eight hours per day in sedentary
pursuits [119]. Furthermore, sedentary behavior appears to be a class of
behaviors that coexist with physical activity and that demonstrates clear behavior
profiles based on the types of activities selected [16]. With increasing recognition
10

of sedentary time as a behavior rather than physical inactivity, researchers have
begun to develop methods for measuring this behavior [190]. Although
accelerometry provides better estimates of time spent in sedentary behavior, it
does not provide any contextual information about the types of activities
performed. However, self-report instruments are valuable tools that provide
contextual information about the activities that constitute ‗sedentary behavior‘
and this data has been used in the justification for treating sedentary behavior as
a unique behavior.
Physical Activity and Individual Demographic Factors
Age. Cross-sectional [30, 206] and longitudinal studies [149] have found
an age-related decline in physical activity. Research suggests that this decline in
activity begins around the start of puberty in girls and shortly after in boys. There
are gender differences in the trajectory of decline, and girls tend to have a
steeper decline than boys [146]. In a cohort of 501 females age 12 at baseline,
there was an annual 4% (-1.76 min/d) decrease of MVPA measured via
accelerometry over two years [149] that was accompanied by a corresponding
increase in time spent in sedentary behavior [203]. It is difficult to determine what
accounts for the age-related decline in physical activity; however the fact that it
occurs across many species points towards potential biological mechanisms [89,
176]. It may be that age is a proxy for puberty onset, which has been related to
decreases in activity [7, 47] and will be discussed later in more detail. This
dissertation contributes to the literature describing the age-related decline both
11

cross-sectionally and longitudinally. The cross-sectional study in Study 1
provides estimates of differences in activity between age groups and the
longitudinal study in Study 2 provides evidence on the influence of age and
puberty on the activity decline.
Gender. It is well known that there exist gender differences in physical
activity levels. Girls have lower levels of MVPA and higher levels of sedentary
behavior than boys [31, 211, 216]. Using NHANES data, accelerometry-
assessed physical activity indicated that female youth ages 12 to 15 spent
approximately 20 minutes fewer min/d than males in MVPA [206], and
significantly more min/d in sedentary behavior [119]. In a longitudinal study of
1,032 youth age 9 years old at baseline, females fell below the physical activity
recommendations during weekdays at 13.2 years compared to 14.9 years for
males [132], indicating that the trajectory of age-related decline in physical
activity is influenced by gender. Socio-environmental factors such as differences
in preferences of leisure time activity and fewer opportunities to be active for
females have also been implicated [116].
Pubertal Status. Pubertal status, independent of age and gender, has
been strongly implicated as a cause for the differences in physical activity. Early
maturation has been associated with decreased physical activity in girls [47, 57].
In a longitudinal study with 143 females, Baker et al. (2007) reported that early
maturing females had lower levels of physical activity than late maturing females.
Furthermore, at the 2-year follow-up, early maturing youth recorded fewer
12

minutes in MVPA than late maturing youth [7], suggesting that timing of initiation
of puberty influences physical activity later in puberty. Among those that mature
at a normal age, females in more advanced Tanner stages spend less time in
MVPA [226].There are biological factors (such as increasing growth hormone
and estrogen levels) that are activated during the start of puberty that can impact
physical activity levels, particularly in females [33]. To date, the majority of
studies have employed cross-sectional study designs. This dissertation will
contribute to the literature through its use of a longitudinal design with frequent
assessment points that can provide a view of the trajectory of decline in activity.
Furthermore, this dissertation attempts to determine the differences in the factors
related to physical activity decline in females according to Tanner stage, which
will inform how pubertal status modifies physical activity as well as the
relationships between predictors of physical activity and the behavior itself.
Race/Ethnicity. There are conflicting findings on the differences in physical
activity across race/ethnic groups. Previous studies using self-report measures of
physical activity indicated that Caucasian White youth were the most physically
active group [4]. However, these differences have not been consistently
observed in studies that used accelerometry to measure activity. In a large cohort
of 1,578 sixth grade females, Caucasian Whites recorded more accelerometer-
measured minutes per day in MVPA than African American and Hispanic youth
[148]. Conversely, Owen et al. (2009) reported that compared to Caucasian
Whites, Black children recorded significantly more counts per minute of activity
13

[139]. Differences in activity levels among race/ethnic groups may be explained
in part by developmental differences. For example, African Americans appear to
begin puberty earlier than Caucasian Whites and Hispanics. In a nationally-
representative sample, non-Hispanic black girls began puberty significantly
earlier than non-Hispanic whites and Mexican American girls (mean ages: 10.3,
11.0, and 11.2 years, respectively) [196]. A similar pattern was observed in boys
(mean ages: 11.5, 11.8, and 12.2, respectively). The start of puberty is strongly
related to a decline in physical activity [57]. Also, environment may account for
lower activity levels in Hispanics and African Americans, who are more likely to
live in urban areas with less neighborhood access to recreational facilities and
more exposure to violence [104]. The first study of this proposal addresses the
discrepancies in findings concerning racial/ethnic differences in activity levels
through its use of objectively measured activity while controlling for other factors
such as SES and diet that have been related to race/ethnicity. This will allow
researchers to gain a better understanding of basic differences in activity across
race/ethnic groups that can inform future research into the potential mechanisms
behind these differences.
Weight Status. Previous cross-sectional studies have found that
overweight and obese youth are less physically active than their normal weight
counterparts [30, 31, 212]. Findings from longitudinal studies report that
increased levels of physical activity are related to decreased adiposity in youth.
Butte et al. (2007) followed a large Hispanic cohort of youth aged 4 to 19 years
14

and found that over one year, increased time spent in sedentary behavior and
decreased time spent in light activity were related to increases in weight [30]. In
an analysis using NHANES data, Mark & Janssen (2009) found a dose-response
relationship between increasing bouts of MVPA and BMI, where longer bouts
(defined as ≥ 10 minutes) of activity were significantly related to lower BMI
percentiles [118]. Metallinos-Katsaras et al. (2007) reported that less time spent
in very vigorous physical activity (defined as > 5331 cpm) was related to higher
odds of overweight and that overweight children participated in two minutes less
per day of very vigorous physical activity than normal weight children [125].
However, the number of minutes and intensity of activity that impact weight
status is not currently defined and some studies have not found inverse
associations between physical activity and adiposity in youth [168, 186, 194,
223].
The abovementioned relationships may be sensitive to the type of
adiposity measure used [60, 109]. Body mass index (kg/m
2
; BMI) is frequently
used as a measure of general growth and health but also to determine weight
status [84]. Multiple studies have investigated the utility of using BMI as a
measure of adiposity in youth. Some have focused on discordance between BMI
and other epidemiological measures such as waist to hip ratio (WHR) and BMI z-
score while others have looked at more specialized measures such as dual-
energy x-ray absorptiometry (DEXA). For example, Lohman et al. (2006)
reported that body fat had a higher correlation than BMI with MVPA in a 6
th
grade
15

sample of females [110]. BMI percentile attempts to address these differences by
taking age and gender into account [102], but shows weaker associations with
body fatness in youth than BMI [62]. There also appear to be differences in the
reliability of BMI in certain weight status groups, and therefore it may not be a
good measure of body composition in those at the extreme ends of the spectrum
(i.e.: the underweight and obese) [75]. Race/ethnic differences in BMI have also
been observed in youth. African American youth have lower body fat estimates
than Caucasian White youth of the same BMI [42, 67]. The estimates of body fat
variations by race/ethnicity vary from 3% to 5% in those of the same BMI.
Subsequent studies indicated that body build accounted for part of the variance
in the percent body fat-BMI differences across race/ethnic groups [51]. Using
data from the third National Health and Nutrition Examination Survey (NHANES),
Mei et al. (2002) found that BMI was adequate for detecting overweight when
total fat mass and body fat were used as the criteria measures [124]. This finding
is consistent with other studies in youth that support the use of BMI as a proxy
measure for adiposity in population studies [14, 50].Despite its limitations, BMI is
a cost-effective proxy measure for adiposity [52] that is feasible in
epidemiological studies and it has been shown to be significantly correlated with
percent body fat in youth [64, 152].
In addition to BMI, growth charts delineating BMI percentiles are
frequently used to categorize youth into groups (e.g.: overweight, obese) by
weight status [102]. The CDC growth charts are age- and gender- specific for
16

children and adolescents ages 2 to 20 years. They were developed using data
from five nationally representative cross-sectional health surveys. The primary
purpose of the percentile zones was for the clinical monitoring of growth and the
secondary purpose was for the definition of overweight and obesity [137]. BMI
values are calculated and plotted on a BMI-for-age growth chart for boys and
girls. Normal weight is considered to be a BMI that falls between the 5
th
and 84
th

percentiles. Overweight is a BMI that falls between the 85
th
and 94
th
percentiles,
and obese is a BMI above the 95
th
percentile [102].
There are several other adiposity measures available that can be
separated into two main categories: region-specific body fat and whole body
composition. Central or abdominal adiposity is an example of region-specific
body fat, and can be measured using specialized techniques such as MRI. Total
body fat mass is an example of whole body composition and can be measured
via methods such as DEXA and subscapular skinfold measurements [71]. DEXA
is considered to be a more precise measure of body composition because it can
distinguish between bone, and fat and lean mass through the tissue-specific
attenuation of the x-rays. DEXA is an increasingly popular measure of total body
fat that is easy to use in a smaller scale study or clinical setting. Its weaknesses
are that it is not highly portable and confers some exposure from the x-rays to the
participants, yet this measure is often used as a criterion measure of body fat in
youth. Because it is considered an ‗objective‘ measure of body composition,
there are not age, gender, and race/ethnic differences in the estimates. Region-
17

specific measures of body composition are typically more specialized than those
used for total body composition. Early studies derived estimates of region-
specific abdominal fat from simple anthropometric measures such as WHR. More
recent studies use MRI scans to estimate abdominal fat content, specifically
subcutaneous and visceral adipose tissue. These estimates provide information
on fat content surrounding organs, and have been associated with type 2
diabetes and obesity-related health outcomes in youth [41]. Because there are
race/ethnic variations in fat distribution, region-specific measures will yield better
estimates of fat distribution and health outcomes in ethnically diverse populations
[78, 108].
Despite the intense research on the topic, there is a lack of consensus in
the field about the ‗best‘ adiposity measure. Choice of measure depends on the
sample size, study budget, and research questions being investigated. In large
epidemiological studies, BMI is a cost-effective adequate tool for measuring
adiposity in most youth populations in that it does display adequate correlations
with other body composition measures. However, BMI may not be the most
appropriate measure to estimate adiposity in obese or underweight samples
therefore other measures of body composition may be more appropriate in these
youth. When investigating the relationship between fat distribution and health
outcomes in diverse populations, region-specific measures are preferable. Study
1 is a large epidemiological study and therefore BMI percentile was used to
categorize participants into weight status groups. Studies 2 and 3 have smaller
18

sample sizes, and are in predominantly overweight race/ethnic minority
populations. Region-specific MRI measures were used in Study 2 and DEXA was
used in Study 3. A strength of this proposal is the strong adiposity measures
used that is appropriate for the study populations being investigated.
Physical Activity and Selected Hormones
Overview of the biological basis of physical activity
There is limited evidence documenting the biological and metabolic
correlates of objectively measured physical activity in youth [177]. While there is
growing evidence of a biological basis for physical activity, it is still usually
viewed as a behavioral choice. In a seminal review, Rowland (1998) proposed
that physical activity is regulated by an internal homeostasis or ‗activity stat‘
center that is synchronized by the central nervous system and governs the level
of activity in an individual [173]. Bouchard and Rankinen (2006) supported the
idea that biology influences physical activity levels because models that don‘t
include biological variables only account for a modest portion of variance in
behavior [20]. Support for these views come from studies on the genetic
influences on activity, which indicates that certain genes may mediate the
response to exercise [26], thus accounting for individual differences in response
to activity. It is estimated that the heritability of physical activity is between 18-
69% [57], with the majority of studies falling at the higher end of the spectrum of
explained variance. Thus, there is evidence to suggest that physical activity is not
merely a behavioral choice, but to a significant extent a biologically based
19

behavior. However, it is also important to consider gene-environment interactions
when investigating the biological basis of activity. Evidence from twin studies
indicates that genetics account for a large part of explained variance in physical
activity behavior. In a sample of 411 Caucasian White twins ages 12 to 25 years,
genetic factors explained 68.4% and 39.8% of the variation sports participation in
males and females, respectively [115]. However, the shared environment
explained only 20.0% and 28.4% of variance in activity in males and females. In
a study of 8,355 adolescent twins (ages 13 to 19 years), the monozygotic twin
correlations were significantly higher than the dizygotic twin correlations,
suggesting that individual differences in exercise behavior chiefly arise from
genetic factors [215]. However, the results also indicated that shared
environment accounted for more variation in exercise in the 13-14 year olds than
the older age group. Other work has also found that the importance of the
environment varies with age [26, 151], suggesting that there are gene-
environment interactions that influence behavior. Although these studies are
valuable, they do not allow for the separation of effects of multiple environmental
exposures and therefore are limited in explaining interactions between genes and
environments.
Several physiological systems are involved in physical activity among
which are energy metabolism and hormone regulation. Physical activity results in
improved insulin action and glucose metabolism by stimulating muscle to
increase glucose uptake and storage during an insulin-mediated increase in
20

blood flow through skeletal muscle [219]. Physical activity also results in
increased insulin-mediated glucose uptake in adipose tissue, insulin sensitivity,
and hepatic storage capacity for glucose [18]. These effects of physical activity
have many health benefits such as increasing fitness and reducing
cardiovascular disease and metabolic syndrome risk [49].
Overview of hormonal energy balance regulation
Figure 1-1: Overview of selected biomarkers as they relate to energy balance

There are physiological factors that regulate energy intake and output
behavior, which are responsible for fat mass and body weight. Figure 1-1
provides a simplified schematic of the role of specific hormones and their role in
energy balance regulation. There is evidence that physical activity is inversely
related to adipose tissue mass [131]. Leptin is secreted in direct proportion to
adipose tissue mass, thus higher amounts of adipose tissue are associated with
21

increased circulating leptin levels [9]. Higher circulating leptin levels are related to
lower physical activity [170]. After energy intake, insulin secretion increases and
stimulates the production of leptin. The combined increase in insulin and leptin
results in decreases in physical activity until the absorption of food is completed
and leptin suppresses insulin secretion through the adipoinsular biofeedback
loop [100]. The influences of insulin and leptin on energy balance are mediated
through brain reward pathways that differ by weight status. In normal weight
individuals, leptin completes a negative feedback loop, that when energy intake
is sufficient, signals to the brain that there is sufficient intake and body fat stores
[5]. In obese individuals, this response is blunted. There is a decreased response
to leptin (termed ‗leptin resistance‘) such that energy intake continues and may
be motivated by external sensory cues to food (such as palatability) rather than
the biological feedback loop [19]. In other words, while the feedback system is
sensitive to maintaining energy balance and fat stores in normal weight
individuals, it is not as sensitive to energy surplus in obese individuals [25].
Insulin & Insulin Resistance. Insulin levels are inversely related to physical
activity in humans [5]. In a sample of 589 children (mean age 9.7 (0.4) years),
accelerometer-measured physical activity counts were inversely correlated with
fasting insulin [23]. Excess fat through energy imbalance can result in sustained
insulin secretion in the body‘s attempt to prevent excess glucose in the blood.
Eventually, this may lead to insulin resistance. Insulin resistance is the process
whereby insulin-responsive tissues are no longer sensitive, forcing the body to
22

produce more insulin to achieve the same magnitude of glucose uptake, and
resulting in elevated levels of circulating insulin [228]. The body‘s insulin
receptors are saturated, and insulin resistance can develop. Several studies
have investigated the relationship between insulin resistance and physical
activity. One year after a focused two-week intervention that included a physical
activity component, Raman et al. (2010) showed decreases in insulin resistance
in African American boys [163]. A longitudinal study found that decreases in
physical activity were associated with insulin resistance (measured via HOMA-
IR) [94]. Onset of puberty also plays a role in insulin dynamics. Insulin resistance
increases and insulin sensitivity (SI) decreases during puberty [73, 87, 101] at
the same time that physical activity decreases. Insulin sensitivity (SI) is a
measure of insulin-stimulated uptake of glucose derived from blood glucose and
insulin concentrations taken during a frequently sampled intravenous glucose
tolerance test (FSIVGTT) [13]. This simultaneous pubertal decline in insulin
sensitivity and physical activity suggests that physical activity may be influenced
by insulin concentrations. The proposed dissertation (Study 3) will investigate
whether insulin indices predict activity levels or mediate the relationship between
meal and activity.
Glucose & Glucose Intolerance. During physical activity, glucose (stored
as glycogen in the muscle) from the bloodstream or muscle tissues is consumed
for energy. Over time, physical activity can help expend excess calories, increase
insulin sensitivity, and decrease blood glucose levels thereby preventing glucose
23

intolerance. Glucose intolerance is an indication of beta cell dysfunction and
insulin resistance whereby beta cells are unable to produce enough insulin to
uptake glucose resulting in prolonged exposure to glucose in the bloodstream
[228]. Glucose intolerance can lead to higher risk of cardiovascular, kidney, and
eye disease [228]. The direct relationship between physical activity and blood
glucose has been assessed in youth, primarily through cross-sectional studies.
Ekelund et al. (2007) reported significant negative correlations between fasting
glucose and accelerometry-measured MPA, VPA, and a significant positive
correlation with sedentary behavior [58]. Davis et al. (2009) reported a significant
decrease in fasting glucose concentrations in 41 Hispanic females after a 16-
week combination aerobic and strength training intervention [45]. This work
illustrated the stimulating influence that physical activity can have on glucose
uptake in the skeletal muscle and liver. However, little work has focused on the
meditational effects of glucose on physical activity and diet. This proposal will
attempt to add to this literature in Study 3 by investigating whether glucose
mediates the relationship between food intake and physical activity or if it directly
affects activity post meal.
Leptin. Leptin is a gut hormone secreted in proportion with adipose tissue
mass and is involved in energy balance regulation [5, 117, 230, 233]. The excess
adipose tissue found in obese individuals produces high circulating leptin levels
that saturate the leptin receptors. The diminished capacity of the leptin receptors
results in a downregulation of leptin transporters and a failure of the system to
24

transport leptin across the blood-brain barrier where it reaches its targets in the
brain [9, 96]. Thus, leptin becomes less effective at influencing behavior in ‗leptin
resistant‘ individuals. Leptin resistance is related to inflammation [191],
cardiovascular disease risk, and obesity in youth [90]. There are observed
gender differences in leptin levels. Evidence suggests that plasma leptin levels
are higher in females than males independent of body fat and Tanner stage
[232]. In a longitudinal weight loss study with 115 youth, leptin levels decreased
with weight loss independent of baseline adiposity in both genders, but females
had a 51% decrease in leptin over 12 weeks whereas males had a 39%
decrease [85]. The authors suggested that body composition differences may
have influenced this gender differences.
The relationship between physical activity and leptin is not well understood
in youth. In a cross-sectional sample of 510 youth aged 8 to 18 years, leptin was
negatively correlated to steps per day in females only [170]. Conversely, Salbe et
al. (1997) found in 125 Pima Indian children that plasma leptin concentrations
were positively associated with physical activity levels [175]. Few studies have
determined the effect of leptin on long-term physical activity in youth. In a
Caucasian White adult sample, Franks et al. (2007) found that individuals with
leptin concentrations below the sex-specific median increased physical activity
levels by 35% more over 5 years than those with above-median leptin levels [66].
However, this relationship was not found in a longitudinal study of 213 healthy
children, where physical activity did not significantly correlate with leptin levels
25

[126]. In summary, obesity is related to energy imbalance, and leptin is a central
hormone that has many direct and peripheral effects related to energy balance
that makes it a salient hormone to study. Leptin influences energy balance
metabolism, the timing of puberty, insulin action in the liver, glucose homeostasis
in the bloodstream, and lipolysis in adipose tissue [117]. Research indicates that
leptin concentrations increase at the onset of puberty, particularly in females,
once fat stores reach a sufficient level and promote higher circulating leptin
concentrations [86]. These higher leptin concentrations during the onset of
puberty may be related to the observed decline in physical activity and increase
in insulin resistance [1, 167]. However, as previously stated, the relationship
between physical activity and leptin in youth is unclear. This may be due to the
use of several different physical activity measures employed in the studies. More
research using valid and reliable measures of physical activity (such as
accelerometry) is needed to understand how an important hormone that plays a
central role in energy balance is associated with physical activity in youth. The
longitudinal data collected in Study 2 using objective physical activity and
adiposity measures will allow me to test if baseline cross-sectional relationships
between leptin and physical activity remain stable in a longitudinal analysis over
one year, which will contribute to the literature gap in this area.
26

Physical Activity and Diet
Along with low levels of physical activity, poor diet, including insufficient
fruit, vegetable and fiber intake, excessive caloric intake and increased added
dietary sugar, is a major contributor to obesity in youth [127]. In an influential
review, Sallis (2000) found that diet and physical activity were correlated in
children [177]. Overweight youth tend to spend less time in physical activity and
eat poorer diets than normal weight youth [150]. While research has been
conducted on the singular contributions of diet and physical activity to obesity,
less work has addressed the direct influence of diet on physical activity levels in
youth. In a sample of predominantly normal weight youth, MVPA measured via
accelerometry was positively associated with total energy intake in boys, but not
in girls [93]. Furthermore, few studies have examined the effects of specific
nutrients on physical activity despite the fact that specific nutrients, specifically
added sugar, have been shown to influence other behaviors [198]. Diets rich in
simple carbohydrates and fats have been associated with lower levels of physical
activity [198]. In a sample of 210 African American females, MVPA measured via
accelerometry was negatively associated with percentage of calories from fats
and carbohydrates after controlling for total caloric intake [91]. Conversely,
Thompson et al. (2004) reported no significant covariation in changes in physical
activity and specific foods in African American girls over 12 weeks [200], but the
authors noted that this relationship may have been influenced by the small
sample size and short study duration.
27

In-lab feeding studies have attempted to elucidate the effects of specific
nutrients on physical activity by offering a controlled environment and
opportunities for using measures that are not feasible in free-living studies. To
our knowledge, only one in-lab feeding study has assessed the relationship
between specific nutrients and physical activity in minority youth. Spruijt-Metz et
al. (2009) reported that a high sugar/low fiber meal resulted in significantly
different patterns of activity over two hours of observation. Compared to the low
sugar/high fiber meal, there was a significant burst of activity in the first 30 to 60
minutes post- high sugar/low fiber meal [189]. In-lab studies conducted in non-
minority youth have found that foods high in refined carbohydrates resulted in
increases in subsequent food intake and insulin and glucose responses [8, 113,
221], but none except our earlier pilot work [189] have assessed the effect on
physical activity.
In conclusion, this dissertation uses the same physical activity measure to
investigate individual demographic, biological, and dietary influences on activity
levels in youth. The three studies allow us to investigate how these factors relate
to hourly, monthly, and yearly differences in activity. Study 1 examines basic
hypotheses about individual-levels factors and activity over the course of
childhood and adolescence. The yearly metabolic measures and quarter-yearly
physical activity measures in Study 2 examines longitudinal influences on
activity. The novel in-lab feeding design in Study 3 investigates the hourly
28

influence of specific nutrients on activity and biomarkers using strong
measurement protocols.
Specific Aims and Hypotheses
Study 1: Physical Activity in US Youth: Impact of Race/Ethnicity, Age, Gender, &
Weight Status
The overall objective of the first study is to report differences in physical
activity by race/ethnicity, age, gender, and weight status in a nationally
representative sample of US youth in order to better understand the relationship
of these different factors to physical activity levels. Specifically, the variations in
physical activity levels by gender, age group, weight status, and race/ethnicity
will be assessed.
Aim 1: To compare physical activity levels in a nationally representative sample
across three race/ethnic groups: non-Hispanic Whites, non-Hispanic Blacks, and
Mexican Americans.
Exploration: As this is the first study to examine activity levels in a nationally-
representative sample of youth using an objective measure of activity,
exploratory hypothesis will test whether there are race/ethnic differences in
activity levels.
Aim 2: To compare physical activity levels across race/ethnic groups by gender.
Hypothesis 2.1: Males will spend more minutes per day (min/d) in MVPA and
fewer min/d in sedentary behavior than females in all race/ethnic groups.
Aim 3: To report on the age-related decline in physical activity.
29

