Citation
Five-year change in resting metabolic rate, respiratory quotient, and adiposity in African American and caucasian adults in the cardia study

Material Information

Title:
Five-year change in resting metabolic rate, respiratory quotient, and adiposity in African American and caucasian adults in the cardia study
Creator:
Sharp, Teresa Ann
Publication Date:
Language:
English
Physical Description:
xvii, 184 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Metabolism -- Measurement ( lcsh )
Respiration -- Measurement ( lcsh )
African Americans ( lcsh )
Whites -- Health and hygiene -- United States ( lcsh )
Obesity ( lcsh )
Body weight ( lcsh )
Body composition ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 157-194).
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Teresa Ann Sharp.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
60404800 ( OCLC )
ocm60404800
Classification:
LD1190.L566 2004d S42 ( lcc )

Full Text
FIVE-YEAR CHANGE IN RESTING METABOLIC RATE, RESPIRATORY
QUOTIENT, AND ADIPOSITY IN AFRICAN AMERICAN AND CAUCASIAN
ADULTS IN THE CARDIA STUDY
by
Teresa Ann Sharp
B.S., Middle Tennessee State University, 1979
M.Ed., Vanderbilt University, 1986
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences
2004


This thesis for the Doctor of Philosophy
degree by
Teresa Ann Sharp
has been approved
David P. Tracer
Tracy J. Horton
[b
Date
Stephen Sidney


Sharp, Teresa Ann (Ph.D., Health and Behavioral Sciences)
Five-year Change in Resting Metabolic Rate, Respiratory Quotient, and
Adiposity in African American and Caucasian Adults in the CARDIA Study
Thesis directed by Associate Professor David P. Tracer
ABSTRACT
Low resting metabolic rate (RMR) and elevated respiratory quotient (RQ)
have been suggested contributors to obesity development. Recent analyses
have shown higher body weight and fat mass in African American versus
Caucasian women, and lower RMR in African American men and women
compared to Caucasians. The current study reports the association of
longitudinal change in RMR and RQ with changes in body weight (BW) and
adiposity.
RMR and body composition were assessed at baseline and 5-year followup
in 72 African American (AA) men, 79 Caucasian (C) men, 72 AA women,
and 74 C women in the Coronary Artery Risk Development in Young Adults


(CARDIA) study who were aged 28-40 years at baseline. All four ethnic and
gender groups increased (p<0.05) body weight (BW) (range 2.8-4.5 kg) and
fat mass (FM)(2.3-4.2 kg) over the 5 years. There was no significant
difference in BW increase (4.5 7.5 AA vs 3.2 6.5 kg C) or FM increase
between the two ethnic groups (4.2 6.9 kg AA, and 3.0 5.2 kg C). AA
men gained more FM than C men (4.2 5.7 vs 2.3 5.7kg, p<0.05), but
there was no difference in the amount of FM gained by women of the two
groups. Change in fat-free mass (FFM) was non-significant in all groups.
Multiple regression models using baseline variables were good predictors of
BW and FM change in C subjects, but not in AA subjects.
Mean unadjusted RMR increased only in AA women (45.12 kcal, p<0.05).
RMR/FFM was lower at both time points in AA compared to C (-2.12 .51
kcal/kg FFM, p<0.0001). When adjusted for BW, RMR change was only
significant in Caucasian women (-0.83 2.27 kcal/d, p<0.05). Change in
RMR adjusted for FFM was only significant in Caucasian men (0.59 2.6
kcal/d, p<0.05). There was no significant change in RQ over 5 years in any
group. In conclusion, changes in RMR or RQ do not explain the greater BW
gain in African American women or the greater gain in FM by African
American men than in Caucasian men over a 5-year period.
hi


This abstract accurately represents the contents of the candidate's thesis. I
recommend its publication.
Signed
David P. Tracer
IV


DEDICATION
I dedicate this thesis to my family, particularly my mother, Mary Margaret,
and my father, John McHugh, who have always provided me the love and
support that has encouraged me to achieve any of the goals that I have
sought to achieve. I am also very appreciative of the support from my
brothers, Mike and Paul, and my sister, Karen, my sisters-in-law, and the
nieces, nephews, aunts, uncles, and cousins, who have encouraged me
(and tolerated me) throughout this process.


ACKNOWLEDGEMENTS
I would first like to thank my committee members; David Tracer (Chair) for
encouraging me, debating with me, and helping me through the process on
countless occasions, Lorna Moore for your support and also your assistance
in catching my statistical mistakes, Steve Sidney for allowing me to be a part
of this project from the beginning and helping me through the CARDIA
process, and finally Tracy Horton for your support and your fine-toothed
comb and assistance with interpretation; you are a godsend.
I would like to thank my cohort members especially Lisa, Bernie, Leigh, and
Mary, and other HBS students Michael, Bonnie, Carol, Mark, and others for
support as you understand and/or have been there, done that. I would
also like to thank my home and work friends for putting up with me and
encouraging me throughout this entire processdont worry, I wont do it
again! The Zaritsky girls (Toba, Kara, and Hannah), Richard, Ken, John,
Michael, Sue, Andra, Ann, Jan, Jere, Ligia, Marsha, Mary, Velma, Debi,
Sharon, Barb, and the list goes on...


Lastly, I would like to thank the faculty and staff of the Health and Behavioral
Sciences Department, but especially Christine Pon for showing me the way
to jump through the institutional hurdles that add fuel to the fire, and doing
so in such a caring manner.
Appreciation is given to the CARDIA Steering Committee, and the Emerging
Science Committee for approval in using data from the Visceral Fat,
Metabolic Rate, and CHD Risk in Young Adults (VIM) Study for the data
analyses included in this thesis. I also want to thank Luisa Hamilton of the
Department of Research of Kaiser Permanente at Oakland, and Nita Webb
and Phil Johnson of the University of Alabama-Birmingham for data
collection assistance. This study was supported by a grant from the
National Institutes of Health; NHLBI R01 HL53359, and contracts HC-
95,095, HC-48,047, HC-48,048, HC-48,049, and HC-48,050.