Hypothesis 3.1: Min/d spent in MVPA will be significantly lower in the older (16-
19 and 12-15 years) versus younger age groups (6-11 years).
Aim 4: To compare physical activity levels across race/ethnic groups by weight
status.
Hypothesis 4.1: Obese and overweight youth will record fewer min/d in MVPA
and more min/d in sedentary behavior than normal weight youth in all race/ethnic
groups.
Study 2: The Longitudinal Effects of Leptin on Physical Activity in Minority
Female Children
The objective of the second study is to investigate the longitudinal
relationship between leptin and objectively measured MVPA, and to determine if
Tanner stage is a moderator of this relationship. This analysis will investigate
whether baseline cross-sectional relationships remain stable in longitudinal
analyses in a sample of 8 to 11 year old Hispanic and African American females.
Aim 1: To determine if baseline leptin levels are associated with baseline MVPA
in this sample.
Hypothesis 1.1: At baseline, leptin levels will be negatively associated with
MVPA.
Hypothesis 1.2: At baseline, Tanner stage will moderate the leptin-MVPA
relationship. Compared to Tanner stage 1, girls in Tanner stage 2 will have
higher leptin levels and lower MVPA levels.
30

Aim 2: To determine if baseline leptin levels predict the longitudinal decline in
MVPA.
Hypothesis 2.1: Over one year, baseline leptin levels will predict the decline in
mean min/day spent in MVPA.
Hypothesis 2.2: Those in Tanner stage 2 will have higher baseline leptin levels
and greater declines in MVPA than those in Tanner stage 1.
Study 3: The Influence of Meal Type on Physical Activity in Minority Adolescents:
The Food Adolescence, Mood, and Exercise (FAME) Study
The overall objective of the third study is to determine whether MVPA and
sedentary behavior vary by meal type (high sugar versus low sugar meals) in an
in-lab crossover design feeding study in 14 to 17 year old Hispanic adolescents.
Specifically, this study aims to test differences in physical activity, sedentary
behavior and insulin indices by meal type over five hours during two in-lab visits.
Aim 1: To compare the effects of a high sugar/low fiber meal versus a low
sugar/high fiber meal on minutes spent in MVPA and sedentary behavior.
Hypothesis 1.1a: The high sugar/low fiber meal will result in decreased total time
spent in MVPA compared to the low sugar/high fiber meal.
Hypothesis 1.1b: The high sugar/low fiber meal will result in increased total time
spent in sedentary behavior compared to the low sugar/high fiber meal.
Hypothesis 1.2a: The high sugar/low fiber meal will result in lower levels of
MVPA at each 30-minute increment over five hours compared to the low
sugar/high fiber meal.
31

Hypothesis 1.2b: The high sugar/low fiber meal will result in higher levels of
sedentary behavior at each 30-minute increment over five hours compared to the
low sugar/high fiber meal.
Aim 2: To compare the effects of a high sugar/low fiber meal versus a low
sugar/high fiber meal on insulin and glucose.
Hypothesis 2.1a: The high sugar/low fiber meal will stimulate greater insulin area
under the curve (AUC) than the low sugar/high fiber meal for the entire five
hours.
Hypothesis 2.1b: The high sugar/low fiber meal will stimulate greater glucose
AUC than the low sugar/high fiber meal for the entire five hours.
Hypothesis 2.2a: The high sugar/low fiber meal will stimulate greater insulin AUC
between the 30- to 60- minute increments compared to the low sugar/high fiber
meal.
Hypothesis 2.2b: The high sugar/low fiber meal will stimulate greater glucose
AUC between the 30- to 60- minute increments compared to the low sugar/high
fiber meal.
Aim 3: To investigate whether insulin and glucose mediate the meal effects on
MVPA and sedentary behavior levels.
Hypothesis 3.1a: For the both high sugar/low fiber and low sugar/high fiber
meals, insulin AUC will mediate the relationship between meal type and MVPA.
Hypothesis 3.1b: For the both high sugar/low fiber and low sugar/high fiber
meals, glucose AUC will mediate the relationship between meal type and MVPA.
32

Data Sources and Study Samples
The proposed dissertation studies draw upon samples from three different
research projects, one conducted at the Centers for Disease Control and
Prevention (CDC), and two conducted at the University of Southern California
(USC). The studies conducted at USC were headed by Dr. Donna Spruijt-Metz,
Ph.D. Some detail of the study methodologies are provided here, and more detail
can be found in the individual paper sections.
Paper 1: Physical Activity in US Youth: Impact of Race/Ethnicity, Age, Gender, &
Weight Status
Paper 1 utilizes data from two cohorts of the National Health and Nutrition
Examination Survey (NHANES). NHANES is a cross-sectional nationally
representative health survey of the US civilian, non-institutionalized population
that is collected yearly by the National Center for Health Statistics (NCHS). Data
is collected year-round using a complex, stratified, multistage probability cluster
sampling design strategy on a variety of health factors. In the 2003-2004
collection year, accelerometry was used to measure physical activity, a first in a
nationally representative sample. Physical activity data from the 2003-2004 and
2005-2006 samples were combined for the purpose of having a large enough
sample size to analyze sub-populations. The analytical sample consisted of
3,106 participants aged 6 to 19 years (1,508 from 2003-2004 and 1,598 from
2005-2006).
33

Paper 2: The Longitudinal Effects of Leptin on Physical Activity in Minority
Female Children
Data for paper 2 come from the Transitions Study: Insulin Resistance and
Declining Physical Activity Levels in African American and Hispanic girls.
Transitions is a longitudinal observational study on the physiological and
psychosocial factors related to the decline in physical activity in 8 to 11 year old
Hispanic and African American females across puberty. Data for this analysis is
from the first five data collection time points. The analytical sample consists of 50
(39 Hispanic and 11 African American) participants.
Paper 3: The Influence of Meal Type on Physical Activity in Minority Adolescents:
The Food Adolescence, Mood, and Exercise (FAME) Study
Paper 3 will analyze data from the FAME Study, Project 2 of the Minority
Health Center of Excellence. The purpose of the FAME Study is to investigate
the behavioral and metabolic associations between dietary intake and its effects
on physical activity in normal weight and overweight/obese Hispanic and African
American 14 to 17 year-old adolescents. FAME is an in-lab feeding study that
employs a cross-over design, employing high sugar/low fiber and low sugar/high
fiber meals with a two to four week washout period. The participants wear
accelerometers in the lab for eight hours and blood draws are performed every
30 minutes for the initial five hours, with a second meal at hour five and three
more hours of observation with an ad-libitum food tray. The analytical sample will
consist of 50 participants: 25 males and 25 females.
34

CHAPTER 2 PHYSICAL ACTIVITY IN US YOUTH: IMPACT OF
RACE/ETHNICITY, AGE, GENDER, & WEIGHT STATUS
Belcher BR, Berrigan D, Dodd KW, Emken BA, Chou CP, Spruijt-Metz, D.
Physical activity in US youth: impact of race/ethnicity, age, gender, & weight
status. Med Sci Sports Exerc, 2010. 42(12): p. 2211-2221.

Introduction
Surveillance results from 2003-6 indicated 31.9% of United States youth
had BMI greater than or equal the 85
th
percentile [135]. Despite recent data
indicating a plateau in the previously upward trend in overweight and obesity in
U.S. youth, obesity still represents a significant health threat to U.S. youth [134,
135]. Elevated BMI is related to an increased risk for type 2 diabetes, elevated
cholesterol, and high blood pressure in youth [92, 174]. These findings coupled
with the fact that obesity has been shown to track into adulthood [225], place
overweight and obese youth at a significant health risk as they progress into
adulthood.
There is evidence of an inverse relationship between weight status and
physical activity in youth [165]. For example, BMI increases were associated with
more television viewing and declines in moderate to vigorous physical activity
(MVPA) in longitudinal studies [15, 123]. There is also evidence for a decline in
physical activity with age, particularly in adolescent females [132, 206].
However, findings on race/ethnic differences in physical activity are less
clear. While a previous study has reported lower levels of physical activity
(measured via self-report) in non-Hispanic Black youth as compared to their non-
35

Hispanic White counterparts [184], another that measured physical activity via
accelerometry has reported that non-Hispanic White youth are less physically
active than other race/ethnic groups [139]. The objective of this study is to report
differences in physical activity by race/ethnicity, age, gender, and weight status in
a nationally representative sample of US youth in order to better understand the
relationship of these different factors to physical activity levels.
Methods
Sample
Data for this study are from the National Health and Nutritional
Examination Survey (NHANES), a cross-sectional representative health interview
survey of a U.S. civilian, non-institutionalized population that is collected by the
National Center for Health Statistics (NCHS) of the Centers for Disease Control
and Prevention (CDC). The data are collected year-round from non-
institutionalized individuals using a complex, stratified, multistage probability
cluster sampling design strategy. Details of sampling and data collection have
been reported previously [158, 206].
The present study focused on youth aged 6 to 19 years. The sample
consisted of 1,508 participants from NHANES 2003-2004 and 1,598 participants
from NHANES 2005-2006 survey cycles in which accelerometer measurement
was included. Preliminary data analysis indicated no statistically significant
differences in the outcome variables of interest between the 2003-2004 and
2005-2006 samples. Therefore, the samples were combined. Participants were
36

included if they had no missing demographic or anthropometric data and at least
four days with 10 or more hours of accelerometer data [206]. The final analytic
sample consisted of 3,106 participants between 6 and 19 years of age with
complete accelerometer, demographic, and anthropometric data. Based on
survey design characteristics, results for non-Hispanic White, non-Hispanic
Black, and Mexican American race/ethnic groups were estimable. The NCHS
ethics review board approved the study and written informed consent was
obtained from all participants. The University of Southern California IRB did not
require review for this analysis.
Measures
Demographics: Age (years) was calculated as the time between birth and
examination date. Race/ethnicity was self-reported by participants & categorized
as non-Hispanic White, non-Hispanic Black, Mexican American, Other Hispanic,
and Other. Age and racial/ethnic groups were categorized according to the
NHANES analytic guidelines [157]. Socioeconomic status was calculated using
the Poverty to Income Ratio (PIR), which is a ratio of the household income to
the Census Bureau poverty threshold that is adjusted for family size and updated
annually to adjust for inflation.
Anthropometrics: Standing height (cm) and weight (kg) were used to
calculate body mass index (BMI; kg∙m
-2
). BMI percentiles were defined for
normal, overweight, and obese categories according to the age- and gender-
specific growth charts published by the CDC [102]. Normal weight is defined as
37

between the 5
th
to less than the 85
th
percentile, overweight is defined as between
the 85
th
to less than the 95
th
percentile, and obese is defined as at or above the
95
th
percentile.
Physical Activity: All ambulatory participants over the age of 6 years were
asked to wear an Actigraph (Actigraph, LLC; Ft. Walton Beach, FL) model 7164
accelerometer on the right hip. The uniaxial accelerometer measures
acceleration intensity as ‗counts‘ in response to body movement. Details on the
accelerometer protocol, data reduction process, and the definition of wear time
have been reported previously [119, 206]. Similar to previous studies using
accelerometer data, a valid day was considered to be 10 or more hours of wear
time and participants with at least 4 valid days were included in the present study
[30, 206]. Accelerometer data is presented here as: a) mean counts per minute;
and b) as estimates of mean minutes per day in sedentary behavior, moderate,
vigorous, and MVPA. The thresholds for moderate (4 METs) and vigorous (7
METs) physical activity were age-adjusted using the criteria from Freedson et al.
(2002) [210] for participants between the ages of 6 and 17. The thresholds for
moderate (3 METs) and vigorous (6 METs) physical activity were used for
participants older than 18, and are defined at 2020 counts per minute and 5999
counts per minute, respectively [24, 68, 107, 206, 231]. A sedentary behavior cut
point of 100 counts per minute was applied to all age groups [119].
38

Statistical Analyses
All analyses were conducted using the SURVEY Procedures in SAS 9.1
(SAS Institute, Inc., Cary, NC) to account for the complex multistage probability
design. The NHANES-provided sample weights were adjusted to account for the
combination of data from 2003-2004 and 2005-2006 and for the sub sampling of
individuals with at least 4 valid days. Time spent in each activity level (i.e.:
moderate, vigorous, and MVPA) was calculated by summing each minute with a
count above the threshold for that activity level. Mean counts per minute were
calculated by dividing the sum of counts each day by the wear time minutes each
day across all valid days. Standard error of the mean (SEM) was calculated in
SAS using the Taylor series linearization method with PIR as a covariate [229].
Planned comparisons of mean counts per minute between each data collection
year and between subpopulations were made with pairwise contrasts.
Significance was set at alpha = .05 and 30 degrees of freedom, which is the
number of primary sampling units [159]. To further identify differences in mean
minutes spent in MVPA within subgroups, the data was stratified by gender and a
3-way race/ethnic-age group-BMI percentile interaction was examined using the
SAS SUVEYREG procedure with an ANOVA statement that adjusted for the
sampling strategy and PIR. Linear regression using age, BMI percentile, and
poverty to income ratio as continuous variables was conducted to understand the
magnitude of the effect of each variable on mean minutes per day spent in
MVPA. To predict the odds of reaching the 2008 Physical Activity Guidelines of
39

60 minutes per day of MVPA for youth ages 6-17 years or 21 minutes per day for
older youth ages 18-19 years (based on the recommendation adults perform 150
minutes/week MVPA, or an average of 21 minutes/day) [214] logistic regression
was conducted using the SURVEYLOGISTIC procedure while controlling for PIR.
Results
General Results
The characteristics of the sample are presented in Table 2-1 by
race/ethnicity and gender. Of the 5,687 participants with accelerometer data,
3,698 met the inclusion criteria of having 4 days (10 hours/day) of valid data. Of
these, 99 underweight youth were excluded, 453 had invalid demographic data,
and 40 had invalid anthropometric data. The participants with invalid physical
activity data significantly differed from those with valid physical activity data on
demographic characteristics; however the sample was re-weighted based on the
inclusion criteria of 4 or more 10-hour days of valid data. The analytical sample
(N = 3,106) was 48.8% female, 69.9% non-Hispanic White, 16.6% non-Hispanic
Black, and 13.5% Mexican American. Similar to previous studies using NHANES
data [135] 16.7% of youth ages 6-19 were obese and 17.8% were overweight.
The highest prevalence of obesity by race/ethnic group and gender was in non-
Hispanic Black females, Mexican American males, and Mexican American
females. Obesity prevalence was highest in 16-19 year olds and lowest in 6-11
year olds.
40

Mean Activity Counts per Minute
Non-Hispanic White youth recorded fewer mean counts per minute than
non-Hispanic Black and Mexican American youth and counts per minute
consistently declined with age. Youth in the 6-11 age group demonstrated higher
mean counts per minute than youth in both the 12-15 and 16-19 age groups
(p<.001 for both); and youth in the 12-15 age group had higher mean counts per
minute than youth in the 16-19 age group (p<.001) (Table 2-2). Females always
had lower mean counts per minute than males independent of race/ethnicity and
weight status (p<.001). Overweight non-Hispanic Black males recorded 67 more
counts per minute per day than overweight Mexican American males (p=.037)
and 104 more counts per minute per day than overweight non-Hispanic White
males (p=.002) (Table 2-3). Normal weight youth of all race/ethnic groups had on
average 87 more counts per minute than obese youth (p<.001). This decline was
constant across all race/ethnic groups with the exception of overweight and
obese Mexican American females who recorded similar counts per minute.
Mean Minutes in Activity Levels
The 6-11 age group recorded 88 minutes of MVPA per day whereas youth
12-15 and 16-19 recorded 33 and 26 minutes of MVPA per day, respectively
(Table 2-4). The 6-11 age group also spent fewer minutes in sedentary behavior,
and more minutes in moderate PA and vigorous PA per day than the 12-15 and
the 16-19 age groups across all race/ethnic groups (p<.001 for both). Non-
Hispanic Black 16-19 year olds were the most inactive race/ethnic subgroup,
41

spending 520 minutes per day in sedentary behavior. Non-Hispanic White youth
spent fewer minutes in vigorous PA than both non-Hispanic Black and Mexican
American youth (p<.001 and p=.004, respectively) (Table 2-5). Females spent
fewer minutes in MVPA than males (p<.001) and 20 more minutes per day in
sedentary behavior than males (p<.001). Normal weight youth spent more
minutes in moderate PA, vigorous PA, and MVPA (p<.001 for all) than obese
youth (Table 2-6). Normal weight non-Hispanic White youth spent 34 fewer
minutes in sedentary behavior per day than normal weight non-Hispanic Black
youth (p=.001). Also, the non-Hispanic White youth were the only race/ethnic
subgroup with differences in sedentary behavior between the normal weight and
overweight and obese groups (p=.008 and p=.026, respectively). Multivariable
linear regression with continuous variables indicated that age was inversely
associated with MVPA in all race/ethnic groups for both genders (p<.001 for all)
(Table 2-7). Similarly, BMI percentile was negatively associated with MVPA in all
race/ethnic groups for both genders (p<.050 for all) however the effect was not
as strong as age. Age, BMI percentile, and PIR accounted for 44% of the
variance in males and 49% of the variance in females.
There was a statistically significant 3-way interaction for mean minutes per
day spent in MVPA between age group, BMI percentile category, and race/ethnic
group for both males and females (p<.001) (Figures 2-1 and 2-2). In the 12-15
year old Mexican American race/ethnic group, male and female overweight youth
spent significantly fewer minutes in MVPA than normal weight youth (p=.022 and
42

p=.027, respectively), however this consistent difference was not seen within the
other race/ethnic groups.
Race/Ethnic Physical Activity Differences
By Age: Table 2-4 presents differences in physical activity levels by age
group. Non-Hispanic White 6-11 (p=.018) and 12-15 (p=.005) year olds spent
fewer minutes in MVPA than non-Hispanic Black youth of the same age groups.
Non-Hispanic Black 6-11 year olds spent 12 more minutes per day in MVPA than
Mexican American 6-11 year olds (p=.007). Non-Hispanic Black 12-15 year olds
spent 25 more minutes in sedentary behavior than non-Hispanic White 12-15
year olds (p<.001), while non-Hispanic Black 6-11 year olds spent 3 more
minutes in vigorous PA than Mexican American 6-11 year olds (p=.007).
Although differences in physical activity were seen in younger age groups,
physical activity levels declined in the oldest age group so that youth ages 16-19
of all race/ethnic groups spent the same amount of time (between 24 and 29
minutes) in MVPA (p>.050).
By Gender: Table 2-5 presents differences in activity levels by gender. In
non-Hispanic White and Mexican American youth, females spent about 20 more
minutes per day in sedentary behavior than males (p=.002 and p=.029
,respectively).Non-Hispanic White males spent fewer minutes per day in vigorous
PA than both non-Hispanic Black (p<.001) and Mexican American males
(p=.004). With the exception of non-Hispanic Black youth, females of all
43

race/ethnic groups spent significantly more time in sedentary behavior (p<.050)
and less time in physical activity behavior (p<.050) than males.
By BMI Percentile: Table 2-6 presents the physical activity differences by
weight status. Normal weight and overweight non-Hispanic Whites spent 9 and
18 fewer minutes in MVPA per day than normal weight and overweight non-
Hispanic Blacks (p=.043 and p=.001, respectively). There we no race/ethnic
differences within the obese group for any activity level. However, the only
differences in minutes spent in moderate PA, vigorous PA, and MVPA between
the overweight and obese groups were seen in the non-Hispanic Black
race/ethnic group, where the overweight group recorded more minutes in each of
the activity levels than obese youth (p<.001 for all activity levels).
Age- Gender-BMI Interaction Effects within Race/Ethnic Groups (Figures
2-1 & 2-2): In 6-11 and 16-19 year old Mexican American males and females,
obese youth spent fewer minutes in MVPA than normal youth (p<.016) but in the
12-15 Mexican American males and females, obese and normal youth spent
statistically equivalent amounts of time in MVPA per day (p>.379). In 6-11 non-
Hispanic Black males and females, obese youth recorded less time in MVPA
than both normal weight and overweight youth. In non-Hispanic White females,
6-11 and 12-15 year old normal weight youth recorded more minutes in MVPA
per day than obese youth (p<.019).
44

Age- Gender-BMI Interaction Effects between Race/Ethnic Groups
(Figures 2-1 & 2-2): In 6-11 year old females, normal weight youth spent more
minutes per day in MVPA than obese youth in all race/ethnic groups (p<.050).
However, in 6-11year old males this difference was seen in the non-Hispanic
Black (p<.001) and Mexican American (p<.001) race/ethnic groups but not in the
non-Hispanic White race/ethnic group. Most of the interaction effects were seen
in the youngest age group. Within the 6-11 year old females, the non-Hispanic
White and Mexican American race/ethnic groups had the greatest differences in
MVPA between the normal and obese groups however the non-Hispanic Black
race/ethnic group had the greatest differences in MVPA between the overweight
and obese groups.
Meeting Physical Activity Guidelines
Table 2-8 presents the results from the logistic regression analysis to
determine which groups were more likely to meet the 2008 physical activity
guidelines. In this sample, 41.4% met the recommendations. Females were less
likely to adhere to the 2008 PA guidelines than males (OR = 0.41; 95% CL: 0.30-
0.54). For males, those most likely to meet the 2008 PA Guidelines were non-
Hispanic Black, ages 6-11, and of normal weight. Because females were found to
be profoundly inactive, race/ethnicity did not predict whether or not they would
meet the guidelines of at least 60 minutes per day of MVPA. Females in the older
age groups who were obese and had a higher poverty level were less likely to
meet the guidelines.
45

Discussion
Based on objectively measured levels of PA, non-Hispanic White youth
were the least active race/ethnic group and non-Hispanic Black youth were the
most active race/ethnic group. Previous findings using self-report measures of
physical activity in U.S. samples showed that non-Hispanic Black youth were the
least physically active race/ethnic group [4]. However, using accelerometers to
measure physical activity Owen et al. (2008) found that Blacks recorded four
more minutes per day in MVPA than Whites [139]. In our sample, non-Hispanic
Black youth spent about 8 more minutes per day in MVPA than non-Hispanic
White youth. These differences may partially explained by differences in activity
preferences between Black and White youth, which suggest that Blacks
participate in more activities better captured by accelerometry (such as social
dancing and basketball) than Whites (who participate more in swimming and
calisthenics) [54, 77].
Troiano et al. (2008) previously reported an age-related decline in physical
activity of U.S. youth based on results from NHANES 2003-4 [206]. For 2003-4
and 2005-5 survey combined samples, we also found an age-related decline in
physical activity. Pate et al. (2009) reported a 4% annual decline in MVPA in
adolescent girls [149]. Broderson et al. (2007) found that physical activity
dropped off between ages 11 and 12 [27]. In our sample, 6-11 year olds
participated in twice as much MVPA than the older age groups, consistent with
46

the observation that the most dramatic age-related decline in physical activity
may occur at the start of puberty.
To better understand the age-related decline in MVPA observed in this
sample, the 3-way interaction between age, BMI category, and race/ethnic group
was examined for each gender. Non-Hispanic Whites and Blacks in higher BMI
categories spent less time in MVPA than normal weight non-Hispanic White and
non-Hispanic Black youth, whereas obese 12-15 year old Mexican American
youth recorded the same amount of time in MVPA per day as normal weight
youth. A previous study in adolescent Hispanic females found a trend for normal
weight Mexican American females spending fewer minutes in MVPA than
overweight females [31]. This suggests that BMI may interact with physical
activity levels differently in Mexican American youth than in other race/ethnic
groups. In the non-Hispanic White race/ethnic group, the largest difference in
MVPA is between the normal and overweight groups, with the overweight and
obese groups participating in approximately the same amount of MVPA.
However, in the non-Hispanic Black race/ethnic group, the largest difference in
MVPA is between the overweight and obese youth. Thus, the findings indicate
that different race/ethnicities have different thresholds of BMI percentile past
which MVPA declines.
The large difference in physical activity between males and females is
particularly striking. Normal weight females of all race/ethnic groups achieved
less physical activity than obese males of all race/ethnic groups. A recent study
47

conducted in adolescents using accelerometers found that while the rates of
decline of physical activity were the same for both genders, females participated
in significantly fewer minutes of MVPA than males and MVPA dropped below 60
min/day one year earlier in females than in males [132]. Non-Hispanic Black
youth had the largest gender differences in counts per minute and minutes spent
in MVPA: females recorded about 140 counts per minute and 27 minutes per day
fewer than males. Using previous prediction equations based on overweight
youth [59], this deficit is broadly similar to 600 kcal per day. The difference in
physical activity levels between males and females in this sample may contribute
to the fact that the non-Hispanic Black females had the highest prevalence of
obesity.
Overall, the inverse association between physical activity levels and BMI
percentile in this sample is consistent with previous findings [174]. Contrary to
our expectations, higher levels of PA were not associated with lower prevalence
of obesity across the race/ethnic groups. Non-Hispanic White youth had lower
mean counts per minute and spent fewer minutes in MVPA than non-Hispanic
Black and Mexican American youth yet had a lower prevalence of obesity than
the other race/ethnic groups. This paradox may be accounted for by the fact that
non-Hispanic White youth may spend more time in activities not captured well by
accelerometry such as swimming or bicycling. These differences could also be
attributed to the higher socioeconomic status (SES) found in the non-Hispanic
White youth since SES has been inversely related to obesity and positively
48

related to physical activity [95]. SES was controlled for in all analyses; other
factors may contribute to the pattern of obesity and physical activity in non-
Hispanic White youth.
Genetic predisposition to obesity, SES, and cultural differences in
behavior may play a role in the race/ethnic differences found in this sample and
elsewhere [204]. Non-Hispanic Black and Mexican American adults have the
highest prevalence of overweight and obesity in the U.S. population [134].
Children of overweight and obese parents have been shown to have higher rates
of obesity than children of normal weight parents [17]. Furthermore, non-Hispanic
Black and Mexican American females have been shown to have lower basal
metabolic rates and expend less activity energy than non-Hispanic White females
which may put them at higher risk for overweight and obesity [56]. Dietary intake
may also account for the differences in the obesity prevalence between non-
Hispanic White youth and the minority race/ethnic groups, particularly given that
there are race/ethnic differences in the consumption of unhealthy foods. Arcan et
al. (2009) found that non-Hispanic Black high school students were more likely to
consume sugar-sweetened beverages and high-fat foods than other race/ethnic
groups [6]. The higher rate of obesity in non-Hispanic Blacks may be explained
by a higher intake of unhealthy foods, particularly in non-Hispanic Black females
who have been found to have the lowest levels of physical activity and highest
intakes of unhealthy foods [103].
49

No other past study that we are aware of has described race/ethnic
differences in objectively measured physical activity in large representative
sample of U.S. youth, however several limitations to the present study merit
discussion. First, this is a cross-sectional analysis, thus we cannot determine
any causal associations. However, the large sample size allows robust estimates
of associations between variables of interest and may help inform future
longitudinal studies. Second, accelerometers do not capture all types of physical
activity [190]. However, accelerometers are considered to be an excellent
objective measurement of physical activity in youth because they minimize self-
report bias and eliminate human error in recalling previous physical activity [48].
Third, the accelerometers do not record the type of PA as do self-report
measures, which prevents us from exploring the frequency of specific behaviors
(i.e.: TV viewing) that could explain the observed differences in physical activity
levels among race/ethnic groups. Fourth, the NHANES survey is designed to
sample the three largest race/ethnic groups in the U.S. and therefore does not
provide data sufficient for a national estimate for other minority groups such as
Asians or other non-Mexican Hispanic populations that comprise a significant
and growing proportion of the U.S. population. Fifth, this analysis did not include
any measures of food intake, which are associated with obesity. However, Gutin
(2008) reported that vigorous physical activity was sufficient to lower percent
body fat in youth without restricting dietary intake, thus physical activity may have
a significant impact on weight status independent of dietary intake [79]. Finally,
50

BMI percentile category is used here as a proxy measure of adiposity. While
some findings indicate that it is not an accurate measure of body fat for all
race/ethnic groups, Flegal et al. (2009) recently demonstrated that it corresponds
well with percent body fat in an adult sample [65] and BMI has been shown to be
significantly correlated with percent body fat in youth [152]. Furthermore, BMI
percentile is a cost efficient and feasible measure in a large population-based
study [52].
As measured by accelerometry, non-Hispanic White youth engaged in
less physical activity than both non-Hispanic Black and Mexican American youth
yet had the lowest prevalence of obesity in this sample. Also, non-Hispanic Black
females are the least physically active and have the highest prevalence of
obesity in this sample. Mexican American 12-15 year old obese and normal
weight youth had the same amount of MVPA. Explanations for differences in
obesity rates between youth of different race/ethnic groups must be influenced by
other factors than variations in physical activity levels.
51

Weighted values account for survey design, sampling strategy, & reweighting for
those with 4+ days of data; unless otherwise noted, the standard error of the
mean is presented in parentheses.
†
Includes both genders & all race/ethnic groups.
*
N represents actual number sampled.
1
Values represent mean (standard error).
2
Values represent percent (standard
error of percent).
BMI= Body Mass Index.
‡
Defined by the 2000 CDC age- and gender- specific cut
points [102].
52

Table 2-1: Characteristics (mean and standard error) of the analyzed sample for
2003-2006

Non-Hispanic White Non-Hispanic Black Mexican American
Total
Sample
†

Variable* Male Female Male Female Male Female
N 433 443 629 530 568 602 3106
Age (yr)
1
12.3 (0.2) 12.5 (0.2) 12.4 (0.2) 12.5 (0.3) 11.9 (0.2) 12.0 (0.2)
12.4
(0.1)
6 – 11
2

159
69.2 (3.6)
173
67.9 (3.4)

190
16.0 (2.3)
174
16.9 (2.4)

197
14.8 (2.4)
223
15.2 (1.8)
1116
42.0
(1.6)
12 – 15
2

146
71.9 (3.6)
134
69.4 (3.8)

222
15.7 (2.3)
188
16.7 (2.4)

204
12.4 (2.1)
221
13.9 (2.4)
1106
33.3
(0.9)
16 – 19
2

111
69.7 (3.5)
121
72.7 (3.7)

193
18.1 (2.8)
160
16.7 (3.2)

156
12.2 (1.7)
143
10.5 (1.7)
884
24.7
(1.5)
Height (cm)

1

155.4
(1.0)
151.4
(0.8)

156.4
(0.8)
151.6
(0.9)

151.2
(0.9)
147.0
(0.9)
153.0
(0.6)
Weight (kg)
1
53.6 (1.1) 50.6 (1.1) 55.6 (0.9) 54.8 (1.6) 51.2 (1.2) 48.9 (0.9)
52.4
(0.8)
BMI
Percentile
1‡

65.1 (1.4) 64.4 (1.8) 64.5 (1.6) 70.6 (1.1) 66.9 (1.2) 68.4 (1.3)
65.6
(0.9)
Normal
2

(5
th
to <85
th
%)
277
66.2 (2.8)
291
67.8 (2.5)

406
66.0 (2.5)
309
58.9 (1.5)

328
60.0 (1.6)
369
63.3 (2.4)
1980
65.5
(1.8)
Overweight
2

(85
th
to <95
th
%)
70
17.6 (2.1)
79
18.9 (1.5)

100
16.9 (1.3)
84
15.9 (1.4)

102
18.3 (1.4)
94
15.7 (1.5)
529
17.8
(1.2)
Obese
2

(≥95
th
%)
69
16.2 (1.9)
58
13.3 (2.2)

99
17.1 (1.8)
129
25.2 (1.7)

118
21.7 (1.9)
124
20.9 (1.9)
597
16.7
(1.2)
53

1
BMI= Body Mass Index; defined by the 2000 CDC age- and gender- specific cut
points [102].
§
6-11 differ from both 12-15 and 16-19 (p<.050) within all activity levels. *12-15
differ from 16-19 (p<.050).
The letters and weight status symbols indicate statistically significant differences
within activity levels by columns.
†
Normal weight and overweight differ by p<.050.
‡
Normal weight and obese differ
by p<.050.
¶
Overweight and obese differ by p<.050.
a
Non-Hispanic White differ from non-Hispanic Black by p<.050.
b
Non-Hispanic
White differ from Mexican American by p<.050.
c
Non-Hispanic Black differ from
Mexican American by p<.050.
54

Table 2-2: Mean (standard error) counts per minute by race/ethnicity, age group,
and BMI percentile category
1

Age Group (yr) 6-11 12-15
§
16-19
§

Normal
(5
th
to <85
th
%)
Non-Hispanic White 652.0 (13.6)
†‡
453.8 (14.0)
*
386.6 (15.9)
*

Non-Hispanic Black 695.7 (16.0)
‡ c
465.3 (14.9)
‡ *
393.7 (12.7)
‡ *

Mexican American 644.7 (17.6)
†‡ c
455.5 (17.0)
† *
404.1 (11.2)
‡ *

All Ethnicities 657.6 (11.1)
†‡
455.9 (10.7)
*
389.5 (12.2)
‡ *

Overweight
(85
th
to <95
th
%)
Non-Hispanic White 561.0 (26.7)
† a
411.6 (19.3)
† * a
343.3 (25.9)
* b

Non-Hispanic Black 668.4 (30.5)
¶ ac
485.2 (23.7)
¶ * ac
381.1 (21.9)
*

Mexican American 570.8 (21.4)
† c
407.0 (16.6)
† c
427.8 (22.1)
¶ b

All Ethnicities 579.5 (21.6)
†¶
421.0 (16.4)
*
361.2 (20.1)
*

Obese
(≥95
th
%)
Non-Hispanic White 523.0 (31.1)
‡
408.2 (35.6) 372.9 (21.1)
Non-Hispanic Black 518.6 (26.3)
‡¶
424.6 (16.5)
‡¶ *
345.1 (13.8)
‡ *

Mexican American 534.4 (16.2)
‡
447.9 (12.6)
*
358.4 (17.9)
‡¶ *

All Ethnicities 524.5 (16.8)
‡¶
416.8 (23.6)
*
364.8 (12.7)
‡ *

55

1
BMI= Body Mass Index; defined by the 2000 CDC age- and gender- specific cut
points [102] .
§
Males and females in all race/ethnic & BMI percentile categories differ by
p<.010.
The letters and weight status symbols indicate statistically significant differences
within activity levels by columns.
†
Normal weight and overweight differ by p<.050.
‡
Normal weight and obese differ
by p<.050.
¶
Overweight and obese differ by p<.050.
a
Non-Hispanic White differ from non-Hispanic Black by p<.050.
b
Non-Hispanic
White differ from Mexican American by p<.050.
c
Non-Hispanic Black differ from
Mexican American by p<.050.

56

Table 2-3: Mean (standard error) counts per minute by race/ethnicity, gender, &
BMI percentile category
1

Gender Male
§
Female Both Genders
Normal
(5
th
to <85
th
%)
Non-Hispanic White 567.6 (12.8)
†‡

a
476.5 (14.7)
†‡
522.8 (9.1)
†‡

Non-Hispanic Black 610.1 (15.3)
‡

a
465.9 (19.2) 542.9 (14.8)
‡

Mexican American 608.2 (18.3)
†‡
454.6 (13.9) 530.7 (13.1)
†‡

All Ethnicities 579.6 (10.6)
†‡
472.1 (11.2)
†‡
527.0 (7.1)
†‡

Overweight
(85
th
to <95
th
%)
Non-Hispanic White 500.0 (21.9)
†

a
403.1 (16.8)
†
451.2 (13.3)
†

ab

Non-Hispanic Black 604.0 (23.7)
¶ ac
449.8 (17.6) 530.2 (16.0)
¶

ac

Mexican American 536.6 (19.5)
†

c
433.4 (20.1) 489.8 (13.0)
†

bc

All Ethnicities 521.4 (19.3)
†¶
413.6 (12.0)
†
468.2 (11.6)
†¶

Obese
(≥95
th
%)
Non-Hispanic White 488.5 (25.6)
‡
378.7 (28.3)
‡
440.5 (21.5)
‡

Non-Hispanic Black 493.2 (23.5)
‡¶
410.1 (23.2) 444.2 (18.4)
‡¶

Mexican American 507.4 (14.5)
‡
435.3 (11.4) 472.7 (11.1)
‡

All Ethnicities 492.5 (17.1)
‡¶
396.7 (15.8)
‡
446.8 (12.6)
‡¶

57

1
Physical activity categories were based on age-specific criteria for moderate and
vigorous intensity for ages 6-17 years [210]; for ages 18 & older: moderate
intensity cut point = 2020 cpm & vigorous intensity cut point = 5999 cpm [206];
sedentary cut point = 100 cpm for all ages [119].
§
6-11 differ from both 12-15 and
16-19 (p<.050) within all activity levels. *12-15 differ from 16-19 (p<.050). The
letters and weight status symbols indicate statistically significant differences
within activity levels by columns.
†
Normal weight and overweight differ by
p<.050.
‡
Normal weight and obese differ by p<.050.
¶
Overweight and obese differ
by p<.050.
a
Non-Hispanic White differ from non-Hispanic Black by p<.050.
b
Non-
Hispanic White differ from Mexican American by p<.050.
c
Non-Hispanic Black
differ from Mexican American by p<.050.
58

Table 2-4: Mean (standard error) minutes per day above specified thresholds
1
for
sedentary behavior, moderate, vigorous, & moderate + vigorous physical activity
by race/ethnicity and age group
Age Group (yr) 6-11
§
12-15 16-19
Sedentary Non-Hispanic White 344.9 (3.5)
ab
456.4 (6.7)
* ab
497.8 (10.0)
*

a
Non-Hispanic Black 372.5 (6.5)
ac
481.1 (6.0)
* a
519.9 (5.9)
* ac
Mexican American 355.1 (3.9)
bc
472.9 (4.4)
b
474.4 (5.6)
c

All Ethnicities 351.0 (2.7) 462.6 (5.5)
*
499.0 (7.1)
*
Moderate Non-Hispanic White 74.1 (2.0) 27.4 (1.2)
* ab
22.0 (1.6)
* a

Non-Hispanic Black 79.7 (2.7)
c
33.5 (1.8)
* a
26.7 (2.1)
*a

Mexican American 71.4 (1.6)
c
32.3 (1.7)
* b
26.2 (1.6)
*

All Ethnicities 74.7 (1.4) 29.0 (1.0)
*
23.3 (1.3)
*

Vigorous Non-Hispanic White 12.4 (0.7)
a
4.0 (0.4)
* ab
2.3 (0.4)
*

Non-Hispanic Black 17.0 (0.8)
ac
5.0 (0.4)
* a
2.1 (0.3)
*

Mexican American 13.6 (0.9)
c
5.0 (0.4)
* b
2.0 (0.2)
*

All Ethnicities 13.3 (0.6) 4.3 (0.3)
*
2.2 (0.3)
*

Moderate +
Vigorous
Non-Hispanic White 86.6 (2.4)
a
31.4 (1.5)
* ab
24.3 (1.7)
*

Non-Hispanic Black 96.7 (3.3)
ac
38.5 (2.1)
* a
28.8 (2.1)
*

Mexican American 85.1 (2.3)
c
37.3 (2.0)
* b
28.2 (1.6)
*

All Ethnicities 88.0 (1.8) 33.3 (1.3)
*
25.5 (1.4)
*

59

1
Physical activity categories were based on age-specific criteria for moderate and
vigorous intensity for ages 6-17 years [210]; for ages 18 & older: moderate
intensity cut point = 2020 cpm & vigorous intensity cut point = 5999 cpm [206];
sedentary cut point = 100 cpm for all ages [119]. The letters and symbol indicate
statistically significant differences within activity levels by columns and rows,
respectively.
§
Males and females differ by p<.050.

60

Table 2-5: Mean (standard error) minutes per day above specified thresholds
1
for
sedentary behavior, moderate, vigorous, & moderate + vigorous physical activity
by race/ethnicity and gender
Gender Male Female Both Genders
Sedentary Non-Hispanic White 410.6 (5.1)
§ a
431.8 (7.0)
a
420.9 (5.1)
a

Non-Hispanic Black 441.3 (6.0)
ac
450.7 (7.7)
a
445.9 (5.6)
ac

Mexican American 407.0 (5.6)
§ c
429.7 (6.1) 418.1 (3.6)
c

All Ethnicities 415.1 (4.2)
§
434.7 (5.2) 424.7 (4.1)
Moderate Non-Hispanic White 52.4 (2.0)
§a
37.8 (1.5) 45.3 (1.2)
a

Non-Hispanic Black 61.1 (2.2)
§a
40.6 (2.2) 51.0 (1.9)
a

Mexican American 57.9 (1.9)
§
40.5 (1.7) 49.3 (1.3)
All Ethnicities 54.6(1.7)
§
38.6 (0.9) 46.8 (1.0)
Vigorous Non-Hispanic White 8.2 (0.6)
§

ab
5.8 (0.4) 7.0 (0.4)
ab

Non-Hispanic Black 12.5 (0.7)
§ a
6.0 (0.6) 9.2 (0.5)
a

Mexican American 11.1 (0.8)
§ b
5.9 (0.6) 8.4 (0.5)
b

All Ethnicities 9.3 (0.5)
§
5.8 (0.2) 7.6 (0.3)
Moderate +
Vigorous
Non-Hispanic White 60.5 (2.4)
§ ab
43.6 (1.8) 52.3 (1.5)
a

Non-Hispanic Black 73.7 (2.8)
§ a
46.9 (2.7) 60.2 (2.4)
a

Mexican American 69.0 (2.5)
§ b
46.5 (2.7) 57.7 (1.8)
All Ethnicities 63.8 (2.0)
§
44.4 (1.1) 54.4 (1.2)
61

1
Physical activity categories were based on age-specific criteria for moderate and
vigorous intensity for ages 6-17 years [210]; for ages 18 & older: moderate
intensity cut point = 2020 cpm & vigorous intensity cut point = 5999 cpm [206];
sedentary cut point = 100 cpm for all ages [119].
2
BMI= Body Mass Index;
defined by the 2000 CDC age- and gender- specific cut points [102]. The letters
and weight status symbols indicate statistically significant differences within
activity levels by columns.
†
Normal weight and overweight differ by p<.050.
‡
Normal weight and obese differ by p<.050.
¶
Overweight and obese differ by
p<.050.