CONTENTS
Figures............................................................xiv
Tables ............................................................xv
Chapter
1. Introduction..................................................1
1.1 Statement of the Problem......................................1
1.2 Purpose and Significance of the Study.........................2
1.3 Hypotheses and Specific Aims..................................5
1.3.1 Hypothesis #1.................................................5
1.3.2 Hypothesis #2.................................................6
1.3.3 Hypothesis #3.................................................7
1.3.4 Hypothesis #4.................................................9
1.4 Chapter Synopses.............................................10
1.4.1 Chapter 2: Literature Review.................................11
1.4.2 Chapter 3: Participants and Methodology.....................11
1.4.3 Chapter 4: Results...........................................12
1.4.4 Chapter 5: Discussion........................................12
2. Literature Review............................................13
2.1 Prevalence of Obesity........................................13
IX


2.2 Use of Ethnic and Gender Terminology.........................15
2.3 Ethnic and Gender Differences in Obesity Prevalence..........16
2.4 Obesity-related Disease......................................22
2.5 Economic Impact of Obesity...................................25
2.6 Cardiovascular Disease, Type 2 Diabetes, and Metabolic
Syndrome....................................................27
2.7 Ethnicity, Obesity, and Disease Risk.........................28
2.8 Metabolic Factors Associated with Body Weight Regulation....31
2.9 Energy Expenditure...........................................31
2.10 Relationship of Gender, Ethnicity, Body Weight and Body
Composition with Resting Metabolic Rate....................34
2.10.1 Gender Differences in Energy Expenditure.....................34
2.10.2 Ethnic Differences in Energy Expenditure....................36
2.10.3 Ethnic and Gender Differences in Resting Metabolic Rate.....38
2.10.4 Body Weight, Body Composition and Energy Expenditure........41
2.11 Impact of Respiratory Quotient on Obesity Development, and
Factors Affecting Respiratory Quotient......................41
2.11.1 Factors Affecting Resting Metabolic Rate and Respiratory
Quotient....................................................43
2.12 Summary of Literature Review.................................44
3. Participants and Methodology.................................46
3.1 CARDIA Background and Methods................................46
3.2 Background on VIM............................................48
x


3.3 VIM Study Methods............................................49
3.4 Participants.................................................50
3.5 Resting Metabolic Rate.......................................51
3.6 Body Weight, Height, and Body Mass Index (BMI)...............53
3.7 Body Composition.............................................54
3.8 Statistics...................................................56
3.8.1 Methods of Data Expression...................................56
3.8.2 Descriptive Data.............................................57
3.8.3 Analyses of Baseline and Year Five Data......................57
3.8.4 Analyses of Change Data......................................58
3.8.5 Ability of Data to Predict Five-year Change..................58
4. Results......................................................60
4.1 Descriptive Data for VIM Baseline and VIM Year Five Time
Points......................................................60
4.1.1 BMI and Weight...............................................63
4.1.2 Body Composition.............................................66
4.1.3 Resting Metabolic Rate and Respiratory Quotient..............67
4.1.4 Resting Metabolic Rate Expressed per Kilogram of Fat-free
Mass, and per Kilogram of Body Weight.......................69
4.2 Results for Hypothesis #1....................................73
4.2.1 Relationship of Baseline Measures of Resting Metabolic Rate,
Body Weight, and Adiposity to Five-year Change in Weight and
Adiposity...................................................74
XI


4.2.2 Baseline Age as a Predictor of Five-year Change in Body
Weight, Adiposity, Resting Metabolic Rate, and Respiratory
Quotient.....................................................78
4.2.3 Multiple Regression Results of Baseline Variables as Predictors
of Five-year Change in Body Weight and Fat Mass.............84
4.2.4 Relationship of Baseline Weight Status (Overweight and Non-
overweight) to Five Year Change in Weight and Fat Mass........99
4.3 Results for Hypothesis #2...................................102
4.4 Results for Hypothesis #3...................................104
4.4.1 Relationship of Baseline Body Weight with Change in Resting
Metabolic Rate..............................................111
4.4.2 Relationship of Baseline BMI with Change in Resting Metabolic
Rate........................................................112
4.4.3 Relationship Between Change in Resting Metabolic Rate and
Change in Body Weight and Adiposity over Five Years.........113
4.5 Results for Hypothesis #4...................................116
4.6 Primary Results.............................................119
5. Discussion................................................. 122
5.1 Current Knowledge...........................................122
5.2 Justification to Conduct the Present Data Analyses..........124
5.3 Main Questions Addressed in the Present Study...............125
5.4 Primary Findings............................................125
5.4.1 Body Weight and Body Composition............................126
5.4.2 Unadjusted Resting Metabolic Rate...........................129
XII


5.4.3 Resting Metabolic Rate Adjusted for Weight or Fat-free
Mass.........................................................131
5.4.4 Change in Resting Metabolic Rate and Change in Body
Weight and Adiposity.........................................133
5.4.5 Change in Respiratory Quotient and Change in Weight and
Adiposity....................................................134
5.5 Ability to Predict Five-year Changes in Body Weight and
Adiposity....................................................139
5.5.1 Resting Metabolic Rate and Respiratory Quotient...............140
5.5.2 Age...........................................................141
5.5.3 Body Weight and Body Composition..............................144
5.5.4 Body Mass Index...............................................145
5.5.5 Prediction Models for Body Weight and Fat Mass Change.......145
5.6 Conclusions...................................................147
5.7 Significance..................................................150
5.8 Limitations...................................................151
5.9 Future Directions.............................................154
References.........................................................157
XIII