62

Table 2-6: Mean (standard error) minutes per day above specified thresholds
1
for
sedentary behavior, moderate, vigorous, & moderate + vigorous physical activity by
race/ethnicity and BMI percentile category
2

BMI Percentile Category Normal
(5
th
to <85
th
%)
Overweight
(85
th
to <95
th
%)
Obese
(≥95
th
%)
Sedentary Non-Hispanic White 412.0 (5.9)
†‡ a
442.4 (11.5)
†
434.5 (11.7)
‡

Non-Hispanic Black 446.2 (7.5)
ac
443.4 (9.3) 446.9 (8.8)
c
Mexican American 415.7 (5.4)
c
428.3 (6.2) 417.2 (6.2)
c
All Ethnicities 417.9 (5.0)
†
440.7 (8.2)
†
434.1 (7.5)
Moderate Non-Hispanic White 48.5 (1.5)
†‡

a
39.5 (2.9)
†

a
38.0 (4.0)
‡

Non-Hispanic Black 55.3 (2.7)
‡

a
52.5 (2.9)
¶ a
37.1 (2.7)
‡¶

Mexican American 52.4 (1.7)
‡
46.8 (2.7) 42.4 (1.7)
‡

All Ethnicities 50.1 (1.2)
†‡
42.4 (2.4)
†¶
38.5 (2.3)
‡¶

Vigorous Non-Hispanic White 8.2 (0.4)
†‡ a
4.5 (0.5)
†

ab
4.5 (0.7)
‡

Non-Hispanic Black 10.8 (0.8)
‡

a
9.0 (0.9)
¶ a
4.9 (0.5)
‡¶

Mexican American 9.5 (0.7)
†‡
7.3 (0.7)
†

b
6.1 (0.4)
‡

All Ethnicities 8.8 (0.3)
†‡
5.6 (0.5)
†¶
4.9 (0.4)
‡¶

Moderate +
Vigorous
Non-Hispanic White 56.8 (1.8)
†‡

a
44.0 (3.2)
†

ab
42.5 (4.7)
‡

Non-Hispanic Black 66.1 (3.4)
‡ a
61.6 (3.5)
¶ a
42.0 (3.2)
‡¶

Mexican American 61.9 (2.3)
†‡
54.1 (3.2)
†

b
48.5 (2.1)
‡

All Ethnicities 58.9 (1.4)
†‡
48.0 (2.8)
†¶
43.4 (2.7)
‡¶

63

Table 2-7: Multivariable linear regression model predicting MVPA (min/day)

Values are beta coefficients with standard error in parentheses. MVPA=
moderate to vigorous physical activity. BMI= Body Mass Index. PIR= Poverty to
Income Ratio.
*
p< .050,
†
p< .010,
‡
p< .001
Males
R-square= .44

Females
R-square= .49
Non-Hispanic
White
Non-Hispanic
Black
Mexican
American
Non-Hispanic
White
Non-Hispanic
Black
Mexican
American
Intercept
178.2 (7.8)
‡
214.6 (8.6)
‡

185.0
(8.3)
‡

147.4 (6.8)
‡

167.9 (13.4)
‡

136.3
(6.3)
‡

Age (yr)
-8.3 (0.5)
‡
-9.4 (0.5)
‡

-7.8 (0.4)

‡

-7.6 (0.4)
‡

-7.9 (0.8)
‡
-6.7 (0.4)
‡

BMI (%)
-0.2 (0.1)
*
-0.3 (0.1)
‡

-0.3 (0.1)

‡

-0.1 (0.1)
*

-0.3 (0.1)
‡
-0.1 (0.0)
†

PIR -1.5 (1.4) -3.1 (1.1)
†
-1.5 (1.2) -0.4 (1.0) -3.0 (1.2)
*
-1.5 (0.7)
*

64

Table 2-8: Logistic regression model predicting meeting the 2008 PA Guidelines
1

by gender

1
Outcome modeled is ‗Yes, meeting 2008 PA Guidelines‘. PA= physical activity.
BMI= Body Mass Index.
2
Defined by the 2000 CDC age- and gender- specific cut
points [102].
*
p< 0.05.
Males Females
Odds Ratio (95% CI) Odds Ratio (95% CI)
Race/Ethnicity (vs. Non-Hispanic White)
Non-Hispanic Black 1.61
*
(1.07 – 2.42) 1.15 (0.66 – 2.01)
Mexican American 1.30 (0.84 – 2.02) 0.74 (0.46 – 1.18)
Age Group (vs. 6 – 11)
12 – 15 0.08
*
(0.06 – 0.14) 0.02
*
(0.01 – 0.03)
16 – 19 0.18
*
(0.12 – 0.27) 0.16
*
(0.10 – 0.24)
BMI Category
2
(vs. Normal (5
th
to <85
th
))
Overweight (85
th
to <95
th
) 0.52
*
(0.29 – 0.95) 0.78 (0.45 – 1.35)
Obese (≥95
th
) 0.42
*
(0.23 – 0.76 0.47
*
(0.28 – 0.79)
Poverty to Income Ratio Category (vs. <100%)
101-199% 1.03 (0.26 – 1.89) 0.71 (0.46 – 1.11)
200-299% 1.32 (0.73 – 2.41) 0.44
*
(0.22 – 0.89)
300-399% 1.08 (0.60 – 1.95) 0.75 (0.44 – 1.27)
≥400% 0.62 (0.34 – 1.16) 0.52
*
(0.29 – 0.92)
65

Figure 2-1: 3-way age group-BMI-race/ethnic interaction of MVPA in males

Within Race/Ethnic & Age Groups:
†Normal differ from overweight p< 0.05
‡Normal differ from obese by p< 0.05
*Overweight differ from obese by p< 0.05
Within Age Groups:
a Non-Hispanic White differ from Non-Hispanic Black p< 0.05
b Non-Hispanic White differ from Mexican American p< 0.05
c Non-Hispanic Black differ from Mexican American p< 0.05

66

Figure 2-2: 3-way age group-BMI-race/ethnic interaction of MVPA in females

Within Race/Ethnic & Age Groups:
†Normal differ from overweight p< 0.05
‡Normal differ from obese by p< 0.05
*Overweight differ from obese by p< 0.05
Within Age Groups:
a Non-Hispanic White differ from Non-Hispanic Black p< 0.05
b Non-Hispanic White differ from Mexican American p< 0.05
c Non-Hispanic Black differ from Mexican American p< 0.05

67

CHAPTER 3 THE LONGITUDINAL EFFECTS OF LEPTIN ON PHYSICAL
ACTIVITY IN MINORITY FEMALE CHILDREN
Introduction
Obesity prevention in children is a primary public health concern. It is
estimated that 33.3% of youth ages 6 to 11 are overweight [135] (defined as
having an age- and sex- specific body mass index (BMI) percentile ≥ 85) [136].
The epidemic is worse in minority youth. Recent national estimates indicate that
among children 6-11 years of age, 42.6% of Hispanics versus 34.5% of non-
Hispanic whites are overweight [133]. Obesity has deleterious health
consequences such as type 2 diabetes [162] and metabolic syndrome [218].
Furthermore, youth who are overweight or obese have an increased risk for all-
cause mortality as adults [114].
Many factors contribute to the development of obesity. Low levels of
physical activity have been associated with increased risk for obesity [29]. Based
on a nationally-representative sample of youth, only 41.4% meet the current
Physical Activity Recommendations that youth participate in at least 60 minutes
or more of moderate to vigorous physical activity (MVPA) per day [12]. Much
research has investigated factors associated with low levels of physical activity in
US youth. There is a well-documented age-related decline in physical activity
[149, 206]. Biological factors have been hypothesized to contribute to the
declining levels of physical activity in youth [57]. One such factor is leptin, an
adipose-derived hormone secreted in direct proportion with adipose tissue mass
68

[232]. Leptin was first studied as a satiety signal [233]. However, subsequent
research indicates that leptin is involved in both energy intake and expenditure
[5]. Leptin signals that fat stores are sufficient and energy intake can decrease
while energy output can increase [100]. In lean individuals, lower circulating
leptin levels are related to higher levels of physical activity [9]. In a sample of 253
normal weight females and 257 males ages 8 to 18 years, fasting leptin
concentrations were negatively related to pedometer-measured steps per day in
females only (p< .001)[170]. In overweight and obese individuals, there is
evidence of ‗leptin resistance‘ where plasma leptin levels are chronically high, yet
do not produce the expected decrease in energy intake and increase in energy
expenditure [83, 105]. The excess adipose tissue found in obese individuals
produces high circulating leptin levels that saturate the leptin receptors. The
diminished capacity of the leptin receptors results in a downregulation of leptin
transporters and a failure of the system to transport leptin across the blood-brain
barrier where it reaches its targets in the brain [9, 96]. Leptin resistance can
develop at an early age. In a sample of children (mean age 10.6 years),
overweight youth had higher fasting leptin levels than normal weight youth at
baseline (p< .0001) and at a one-year follow-up (p< .025). The authors noted that
the leptin levels remained consistently higher in the overweight children, which
may indicate some degree of leptin resistance [34]. Leptin resistance can lead to
excess energy intake and deficient energy expenditure [5]. However, little
69

research has determined the effect of leptin resistance on physical activity in
youth.
Physical activity declines as youth enter puberty [176]. Lower levels of
activity have been observed in relation to increasing pubertal Tanner stage [47,
169]. Furthermore, several studies have observed differences in leptin levels
among Tanner stages [2, 86]. In a sample of 102 youth ages 6 to 19 years,
circulating leptin concentrations differed by Tanner stage; increasing in females
from stages I to V and increasing in males from stages I to III after which they
declined [86]. Due to its central role in energy regulation and differences by
pubertal stage, leptin has been hypothesized to be related to physical activity
declines observed as youth begin puberty [9].
The central role leptin has in energy balance and its effects on health
make it an important hormone to study in obesity prevention research in youth.
The aims of this study are twofold: 1) to determine if leptin is related to the
decline in physical activity, and whether this relationship is influenced by Tanner
stage; and 2) to determine if baseline cross-sectional relationships between
leptin and physical activity remain stable in a longitudinal analysis over one year.
This analysis was driven by the following hypotheses:
1a) at baseline, leptin will be negatively associated with MVPA;
1b) at baseline, Tanner stage will moderate the leptin-MVPA relationship;
those in Tanner stage 2 will have higher baseline leptin concentrations and lower
levels MVPA than Tanner stage 1.
70

2) Over one year, baseline leptin values will predict the decline in mean
minutes per day (min/day) spent in MVPA.
Methods
Participants
Participants were recruited from participating clinics, churches, schools,
and community centers in the Los Angeles area from June 2006 to August 2009.
The study inclusion criteria were as follows: Hispanic or African American
race/ethnicity, non-menstruating females, between the ages of 8 and 11, who
were at pubertal Tanner breast stage 1 or 2 [197]. Race/ethnicity was self-
reported, and to be included in the study the participants were required to report
maternal and paternal grandparents and both parents as either Hispanic or
African American. Additional inclusion criteria were determined at a preliminary
screening visit conducted at the Clinical Trials Unit (CTU, formerly known as
GCRC). These criteria were: having a BMI 85
th
percentile (overweight or
obese); or a BMI between the 5
th
and 85
th
percentiles (normal weight) with one
parent meeting adult criteria for overweight (BMI 25) or diagnosed with type 2
diabetes (defined as a fasting plasma glucose 126 mg/dl). Participants were
excluded from the study if they had diabetes at the screening visit (defined as a
fasting plasma glucose 126 mg/dl), or diagnosis of a condition that could affect
body composition, fat distribution, or insulin action or secretion. Informed written
parental consent and participant assent (in the primary language of each) was
71

obtained prior to participation. The study was approved by the Institutional
Review Board of the University of Southern California.
Study design
The objective of the Transitions Study was to determine the psychosocial
and physiological determinants of physical activity and insulin dynamics in
minority females as they traversed early puberty. The longitudinal study
assessed insulin dynamics, selected gut hormones, body composition, and
pubertal stage each at inpatient visits each year at the University of Southern
California CTU. In addition to the yearly visits, objectively measured physical
activity was assessed every three months. Data presented include the first year
inpatient visit, three quarterly visits, and the second year inpatient visit.
Procedures
At the yearly inpatient visits, participants arrived at the CTU in the
afternoon on Day 1. The participants completed demographic and psychosocial
questionnaires. The participants also completed anthropometric and body
composition measures before receiving a complete medical examination and
health history conducted by a licensed pediatric health care provider. After the
examination, participants were given a standard meal and snack before 2000h,
after which they were only permitted to drink water. After spending the night, on
Day 2 at 0800h, the Frequently Sampled Intravenous Glucose Tolerance Test
(FSIVGTT) was performed over the course of two hours to assess insulin action
and secretion. During this procedure, additional fasting blood samples were
72

taken to measure leptin and other hormones. Participants were instructed on how
to wear an accelerometer for the seven days following the inpatient visit. After
instruction and lunch, participants were given lunch and discharged from the
CTU. A study team member subsequently contacted the participant to retrieve
the accelerometer.
In between the yearly visits, quarterly home visits were conducted about
every three months. At these visits, a study team member would provide the
participant with an accelerometer. The study team member contacted the
participant several times throughout the week to remind her to wear the
accelerometer and collected the device at the end of the quarterly visit. The
accelerometer data was downloaded and checked for completeness. If the
participant had fewer than four 10 hour days then the study team member
contacted the participant and asked her to wear the device a second time.
Accelerometer data from the Year 1 (baseline) inpatient visit, three subsequent
quarterly assessments, and the Year 2 inpatient visit are included in this analysis,
for a total of five time points.
Measures
Demographics & Physical Exam
Baseline age (in years) was recorded on the protocol flowsheets at the
Year 1 baseline visit. Child race/ethnicity was self-reported by the parent.
Pubertal Tanner breast stage [197]was assessed via palpation by the licensed
pediatric medical care provider at the screening visit.
73

Body Composition
Subcutaneous abdominal adipose tissue (SAAT; in L) and visceral
adipose tissue (VAT; in L) were measured by magnetic resonance imaging (MRI)
at Year 1 baseline visit.
Clinical Measures
The insulin-modified FSIVGTT procedure consisted of collection of blood
samples at time points -15, -5, 2, 4, 8, 19, 22, 30, 40, 50, 70, 100, and 180
minutes. At time 0, glucose (25% dextrose, 0.3 g/kg body weight) was
administered intravenously. Insulin [0.02 units/kg body weight, Hamelin R
(regular insulin for human injection); Eli Lilly, Indianapolis, IN] was injected
intravenously at 20 minutes. Insulin Sensitivity (SI), a measure of insulin-
stimulated glucose uptake, was calculated using the minimal model from the
FSIVGTT results [13] [MINMOD MILLENIUM 2002 computer program, v5.16;
Richard Bergman, Los Angeles, CA [143]].
Leptin was assessed at the Year 1 baseline visit. Fasting plasma leptin
(ng/ml) levels were determined in duplicate using a double antibody Radio
Immune Assay (RIA) kit (Millipore, St. Charles, MO) with a 0.5 ng/ml limit of
sensitivity. Leptin intra- and inter-assay CV were up to 6.2% and up to 5.3%,
respectively. All samples were measured in duplicate.
Physical Activity
Physical activity and sedentary behavior were assessed via accelerometry
at all five visits. The participants wore a uniaxial Actigraph accelerometer for at
74

least four 10-hour days. The monitor was programmed to collect data in 15-
second epochs. The data were then compounded into 60-second epochs for use
with the current activity thresholds. Data were downloaded after each visit and
reviewed to ensure the monitor was functioning correctly. The data was
processed using SAS code developed by the National Cancer Institute for use
with NHANES data (available at: http://riskfactor.cancer.gov/tools/nhanes_pam).
Non-wear time is defined as 60 minutes or more of 0 intensity counts, with
allowance for up to 2 consecutive minutes of counts between 1 and 100. The
non-wear period ends when at least one of the following conditions are met: one
minute of an intensity above 100; one minute with a missing intensity count; 3 or
more minutes with intensity counts between 1 and 100 counts; or the last minute
of the day. Wear time is defined by subtracting non-wear time from the total
observation time for each day.
Accelerometer data is presented as mean minutes per day spent in
MVPA. Time spent in the activity levels is calculated by summing each minute
spent above the user-defined thresholds for MVPA. The cut points for moderate
(4 METs) and vigorous (7 METs) activity were age-adjusted using the criteria
from the Freedson group [68] for participants between the ages of 6 and 17 years
in order to adjust for the higher resting energy expenditure in this age range [80,
169].
75

Data Analysis
Time spent in MVPA was calculated for each participant by summing each
minute that had a count above the age-adjusted threshold. Mean estimates and
standard deviations used to describe the sample. T-tests and chi-square tests
were used to assess Tanner stage differences in these descriptive
characteristics. Generalized linear regression was used to determine the cross-
sectional association between baseline leptin and MVPA, controlling for selected
covariates. Covariates included: age (years), Tanner stage (1 vs. 2),
race/ethnicity (Hispanic vs. African American) insulin sensitivity (SI),
subcutaneous abdominal adipose tissue (SAAT; L), and visceral adipose tissue
(VAT; L). A leptin-Tanner interaction term was included to assess whether the
relationship between leptin and MVPA differed in Tanner stage 1 compared to
Tanner stage 2. Should this interaction be significant, separate generalized linear
models would be run for each Tanner stage. The a priori planned mixed model
analysis would also contain the interaction term if it was significant in the cross-
sectional model.
Mixed models were used to assess the longitudinal effects of baseline
leptin concentrations on MVPA over one year. The models used all available
data for each participant and a compound symmetric covariance matrix. Variance
components were estimated using the restricted maximum likelihood estimate
method to control for within-individual correlations over the five visits. The
dependent variable was mean min/day spent in MVPA. Baseline variables were
76

treated as time-invariant variables and were baseline values for age,
race/ethnicity, Tanner stage, SAAT, and VAT. Visit number (0 to 4) was treated
as a repeated measure, and model-based estimates of mean differences
between baseline and subsequent visits were calculated. All analyses were
conducted using SAS v9.2 (SAS Institute, Inc., Cary, NC) with statistical
significance set at p= 0.05.
Results
Participants with missing data were excluded from the analyses. The
original sample size was 79 participants. Of these, 14 had no physical activity
data, 9 did not have a sufficient number of days of physical activity data, 1 had
missing leptin data, and 5 had missing MRI data. The final sample size for these
analyses was 50 participants. Excluded participants did not significantly differ on
baseline age, % Hispanic, BMI percentile, and body composition from those with
complete data.
Table 3-1 presents baseline sample characteristics as a whole and by
Tanner stage. The majority of the analyzed sample was Hispanic (78%) with a
mean age of 9.4 (±0.9) years. The sample recorded on average 47.4 (±28.0)
min/day of MVPA. Participants in Tanner stage 1 had significantly lower SAAT,
VAT, and leptin levels (p< .010 for all) compared to those in Tanner stage 2.
The cross-sectional baseline linear regression results are presented in
Table 3-2. The model explained 48% of the variance in mean MVPA (min/d). Age
77

(p= .014) and leptin (p= .013) were negatively associated with mean minutes/day
spent in MVPA independent of central adiposity. African Americans participated
in approximately 23 minutes more min/day of MVPA than Hispanic females (p<
.001). The leptin-Tanner stage interaction included in the cross-sectional analysis
was not significant, and therefore was not included in the mixed models.
Table 3-3 presents the mixed model results. Over one year, model-based
estimates indicated that MVPA decreased by 12.3%, from 52.4 min/day at
baseline to 46.0 min/day at the Year 2 visit (Figure 3-1). The effect of baseline
leptin on the overall change in MVPA was investigated. Baseline leptin
concentrations significantly predicted the decline in MVPA over all five visits (β ±
SE= -1.1 ± 0.4, p= .017. African American females averaged 21.7 more min/day
of MVPA than Hispanic females per visit (p= .006). Those in Tanner stage 2 had
7.2 fewer min/day in MVPA per visit than Tanner stage 1, however this difference
was not statistically significant (p= .292).
Discussion
The purpose of these analyses was to evaluate the cross-sectional and
longitudinal effects of leptin on physical activity in a sample of minority female
youth. The main finding of this study was that the inverse relationship between
leptin concentrations and physical activity at baseline remained stable in a
longitudinal model. The first hypothesis, that leptin would be negatively
associated with MVPA in the cross-sectional analyses, was supported. However,
78

Tanner stage did not moderate this relationship. Physical activity decreased at
each visit. In the longitudinal model, high baseline leptin concentrations were
related to a decline in physical activity levels over one year.
Pate et al. reported a similar decline in females of 2.1% per year in MVPA
from 6
th
to 8
th
grade [149]. Findings from previous studies on the relationship
between leptin and physical activity have been inconsistent. Romon et al.
reported leptin was negatively correlated to pedometer-calculated steps per day
in females only [170] in a cross-sectional sample of 510 White youth aged 8 to 18
years. Conversely, in a sample of 125 Pima Indian children Salbe et al. (1997)
found that plasma leptin concentrations were positively associated with physical
activity levels [175]. However, this relationship was not found in a longitudinal
study of 213 healthy children, where accelerometer-measured physical activity
did not significantly correlate with leptin levels [126]. Our results support the
findings of Romon et al., suggesting that higher circulating leptin concentrations
are associated with lower levels of physical activity in girls. There may be
correspondence in findings from Romon et al. et al. in part because objective
measures of physical activity were used in both studies, providing more robust
estimates of activity in youth than other measures. Also, the female samples
between these two studies are more similar than that in Sable et al. and Metcalf
et al. who did not stratify by gender. This may have influenced their findings
because there are gender differences in leptin levels during puberty [86]. Taken
together, these findings indicate that high leptin concentrations act to suppress
79

physical activity in females. This is counterintuitive to the typical leptin
mechanism. In the normal response, high leptin levels signal the brain that there
is excessive intake, which results in decreased appetite and increased energy
expenditure [183]. However, the above-normal leptin levels (normal average: 6.5
ng/ml vs. sample average: 15.2 ng/ml) [99] in this sample indicate a degree of
leptin resistance whereby the body has a blunted response to leptin and less
resistance to obesity [5]. This blunted response to leptin includes decreased
energy expenditure, particularly in obese individuals. Although a minority of this
sample was obese, the initiation of puberty may act to raise leptin levels to
above-normal concentrations in females and account for the observed decrease
in physical activity.
Previous findings indicate that leptin concentrations increase at the onset
of puberty in females, once fat stores reach a sufficient level [86]. These higher
leptin concentrations during the onset of puberty have been related to the
observed decline in physical activity and increase in insulin resistance [1, 167].
However, Tanner stage was not significant in either the cross-sectional or
longitudinal models. This finding may be attributed to the use of baseline data
and the narrow range of stages in this sample (stages 1 and 2 only). A longer
follow-up and multiple waves of data may be needed to determine the effect of
Tanner stage on the leptin-physical activity relationship.
To our knowledge, this is the first study to demonstrate the stability of this
relationship in Hispanic and African American females. The majority of studies
80

have assessed only the cross-sectional relationship in predominantly non-
minority populations, and none have used accelerometry to measure physical
activity. There are some study limitations that must be addressed. This analysis
included only baseline leptin concentrations and therefore prevented the
assessment of the simultaneous change in leptin and physical activity over one
year. Collecting leptin at more frequent intervals may help improve the ability to
detect simultaneous changes in leptin and physical activity as the females initiate
puberty. The small sample size precluded stratified analysis based on race/ethnic
group. There are race/ethnic differences in physical activity levels and future
studies should attempt to discern whether these differences are a result of
variations in leptin levels between groups.
In conclusion, high baseline leptin levels were predictive of decreased
physical activity in minority females over one year. Based on expected biological
mechanisms, this result is opposite to what would be expected as the normal
response to high leptin levels is to decrease intake and increase energy output.
However, the high baseline leptin levels indicated that that this sample may be
leptin resistant, resulting in decreasing activity levels. Although there are many
target areas for leptin in the body, the arcuate nucleus region in the brain has
been proposed to mediate the effect of leptin on physical activity [38, 129]. The
diminished ability of the leptin receptors to reach these target areas influence
activity levels [96]. Thus, high baseline leptin levels like those seen in this study
may indicate a degree of central leptin resistance in the brain that leads to an
81

abnormal response (reduced physical activity). However, previous studies have
also shown this inverse relationship between leptin and physical activity. To our
knowledge this is the first study to find this relationship in a longitudinal model
with minority youth. These findings add to the growing support for the biological
basis of declining activity levels in pubertal females and may explain in part why
theory-based interventions to increase physical activity have shown little success
to date.
82

Table 3-1: Baseline descriptive characteristics (mean (SD)) of analytical sample
by Tanner Stage (N=50)

Variable
All
(N=50)
Tanner 1
(n= 26)

Tanner 2
(n= 24)

p value

Ethnicity (% Hispanic)
1
78.00 84.62 70.84 .240
Weight Status (% Obese)
2
38.00 23.08 54.17 .123
Age (years) 9.38 (0.90) 9.15 (0.97) 9.63 (0.77) .064
Insulin Sensitivity (SI) 3.23 (1.95) 3.75 (1.97) 2.67 (1.79) .048
Leptin (ng/ml)
15.21 (10.11) 11.37
(7.64)

19.37
(10.93)
.004
SAAT (L) 4.46 (3.22) 3.04 (1.88) 6.00 (3.67) .001
VAT (L) 0.69 (0.60) 0.46 (0.37) 0.94 (0.71) .005
Mean MVPA (min/day)
47.40 (27.98) 50.37
(27.67)

44.18
(28.55)
.440
SD= standard deviation; SAAT= subcutaneous abdominal adipose tissue; VAT=
visceral adipose tissue; MVPA= moderate to vigorous physical activity
1
frequencies
2
body mass index (BMI) ≥ 95
th
as defined by the 2000 CDC age- and gender-
specific cut points [102].

83

Table 3-2: Baseline Generalized Linear Model (N=50)
Variable
MVPA
Model R
2
= 0.48

β (SE) p value
Intercept 185.50 (41.20) < .001
Ethnicity (Hispanic vs. AA) -23.27 (8.56) .010
Tanner (1 vs. 2) -11.32 (13.38) .403
Age (yrs) -9.99 (3.87) .014
SAAT (L) 0.05 (2.20) .982
VAT (L) -2.07 (10.82) .849
SI 0.29 (2.44) .907
Leptin (ng/ml) -1.43 (0.55) .013
Leptin*Tanner 0.30 (0.74) .692
MVPA= moderate to vigorous physical activity; SE= Standard Error; AA= African
American; SAAT= subcutaneous abdominal adipose tissue; VAT= visceral
adipose tissue; SI= insulin sensitivity
84

Table 3-3: Longitudinal mixed model predicting change in moderate to vigorous
physical activity over 5 visits (N=50)
Baseline Variable MVPA
β SE t p
Intercept 139.07 36.46 3.81 <.001
Ethnicity (vs. AA) -21.75 7.42 -2.93 .006
Tanner (vs. 2) -7.19 6.73 -1.07 .292
Age (yrs) -6.30 3.41 -1.84 .072
SAAT (L) 0.08 1.90 0.04 .968
VAT (L) -6.36 9.23 -0.69 .494
SI 0.71 2.13 0.33 .741
Leptin (ng/ml) -1.09 0.44 -2.50 .017
Visit (vs. 0)
1 -0.09 3.00 -0.03 .976
2 -1.72 3.27 -0.52 .601
3 -9.12 3.75 -2.43 .017
4 -6.42 3.47 -1.85 .068
MVPA= moderate to vigorous physical activity; SE=standard error; AA= African
American; SAAT= subcutaneous abdominal adipose tissue; VAT= visceral
adipose tissue; SI= insulin sensitivity
Visit has been coded as 0-4

85

Figure 3-1: Model-based estimates of mean MVPA (min/day) by visit

MVPA= moderate to vigorous physical activity
* visit 3 is statistically significantly different from visit 0 (p= .017)

*
86

CHAPTER 4 THE INFLUENCE OF MEAL TYPE ON PHYSICAL ACTIVITY IN
MINORITY ADOLESCENTS: THE FOOD ADOLESCENCE, MOOD, AND
EXERCISE 2 (FAME) STUDY
Introduction
Youth in the United States are increasingly overweight [133]. Poor diet
and low levels of physical activity are contributing to the observed rise in obesity
prevalence. Recent national estimates of physical activity indicated that
overweight and obese youth spend fewer minutes per day in moderate to
vigorous physical activity (MVPA) than normal weight youth [12]. Also, data from
1998 to 2002 of the National Health and Nutrition Examination Survey (NHANES)
showed that youth who met the criteria for central adiposity (waist circumference
≥ 85
th
percentile) reported less dairy, grain, and fruit and vegetable intake [21].
This problem is more severe in minority youth than White youth. According to
recent national data, 21.1% of Hispanic youth are obese (age- and gender-
specific body mass index (BMI) ≥95
th
percentile), compared to only 16.0% of
non-Hispanic Whites [135].
Hispanic youth have diets high in added sugar [217]. In a population-
based sample of adolescents, Hispanic boys and girls consumed more sugar
sweetened beverages than Caucasian White youth [164]. Evidence suggests that
high sugar meals result in elevated glucose, insulin, and leptin responses that
are a-typical when compared to the physiologic response to low carbohydrate
and sugar meals [10, 112, 172]. These a-typical physiological responses put
87

youth at increased risk for type 2 diabetes and metabolic syndrome [217, 218].
Understanding the mechanisms that underlie the increased disease risk may
help inform future interventions and public policies.
While research has been conducted on the independent contributions of
diet and physical activity to obesity, less work has addressed the direct influence
of diet on physical activity levels in youth. Physical activity has been related to
higher caloric intake. In a sample of predominantly normal weight youth, MVPA
(MVPA) measured via accelerometry was positively associated with total energy
intake in boys, but not in girls [93]. Similar results were found in a sample of 472
youth ages 10 to 14 years, MVPA and energy intake were significantly positively
correlated after controlling for age [69]. However, intake of specific nutrients
might influence physical activity. For instance, diets high in simple carbohydrates
have been associated with poor glycemic control [122, 130, 141, 142], poor
mood, feelings of fatigue, and low levels of physical activity [198, 199]. To better
understand the influence of diet on physical activity, research should address
how specific nutrients influence activity behavior in youth.
Few studies have examined the effects of specific nutrients on physical
activity despite the fact that specific nutrients, specifically added sugar, have
been shown to influence other behaviors [198]. In-lab feeding studies have
attempted to elucidate the effects of these nutrients on physical activity by
offering a controlled environment and opportunities for using measures that are
not feasible in free-living studies. In-lab studies conducted in non-minority youth
88

have found that foods high in refined carbohydrates resulted in increases in
subsequent food intake and insulin and glucose responses [8, 113, 221]. Jago et
al. (2004) found in a sample of 210 African American females that MVPA
measured via accelerometry was negatively associated with percentage of
calories from fats and carbohydrates after controlling for total caloric intake [91].
Conversely, Thompson et al. (2004) reported no significant covariation in
changes in physical activity and specific foods in African American girls over 12
weeks [200], but the authors noted that this relationship may have been
influenced by the small sample size and short study duration.
The relationship between specific nutrients commonly found in adolescent
diets and physical activity should be elucidated because if diet adversely
influences activity levels, interventions targeting increasing physical activity will
also need to take these nutrients into account. To our knowledge, only one in-lab
feeding study has assessed the relationship between specific nutrients and
physical activity in minority youth. Spruijt-Metz et al. (2009) reported that a high
sugar/low fiber meal resulted in significantly different patterns of activity over two
hours of observation. Compared to the low sugar/high fiber meal, there was a
significant burst of activity in the first 30 to 60 minutes post- high sugar/low fiber
meal [189]. We tested the effect of a high-sugar/low fiber versus a low sugar/high
fiber meal on metabolic profiles and behavior in a population of overweight
Hispanic youth using a laboratory protocol and a crossover design. The objective
of the FAME 2 study is to determine how diet impacts insulin indices, physical
89

activity, and sedentary behavior in minority adolescents. These analyses is
driven by the following hypotheses which in part follow from the findings from our
pilot study [189]:
1. Activity Levels
1.1a): The high sugar/low fiber meal will result in decreased total time spent in
MVPA compared to the low sugar/high fiber meal.
1.1b): The high sugar/low fiber meal will result in increased total time spent in
sedentary behavior compared to the low sugar/high fiber meal.
1.2a): The high sugar/low fiber meal will result in lower levels of MVPA at each
30-minute increment over five hours compared to the low sugar/high fiber
meal.
1.2b): The high sugar/low fiber meal will result in higher levels of sedentary
behavior at each 30-minute increment over five hours compared to the low
sugar/high fiber meal.
2. Biomarkers
2.1a) The high sugar/low fiber meal will stimulate greater insulin area under
the curve (AUC) than the low sugar/high fiber meal for the entire five hours.
2.1b) The high sugar/low fiber meal will stimulate greater glucose AUC than
the low sugar/high fiber meal for the entire five hours.
2.2a) The high sugar/low fiber meal will stimulate greater insulin AUC between
the 30- to 60- minute increments compared to the low sugar/high fiber meal.
90

2.2b) The high sugar/low fiber meal will stimulate greater glucose AUC
between the 30- to 60- minute increments compared to the low sugar/high
fiber meal.
3. Mediation
3.1a): Insulin AUC will mediate the relationship between meal type and MVPA.
3.1b): Glucose AUC will mediate the relationship between meal type and
MVPA.
Methods
Participants
Participants were recruited from participating clinics, hospitals, churches,
schools, and community centers in the Los Angeles area. The study inclusion
criteria were as follows: Hispanic or African American ethnicity, in 9
th
to 11
th

grade (approximately 14 to 17 years old), who were at Tanner stage 5 [197] and
had a BMI ≥ 85
th
percentile. Because we have shown metabolic differences in
insulin resistance and compensatory responses between African American and
Hispanic youth [72], this analysis includes data collected at the free-living and in-
lab visits for Hispanic youth only. Informed written parental consent and
participant assent (in the primary language of each) was obtained prior to
participation in any of the abovementioned procedures. The study was approved
by the Institutional Review Board of the University of Southern California.
91

Procedures
Free-living physical activity measurement: The preliminary data collection
included a two-week free-living measurement time period in which activity levels
and psychosocial measurements were taken. To determine baseline activity
levels, free-living data collection took place approximately two weeks prior to the
first in-lab visit. The participants completed demographic and psychosocial
questionnaires and wore an accelerometer. A study team member contacted the
participants several times during the week to remind them to wear the
accelerometers. At the end of the collection period, the accelerometer data was
downloaded and checked for completeness. If there were less than four 10-hour
days of data, the participants were asked to wear the accelerometer a second
time. The data from both accelerometer wears was combined to ensure the
participants had adequate data.
Overnight hospital visit: The participants then came to the Clinical Trials
Unit (CTU) and if they met the eligibility criteria, they completed an inpatient stay
where diet, body composition, and psychosocial variables were assessed.
Participants were excluded at the CTU visit if they met any of the following
criteria: evidence of diabetes, currently in a weight loss or exercise program, use
of medications that influence body composition and insulin sensitivity, or
conditions that affect body composition. After the CTU visit, eligible participants
were scheduled to complete two in-lab feeding visits within approximately two
months at the Physical Activity Lab (PAL). The in-lab feeding study assessed
92

insulin indices, selected gut hormones, psychosocial measures of mood, and
activity levels via accelerometry and direct observation.
In-lab feeding study: The in-lab feeding study employed a cross-over
design, consisting of two separate visits to the Physical Activity Lab (PAL)
separated by a 2- to 4- week wash-out period. Each participant received both in-
lab meals (high sugar and high fiber) in random order. A different standardized
test meal (high sugar or high fiber) was given at each visit. The PAL is supplied
with several opportunities for active and sedentary activities. For example, there
is a treadmill, Wii Fit, and Dance Dance Revolution for active options and a
television, movies, and books for sedentary options. The participants arrived at
the lab at approximately 0700 hours after a 10-hour overnight fast. The
participant was fitted on the left hip with an Actigraph GT1M uniaxial
accelerometer. A small saline lock intravenous catheter was inserted into the
forearm by a registered nurse. The first baseline blood draw was taken 5 minutes
prior to the breakfast meal (either high sugar/low fiber or low sugar/high fiber).
Participants had 15 minutes to consume the meal in its entirety. Blood samples
were taken at 30-minute intervals for five hours. After five hours, the saline lock
was removed and the participants were served a second test meal identical to
the first one that was completed in 15 minutes. The participants remained in the
lab for an additional three hours of observation of activity levels. They had
access to an ad libitum food tray with a variety of healthy and unhealthy options.
93

The second in-lab visit was scheduled approximately two to four weeks later and
the process was repeated with the other test meal.
Measures
Demographics & Physical Exam: Age (in years) was recorded from the
protocol flowsheets at the inpatient CTU visit. Race/ethnicity was self-reported by
the participants. Tanner stage (breast development for females and testicular
volume for males) was assessed by the licensed pediatric medical care provider
at the inpatient CTU visit. Height (cm) and weight (kg) were taken in triplicate by
a registered nurse at the CTU. Body mass index (BMI) and BMI percentile were
calculated in triplicate from the height and weight measurements.
Body Composition: Fat and lean mass (kg), percent fat, and trunk fat (kg)
was measured via dual-energy x-ray absorptiometry (DEXA) using a Hologic
QDR 4500 densitometer (Hologic Inc, Bedford, MA).
Clinical Measures: At the CTU, the Frequently Sampled Intravenous
Glucose Tolerance Test (FSIVGTT) consisted of collection of blood samples at -
15, -5, 2, 4, 8, 19, 22, 30, 40, 50, 70, 100, and 180 minutes. At time 0, glucose
(25% dextrose, 0.3g/kg body weight) was administered intravenously. Insulin
(0.02 units/kg body weight) was injected at 20 minutes. Insulin sensitivity (SI)
was calculated using the minimal model from data from the FSIVGTT [13]
[MINMOD MILLENIUM 2002 computer program, v5.16; Richard Bergman, Los
Angeles, CA [143]].
94

Table 4-1 presents the in-lab timing of measures. Blood samples were
collected at -5 minutes and at 30-minute increments starting with the meal (Time
0) for five hours. Samples were centrifuged in microfuge on-site within one hour
of blood draw, placed on ice, and transported on dry ice to the core lab where all
metabolic assays were performed. Samples were stored at –70 C until assayed.
Insulin was assayed with an automated random access enzyme immunoassay
system Tosoh AIA 600 II analyzer (Gibbco Scientific, Inc. Coon Rapids, MN)
using an immunoenzymemetric assay (IEMA) method. Blood samples for
glucose were analyzed on a Dimension Clinical Chemistry system and an in vitro
Hexokinase method (Dade Behring, Deerfield, IL). Insulin area under the curve
(IAUC) and glucose area under the curve (GAUC) were calculated from the in-lab
data. Physical Activity: Physical activity and sedentary behavior were assessed
via accelerometry for the free-living and in-lab visits. During the free-living data
collection, the participants wore a uniaxial Actigraph GT1M accelerometer for at
least four 10-hour days. The monitor was programmed to collect data in 15-
second epochs. The data was processed using a modification of the SAS code
developed by the National Cancer Institute for use with NHANES data (available
at: http://riskfactor.cancer.gov/tools/nhanes_pam) where the 15-second epochs
were compounded into 60-second epochs for current activity thresholds. Non-
wear time was defined as 60 minutes or more of 0 intensity counts, with
allowance for up to 2 consecutive minutes of counts between 1 and 100. The
non-wear period ended when at least one of the following conditions were met:
95

one minute of an intensity above 100; one minute with a missing intensity count;
3 or more minutes with intensity counts between 1 and 100 counts; or the last
minute of the day. Wear time was defined by subtracting non-wear time from the
total observation time for each day.
A different accelerometer protocol was used for the in-lab visits. The start
time and end time for each visit was recorded from the study flowsheets and the
data was processed using a modified version of the NHANES program. This
modified program does not remove supposed ‗non-wear time‘ since that was not
applicable to the in-lab visits where the participant was wearing the device
continuously. Accelerometer data before and after the start/end points was
removed so that only data for the observation period was retained. We aimed to
collect data in 1-second epochs, however some participants had data collected in
15-second (N= 18) and 60-second (N= 1) epochs. The data were collapsed into
60-second epochs and time spent in MVPA and sedentary behavior was
partitioned into 30-minute increments. Total time spent in MVPA and sedentary
behavior over the five hours was calculated by summing the seconds with counts
above and below the user-defined thresholds. Incremental time spent in MVPA
and sedentary behavior for each 30-minute interval was calculated by summing
the seconds with counts above and below the thresholds for each interval over
the five hours.
Free-living accelerometer data is presented as mean minutes per day
spent in MVPA and sedentary behavior. Time spent in the activity levels was
96

calculated by summing each minute spent above the user-defined thresholds for
MVPA and below the user-defined threshold for sedentary activity. The cut points
for moderate (4 METs) and vigorous (7 METs) activity were age-adjusted using
the criteria from the Freedson group [68]. The sedentary cut point of 100 counts
was previously defined by Matthews et al. [119]. This cut point was validated in
adolescents [205] and adults and found to be a good estimation of time spent in
sedentary behavior [82].
Experimental meals: All meals were prepared under the supervision of a
registered dietician at the University of Southern California. All meals were
prepared under the supervision of the research registered dietitian at the
University of Southern California Institute for Prevention Research. Meals were
developed using data from focus groups conducted with African American and
Hispanic youth. In all, 31 youth (9 Hispanic boys, 10 Hispanic girls, 5 African
American boys and 7 African American girls) participated in these groups. Youth
were also asked to identify favorite breakfast and snack foods. Two intervention
meals (a high sugar, low fiber meal and a low sugar, high fiber meal) were
developed using data from the focus groups. Intervention meals are similar in
calorie content and percentage of all nutrients except the nutrients under study
(i.e. sugar and fiber). Nutrient compositions were determined using the Nutrient
Data System for Research (NDS-R 2010, University of Minnesota, Minneapolis,
MN). Table 4-2 presents the meal composition information for each meal. The
high sugar meal consisted of a regular Poptart (Kellogg NA Co., Battle Creek.
97

MI), sticks of calcium enriched string cheese (Sargento Mootown Light String
Cheese, Sargento Foods Inc., Plymouth, WI), and Tampico juice (Tampico
Beverages, Chicago, IL). The high fiber meal consisted of a whole wheat bagel,
margarine (I Can‘t Believe It‘s Not Butter Light, Unilever PLC/Unilever N.V.,
Englewood Cliffs, NJ), and water treated with Benefiber Powder (Novartis
Consumer Health, Inc., Parsippany, NJ). The size of each of the test meals is
determined individually for each subject as ~20.0% of predicted resting metabolic
rate (RMR). This is calculated from gender, age, height, and body weight using
the Dietary Reference Intakes Guidelines Estimation of Energy Expenditure for
overweight children ages 3-18 [138]. The amounts of the items in the two meals
are adjusted based on these characteristics so that the % kcal from each nutrient
remains the same across participants. Food items were served to the nearest
half piece due to food shapes and to preserve food products from degrading.
Each test meal was given twice (Table 4-1). The first meal was provided at
Time 0, and participants were instructed to eat the meal within 15 minutes. The
same test meal was given a second time at 240 minutes. After 310 minutes, ad
libitum platters with pre-weighed items were provided for the participants and
included both high sugar and high fiber choices. Plate waste from the platters
was weighed to determine the amount of food consumed.
98

Data Analysis
For the purpose of this paper, data from the initial four hours of
observation were analyzed. Significance level was set at α=.05 for all the
statistical tests performed in this study. T-tests were used to assess gender
differences in the baseline measures. For hypotheses 1 and 2, mixed models
were used to assess the influence of each meal type on MVPA, sedentary
behavior, insulin, and glucose over the entire five hours (hypotheses 1.1a, 1.1b,
2.1a, and 2.1b) and at each 30-minute segment (hypotheses 1.2a, 1.2b, 2.2a,
and 2.2b) with a total of 9 repeated measures for each participant. Time-invariant
variables considered in this analysis were the baseline values for age, SI, body
composition, meal order, Tanner stage, and gender. Time for each repeated
measure being observed was also included as a covariate. Data were analyzed
using PROC MIXED in SAS v9.2 [SAS v9.2 (SAS Institute, Cary, NC)] with a
REPEATED statement. Variance components were estimated using the
restricted maximum likelihood (REML) estimate method to control for within-
individual correlations and an autoregressive covariance structure. To evaluate
differences in the dependent variables at each time point, planned comparisons
of model-based mean differences between the specified time points were
calculated with the Bonferroni adjustment.
Mixed models were also used to analyze hypotheses 3.1a and 3.1b, to
determine if insulin and glucose mediate the relationship between meal type and
MVPA. There are four criteria for mediation according to Baron and Kenny [11].
99

Criterion one: Meal type must predict insulin or glucose AUC (path 1). Criterion
two: insulin or glucose IAUC must predict MVPA while controlling for meal type
(path 2). Criterion three: meal type must significantly predict MVPA (path 3).
Criterion four: the mediation effect (calculated as the product of the regression
coefficients from criteria one and two) must be statistically significant by the
Sobel test [187]. These models assessed the effects of the independent
variables on the dependent variables over the entire observation time, controlling
for nesting within meal type. All models controlled for visit order to correct for
potential effects of participant familiarity with the lab surroundings and
procedures at the second visit. Covariates are baseline age, SI, body
composition, meal order, Tanner stage, and gender. The cross-over design
allows for the examination of within-individual comparisons of meal responses
and within-individual longitudinal effects. This design removes the between-
individual variation from diet-related differences.
Results
Descriptive Statistics
The characteristics of the sample are presented in Table 4-3. The sample
consisted of 50 Hispanic participants (25 male, 25 female). Of these, one was
excluded due to missing in-lab accelerometry data, two were excluded due to
missing body composition measures, and three were excluded due to missing SI
data. The analytical sample consisted of 44 participants. There were no
statistically significant differences in demographic variables between those
100

included and excluded from the analysis. Seventy-five percent of the participants
were categorized as obese according to the CDC age- and gender- specific
growth charts [102]. Females had significantly lower lean mass and fasting
glucose levels than males (p< .05 for both). Males spent approximately 7 more
minutes per day in MVPA (21.9 versus 14.8 minutes) and 24 more minutes per
day in sedentary behavior (571.1 versus 547.0 minutes) (p> .05 for both) than
females.
Effect of Meal Type on Activity Levels
Figures 4-1 and 4-2 present the mean activity levels by meal type for
MVPA and sedentary behavior, respectively. Results from the models for activity
levels are presented in Table 4-4. The ‗time*meal type‘ interaction term indicates
that meal type did not statistically significantly predict change in MVPA (F
7, 301
=
1.72, p= .104) and sedentary behavior (F
7, 301
= 1.40, p= .204) over the
observation period. The non-significant ‗meal‘ term in the model indicated that
there were no significant mean differences in overall mean MVPA or sedentary
behavior by meal (β±SE= -0.02 (0.48), p=.967 for MVPA; β±SE= -0.74 (0.77), p=
.340 for sedentary behavior). The only significant difference in mean minutes
spent in MVPA between meals was at 180 minutes. At this time point, the MVPA
was significantly higher in the high fiber condition (p= .029). Also at 180 minutes,
there was a trend for the high sugar meal promoting more time spent in
sedentary behavior (p= .058).
101

Effect of Meal Type on Insulin & Glucose Levels
Figures 4-3 and 4-4 present the mean insulin and glucose levels by meal
type. Results from the models for insulin and glucose are presented in Table 4-5.
The ‗time*meal type‘ interaction term indicates that meal type significantly
predicted change in insulin IAUC (F= 9.39, p< .001) and glucose IAUC (F= 3.94,
p< .001) over the observation period. However, the non-significant ‗meal‘ term in
the model indicated that there were no significant mean differences in overall
insulin IAUC (p= .740) and glucose IAUC (p=.779) by meal type. As
hypothesized, the high sugar meal resulted in higher insulin IAUC at the 30- and
60- minute time points (p< .001 for both) than the high fiber meal. Also consistent
with hypotheses 2.2a and 2.2b, the high sugar meal produced higher glucose
IAUC (p< .001) at 30- and 60- minutes post-meal compared to the high fiber
meal. The only other significant difference in mean glucose levels was seen at
120 minutes, where the levels were significantly higher in the high sugar meal
(p= .036).
Mediation Model
Insulin and glucose were examined as potential mediators in the meal
type effect on MVPA over the entire observation period. Meal type did not predict
overall mean differences in insulin and glucose IAUC over the observation
period. Insulin IAUC predicted lower mean minutes spent in MVPA in the high
sugar condition, however meal type did not predict MVPA levels. Thus, the
criteria for mediation were not met.
102