FIGURES
Figure 2.1. Baseline (1985-86) and Year 10 (1995-96) Body Weights
for CARDIA Participants (N=5115, mean sd).............21
Figure 4.1. Mean ( sd) Body Mass Index (BMI) for the Total Sample
at Baseline and Year Five...............................65
Figure 4.2. Mean BMI ( sd) at Baseline and Year Five for each
Ethnic and Gender Group.................................66
XIV


TABLES
Table 2.1.
Table 4.1.
Table 4.2.
Table 4.3.
Table 4.4.
Table 4.5.
Table 4.6.
Table 4.7.
Table 4.8.1.
Mortality Ratios from Diabetes, Coronary Heart Disease
(CHD), and Cancer by BMI................................25
Participant Characteristics: Baseline Age, and Baseline
and Year Five Body Weight (Wt kg), Body Mass Index
(BMI), and Body Composition (FFM kg, FM kg, and % fat)
Mean (+ SD).............................................62
Participant Characteristics: Baseline and Year Five Resting
Metabolic Rate (RMR, kcal/day) and Respiratory Quotient
(RQ) Mean (+SD).........................................68
Scheffe' Multiple Comparisons Test of Differences in Mean
Values of Resting Metabolic Rate per Kilogram of Body
Weight (Kcal/wtkg) at VIM Baseline and VIM Year Five
Between African Americans and Caucasians................71
Scheffe' Multiple Comparisons Test of Differences in Mean
Values of Resting Metabolic Rate per Kilogram of Fat-free
Mass (Kcal/FFM) at VIM Baseline and VIM Year Five
Between African Americans and Caucasians.............72
Relationship of Baseline Measures of Age, Resting Metabolic
Rate (RMR), Body Weight (WT), Fat-free Mass (FFM), Fat
Mass (FM), or Percent Body Fat (% fat) to the Change in
Weight or Adiposity over Five Years....................77
Association of Age at VIM Baseline with Change in Resting
Metabolic Rate (RMR), Respiratory Quotient (RQ), and Fat-
free Mass (FFM).........................................81
Association of Stratified Age (younger and older) at VIM
Baseline with Change in Resting Metabolic Rate (RMR),
Respiratory Quotient (RQ), Weight, and Adiposity........83
Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in the Total Sample.... 85
xv


Table 4.8.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in the Total Sample......86
Table 4.9.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in African American
Participants.........................................................87
Table 4.9.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in African American
Participants.........................................................88
Table 4.10.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in Caucasian
Participants.........................................................89
Table 4.10.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in Caucasian
Participants.........................................................90
Table 4.11.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in African American
Men..................................................................92
Table 4.11.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in African American
Men..................................................................93
Table 4.12.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in Caucasian Men......94
Table 4.12.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in Caucasian Men.........95
Table 4.13.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in African American
Women................................................................96
Table 4.13.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in African American
Women................................................................97
XVI


Table 4.14.1. Multiple Regression Analyses of Baseline Variables as
Predictors of Body Weight Change in Caucasian
Women...............................................................98
Table 4.14.2. Multiple Regression Analyses of Baseline Variables as
Predictors of Fat Mass Change in Caucasian Women.....99
Table 4.15 Difference in the Amount of Weight and Fat Mass
Change over Five Years by Overweight and Non-
overweight Status (Mean SD)......................................101
Table 4.16. Prediction of Weight and Fat Mass Change over Time by
Baseline Respiratory Quotient (RQ) in Total Sample and
within each Ethnic and Gender Group................................103
Table 4.17. Magnitude and Percent Change of Baseline Value for
Body Weight (Wtkg), Fat-free Mass (FFM), Fat Mass
(FM), Percent Body Fat (%fat), Resting Metabolic Rate
(RMR), and Respiratory Quotient (RQ) Mean (+ SD)..................107
Table 4.18. Change from Baseline in Resting Metabolic Rate (RMR)
per Kilogram of Body Weight (Wt), per Kilogram of Fat-
free Mass (FFM), and per Kilogram of Fat Mass (FM)
(Mean+ SD)............................................110
Table 4.19. Relationship of Baseline Body Weight to Five-year
Change in Resting Metabolic Rate (RMR)..............112
Table 4.20. Relationship of Baseline Body Mass Index (BMI) to Five-
year Change in Resting Metabolic Rate (RMR)..........113
Table 4.21. Association of Five-year Change in Resting Metabolic Rate
(RMR) with Change in Body Weight, Fat Mass, and Fat-
free Mass..........................................................115
Table 4.22. Association Between Five-year Change in Respiratory
Quotient and Five-year Change in Body Weight.........117
Table 4.23 Association Between Five-year Change in Respiratory
Quotient and Five-year Change in Fat Mass.............118
XVII


1.
Introduction
1.1 Statement of the Problem
Obesity has reached epidemic proportions worldwide, and it is believed that
obesity has a strong association with at least 17 chronic diseases (Booth,
Gordon, Carlson, and Hamilton 2000). In the United States, African
American adults have a 7.3% higher prevalence of overweight (69.9 vs 62.3,
n.s.), and an 11.2% (39.9 vs 28.7, p<0.05) higher prevalence of obesity than
Caucasian adults (Flegal, Carroll, Ogden, and Johnson 2002). Therefore,
African Americans are at greater risk for obesity-related diseases than are
Caucasians, and are known to have the highest coronary artery disease
mortality compared to other ethnic groups (Hall, et al. 2003). At present, the
causes of this disparity are not fully known, but may include ethnic
differences in biological, metabolic, behavioral/psychosocial, and
environmental factors (Dubbert, et al. 2002). There is a need to address the
etiology of the body weight regulation process in different ethnic groups to
gain information for strategies targeting the prevention and treatment of
obesity.
1