Discussion
In this randomized cross-over study design, meal type (high sugar or high
fiber) did not have an effect on activity levels or rate of change across time points
over four hours of observation. The only significant meal differences in activity
were seen at 180 minutes where the high sugar condition resulted in less time
spent in MVPA and more time spent in sedentary behavior. Based on findings
from the pilot study [189], we hypothesized that there would be differences in
activity levels by meal condition at 30- and 60- minutes post-meal. The high fiber
condition did show a non-significant increase in MVPA from 30- to 60-minutes,
concurrent with the increase in insulin and glucose during this time period.
Although not significant, Figure 4-1 shows that MVPA increased in the 120- to
150- minute time block in the high sugar condition, which was concurrent with the
second drop in insulin levels. This is consistent with previous research that
shows an inverse relationship between insulin and physical activity in youth [22,
178]. The absence of an effect of meal type on activity levels may be due to the
very small fluctuations in activity levels over the observation period. Sedentary
behavior remained nearly stable and changes in MVPA were on the magnitude of
half a minute between time periods. Given the small changes in activity,
compounding the accelerometer output into 60-second epochs may have
prevented us from detecting significant changes in activity. Furthermore, the
design of the accelerometers makes them less sensitive to changes in activity in
sedentary behaviors and therefore may have influenced our estimates [190].
103

Compared to the high fiber condition, the high sugar condition resulted in
higher insulin and glucose IAUC levels and different rates of change over the four
hours. Insulin and glucose lAUC were significantly higher at the 30- and 60-
minute time points in the high sugar condition. This finding is consistent with the
findings from the pilot study [189]. These differences are a result of the body‘s
compensatory response to the sucrose in the high sugar condition. When food is
consumed, there is a sharp increase in circulating glucose levels as energy is
absorbed into the bloodstream. This increase in blood glucose stimulates insulin
secretion by the pancreas, which promotes glucose uptake in muscle and
adipose tissue [112]. The excess sugar in the high sugar condition resulted in an
a-typical physiologic response (high insulin and glucose levels) when compared
to the low carbohydrate and sugar meals [10, 112, 172]. Based on the fact that
youth consume diets high in excess sugar, the body‘s continual a-typical
response to food stresses the system and may result in disease over the lifetime.
A major finding of this study was that insulin and glucose did not mediate
the relationship between meal type and activity levels over the initial four hours of
observation. To our knowledge, this is the first study to investigate this pathway
using an in-lab feeding study design. Diets high in simple carbohydrates (which
are rapidly metabolically transformed into sugar) have been associated with low
levels of physical activity [198, 199]. We expected that the high sugar condition
would have resulted in lower overall physical activity than the high fiber condition,
which elicits a lower glucose response that has been associated with lower
104

fatigue scores [144]. However, meal composition did have an effect on insulin
and glucose IAUC over the entire observation period and at specific timepoints.
Compared to the high fiber condition, the high sugar condition elicited higher
insulin and glucose IAUC at 30 and 60 minutes post-meal. Insulin IAUC predicted
physical activity levels: higher insulin IAUC predicted lower MVPA in the high
sugar condition. It is possible that the overweight youth were accustomed to the
high sugar content of the meal and these youth were so profoundly inactive, that
the acute effect of meal type was not sufficient to change habitual ingrained
behavior. Future research should assess whether long-term consumption of high
fiber foods impact activity levels in this population.
Several limitations of this study bear discussion. Participants were aware
that they were being observed during the study period and they were limited to
the activities offered in the lab. This may have influenced their behavioral
choices, however the statistical control for visit and randomization accounted for
systemic variation. The in-lab acute feeding study design limits generalizability to
other settings. As previously mentioned, the epoch length on the accelerometers
may have been insensitive to categorization of activity into the two levels. Direct
observation methods may provide better estimates of time spent in activity levels.
Therefore, the in-lab video is currently being analyzed and future analyses will
include this as a measure of activity levels.
Diets high in fiber have been associated with improved insulin sensitivity
[193] and lower visceral fat [44] in youth. Conversely, diets high in sugar have
105

been associated with worse metabolic outcomes such as higher insulin and
glucose loads [88]. This study adds to the literature showing that meal effects on
insulin and glucose have not only long-term, but also immediate effects.
Deleterious effects of high sugar meals are acute and may be exacerbated by
obesity status. Health has been shown to be a prevalent worry in Hispanic youth
but worries do not predict weight status [12] and knowledge of potential health
consequences is only weakly associated with dietary sugar consumption in youth
[120]. Future research should attempt to incorporate findings from this and other
in-lab studies into interventions aimed toward reducing obesogenic behaviors in
youth.
106

Table 4-1: Timing of in-lab measures

107

Table 4-2: Meal compositions
Macronutrient
g (% kcal)
High Sugar Meal

High Fiber Meal
54.0g Poptart 61.0g whole wheat bagel
42.0 string cheese 14.0g margarine
247.0g juice 10.5g fiber supplement
Fat 11.0 (23.6%) 9.5 (23.8%)
Carbohydrate 64.0 (61.0%) 61.0 (67.8%)
Protein 14.0 (13.3%) 10.0 (11.1%)
Sugar 41.0 (39.1%) 7.0 (7.8%)
Fiber 1.0 (1.0%) 16.0 (17.8%)

108

Table 4-3: Baseline sample descriptive statistics (N=44)
Variable
Mean (SD)
or Frequency
Males
(N= 22)
Females
(N= 22)
p value
Age (years) 15.34 (1.06) 15.64 (1.14) 15.05 (0.90) 0.063
BMI (kg/m
2
) 32.87 (5.65) 33.15 (6.86) 32.62 (4.43) 0.770
Tanner Stage (%)
1
Stage 4
Stage 5

13.64%
86.4%

13.64%
86.4%

13.64%
86.4%

1.00
1.00
Total Fat Mass (kg) 32.57 (10.30) 32.28 (13.22) 32.86 (6.50) 0.854
Total Lean Mass (kg) 56.28 (10.65) 62.50 (9.28) 50.05 (8.06) <0.001
Glucose (mg/dl) 93.18 (7.02) 95.91 (5.65) 90.45 (7.30) 0.008
Insulin (μU/ml) 26.57 (21.32) 28.64 (25.39) 24.50 (16.65) 0.526
Insulin Sensitivity (SI) 1.67 (0.95) 1.78 (1.06) 1.57 (0.83) 0.465
Mean MVPA
(min/day)
3
18.99 (15.23)
21.88 (17.69) 14.75 (9.87) 0.156
Mean SB (min/day)
3
561.33 (77.33)
571.10 (86.15) 547.04 (62.80) 0.396
SD= standard deviation; BMI= body mass index; MVPA= moderate to vigorous
physical activity; SB= sedentary behavior
1
frequencies
2
BMI ≥ 95
th
as defined by the 2000 CDC age- and gender-specific cut points
[102].
3
N=32 (12 participants had fewer than four 10-hour days of free-living
accelerometer data)
109

Table 4-4: Mixed model to examine the effect of meal type on activity levels

MVPA= moderate to vigorous physical activity; SB=sedentary behavior; SE=
standard error.
1
F statistic with df= 7, 301
MVPA (min) SB (min)
Variable β SE p value β SE p value
Intercept -2.21 1.74 .213 35.52 2.90 <.001
Gender (Male vs. Female) -3.32 0.28 .238 0.53 0.46 .256
Tanner stage (4 vs. 5) -0.24 0.31 .443 1.19 0.52 .027
Age (years) 0.20 0.11 .073 -0.43 0.18 .026
Total Fat Mass (kg) 0.01 0.01 .366 -0.02 0.02 .323
Total Lean Mass (kg) -0.01 0.01 .378 0.00 0.02 .974
Insulin Sensitivity (SI) 0.16 0.13 .224 -0.33 0.22 .146
Meal order (1=HS, HF vs. 2= HF,
HS)
-0.12 0.20 .559 -0.75 0.33 .029
Meal type (HS vs. HF) -0.02 0.48 .967 -0.74 0.77 .340
Time
1
1.10 - .363 0.73 - .645
Meal type*Time
1

HS vs. HF meal:
30 min
60 min
90 min
120 min
150 min
180 min
210 min
240 min
1.72

0.40
0.05
0.52
0.62
0.85
-1.05
0.03
0.02
-

0.48
0.48
0.48
0.48
0.48
0.48
0.48
0.48
.104

.404
.924
.276
.199
.078
.029
.949
.967
1.40

-0.65
-0.03
-0.91
0.51
-0.64
1.46
0.30
0.74
-

0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
.204

.398
.965
.235
.511
.403
.058
.696
.336
110

Table 4-5: Mixed model to examine the effect of meal type on insulin and glucose
IAUC
Insulin (μU/ml) Glucose (mg/dl)
Variable β SE p value β SE p value
Intercept 78.56 87.84 .377 3.87 16.84 .819
Gender (Male vs. Female) -10.89 14.15 .446 1.14 2.71 .678
Tanner stage (4 vs. 5) 5.30 15.82 .740 -2.44 3.03 .463
Age (years) -5.51 5.65 .335 0.38 1.08 .726
Total Fat Mass (kg) 1.50 0.68 .034 0.02 0.13 .894
Total Lean Mass (kg) 0.26 0.75 .731 -0.08 0.14 .600
Insulin Sensitivity (SI) -16.55 6.81 .020 0.69 1.31 .604
IAUC Insulin (μU/ml) - - - 0.07 0.01 <.001
Meal order (1=HS, HF vs. 2= HF,
HS)
5.41 6.13 .383 0.49 1.18 .680
Meal type (HS vs. HF) -4.47 13.39 .740 0.73 2.57 .779
Time
1
66.68 - <.001 72.39 - <.001
Meal type*Time
1

HS vs. HF meal:
-5 min
30 min
60 min
90 min
120 min
150 min
180 min
210 min
240 min
9.39

-0.25
59.28
48.62
-10.52
13.39
15.72
2.25
0.27
4.47
-

13.39
13.39
13.39
13.39
13.39
13.39
13.39
13.39
13.39
<.001

.985
<.001
<.001
.433
.318
.241
.866
.984
.739
3.94

-0.02
8.85
9.62
3.99
5.42
3.58
1.51
-0.29
-0.73
-

2.57
2.57
2.57
2.57
2.57
2.57
2.57
2.57
2.57
<.001

.993
<.001
<.001
.122
.036
.166
.557
.911
.778
IAUC= incremental area under the curve; SE= standard error.
1
F statistic with
df= 7, 301
111

Figure 4-1: Mean moderate to vigorous physical activity levels (min) by meal type

Figure 4-2: Mean sedentary behavior levels (min) by meal type

112

Figure 4-3: Mean insulin values (μU/ml) by meal type

Figure 4-4: Mean glucose values (mg/dl) by meal type

113

CHAPTER 5 SUMMARY AND CONCLUSIONS
Summary of Findings
The overall goal of this dissertation was to examine the effects of
individual demographic, biological, and dietary factors on objectively measured
physical activity levels in youth. The first objective was to describe activity levels
across race/ethnic, weight status, age, and gender groups in a large nationally-
representative sample of youth. The second objective was to examine cross-
sectional and longitudinal relationships between leptin and physical activity levels
in a sample of minority female children. The third objective was to assess the
effects of a high sugar versus a high fiber meal on activity levels in an in-lab
setting in minority adolescents.
The results from Study 1 demonstrated that activity levels differ across
race/ethnic group, by weight status, by age, and by gender. The findings
indicated that non-Hispanic White youth engaged in the lowest levels of activity
compared to other race/ethnic groups, yet had the lowest prevalence of obesity
in this sample. The highest prevalence of obesity in this sample was seen in non-
Hispanic Black females, the least physically active group. A different association
between weight status and activity was seen in Mexican American 12-15 year old
females, where obese and normal weight youth recorded the same levels of
physical activity.
114

The results from Study 2 show that leptin is inversely related to physical
activity, independent of central adiposity and pubertal status. In the baseline
cross-sectional model, higher leptin levels were associated with lower physical
activity levels, but not pubertal status. This relationship remained stable in the
longitudinal model. Specifically, higher baseline leptin levels predicted a decline
in physical activity over one year. These findings support a biological basis for
the declines in physical activity observed in adolescents as they begin puberty.
In Study 3, meal composition had little effect on time spent in physical
activity or sedentary behavior. Meal composition did predict change in insulin and
glucose IAUC during the observation period, although there were no overall
mean differences in these values. In addition, meal composition did have an
effect on insulin and glucose IAUC at specific time points. Compared to the high
fiber condition, the high sugar condition elicited higher insulin and glucose IAUC
at 30 and 60 minutes post-meal. Insulin IAUC predicted physical activity levels
over the entire observation period: higher insulin IAUC predicted lower MVPA in
the high sugar condition. When insulin was investigated as a potential mediator
between the effects of meal type on physical activity, no evidence of mediation
was found. Although overall insulin and glucose did not mediate the relationship
between meal type and physical activity over the four hours, mean insulin and
glucose values at specific time points may mediate changes in physical activity at
specific time points. These more detailed relationships remain to be explored.
115

Taken together, the findings from the three studies suggest some
overarching conclusions and directions for future research. The results from
Study 1 indicated that physical activity was influenced by individual demographic
factors. The anomalous finding that activity levels were similar in obese and
normal weight Mexican American youth is supported by previous research from
our group. Byrd-Williams et al. (2007) found a similar result in a sample of 169
youth (mean age 9.4 years, 50% female, 43% overweight) where overweight and
normal weight Hispanic females engaged in similar levels of physical activity [31].
The results from Study 1 support the need for further research in minority youth.
African American and Hispanic youth are progressively inactive as they age but
record less time in sedentary behavior than White youth. However, these minority
groups have higher prevalence of obesity, and diet did not confound the PA-
weight status relationship in this sample. There are genetic differences in
predisposition to obesity that may play a role in these observed race/ethnic
differences [204], and emerging studies suggest that there might be genetic
predisposition factors that influence habitual physical activity levels [202]. Also,
certain genes may mediate the response to exercise [26]. Taken together, these
findings suggest that physical activity is not merely a behavioral choice, but to a
significant extent a biologically based behavior.
Because certain individual factors such as race/ethnicity and genes are
not modifiable, the findings from Study 1 highlight the need for further research in
minority youth to find modifiable factors on which to intervene. The paradoxical
116

findings from Study 1 led us to investigate Hispanic youth further in Studies Two
and Three. Study 2 investigated the effect of one biomarker on physical activity,
and supported a biological basis for physical activity behavior in Hispanic per
pubertal females. The inclusion of a biomarker (leptin) in the cross-sectional
model increased the explained variance in activity behavior from 39% to 48%
after race, body composition, and age were included in the model. The
substantial explained variance in behavior with the addition of a single biomarker
lends support for the effect of biology on activity. Our results concur with a
previous longitudinal study in adults that found that baseline leptin levels
inversely predicted follow-up physical activity energy expenditure (p= .033) five
years later [66].
Study 3 added further support to biologically driven activity in overweight
Hispanic youth. The results were similar to Study 2 in that we found a biomarker
(insulin) predicted physical activity. Additionally, Study 3 included strong dietary
measures that allowed us to determine if specific nutrients commonly found in
diets of minority youth influenced activity and the biomarker-activity relationship.
There was support for the acute influence of high sugar meals on insulin and
glucose levels, and for insulin predicting physical activity. There was not a direct
relationship between diet and activity levels. We conclude that it is possible that
certain biomarkers may mediate the relationship between diet and physical
activity. The results from the three studies provide support for relationships
between personal and biological, but not dietary, factors and physical activity.
117

Implications
There have been several studies on race/ethnic differences in activity
levels in youth. Using self-report measures, previous findings have indicated that
White youth were the most active, which explained the lower prevalence of
obesity in this group. However, more recent studies using objective measures of
activity show that African American and Hispanic youth are more active than their
White counterparts, although they experience equally profound declines in
activity as they enter puberty. As discussed earlier, increasing physical activity
has been shown to be an effective means to both treat and prevent obesity. We
and others have shown ethnic differences in metabolic reactions to specific
stimuli including FSIVGTT [72] and exercise [81]. It is possible that the
dose/response ratio could differ according to race/ethnicity, and that
dose/response could be influenced by specific mechanisms and metabolic
markers, such as insulin or leptin. For example, Johnson et al. (2001) showed
that changes in leptin levels over 6 years were significantly positively related to
rate of change in fat mass in African American (p= .008) but not White (p= .490)
youth [97]. Study 2 found that leptin was inversely related to PA over one year in
a sample of African American and Hispanic females. Future studies with larger
sample sizes and diverse race/ethnic groups are needed to further understand if
there are race/ethnic differences in these relationships.
It is important to investigate other factors that are contributing to this
decline in order to compose a comprehensive picture of what is driving low
118

physical activity behavior in youth. The majority of interventions have focused on
modifying psychosocial and environmental influences of physical activity [188].
To date, most interventions have been unsuccessful in producing changes in
physical activity that have meaningful long-term health outcomes [195]. It was
hypothesized that these interventions may not work because they do not take
into account the biological aspects of physical activity [20]. An increasing body of
evidence supports the idea of a biological basis of activity levels in youth [57], yet
there is room for growth in this area. Rowland (1998) and subsequently
Eisenmann & Wickel (2009) have reviewed the evidence on the biological basis
of physical activity in children, concluding that more research on this topic is
needed to provide a complete picture of determinants of physical activity in
children during puberty [57, 173]. The results from Studies Two and Three add to
this literature by providing support for the idea that high leptin and insulin, which
are related to obesity, negatively influence physical activity behavior in minority
youth, suggesting a spiral of negative metabolic outcomes and low physical
activity.
Perhaps the most widely-supported variable associated with physical
activity is age. Several studies have reported age-related declines in physical
activity across several diverse populations of youth [7, 12, 30, 89, 94, 149, 177,
180, 206, 210, 216]. Age is commonly a covariate in analyses and all three
studies in this dissertation found that age was related to physical activity. It is
important to ask: to what extent is age a ‗proxy‘ for other factors? Age is likely
119

acting as a marker for aspects of physical maturation. There is also support for
declining physical activity levels as youth traverse puberty [7, 47, 169, 226].
There were differences in physical activity levels by Tanner stage in Study 2, but
this variable was not significant in Study 3. This may be accounted for by the fact
that the Study 3 was conducted in a lab where physical activity behavior may
have been constrained by the laboratory environment. This may also be because
youth in Study 3 were in the final stages of puberty, while youth in Study 2 were
entering puberty. The finding that physical activity levels decline during the
puberty in multiple species suggests that there is a biological mechanism
underlying this variable that explains this behavior [176]. The age at which the
initiation of puberty takes place differs between individuals and may account for
some of the age-related differences in activity. In a sample of 143 White females
ages 11 to 13 years, early maturing youths had lower levels of physical activity at
a two year follow-up than late maturing youths, irrespective of chronological age
[7]. This suggests the idea that the initiation of puberty somehow triggers a
decline in physical activity levels. Age and pubertal stage may not represent the
same factors, but their individual effects on physical activity levels are difficult to
disentangle because they are both related to maturational development. Age
may represent psychological and social development as adolescents begin to
gain autonomy, whereas pubertal stage may represent the internal biological
changes that take place during growth [53]. Both variables are related to
development, but they may interact differently with specific behaviors. Past
120

interventions have been tailored to age groups to some extent, however to date
no interventions have been tailored to biological development.
By studying three major hypothesized components of activity (namely
individual demographic, biological, and dietary factors), this dissertation
attempted to provide evidence supporting that all three factors should be taken
into account when designing interventions aimed at increasing physical activity
levels in youth. Taken together, findings from the three studies that comprise this
work provide the strongest evidence for individual demographic and biological
factors being related to physical activity levels.
Future Research
This dissertation aimed to fill a substantial gap in the literature by
exploring possible biological determinants of physical activity in minority youth.
Similar to findings from previous studies, insulin levels were inversely related to
lower levels of physical activity [3, 5, 23]. Furthermore, we found that high sugar
meals were related to higher insulin levels. Sustained high insulin levels, perhaps
exacerbated by high sugar diets typically consumed by youth today, may predict
sustained low levels of physical activity in youth. The single high sugar or high
fiber meal approach used in Study 3 may not be sufficient to elicit changes in
activity levels, and the limitations of an in-lab study to detect real-life physical
activities cannot be overlooked. Therefore, long-term influences of specific
nutrients on insulin and physical activity levels in this population should be
studied in order to further investigate meditational effects.
121

Several biological variables remain to be investigated as potential
mediators in the diet-physical activity relationship. For instance, leptin has been
associated with food intake and energy expenditure [5, 9, 43]. The influence of
diet on postprandial leptin concentrations is not yet entirely clear. Results from
animal studies indicate that diet-induced obesity causes the development of
central leptin resistance that leads to low levels of activity [61, 128]. Previous
studies have reported a rise in leptin levels following high carbohydrate and high
sugar meals [171, 189, 222]. Findings from Study 2 and other studies [66, 170]
indicate that leptin is likely negatively related to physical activity in children. Thus,
a high leptin response from a high sugar meal may lead to a decrease in physical
activity. These pathways can be examined in the future using data from
Transitions to examine the influences of self-reported dietary sugar intake on
leptin and objectively measured physical activity in the Hispanic female sample.
The studies in this dissertation focused primarily on minority youth. The
classification of individuals into race/ethnic categories for use in research bears
further discussion. ‗Race‘ is used to categorize people on the basis of shared
biological characteristics (e.g.: genes), while ‗ethnicity‘ is used to categorize
people on the basis of cultural characteristics (e.g.: language) [35]. The
constructs ‗race‘ and ‗ethnicity‘ are inexact variables in epidemiological research,
but are frequently employed to classify participants and explain disparities in
disease outcomes [70, 76, 227]. While disparities have been noted, there is little
understanding as to why they exist. Race/ethnicity are inextricably tied in science
122

to variables such as socioeconomic status (SES), acculturation, genetics, and
the built environment, which have been proposed to account for differences in
behavior and health outcomes. Yet, none of the proposed underlying variables
account for all the variation in health behaviors and outcomes across race/ethnic
groups.
Perhaps the best way to conceptualize race/ethnicity in analyses is to
include several related variables such as SES, race/ethnicity classification (i.e.:
White, Black, etc.), proportion of genetic variation (e.g.: genetic admixture), and
sociocultural variables in the same model. When study design and/or sample
size preclude using all of these variables in a model with other predictors of
health outcomes and behaviors, self-reported ethnicity may be the best measure
to include in analyses because reflects self-perception [179]. It is important to
remember that the ‗best‘ measure of race/ethnicity depends on the research
question and the field of study. Biomedical research favors the use of genetic
admixture to measure race/ethnicity while social science research favors
sociocultural and environmental measures. What can be concluded is that this
particular variable is difficult to directly measure, if a true measure exists. For
now, researchers should use ‗proxy‘ measures such as self-classification, SES,
skin color, and genetic admixture with careful consideration of what has been
used in their field and what will give the best estimate of race/ethnicity for their
specific research question.
123

Finally, psychosocial and environmental variables were not investigated in
the three studies presented here. Next steps would be to establish the unique
contributions of demographic, biological, genetic, psychosocial and
environmental components to physical activity and sedentary behavior. Data
collected from minority samples by the research group at USC could be used to
investigate these relationships in the future. Only through this type of
transdisciplinary research will effective interventions that address multiple
influences on behavior be developed.
Limitations
The three studies that comprise this dissertation have several limitations
that must be addressed. First, the cross-sectional nature of Study 1 precludes
the assessment of temporal relationships. However, the large sample size
allowed for the description of activity levels across race/ethnic groups that helped
inform Studies 2 and 3, which do allow for the investigation of longitudinal
relationships. Second, the use of accelerometers in all three studies may not
have captured certain aspects of activity. They do not record the type of physical
activity as do self-report measures, which prevents the exploration of the
frequency of specific behaviors. Furthermore, accelerometers do not capture all
types of physical activity [190]. However, accelerometers are considered to be an
excellent objective measurement of physical activity in youth because they
minimize self-report bias and eliminate human error in recalling previous physical
activity [48]. Third, BMI percentile category was used as a proxy measure of
124

adiposity in Study 1. Research supports that it corresponds well with percent
body fat in youth [152] and is a cost-efficient and feasible measure in a large
population-based study [52]. Fourth, with the exception of the range of
race/ethnic groups in Study 1, this dissertation focused predominantly on Los
Angeles Hispanic youth, an understudied population. The homogenous study
samples in Studies 2 and 3 may prevent the generalization of the findings to
other race/ethnic groups or geographical regions. Fifth, the selected hormones
studied in this dissertation comprise a small representation of potential
biomarkers that may influence physical activity levels. They were chosen based
on previous research, however future research should examine other potential
agents of change. Despite the limitations of this work, the findings inform the
understanding of the individual, biological, and dietary factors that are related to
the decline in physical activity in youth.
Contribution to the Literature
This dissertation makes an important contribution to the youth physical
activity literature by: 1) providing the first description of physical activity levels by
race/ethnic, age, weight status, and gender groups in the US; 2) demonstrating
that cross-sectional relationships between leptin and physical activity remain
stable over one year; and 3) demonstrating that high sugar meals influence
insulin and glucose but not physical activity in overweight youth in a controlled
setting. The findings from this dissertation leave us with several directions for
future research, and prompt one very difficult question: How can biological
125

determinants of physical activity across puberty be best taken into account to
successfully intervene on physical activity and sedentary behaviors in minority
youth?
126

BIBLIOGRAPHY

1. Ahmed, M.L., K.K. Ong, and D.B. Dunger, Childhood obesity and the timing of
puberty. Trends Endocrinol Metab, 2009. 20(5): p. 237-42.

2. Ahmed, M.L., et al., Longitudinal Study of Leptin Concentrations during Puberty: Sex
Differences and Relationship to Changes in Body Composition. J Clin Endocrinol
Metab, 1999. 84(3): p. 899-905.

3. Alhassan, S. and T.N. Robinson, Objectively measured physical activity and
cardiovascular disease risk factors in African American girls. Ethn Dis, 2008.
18(4): p. 421-6.

4. Anderson, S., C. Economos, and A. Must, Active play and screen time in US children
aged 4 to 11 years in relation to sociodemographic and weight status
characteristics: a nationally representative cross-sectional analysis. BMC Public
Health, 2008. 8(1): p. 366.

5. Anubhuti and S. Arora, Leptin and its metabolic interactions: an update. Diabetes
Obes Metab, 2008. 10(11): p. 973-93.

6. Arcan, C., et al., Sociodemographic Differences in Selected Eating Practices among
Alternative High School Students. Journal of the American Dietetic Association,
2009. 109(5): p. 823-829.

7. Baker, B.L., et al., Advanced pubertal status at age 11 and lower physical activity in
adolescent girls. J Pediatr, 2007. 151(5): p. 488-93.

8. Ball, S.D., et al., Prolongation of satiety after low versus moderately high glycemic
index meals in obese adolescents. Pediatrics, 2003. 111(3): p. 488-94.

9. Banks, W.A., The many lives of leptin. Peptides, 2004. 25(3): p. 331-8.

10. Barkoukis, H., et al., A high glycemic meal suppresses the postprandial leptin
response in normal healthy adults. Ann Nutr Metab, 2007. 51(6): p. 512-8.

11. Baron, R. and D. Kenny, The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations.
Journal of personality and social psychology, 1986. 51(6): p. 1173-1182.

12. Belcher, B.R., et al., Physical Activity in US Youth: Impact of Race/Ethnicity, Age,
Gender, & Weight Status. Med Sci Sports Exerc, 2010. In press.

13. Bergman, R.N., et al., Quantitative estimation of insulin sensitivity. Am J Physiol,
1979. 236(6): p. E667-77.

127

14. Berkey, C.S. and G.A. Colditz, Adiposity in Adolescents: Change in Actual BMI
Works Better Than Change in BMI z Score for Longitudinal Studies. Annals of
Epidemiology, 2007. 17(1): p. 44-50.

15. Berkey, C.S., et al., Activity, dietary intake, and weight changes in a longitudinal
study of preadolescent and adolescent boys and girls. Pediatrics, 2000. 105(4):
p. E56.

16. Biddle, S.J., et al., Physical activity and sedentary behaviours in youth: issues and
controversies. The Journal of the Royal Society for the Promotion of Health,
2004. 124(1): p. 29-33.

17. Boney, C.M., et al., Metabolic Syndrome in Childhood: Association With Birth
Weight, Maternal Obesity, and Gestational Diabetes Mellitus. Pediatrics, 2005.
115(3): p. e290-296.

18. Booth, F.W., et al., Waging war on modern chronic diseases: primary prevention
through exercise biology. J Appl Physiol, 2000. 88(2): p. 774-87.

19. Borer, K.T., Nonhomeostatic control of human appetite and physical activity in
regulation of energy balance. Exerc Sport Sci Rev. 38(3): p. 114-21.

20. Bouchard, C. and T. Rankinen, Are people physically inactive because of their
genes? Pres. Council Phys. Fit Sports Research Digest, 2006. 7: p. 1-8.

21. Bradlee, M.L., et al., Food group intake and central obesity among children and
adolescents in the Third National Health and Nutrition Examination Survey
(NHANES III). Public Health Nutr, 2010. 13(6): p. 797-805.

22. Brage, S., et al., Objectively measured physical activity correlates with indices of
insulin resistance in Danish children. Int J Obes Relat Metab Disord, 2004.
28(11): p. 1503-1508.

23. Brage, S., et al., Objectively measured physical activity correlates with indices of
insulin resistance in Danish children. The European Youth Heart Study (EYHS).
Int J Obes Relat Metab Disord, 2004. 28(11): p. 1503-8.

24. Brage, S., et al., Reexamination of validity and reliability of the CSA monitor in
walking and running. Med Sci Sports Exerc, 2003. 35(8): p. 1447-54.

25. Braun, B., Why doesn't weight gain blunt appetite and increase movement?
Nonhomeostatic responses to energy surplus in humans. Exerc Sport Sci Rev.
38(3): p. 103-4.

26. Bray, M.S., Genomics, genes, and environmental interaction: the role of exercise. J
Appl Physiol, 2000. 88(2): p. 788-92.

128

27. Brodersen, N.H., et al., Trends in physical activity and sedentary behaviour in
adolescence: ethnic and socioeconomic differences * COMMENTARY. Br J
Sports Med, 2007. 41(3): p. 140-144.

28. Bureau, U.S.C., Current Population Reports: Population Projections of the United
States by Age, Sex, Race, and Hispanic Origin: 1995 to 2050, P. Division, Editor.

29. Butte, N.F., E. Christiansen, and T.I. Sorensen, Energy imbalance underlying the
development of childhood obesity. Obesity (Silver Spring), 2007. 15(12): p. 3056-
66.

30. Butte, N.F., et al., Physical activity in nonoverweight and overweight Hispanic
children and adolescents. Med Sci Sports Exerc, 2007. 39(8): p. 1257-66.

31. Byrd-Williams, C., et al., Influence of gender, BMI and Hispanic ethnicity on physical
activity in children. Int J Pediatr Obes, 2007. 2(3): p. 159-66.

32. Byrd-Williams, C.E., et al., Increased physical activity and reduced adiposity in
overweight Hispanic adolescents. Med Sci Sports Exerc, 2010. 42(3): p. 478-84.

33. Calle, E.E. and R. Kaaks, Overweight, obesity and cancer: epidemiological evidence
and proposed mechanisms. Nat Rev Cancer, 2004. 4(8): p. 579-91.

34. Cambuli, V.M., et al., Assessment of Adiponectin and Leptin as Biomarkers of
Positive Metabolic Outcomes after Lifestyle Intervention in Overweight and
Obese Children. J Clin Endocrinol Metab, 2008. 93(8): p. 3051-3057.

35. Caprio, S., et al., Influence of Race, Ethnicity, and Culture on Childhood Obesity:
Implications for Prevention and Treatment. Diabetes Care, 2008. 31(11): p. 2211-
2221.

36. Chaloupka, F.J. and L.M. Powell, Price, availability, and youth obesity: evidence
from Bridging the Gap. Prev Chronic Dis, 2009. 6(3): p. A93.

37. Chinapaw, M.J., et al., Physical activity questionnaires for youth: a systematic
review of measurement properties. Sports Med. 40(7): p. 539-63.

38. Coppari, R., et al., The hypothalamic arcuate nucleus: a key site for mediating
leptin's effects on glucose homeostasis and locomotor activity. Cell metabolism,
2005. 1(1): p. 63-72.

39. Corder, K., et al., Is it possible to assess free-living physical activity and energy
expenditure in young people by self-report? The American Journal of Clinical
Nutrition, 2009. 89(3): p. 862-870.

40. Crow, S., et al., Psychosocial and behavioral correlates of dieting among overweight
and non-overweight adolescents. Journal of Adolescent Health, 2006. 38(5): p.
569-574.
129

41. D'Adamo, E., et al., Ethnic differences in lipoprotein subclasses in obese
adolescents: importance of liver and intraabdominal fat accretion. Am J Clin Nutr.
92(3): p. 500-508.

42. Daniels, S.R., P.R. Khoury, and J.A. Morrison, The Utility of Body Mass Index as a
Measure of Body Fatness in Children and Adolescents: Differences by Race and
Gender. Pediatrics, 1997. 99(6): p. 804-807.

43. Davis, J.F., D.L. Choi, and S.C. Benoit, Insulin, leptin and reward. Trends
Endocrinol Metab, 2010. 21(2): p. 68-74.

44. Davis, J.N., et al., Inverse relation between dietary fiber intake and visceral adiposity
in overweight Latino youth. The American Journal of Clinical Nutrition, 2009.
90(5): p. 1160-1166.

45. Davis, J.N., et al., Aerobic and strength training reduces adiposity in overweight
Latina adolescents. Med Sci Sports Exerc, 2009. 41(7): p. 1494-503.

46. Davis, J.N., et al., The relation of sugar intake to β cell function in overweight Latino
children. The American Journal of Clinical Nutrition, 2005. 82(5): p. 1004-1010.

47. Davison, K.K., et al., Why are early maturing girls less active? Links between
pubertal development, psychological well-being, and physical activity among girls
at ages 11 and 13. Soc Sci Med, 2007. 64(12): p. 2391-404.

48. de Leeuw, E., N. Borgers, and A. Smits, Pretesting Questionnaires for Children and
Adolescents, in Methods for Testing and Evaluating Survey Questionnaires, S.
Presser, et al., Editors. 2004, Wiley-Interscience. p. 409-429.

49. Dencker, M. and L.B. Andersen, Health-related aspects of objectively measured
daily physical activity in children. Clin Physiol Funct Imaging, 2008. 28(3): p. 133-
44.

50. Dencker, M., et al., BMI and objectively measured body fat and body fat distribution
in prepubertal children. Clinical Physiology and Functional Imaging, 2007. 27(1):
p. 12-16.

51. Deurenberg, P., et al., The impact of body build on the relationship between body
mass index and percent body fat. International Journal of Obesity, 1999. 23(5): p.
537-542.

52. Dietz, W.H. and T.N. Robinson, Use of the body mass index (BMI) as a measure of
overweight in children and adolescents. Journal of Pediatrics, 1998. 132(2): p.
191.

53. Dorn, L.D. and F.M. Biro, Puberty and Its Measurement: A Decade in Review.
Journal of Research on Adolescence, 2011. 21(1): p. 180-195.
130

54. Dowda, M., et al., Physical activities and sedentary pursuits in African American and
Caucasian girls. Res Q Exerc Sport, 2004. 75(4): p. 352-60.

55. Dowda, M., et al., Associations among Physical Activity, Health Indicators, and
Employment in 12th Grade Girls. Journal of Women's Health, 2007. 16(9): p.
1331-1339.

56. Dugas, L.R., et al., Very low levels of energy expenditure among pre-adolescent
Mexican-American girls. Int J Pediatr Obes, 2008. 3(2): p. 123-6.

57. Eisenmann, J.C. and E.E. Wickel, The biological basis of physical activity in
children: revisited. Pediatr Exerc Sci, 2009. 21(3): p. 257-72.

58. Ekelund, U., et al., Independent associations of physical activity and
cardiorespiratory fitness with metabolic risk factors in children: the European
youth heart study. Diabetologia, 2007. 50(9): p. 1832-40.

59. Ekelund, U., et al., Does physical activity equally predict gain in fat mass among
obese and nonobese young adults? Int J Obes, 2006. 31(1): p. 65-71.

60. Ekelund, U., et al., Does physical activity equally predict gain in fat mass among
obese and nonobese young adults? Int J Obes (Lond), 2007. 31(1): p. 65-71.

61. Enriori, P.J., et al., Leptin resistance and obesity. Obesity, 2006. 14: p. 254S-258S.

62. Field, A.E., et al., Which metric of relative weight best captures body fatness in
children? Obes Res, 2003. 11(11): p. 1345-52.

63. Finkelstein, E.A., C.J. Ruhm, and K.M. Kosa, Economic causes and consequences
of obesity. Annu Rev Public Health, 2005. 26: p. 239-57.

64. Flegal, K.M., et al., High adiposity and high body mass index-for-age in US children
and adolescents overall and by race-ethnic group. Am J Clin Nutr, 2010. 91(4): p.
1020-6.

65. Flegal, K.M., et al., Comparisons of percentage body fat, body mass index, waist
circumference, and waist-stature ratio in adults. Am J Clin Nutr, 2009. 89(2): p.
500-508.

66. Franks, P.W., et al., Physical activity energy expenditure may mediate the
relationship between plasma leptin levels and worsening insulin resistance
independently of adiposity. J Appl Physiol, 2007. 102(5): p. 1921-6.

67. Freedman, D.S., et al., Racial/ethnic Differences in Body Fatness Among Children
and Adolescents. Obesity, 2008. 16(5): p. 1105-1111.

131

68. Freedson, P.S., E. Melanson, and J. Sirard, Calibration of the Computer Science
and Applications, Inc. accelerometer. Med Sci Sports Exerc, 1998. 30(5): p. 777-
81.

69. Fulton, J.E., et al., Physical activity, energy intake, sedentary behavior, and
adiposity in youth. Am J Prev Med, 2009. 37(1 Suppl): p. S40-9.

70. Goodman, A.H., Why genes don't count (for racial differences in health). Am J
Public Health, 2000. 90(11): p. 1699-702.

71. Goodpaster, B.H., Measuring body fat distribution and content in humans. Curr Opin
Clin Nutr Metab Care, 2002. 5(5): p. 481-7.

72. Goran, M.I., et al., Insulin resistance and associated compensatory responses in
african-american and Hispanic children. Diabetes Care, 2002. 25(12): p. 2184-
90.

73. Goran, M.I. and B.A. Gower, Longitudinal study on pubertal insulin resistance.
Diabetes, 2001. 50(11): p. 2444-50.

74. Gordon-Larsen, P., L.S. Adair, and B.M. Popkin, The relationship of ethnicity,
socioeconomic factors, and overweight in US adolescents. Obes Res, 2003.
11(1): p. 121-9.

75. Goulding, A., et al., DEXA supports the use of BMI as a measure of fatness in
young girls. Int J Obes Relat Metab Disord, 1996. 20(11): p. 1014-21.

76. Gravlee, C.C., How race becomes biology: embodiment of social inequality. Am J
Phys Anthropol, 2009. 139(1): p. 47-57.

77. Grieser, M., et al., Physical activity attitudes, preferences, and practices in African
American, Hispanic, and Caucasian girls. Health Educ Behav, 2006. 33(1): p. 40-
51.

78. Guerrero, R., et al., Ethnic differences in hepatic steatosis: An insulin resistance
paradox? 2009, Wiley Subscription Services, Inc., A Wiley Company. p. 791-801.

79. Gutin, B., Child obesity can be reduced with vigorous activity rather than restriction
of energy intake. Obesity (Silver Spring), 2008. 16(10): p. 2193-6.

80. Harrell, J.S., et al., Energy costs of physical activities in children and adolescents.
Med Sci Sports Exerc, 2005. 37(2): p. 329-36.

81. Hasson, R.E., et al., Effects of a Single Exercise Bout on Insulin Sensitivity in Black
and White Individuals. J Clin Endocrinol Metab, 2010.

132

82. Healy, G.N., et al., Objectively measured light-intensity physical activity is
independently associated with 2-h plasma glucose. Diabetes Care, 2007. 30(6):
p. 1384-9.

83. Heptulla, R., et al., Temporal Patterns of Circulating Leptin Levels in Lean and
Obese Adolescents: Relationships to Insulin, Growth Hormone, and Free Fatty
Acids Rhythmicity. J Clin Endocrinol Metab, 2001. 86(1): p. 90-96.

84. Hlaing, W.M., et al., Body mass index growth in a sample of U.S. children: repeated
measures data analysis of the Minneapolis Children's Blood Pressure Study. Am
J Hum Biol, 2001. 13(6): p. 821-31.

85. Holm, J.C., et al., Time course and determinants of leptin decline during weight loss
in obese boys and girls. Int J Pediatr Obes, 2007. 2(1): p. 2-10.

86. Horlick, M.B., et al., Effect of puberty on the relationship between circulating leptin
and body composition. J Clin Endocrinol Metab, 2000. 85(7): p. 2509-18.

87. Hosking, J., et al., Changes in resting energy expenditure and their relationship to
insulin resistance and weight gain: a longitudinal study in pre-pubertal children
(EarlyBird 17). Clin Nutr. 29(4): p. 448-52.

88. Hu, F.B. and V.S. Malik, Sugar-sweetened beverages and risk of obesity and type 2
diabetes: epidemiologic evidence. Physiol Behav, 2010. 100(1): p. 47-54.

89. Ingram, D.K., Age-related decline in physical activity: generalization to nonhumans.
Medicine & Science in Sports & Exercise., 2000. 32(9): p. 1623-9.

90. Ischander, M., et al., Physical activity, growth, and inflammatory mediators in BMI-
matched female adolescents. Med Sci Sports Exerc, 2007. 39(7): p. 1131-8.

91. Jago, R., et al., Relationship between Physical Activity and Diet among African-
American Girls. Obesity, 2004. 12(S9): p. 55S-63S.

92. Jago, R., et al., Prevalence of abnormal lipid and blood pressure values among an
ethnically diverse population of eighth-grade adolescents and screening
implications. Pediatrics, 2006. 117(6): p. 2065.

93. Jago, R., et al., Obesogenic diet and physical activity: independent or associated
behaviours in adolescents? Public Health Nutr, 2010. 13(5): p. 673-81.

94. Jago, R., et al., Six-year change in youth physical activity and effect on fasting
insulin and HOMA-IR. Am J Prev Med, 2008. 35(6): p. 554-60.

95. Janz, K.F., T.L. Burns, and S.M. Levy, Tracking of Activity and Sedentary Behaviors
in Childhood: The Iowa Bone Development Study. American Journal of
Preventive Medicine, 2005. 29(3): p. 171-178.

133

96. Jéquier, E., Leptin signaling, adiposity, and energy balance. Annals of the New York
Academy of Sciences, 2002. 967(1): p. 379-388.

97. Johnson, M.S., et al., Influence of Leptin on Changes in Body Fat during Growth in
African American and White Children. Obesity, 2001. 9(10): p. 593-598.

98. Jørgensen, T., L. Andersen, and K. Froberg, Position statement: Testing physical
condition in a population–how good are the methods? European Journal of Sport
Science, 2009. 9(5): p. 257-267.

99. Juerimaee, J., A.P. Hills, and T. Juerimaee, Cytokines, Growth Mediators and
Physical Activity in Children During Puberty. 2010: Karger Publishers.

100. Kieffer, T.J. and J.F. Habener, The adipoinsular axis: effects of leptin on pancreatic
beta-cells. Am J Physiol Endocrinol Metab, 2000. 278(1): p. E1-E14.

101. Ku, C.Y., et al., Racial differences in insulin secretion and sensitivity in prepubertal
children: role of physical fitness and physical activity. Obes Res, 2000. 8(7): p.
506-15.

102. Kuczmarski, R.J., et al., 2000 CDC Growth Charts for the United States: methods
and development. Vital Health Stat 11, 2002(246): p. 1-190.

103. Kumanyika, S., Ethnic minorities and weight control research priorities: where are
we now and where do we need to be? Prev Med, 2008. 47(6): p. 583-6.

104. Kumanyika, S.K., Environmental influences on childhood obesity: Ethnic and
cultural influences in context. Physiology & Behavior, 2008. 94(1): p. 61-70.

105. Lahlou, N., et al., Circulating leptin in normal children and during the dynamic
phase of juvenile obesity: relation to body fatness, energy metabolism, caloric
intake, and sexual dimorphism. Diabetes, 1997. 46(6): p. 989-993.

106. Lakka, T.A. and D.E. Laaksonen, Physical activity in prevention and treatment of
the metabolic syndrome. Appl Physiol Nutr Metab, 2007. 32(1): p. 76-88.

107. Leenders, N.Y., et al., Evaluation of methods to assess physical activity in free-
living conditions. Med Sci Sports Exerc, 2001. 33(7): p. 1233-40.

108. Liska, D., et al., Interethnic differences in muscle, liver and abdominal fat
partitioning in obese adolescents. PLoS One, 2007. 2(6): p. e569.

109. Lohman, T.G., et al., Relationships among fitness, body composition, and physical
activity. Med Sci Sports Exerc, 2008. 40(6): p. 1163-70.

110. Lohman, T.G., et al., Associations of body size and composition with physical
activity in adolescent girls. Med Sci Sports Exerc, 2006. 38(6): p. 1175-81.

134

111. Lovasi, G.S., et al., Built Environments and Obesity in Disadvantaged Populations.
Epidemiologic Reviews, 2009.

112. Ludwig, D.S., The Glycemic Index: Physiological Mechanisms Relating to Obesity,
Diabetes, and Cardiovascular Disease. Jama, 2002. 287(18): p. 2414-2423.

113. Ludwig, D.S., et al., High glycemic index foods, overeating, and obesity. Pediatrics,
1999. 103(3): p. E26.

114. Maffeis, C. and L. Tato, Long-term effects of childhood obesity on morbidity and
mortality. Horm Res, 2001. 55 Suppl 1: p. 42-5.

115. Maia, J.A.R., M. Thomis, and G. Beunen, Genetic factors in physical activity
levels:: A twin study. American Journal of Preventive Medicine, 2002. 23(2): p.
87-91.

116. Manios, Y., A. Kafatos, and C. Codrington, Gender differences in physical activity
and physical fitness in young children in Crete. J Sports Med Phys Fitness, 1999.
39(1): p. 24-30.

117. Margetic, S., et al., Leptin: a review of its peripheral actions and interactions. Int J
Obes Relat Metab Disord, 2002. 26(11): p. 1407-33.

118. Mark, A.E. and I. Janssen, Influence of bouts of physical activity on overweight in
youth. Am J Prev Med, 2009. 36(5): p. 416-21.

119. Matthews, C.E., et al., Amount of Time Spent in Sedentary Behaviors in the United
States, 2003-2004. Am. J. Epidemiol., 2008. 167(7): p. 875-881.