1.2 Purpose and Significance of the Study
Obesity develops overtime, and is related to a combination of behavioral,
environmental, and physiological factors. Two metabolic factors that could
play a role in body weight changes include metabolic rate and nutrient
oxidation. Indeed, cross-sectional data suggest that a reduced initial resting
metabolic rate, and a higher resting respiratory quotient (reflecting a greater
preference for carbohydrate over fat oxidation), predict weight gain over time.
Only one study (Nicklas, Berman, Davis, Dobrovolny, and Dennis 1999) has
addressed potential ethnic and gender differences in the relationship
between initial resting metabolic rate, respiratory quotient, and subsequent
weight change. To date, no investigation has examined longitudinally the
change in measured resting metabolic rate and respiratory quotient over time
in African American and Caucasian adults and related this to longitudinal
changes in body weight and body composition. Whereas cross-sectional
studies allow only a theoretical cause and effect of relationships based upon
one-time measures, the current study provides longitudinal data that allow
for the study of direct comparisons within these relationships.
Since resting metabolic rate represents 65-80% of total daily energy
expenditure in moderately active or sedentary populations (Ravussin 1995),
a lesser increase in metabolic rate for a given increase in body weight over
2


time, or a reduction in resting metabolic rate, would likely result in an
increase in body weight and adiposity unless energy intake or energy
expended in physical activity were altered to compensate for this change.
Additionally, changes in substrate oxidation over time, particularly reductions
in fat oxidation, could be related to changes in body weight. Since African
Americans have been shown to have a higher prevalence of excess body
weight when compared to Caucasians, we hypothesize that African
Americans will have a lesser increase in the amount of energy expended at
rest (over the course of a five-year period) than Caucasians, which will
contribute to their greater increase in body weight and adiposity compared
to Caucasians over the five-year period. We also hypothesize that African
Americans will show a greater increase in respiratory quotient from baseline
to five-year follow-up compared to Caucasians. This suggests a decreased
reliance on fat as a fuel source resulting in the potential for further weight
gain, but especially fat mass gains.
The purpose of this study, therefore, is to assess the relationship between
changes in resting metabolic rate and whole-body respiratory quotient over
a five-year period with change in body weight and adiposity over the same
time period. In addition, potential ethnic and gender differences in these
variables will be compared to determine if gender and/or ethnic variation in
3


resting metabolic rate, and/or the respiratory quotient, either at baseline or
the change over time, contribute to potential gender and ethnic differences
in changes in adiposity.
The results of this study can significantly contribute to the literature by
providing information relevant to the potential contributors to increasing body
weight and adiposity over time that is not available in the cross sectional
studies currently published. In addition, it can also provide information on
predictors of excess body weight that can be used as target areas for
interventional strategies that attempt to prevent the escalating incidence of
obesity-related disease, particularly in African American groups.
4


1.3 Hypotheses and Specific Aims
This study addresses the following hypotheses. Specific aims are provided
for each hypothesis.
1.3.1 Hypothesis #1
Baseline age and baseline measures of body weight, and adiposity will be
positively associated with changes in body weight and adiposity in African
American and Caucasian subjects.
Baseline resting metabolic rate will be inversely associated with changes in
body weight and adiposity in African American and Caucasian subjects.
Rationale: Body weight and fat mass increase with age, and it is therefore
expected that individuals who are older at baseline will have greater
increases in body weight and adiposity after five years than individuals who
are younger at baseline. A low resting metabolic rate has been suggested
as a risk factor for body weight and fat mass gain. Therefore, it is expected
that individuals with a low baseline resting metabolic rate will have greater
gains in body weight and fat mass over the five-year period.
5


Specific aims for Hypothesis #1:
1. To determine if "older" or "younger" age at baseline is
predictive of five-year weight or fat mass change in African
American and Caucasian adults.
2. To determine if there are ethnic and gender differences in the
relationship between baseline age and change in weight or fat
mass over time.
3. To determine if baseline measures of resting metabolic rate,
body weight and body composition are predictive of five-year
weight or fat mass change in male and female African
American and Caucasian adults.
1.3.2 Hypothesis #2
Baseline respiratory quotient will be positively related to change in body
weight and change in fat mass.
Rationale: It has been suggested that individuals who have an elevated
respiratory quotient, indicating a preference for carbohydrate versus fat as a
fuel source, will have increased fat storage over time.
6


Specific Aim for Hypothesis #2.
1. To determine if baseline respiratory quotient is predictive of
five-year change in body weight and/or adiposity in male and
female African American and Caucasian adults.
1.3.3 Hypothesis #3
For the same increase in body weight, African Americans will have a lower
proportion of that increase in fat-free mass and a higher proportion of that
increase in fat mass.
As fat-free mass is the major determinant of resting metabolic rate, this
predicts that for the same increase in absolute body weight, the increase in
resting metabolic rate will be less in African American than in Caucasian
subjects.
Rationale: African Americans have a higher prevalence of obesity
compared to Caucasians, hence higher fat mass. Previous literature
suggests that African American women have greater fat mass gains than
Caucasian women, but there is currently little data related to adiposity
changes in men of the two ethnic groups. It is expected that increases seen
7


in body weight over the five year period will be primarily due to increases in
fat mass, and not due to fat-free mass increases. Since fat-free mass is the
primary contributor, and fat mass minimally contributes (less than three
percent), to resting metabolic rate, a lesser increase in resting metabolic
rate over the five-year period would be expected in individuals who gain
primarily fat mass over the same time period.
Specific aims for Hypothesis #3:
1. To determine if there is a difference in the change in resting
metabolic rate over a five-year period between African
American and Caucasian subjects.
2. To determine if baseline weight status is predictive of five-year
change in resting metabolic rate.
3. To determine if five-year change in resting metabolic rate is
associated with five-year change in body weight and adiposity.
4. To determine if there are differences in the relative contribution
of fat versus fat-free mass to the five-year increase in body
weight, for the different ethnic and gender groups
8


5. To determine if five-year change in resting metabolic rate, and
five-year change in body weight and adiposity vary by ethnicity
and gender.
1.3.4 Hypothesis #4
Five-year change in respiratory quotient will be positively related to change
in body weight, and change in fat mass.
Rationale: A higher respiratory quotient is indicative of a greater preference
for carbohydrate versus fat as a fuel source. This has been suggested to
result in a preference for fat storage. Thus, an increase in respiratory
quotient over the five-year period could be a factor contributing to a greater
increase in fat mass and body weight over time.
9