120. McClain, A.D., et al., Psychosocial correlates of eating behavior in children and
adolescents: a review. International Journal of Behavioral Nutrition and Physical
Activity, 2009. 6(1): p. 54.

121. McClain, A.D., et al., Physical Activity, but Not Sedentary Behavior or Energy
Intake, is Associated with Higher Fat Mass in Latina and African American Girls.
International Journal of Pediatric Obesity, Submitted.

122. McCrimmon, R.J., B.M. Frier, and I.J. Deary, Appraisal of mood and personality
during hypoglycemia in human subjects. Physiology and Behavior, 1999. 67(1):
p. 27-33.

123. McMurray, R.G., et al., Influence of physical activity on change in weight status as
children become adolescents. Int J Pediatr Obes, 2008. 3(2): p. 69-77.

124. Mei, Z., et al., Validity of body mass index compared with other body-composition
screening indexes for the assessment of body fatness in children and
adolescents. Am J Clin Nutr, 2002. 75(6): p. 978-985.

135

125. Metallinos-Katsaras, E.S., et al., The association between an objective measure of
physical activity and weight status in preschoolers. Obesity (Silver Spring), 2007.
15(3): p. 686-94.

126. Metcalf, B.S., et al., Objectively measured physical activity and its association with
adiponectin and other novel metabolic markers: a longitudinal study in children
(EarlyBird 38). Diabetes Care, 2009. 32(3): p. 468-73.

127. Moreno, L.A., et al., Trends of dietary habits in adolescents. Crit Rev Food Sci
Nutr. 50(2): p. 106-12.

128. Munzberg, H., J.S. Flier, and C. Bjorbaek, Region-specific leptin resistance within
the hypothalamus of diet-induced obese mice. Endocrinology, 2004. 145(11): p.
4880.

129. Munzberg, H. and M.G. Myers, Molecular and anatomical determinants of central
leptin resistance. Nat Neurosci, 2005. 8(5): p. 566-570.

130. Musselman, D.L., et al., Relationship of depression to diabetes types 1 and 2:
epidemiology, biology, and treatment. Biological Psychiatry., 2003. 54(3): p. 317-
29.

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

132. Nader, P.R., et al., Moderate-to-vigorous physical activity from ages 9 to 15 years.
JAMA, 2008. 300(3): p. 295-305.

133. Ogden, C.L., et al., Prevalence of high body mass index in US children and
adolescents, 2007-2008. JAMA, 2010. 303(3): p. 242-9.

134. Ogden, C.L., et al., Prevalence of overweight and obesity in the United States,
1999-2004. JAMA, 2006. 295(13): p. 1549-55.

135. Ogden, C.L., M.D. Carroll, and K.M. Flegal, High Body Mass Index for Age Among
US Children and Adolescents, 2003-2006. JAMA, 2008. 299(20): p. 2401.

136. Ogden, C.L. and K.M. Flegal, Changes in terminology for childhood overweight
and obesity. Natl Health Stat Report, 2010(25): p. 1-5.

137. Ogden, C.L., et al., The epidemiology of obesity. Gastroenterology, 2007. 132(6):
p. 2087-2102.

138. Otten, J.J., J.P. Hellwig, and L.D. Meyers, DRI, dietary reference intakes: the
essential guide to nutrient requirements. 2006: Natl Academy Pr.

136

139. Owen, C.G., et al., Ethnic and gender differences in physical activity levels among
9-10-year-old children of white European, South Asian and African-Caribbean
origin: the Child Heart Health Study in England (CHASE Study). Int J Epidemiol,
2009. 38(4): p. 1082-93.

140. Owen, N., et al., Environmental determinants of physical activity and sedentary
behavior. Exerc Sport Sci Rev, 2000. 28(4): p. 153-8.

141. Owens, D.S. and D. Benton, The impact of raising blood glucose on reaction times.
Neuropsychobiology., 1994. 30(2-3): p. 106-13.

142. Owens, D.S., P.Y. Parker, and D. Benton, Blood glucose and subjective energy
following cognitive demand. Physiology & Behavior., 1997. 62(3): p. 471-8.

143. Pacini, G. and R. Bergman, MINMOD: a computer program to calculate insulin
sensitivity and pancreatic responsivity from the frequently sampled intravenous
glucose tolerance test. Computer Methods and Programs in Biomedicine, 1986.
23(2): p. 113-122.

144. Pasman, W.J., et al., Effect of two breakfasts, different in carbohydrate
composition, on hunger and satiety and mood in healthy men. Int J Obes Relat
Metab Disord, 2003. 27(6): p. 663-668.

145. Pate, R.R., et al., Validation and calibration of an accelerometer in preschool
children. Obesity (Silver Spring), 2006. 14(11): p. 2000-6.

146. Pate, R.R., B.J. Long, and G. Heath, Descriptive epidemiology of physical activity
in adolescents. Pediatric exercise science, 1994. 6(4): p. 434-47.

147. Pate, R.R., et al., Validation of a 3-Day Physical Activity Recall Instrument in
Female Youth. PEDIATRIC EXERCISE SCIENCE, 2003. 15(3): p. 257-265.

148. Pate, R.R., et al., Objectively measured physical activity in sixth-grade girls. Arch
Pediatr Adolesc Med, 2006. 160(12): p. 1262-8.

149. Pate, R.R., et al., Age-related change in physical activity in adolescent girls. J
Adolesc Health, 2009. 44(3): p. 275-82.

150. Patrick, K., et al., Diet, physical activity, and sedentary behaviors as risk factors for
overweight in adolescence. Arch Pediatr Adolesc Med, 2004. 158(4): p. 385-90.

151. PÈRUSSE, L., et al., GENETIC AND ENVIRONMENTAL INFLUENCES ON
LEVEL OF HABITUAL PHYSICAL ACTIVITY AND EXERCISE PARTICIPATION.
American Journal of Epidemiology, 1989. 129(5): p. 1012-1022.

152. Pietrobelli, A., et al., Body mass index as a measure of adiposity among children
and adolescents: a validation study. Journal of Pediatrics, 1998. 132(2): p. 204.

137

153. Plachta-Danielzik, S., et al., Energy gain and energy gap in normal-weight children:
longitudinal data of the KOPS. Obesity (Silver Spring), 2008. 16(4): p. 777-83.

154. Plasqui, G. and K.R. Westerterp, Physical activity assessment with
accelerometers: an evaluation against doubly labeled water. Obesity (Silver
Spring), 2007. 15(10): p. 2371-9.

155. Powell, L.M., F.J. Chaloupka, and Y. Bao, The availability of fast-food and full-
service restaurants in the United States: associations with neighborhood
characteristics. Am J Prev Med, 2007. 33(4 Suppl): p. S240-5.

156. Powell, L.M., et al., Availability of physical activity-related facilities and
neighborhood demographic and socioeconomic characteristics: a national study.
Am J Public Health, 2006. 96(9): p. 1676-80.

157. Prevention, C.f.D.C.a. NHANES Analytic Guidelines. 2004 [cited 2008 June 1];
Available from:
http://www.cdc.gov/nchs/data/nhanes/nhanes_general_guidelines_june_04.pdf.

158. Prevention, C.f.D.C.a. NHANES Analytic and Reporting Guidelines. 2005 [cited
2008 June 1]; Available from:
http://www.cdc.gov/nchs/data/nhanes/nhanes_03_04/nhanes_analytic_guideline
s_dec_2005.pdf.

159. Prevention., U.S.C.f.D.C.a. Preventing Obesity and Chronic Diseases through
Good Nutrition and Physical Activity.; Available from:
http://www.cdc.gov/nccdphp/publications/factsheets/prevention/pdf/obesity.pdf.

160. Puyau, M.R., et al., Validation and calibration of physical activity monitors in
children. Obes Res, 2002. 10(3): p. 150-7.

161. Puyau, M.R., et al., Prediction of activity energy expenditure using accelerometers
in children. Med Sci Sports Exerc, 2004. 36(9): p. 1625-31.

162. Qin, L., et al., Does physical activity modify the risk of obesity for type 2 diabetes: a
review of epidemiological data. Eur J Epidemiol, 2010. 25(1): p. 5-12.

163. Raman, A., et al., Insulin resistance is improved in overweight African American
boys but not in girls following a one-year multidisciplinary community intervention
program. J Pediatr Endocrinol Metab, 2010. 23(1-2): p. 109-20.

164. Ranjit, N., et al., Dietary and Activity Correlates of Sugar-Sweetened Beverage
Consumption Among Adolescents. Pediatrics, 2010. 126(4): p. e754-761.

165. Reichert, F.F., et al., Physical activity as a predictor of adolescent body fatness: a
systematic review. Sports Med, 2009. 39(4): p. 279-94.

138

166. Reilly, J.J., et al., Objective measurement of physical activity and sedentary
behaviour: review with new data. Arch Dis Child, 2008. 93(7): p. 614-9.

167. Roa, J., et al., Metabolic control of puberty onset: new players, new mechanisms.
Mol Cell Endocrinol. 324(1-2): p. 87-94.

168. Robinson, T., et al., Dance and reducing television viewing to prevent weight gain
in African-American girls: the Stanford GEMS pilot study. Ethnicity and Disease,
2003. 13(1; SUPP/1): p. 1-65.

169. Roemmich, J.N., et al., Pubertal alterations in growth and body composition. V.
Energy expenditure, adiposity, and fat distribution. Am J Physiol Endocrinol
Metab, 2000. 279(6): p. E1426-36.

170. Romon, M., et al., Relationships between physical activity and plasma leptin levels
in healthy children: the Fleurbaix-Laventie Ville Sante II Study. Int J Obes Relat
Metab Disord, 2004. 28(10): p. 1227-32.

171. Romon, M., et al., Postprandial Leptin Response to Carbohydrate and Fat Meals in
Obese Women. J Am Coll Nutr, 2003. 22(3): p. 247-251.

172. Romon, M., et al., Leptin response to carbohydrate or fat meal and association
with subsequent satiety and energy intake. Am J Physiol Endocrinol Metab,
1999. 277(5): p. E855-861.

173. Rowland, T.W., The biological basis of physical activity. Med Sci Sports Exerc,
1998. 30(3): p. 392-9.

174. Rubin, D., et al., The association between insulin resistance and cytokines in
adolescents: the role of weight status and exercise. Metabolism: clinical and
experimental, 2008. 57(5): p. 683.

175. Salbe, A.D., M. Nicolson, and E. Ravussin, Total energy expenditure and the level
of physical activity correlate with plasma leptin concentrations in five-year-old
children. Journal of Clinical Investigation, 1997. 99(4): p. 592-5.

176. Sallis, J.F., Age-related decline in physical activity: a synthesis of human and
animal studies. Medicine & Science in Sports & Exercise., 2000. 32(9): p. 1598-
600.

177. Sallis, J.F., J.J. Prochaska, and W.C. Taylor, A review of correlates of physical
activity of children and adolescents. Med Sci Sports Exerc, 2000. 32(5): p. 963-
75.

178. Schmitz, K.H., et al., Association of physical activity with insulin sensitivity in
children. International journal of obesity, 2002. 26(10): p. 1310-1316.

139

179. Senior, P.A. and R. Bhopal, Ethnicity as a variable in epidemiological research.
BMJ, 1994. 309(6950): p. 327-30.

180. Sherar, L.B., et al., Age and gender differences in youth physical activity: does
physical maturity matter? Med Sci Sports Exerc, 2007. 39(5): p. 830-5.

181. Shisslak, C.M., et al., Eating and weight control behaviors among middle school
girls in relationship to body weight and ethnicity. J Adolesc Health, 2006. 38(5):
p. 631-3.

182. Shrewsbury, V. and J. Wardle, Socioeconomic status and adiposity in childhood: a
systematic review of cross-sectional studies 1990-2005. Obesity (Silver Spring),
2008. 16(2): p. 275-84.

183. Sinha, M.K. and J.F. Caro, Clinical aspects of leptin. Vitam Horm, 1998. 54: p. 1-
30.

184. Sirard, J.R., et al., Race differences in activity, fitness, and BMI in female eighth
graders categorized by sports participation status. Pediatr Exerc Sci, 2008. 20(2):
p. 198-210.

185. Slootmaker, S., et al., Disagreement in physical activity assessed by
accelerometer and self-report in subgroups of age, gender, education and weight
status. International Journal of Behavioral Nutrition and Physical Activity, 2009.
6(1): p. 17.

186. Snitker, S., et al., Association of physical activity and body composition with insulin
sensitivity in a community sample of adolescents. Arch Pediatr Adolesc Med,
2007. 161(7): p. 677-83.

187. Sobel, M.E., Direct and indirect effects in linear structural equation models.
Sociological Methods & Research., 1987. 15: p. 151-176.

188. Spruijt-Metz, D., Etiology, Treatment, and Prevention of Obesity in Childhood and
Adolescence: A Decade in Review. Journal of Research on Adolescence, 2011.
21(1): p. 129-152.

189. Spruijt-Metz, D., et al., A high-sugar/low-fiber meal compared with a low-
sugar/high-fiber meal leads to higher leptin and physical activity levels in
overweight Latina females. J Am Diet Assoc, 2009. 109(6): p. 1058-63.

190. Spruijt-Metz, D., et al., Measures of physical activity and exercise, in Handbook Of
Assessment Methods For Eating Behaviors And Weight Related Problems:
Measures, Theory And Research. D. Allison and M. Baskin, Editors. 2010, Sage
Publications Inc.: Thousand Oaks.

191. Spruijt-Metz, D., et al., CRP is related to higher leptin levels in minority peri-
pubertal females regardless of adiposity levels. Accepted.
140

192. Stallmann-Jorgensen, I.S., et al., General and visceral adiposity in black and white
adolescents and their relation with reported physical activity and diet. Int J Obes
(Lond), 2007. 31(4): p. 622-9.

193. Steffen, L.M., et al., Whole Grain Intake Is Associated with Lower Body Mass and
Greater Insulin Sensitivity among Adolescents. American Journal of
Epidemiology, 2003. 158(3): p. 243-250.

194. Stevens, J., et al., Physical activity as a predictor of body composition in American
Indian children. Obes Res, 2004. 12(12): p. 1974-80.

195. Summerbell, C., et al., Interventions for preventing obesity in children. status and
date: Edited (no change to conclusions), published in, 2005. 1.

196. Sun, S.S., et al., National estimates of the timing of sexual maturation and racial
differences among US children. Pediatrics, 2002. 110(5): p. 911-9.

197. Tanner, J.M., Normal growth and techniques of growth assessment. Clin
Endocrinol Metab, 1986. 15(3): p. 411-51.

198. Thayer, R., Calm Energy: How People Regualte Mood with Food and Exercise.
2001, Oxford: Oxford University Press.

199. Thayer, R.E., The biopsychology of mood and arousal. 1989: Oxford University
Press, USA.

200. Thompson, D., et al., Covariability in diet and physical activity in African-American
girls. Obes Res, 2004. 12 Suppl: p. 46S-54S.

201. Thompson, J.K., et al., Relations Among Multiple Peer Influences, Body
Dissatisfaction, Eating Disturbance, and Self-Esteem: A Comparison of Average
Weight, At Risk of Overweight, and Overweight Adolescent Girls. J. Pediatr.
Psychol., 2007. 32(1): p. 24-29.

202. Thorburn, A. and J. Proietto, Biological determinants of spontaneous physical
activity. Obesity Reviews, 2000. 1(2): p. 87-94.

203. Treuth, M., et al., A Longitudinal Study of Sedentary Behavior and Overweight in
Adolescent Girls. Obesity, 2009. 17(5): p. 1003-1008.

204. Treuth, M.S., et al., Weekend and weekday patterns of physical activity in
overweight and normal-weight adolescent girls. Obesity (Silver Spring), 2007.
15(7): p. 1782-8.

205. Treuth, M.S., et al., Defining accelerometer thresholds for activity intensities in
adolescent girls. Med Sci Sports Exerc, 2004. 36(7): p. 1259-66.

141

206. Troiano, R.P., et al., Physical activity in the United States measured by
accelerometer. Med Sci Sports Exerc, 2008. 40(1): p. 181-8.

207. Trost, S.G., Objective measurement of physical activity in youth: current issues,
future directions. Exerc Sport Sci Rev, 2001. 29(1): p. 32-6.

208. Trost, S.G., State of the Art Reviews: Measurement of Physical Activity in Children
and Adolescents. American Journal of Lifestyle Medicine, 2007. 1(4): p. 299-314.

209. Trost, S.G., et al., Using objective physical activity measures with youth: how many
days of monitoring are needed? Med Sci Sports Exerc, 2000. 32(2): p. 426-31.

210. Trost, S.G., et al., Age and gender differences in objectively measured physical
activity in youth. Med Sci Sports Exerc, 2002. 34(2): p. 350-5.

211. Trost, S.G., et al., Correlates of objectively measured physical activity in
preadolescent youth. Am J Prev Med, 1999. 17(2): p. 120-6.

212. Trost, S.G., et al., Physical activity in overweight and nonoverweight preschool
children. Int J Obes Relat Metab Disord, 2003. 27(7): p. 834-9.

213. Trost, S.G., et al., Validity of the computer science and applications (CSA) activity
monitor in children. Med Sci Sports Exerc, 1998. 30(4): p. 629-33.

214. USDHHS, 2008 Physical Activity Guidelines for Americans. 2008: Washington,
D.C.

215. van der Aa, N., et al., Genetic Influences on Individual Differences in Exercise
Behavior during Adolescence. Int J Pediatr, 2010. 2010: p. 138345.

216. Van Der Horst, K., et al., A brief review on correlates of physical activity and
sedentariness in youth. Med Sci Sports Exerc, 2007. 39(8): p. 1241-50.

217. Ventura, E., et al., Reduction in Risk Factors for Type 2 Diabetes Mellitus in
Response to a Low-Sugar, High-Fiber Dietary Intervention in Overweight Latino
Adolescents. Arch Pediatr Adolesc Med, 2009. 163(4): p. 320-327.

218. Ventura, E.E., et al., Persistence of the metabolic syndrome over 3 annual visits in
overweight Hispanic children: association with progressive risk for type 2
diabetes. J Pediatr, 2009. 155(4): p. 535-41.

219. Vuori, I.M., Health benefits of physical activity with special reference to interaction
with diet. Public Health Nutr, 2001. 4(2B): p. 517-28.

220. Wang, Y.C., et al., Estimating the energy gap among US children: a counterfactual
approach. Pediatrics, 2006. 118(6): p. e1721-33.

142

221. Warren, J.M., C.J. Henry, and V. Simonite, Low glycemic index breakfasts and
reduced food intake in preadolescent children. Pediatrics, 2003. 112(5): p. e414.

222. Weigle, D.S., et al., Roles of Leptin and Ghrelin in the Loss of Body Weight
Caused by a Low Fat, High Carbohydrate Diet. J Clin Endocrinol Metab, 2003.
88(4): p. 1577-1586.

223. Weintraub, D.L., et al., Team sports for overweight children: the Stanford Sports to
Prevent Obesity Randomized Trial (SPORT). Arch Pediatr Adolesc Med, 2008.
162(3): p. 232-7.

224. Weston, A.T., R. Petosa, and R.R. Pate, Validation of an instrument for
measurement of physical activity in youth. Medicine & Science in Sports &
Exercise, 1997. 29(1): p. 138.

225. Whitaker, R.C., et al., Predicting Obesity in Young Adulthood from Childhood and
Parental Obesity. N Engl J Med, 1997. 337(13): p. 869-873.

226. Wickel, E.E., J.C. Eisenmann, and G.J. Welk, Maturity-related variation in
moderate-to-vigorous physical activity among 9-14 year olds. J Phys Act Health,
2009. 6(5): p. 597-605.

227. Witzig, R., The Medicalization of Race: Scientific Legitimization of a Flawed Social
Construct. Annals of Internal Medicine, 1996. 125(8): p. 675-679.

228. Wood, P., How fat works. 2006: Harvard Univ Pr.

229. Woodruff, R.S., A Simple Method for Approximating the Variance of a Complicated
Estimate. Journal of the American Statistical Association, 1971. 66(334): p. 411-
414.

230. Woods, S.C. and R.J. Seeley, Adiposity signals and the control of energy
homeostasis. Nutrition, 2000. 16(10): p. 894-902.

231. Yngve, A., et al., Effect of monitor placement and of activity setting on the MTI
accelerometer output. Med Sci Sports Exerc, 2003. 35(2): p. 320-6.

232. Zhang, S., et al., Association of plasma leptin concentrations with adiposity
measurements in rural Chinese adolescents. J Clin Endocrinol Metab, 2009.
94(9): p. 3497-504.

233. Zhang, Y., et al., Positional cloning of the mouse obese gene and its human
homologue. Nature, 1994. 372(6505): p. 425-32.

Abstract (if available)

Abstract Obesity prevalence is increasing and physical activity levels are declining in US youth. The overall goal of this dissertation was to examine the effects of individual demographic, biological, and dietary factors on objectively measured physical activity levels in youth. The objectives of this dissertation were: 1) to describe activity levels across race/ethnic, weight status, age, and gender groups in a large nationally representative sample of youth

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Influences of specific environmental domains on childhood obesity and related behaviors

PDF

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

PDF

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

PDF

The acute relationship between affective states and physiological stress response, and the moderating role of moderate-to-vigorous physical activity

PDF

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

PDF

System dynamics of in-home family eating behavior: insights from intensive longitudinal data using Ecological Momentary Assessment and wearable sensors

PDF

Sleep health and variability in youth: a real-time data capture study to examine influences on daily dietary intake patterns and longitudinal weight trajectories

PDF

Social interaction moderates enjoyment and perception of physical activity during exergame play in young adults with autism spectrum disorders

PDF

Effects of parent stress on weight-related parenting practices and child obesity risk

PDF

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

PDF

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

PDF

Mixed methods investigation of user engagement with a smoking cessation app

PDF

Understanding the dynamic relationships between physical activity and affective states using real-time data capture techniques

PDF

Effects of near-roadway air pollution exposure on obesity, obesity-related behavior, and neurobehavioral deficits during peripuberty

PDF

Effects of progestin-only long-acting reversible contraception on metabolic markers in obese women

PDF

Physical activity and sex hormone levels in postmenopausal women

PDF

Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California

PDF

Prenatal sleep health, cortisol, and gestational weight gain

PDF

Effects of the perceived and objectively assessed environment on physical activity in adults and children

Asset Metadata

Creator Belcher, Britni Ryan (author)

Core Title Objectively measured physical activity and related factors in minority youth

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 04/29/2011

Defense Date 03/10/2011

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

Tag accelerometry,gut hormones,minority youth,OAI-PMH Harvest,obesity,physical activity

Place Name USA (countries)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Spruijt-Metz, Donna (committee chair), Unger, Jennifer B. (committee chair), Cermak, Sharon (committee member), Chou, Chih-Ping (committee member), McConnell, Robert (committee member)

Creator Email bbelcher@usc.edu,britni01@aol.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3805

Unique identifier UC150655

Identifier etd-Belcher-4553 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-463218 (legacy record id),usctheses-m3805 (legacy record id)

Legacy Identifier etd-Belcher-4553.pdf

Dmrecord 463218

Document Type Dissertation

Rights Belcher, Britni Ryan

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