Specific aims for Hypothesis #4:
1. To determine if there is a change in respiratory quotient over
the five-year period in male and female African American and
Caucasian adults.
2. To determine if a change in respiratory quotient over a five-
year period is associated with a change in body weight and fat
mass.
3. To determine if a positive change in respiratory quotient and a
positive change in body weight and adiposity over a five-year
period is higher in African Americans as compared to
Caucasians.
1.4 Chapter Synopses
This thesis is organized into the following five chapters: Introduction,
Literature Review, Participants and Methodology, Results, and Discussion.
A synopsis is provided below for each of the chapters that follow this
Introduction.
10


1.4.1 Chapter 2: Literature Review
Chapter 2 provides background and a literature review relevant to the
prevalence of obesity and obesity-related diseases. This chapter also
provides background on the terminology to be used in this thesis regarding
ethnicity, gender, daily energy expenditure, substrate oxidation, and body
composition. Information is also given about the relationship between each
of these topics and obesity development.
1.4.2 Chapter 3: Participants and Methodology
Chapter 3 provides information related to the participants whose data is
examined in the current thesis project as well as a description of the
background and methods. This chapter contains details of the Coronary
Artery Risk Development in Young Adults (CARDIA) study, which is the
parent study from which an ancillary study, the Visceral Fat, Metabolic Rate,
and CHD Risk in Young Adults (VIM) study was derived. It is the ancillary
study, the VIM study, conducted during two of the follow-up examination
periods of the CARDIA trial (Years 10 and 15), from which the subject
population examined in this thesis was drawn. Finally, descriptions are
provided for each of the measurement methods used in the current thesis
project: resting metabolic rate, body weight and adiposity measures, and
statistical analyses.
11


1.4.3 Chapter 4: Results
This chapter includes descriptive data, the results of statistical analyses of
Baseline and Year 5 VIM data, and the analyses of data used to address
each of the five hypotheses presented above. Graphical illustrations and
data tables are also included in this chapter.
1.4.4 Chapter 5: Discussion
This chapter provides a brief description of current knowledge related to
ethnic and gender considerations in the prevalence of obesity, and
considerations related to the development of obesity, and obesity-related
diseases. This is followed by justification for initiating the current project,
and a discussion of the results of the analyses addressing the specific aims
and hypotheses of this thesis project. The significance and limitations of the
present study are also provided. Finally, directions for future research are
suggested.
12


2.
Literature Review
2.1 Prevalence of Obesity
Body weights of the U.S. population have increased steadily and significantly
from 1980 to the present. The National Health and Nutrition Examination
Survey (NHANES) III report (1988-1994) has indicated an increasing
prevalence of overweight in the U.S; population after staying fairly stable
between 1960 and 1980 (Flegal, Carroll, Kuczmarski, and Johnson 1998). In
a recent paper by Flegal and colleagues (2002), the severity of this epidemic is
illustrated by comparing data on body mass index (BMI) from the National
Health Examination Survey (NHES) I (1960-1962) through a recent survey as
part of the National Health and Nutrition Examination Survey (NHANES)
conducted in 1999 and 2000 (Flegal, et al. 2002). Over this time frame, the
percent of US adults classified as overweight (BMI 25.0 to 29.9 kg/m2)
increased from 46.0% to 64.5%, the prevalence of obesity (BMI 30.0 to 39.9
kg/m2) rose from 13.4% to 30.9%, and severe obesity (BMI > 40 kg/m2) rose
from 2.9% to 4.7%. Remarkably, these changes occurred in just 20 years!
Using data from 184,450 adults responding to a telephone survey conducted
as part of the Behavioral Risk
13


Factor Surveillance System (BRFSS), Mokdad and colleagues (2001) report a
19.8% prevalence of obesity in US adults, which represents a 61% increase
since 1991! Although the percentages differ between the surveys used, it is
still evident that increasing levels of obesity are a critical problem within the
United States.
Data from international surveys illustrate that the increasing prevalence of
obesity and overweight is a global problem (Siddiqui 2003). Both developed
and many developing nations are realizing trends similar to those seen in the
United States (Hodge, Dowse, Zimmet, and Collins 1995, Seidell 1995).
There have been increases in the percent of the adult population that are
overweight in Brazil, Canada, the United Kingdom, Australia, Thailand, China,
Japan, Finland, New Zealand, Western Samoa, and Mauritius (Taubes 1998,
Flegal 1999, Torrance, Hooper, and Reeder 2002). Although there was only a
slight increase in obesity prevalence in some of the Scandinavian countries
(Netherlands, Sweden, and Denmark) between 1980 and 1991, there were
increases nonetheless (Flegal 1999). It has been further reported that the
prevalence of overweight in Lithuanian and Russian women is approaching
70%, and populations of Australian Aborigines and some Pacific Islanders
have reached up to 100% (Bjorntorp 1998). Taubes (1998) suggests that
even though some of the smaller countries may have limited and potentially
14


poor quality data, there is still reason to be concerned about these worldwide
trends.
2.2 Use of Ethnic and Gender Terminology
There have been inconsistencies in the terminology used in the research
literature when referring to race or ethnicity, and sex or gender. Yu and Liu
(1992) suggest that this is because minority classifications do not follow the
conventional definitions of race, which is assumed to be biology based,
versus ethnicity, which is assumed to be socio-behavioral. They suggest
that race or ethnicity are not mutually exclusive in terms of definition, and
can have some overlap within the definitions. For example, race tends to be
primarily biological; including a somewhat constant set of genetically
determined physical traits. Ethnicity typically includes both self-identification
of race along with common cultural norms and values. In relation to obesity-
related research, Brown (1993) has defined ethnicity as, "...the cultural
commonalities of members of a group who claim reference to common origins
and who operate in the context of a wider social system." We are not only
interested in biological traits, but are primarily interested in the combination of
biological, behavioral and cultural impacts on obesity development, and
therefore will use ethnicity in this thesis.
15


In an Institute of Medicine report (Wizeman and Pardue 2001), sex refers to
the classification of living things as male or female according to their
reproductive organs and functions assigned by chromosomal complement.
Gender refers to a person's self-representation as male or female, and is
rooted in biology and shaped by environment and experience.
Since there is no genetic information available for the participants in the study
on which this thesis is based, and because these participants self-reported
ethnicity and gender, the terms "ethnicity" and "gender" are used in this thesis.
2.3 Ethnic and Gender Differences in Obesity
Prevalence
It has been suggested that there is a potential evolutionary compensation
that may lead to obesity development given todays environment of readily
available food sources, and technology promoting physical inactivity. Neel
(1962) suggested that humans throughout history have lived in states of
feast and famine, during which they may have developed a thrifty genotype
that allowed for storage of energy as fat during periods of feast that allowed
them to survive during periods of famine. Following Neels premise, Brown
(1993) has suggested that individuals with an ancestry more recently
progressed from hunting and gathering societies may have a genetic
predisposition for energy storage. The current obesity epidemic seen in
16


different ethnic groups could actually stem from a predisposition for energy
storage, which is exacerbated by environmental factors in todays society.
When food supplies and mechanized transportation are as steady and
abundant as they are to most adults living in the United States today, the
ability for increased fat storage is detrimental, and may lead to obesity and
related disease. It appears that individuals are living in todays environment
with a genetic background that has not had the opportunity to evolve to cope
with this relatively new environment.
Some of the variability that we see in obesity development with these readily
available food sources may actually stem from substantial genotypic
variation among individuals within, and between, the ethnic groups. For
example, there may be a great amount of variation in the proportion of the
genome that promotes energy storage within different individuals that
actually descend from Africa, and this could be dependent upon the location
of the origination of groups with regard to food type and availability. If some
groups originated from areas that had greater fluctuations in food availability
(ie., feast and famine in arid or semi-arid environments), whereas other
groups originated from areas of more stable food availability (ie., coastal or
river basin areas), it might be expected that there would be variation in the
development of a thrifty genotype between the groups. In other words,
17


there may be more, or less, of the thrifty genotype that may explain some
of the variation that we see in obesity development within, and between,
ethnic groups (Allison, et al. 2003). Although further investigation into this
theory is beyond the scope of this thesis, the possibility of ethnic differences
in the development of a thrifty genotype does provide some reasoning
related to the interest in differences in obesity development between ethnic
groups.
As mentioned previously, the increasing prevalence of obesity worldwide is
apparent in a broad range of ethnic and gender groups. In the United
States, obesity disproportionately affects African American women. African
American women have approximately a two-fold greater prevalence of
overweight and obesity than Caucasian women (Najjar and Rowland 1987,
Kumanyika 1987, Winkelby, Kraemer, Ahn, and Varady 1998). Kumanyika
(1987) combined datasets from the National Health Examination Survey
(NHES), the National Health and Nutrition Examination Survey (NHANES),
the National Health Interview Survey (NHIS), and data from the National
Center for Health Statistics (NCHS), and found prevalence ratios of
overweight between 1.6 and 2.5 times more frequent in African American
women compared to Caucasian women. Winkelby and colleagues (1998)
18


found African American women were, on average, 16.8 pounds heavier than
Caucasian women of comparable socio-economic status and age.
To date, there are limited data indicating a difference in the prevalence of
obesity between men of the two ethnic groups (Kumanyika 1994, Flegal, et
al. 1998, Flegal, et al. 2002). These studies report a trend towards a higher
prevalence of overweight (body mass index (BMI) 25-29.9 kg/m2), but a
lower prevalence of obesity (BMI > 30 kg/m2), in Caucasian men as
compared to African American men, but the differences are not statistically
significant. The reason for this gender difference in the prevalence of
obesity between Caucasians and African Americans is not fully understood.
It is worth noting that obesity prevalence in other ethnic groups is also
elevated when compared to Caucasians. Based upon data collected
between 1982 and 1984 as a part of the Hispanic Health and Nutrition
Examination Survey (HHANES), Pawson, Martorell, and Mendoza (1991)
confirmed that Hispanics living in the United States exhibit high prevalences
of overweight (33.5% of males, 42.3% of females) and obesity (10.6% of
males, 15.1% of females) based upon BMI. More recent data indicate that
the prevalence of obesity in Hispanic men and women are following the
national trend; obesity prevalence has increased in Hispanic men by 5%,
19


and by 4.4% in Hispanic women over the last 20 years (Flegal, et al. 2002).
Although these increases are of concern, high obesity prevalence values are
evident in African American (29.3%) and Caucasian adults (18.5%)
(Mokdad, et al. 2001). Therefore, the focus of this thesis will be on these
two ethnic groups.
African American women have a 50% greater incidence of major weight gain
( 10 kg) during middle age (30-55 years) than Caucasian women when data
were tracked over a ten-year period (Williamson, Kahn, and Byers 1991).
Using data from the CARDIA study spanning a ten-year period, Lewis and
colleagues (2000), reported substantial increases in body weight in each
ethnic and gender group (African American and Caucasian males and
females) (see Figure 2.1). On average, African American women gained the
most weight over the ten-year period (11.9 kg) followed by African American
men (10.6 kg), Caucasian men (7.8 kg) and Caucasian women (6.9 kg).
20


Figure 2.1 Baseline (1985-86) and
Year 10 (1995-96) Body Weights for
CARDIA Participants (N=5115, mean
sd).
i i Baseline Wt kg
Ethnic and Gender Group
Even after controlling for income and education levels, African American
women are more than twice as likely to be obese than Caucasian women
(Harrell and Gore 1998). There is clearly a critical need to determine the
underlying mechanisms associated with the higher prevalence of obesity in
African Americans, particularly since obesity-related diseases such as
cardiovascular disease and Type 2 diabetes are more prevalent in African
American adults relative to their Caucasian counterparts (Cowie, Harris,
Silverman, Johnson, and Rust 1993, Winkleby, et al. 1998, Hall, et al. 2003,
21


Dubbert, et al. 2002, Sowers and Sowers 1999, Sowers, Ferdinand, Bakris,
and Douglas 2002).
2.4 Obesity-related Disease
In 1985, a panel of experts at an NIH conference concluded that obesity is a
disease due to its association with many comorbidities (National Institutes of
Health [NIH], 1985). More recently, obesity has been classified as a chronic
disease (Bjorntorp 1998), based on the worldwide trend of an increase in body
weight, with no corresponding change in stature. The health implications of
increasing body weight have been previously reported (Barrett-OConnor
1985, Willett, et al. 1985, Colditz, et al. 1990, Shaper, Wannamethee, and
Walker 1997, Giovannucci, et al. 1995, Pi-Sunyer 1991, Folsom, et al. 1989,
Must, et al. 1999, Kenchaiah, et al. 2002, Michaud, et al. 2001, Stoll 2002).
The disease conditions positively associated with obesity include diabetes,
cardiovascular disease, arthritis, hypertension, hyperlipidemia, gout, impaired
pulmonary function, gallbladder disease, and various cancers.
Calle, Rodriguez, Walker-Thurmond, and Thun (2003) prospectively studied
900,000 U.S. adults who entered the Cancer Prevention Study II free of
diagnosed cancer in 1982 and followed them for 16 years. They determined
that overweight status (BMI >25 kg/m2) was associated with 20% of cancer
deaths in women, and 14% of cancer deaths in men. Excess weight was
22


linked to cancers of the uterus, kidney, liver, pancreas, esophagus, gallbadder,
colon and rectum, and breast. Using data from 1991 that was collected in six
longitudinal studies within the United States (Alameda Community Health
Study, Framingham Heart Study, Tecumseh Community Health Study,
American Cancer Society Cancer Protection Study, National Health and
Nutrition Survey, and the Nurses Health Study), Allison, Fontaine, Manson,
Stevens, and Vanltallie (1999) calculated that the annual number of deaths
attributable to obesity in the United States was 280,000, or approximately 8%
of the total number of deaths.
Of particular concern, is the relationship between obesity and the risk of
associated chronic diseases. In an editorial summarizing the highlights of the
International Obesity Task Force (IOTF) report to the World Health
Organization (WHO), Bjorntorp (1998) stresses the profound health and
economic consequences of this worldwide epidemic. In 1998, the American
Heart Association officially added obesity to the list of risk factors for
cardiovascular disease (CVD); joining smoking, high cholesterol, high blood
pressure, and sedentary lifestyle (Eckel and Krauss 1998). Obesity is also an
independent risk factor for Type 2 diabetes (Ford, Williamson, and Liu 1997,
Resnick, Valsania, Halter, and Lin 2000). Both CVD and Type 2 diabetes are
major causes of obesity-related mortality in the United States. In 1999, over
23


725,000 U.S. adults died of cardiovascular disease, and over 68,000 died of
diabetes mellitus (Hoyert, Arias, Smith, Murphy, and Kochanek 2001).
Perhaps most striking is the information provided in Table 2.1, which presents
mortality ratios by body mass index in non-smoking 50-59 year old women
(Vanltallie and Lew 1993). The table shows that with increasing BMI, mortality
ratios for diabetes, coronary heart disease (CHD) and cancer also increase.
That is, higher BMI levels were associated with higher death rates from these
diseases than were lower BMI levels. Values for men are similar, although for
any fixed value of BMI, the mortality ratios are higher for women.
24


Table 2.1. Mortality Ratios from Diabetes, Coronary Heart Disease
(CHD), and Cancer by BMI.
Mortality Ratios (%) *
BMI (kg/m*) Diabetes CHD Cancer
20.0-22.5 64 82 92
22.5-27.5 100 100 100
27.5-30.0 228 145 114
30.0-32.5 435 164 116
32.5-35.0 712 216 112
>35.0 1232 311 182
* In relation to death rates among those 90 to 109% of average weight.
Death rates among women of average weight exceeded those among
women with below-average ("optimal") weights.
2.5 Economic Impact of Obesity
Not only do increased levels of obesity elevate chronic disease risk and
mortality, they also result in rising health care costs. Obesity-related health
issues result in nearly 3-8% of all of the health care expenditures for the United
States and other countries (Bjorntorp 1997). Moreover, the potential for lost
productivity due to obesity-related illness also contributes to the economic
costs of obesity. In 1998, the U.S. Institute of Medicine estimated an annual
25


burden of $70 billion attributable to direct and indirect health care costs
associated with obesity (Wickelgren 1998). Since 1998, both direct and
indirect health care costs attributable to obesity have profoundly increased.
Estimates have ranged from $100 billion per annum (Centers for Disease
Control [CDC], 2001), to nearly $123 billion per annum (Wolf, Manson, and
Colditz 2003).
Data are also available indicating that not just obesity per se, but the
anatomical location of the body fat can affect health risk and health care costs.
Cornier, Tate, Grunwald, and Bessesen (2002) determined that abdominal (or
central) obesity, independent of smoking, ethnicity, age, and gender, was
associated with higher total health care expenditures, especially the costs of
inpatient care. Central adiposity has been strongly associated with
cardiovascular risk, diabetes, hypertension, and a host of other diseases and
metabolic abnormalities (Poirier and Despres 2003, Vanltallie and Lew 1993).
Thus, obesity and its associated diseases present a major burden to
individuals and to the health care system.
26


2.6 Cardiovascular Disease, Type 2 Diabetes,
and Metabolic Syndrome
Research has shown that risk factors for cardiovascular disease and Type 2
diabetes, such as hypertension, insulin resistance, and abnormal blood lipids,
are influenced by body weight. Data from 35 to 75 year old participants in the
Framingham Heart study have shown that being overweight contributes to
hypertension and increased cardiovascular disease risk (Wilson, DAgostino,
Sullivan, Pariase, and Kannel 2002, Levy 1986). Further, an evaluation of
data from the Bogalusa Heart Study (Myers, Coughlin, Webber, Srinivasan,
and Berenson 1995) indicated that adults who had clustering of cardiovascular
disease risk factor variables were found to be not only more obese, but they
also had higher levels of total triglycerides, total, low-density lipoprotein and
very low-density lipoprotein cholesterol, systolic and diastolic blood pressure,
and increased circulating insulin levels. This shows an association between
body weight and what is termed "Metabolic Syndrome"; a syndrome that has
an estimated prevalence rate of 25-35% in western countries (Eriksson,
Taimela, and Koivisto 1997).
Metabolic Syndrome is characterized by a constellation of risk factors that are
associated with the development of cardiovascular disease (Vega 2002) and
Type 2 diabetes (Pi-Sunyer 2002). Metabolic Syndrome consists of glucose
27


intolerance, hyperinsulinemia, insulin resistance, high blood pressure, and
dyslipidemia, existing in an individual concurrently (Reaven 1988, Reaven
1993). Although various symptoms of the Metabolic Syndrome can be found
in non-obese individuals, evidence by Bjorntorp (1997) suggests that central
adiposity is often associated with this syndrome in adults. Thus, a greater
distribution of body fat in the central body region can present a greater health
risk than when a similar amount of body fat is distributed in the peripheral
regions (Rosenbaum, Leibel, and Hirsch 1997). Nevertheless, increased
central adiposity most commonly occurs concurrently with increased total
adiposity, therefore the identification of factors promoting overall obesity
remain important.
2.7 Ethnicity, Obesity, and Disease Risk
The association between increased body weight and chronic disease is even
more evident when considered within different ethnic groups. Winkleby,
Robinson, Sundquist, and Kraemer (1999) reported analyses of NHANES III
data on ethnic differences in cardiovascular disease risk in children and young
adults. Other than smoking, all risk indicators (BMI, dietary fat intake, blood
pressure and glycosylated hemoglobin) were higher in non-whites than in
whites. These differences were significant even after adjustment for
socioeconomic status and age, magnifying the already evident increased risk
28


of premature death due to obesity related diseases, particularly for non-white
groups. Combining datasets from the CARDIA and ARIC (Atherosclerosis
Risk in Communities) studies, Folsom et al. (1991) found an increased
incidence of hypertension and hypercholesterolemia with increasing adiposity
in younger (18-30 years) and older (45-65 years) African American males and
females, and an increased incidence of diabetes mellitus with increasing
adiposity in African American women.
Lipton, Liao, Cao, Cooper, and McGee (1993) identified an increased
prevalence of Type 2 diabetes with increasing adiposity in both African
American and Caucasian adults. Using subscapular and triceps skinfold data
from NHANES 1, the incidence of Type 2 diabetes was substantially higher at
the same level of adiposity in African American women as compared to
Caucasian women, whereas the incidence in African American men was only
slightly higher than in Caucasian men. Using data from NHANES II, Cowie et
al. (1993) found a 60% higher prevalence of type 2 diabetes in African
Americans compared to Caucasians, with the highest prevalence in African
American women (9.9%), followed by African American men (8.1%),
Caucasian women (7.1%), and Caucasian men (5.3%). Even after the authors
used logistic regression to adjust for type 2 diabetes risk factors (age, socio-
economic status, family history of diabetes, body weight, and adiposity) the
29


odds of developing diabetes were 70% higher for overweight African
Americans than for overweight Caucasians.
Hall et al. (2003), as part of a report from the African-American Lipid and
Cardiovascular Council, highlight the substantial occurrence of elevated blood
pressure, obesity and diabetes in African Americans in the United States.
They also report that African Americans have the highest coronary heart
disease mortality of any ethnic group within the United States.
The results of the studies cited above indicate that elevated body weight due
to increased body fat, increases the potential risk for cardiovascular disease,
as well as Type 2 diabetes in adults, particularly if the fat is predominantly
centrally located. These studies also illustrate the disproportionate occurrence
of obesity-related disease in African Americans compared to Caucasians. As
a result of their economic costs, these diseases are not only a problem for
individuals, but for society as a whole.
30


2.8 Metabolic Factors Associated with Body
Weight Regulation
Obesity occurs when there is a chronic excess of energy intake relative to
energy expenditure. There are many complicated mechanisms that can
influence the components of energy intake and expenditure, and thus body
weight regulation. Factors reported to be predictors of body weight gain
include: a low resting metabolic rate, a low rate of fat oxidation, and a low
physical activity level (Ravussin, et al. 1998, Schutz 1995, Zurlo, et al. 1992).
The focus of this thesis is on resting metabolic rate, and substrate oxidation.
2.9 Energy Expenditure
There are three major components of daily energy expenditure:
1) Resting metabolic rate is the energy expended in maintaining
normal bodily functions and homeostasis. Resting metabolic
rate is considered the sum of sleeping metabolic rate and the
energy cost of arousal, constituting approximately 50 80% of
daily energy expenditure in moderately active or sedentary
individuals.
2) The thermic effect of food (also called dietary-induced
thermogenesis) is the energy required to digest, metabolize, and
31


store consumed food, and represents approximately 7 -10% of
daily expenditure for the typical US diet. This can vary, however,
depending on individual meal size and nutrient composition.
3) The energy expended in physical activity includes the sum of
spontaneous and purposeful physical activity and is the most
variable component, representing approximately 20 40% of
daily energy expenditure for the majority of individuals.
Major determinants have been identified for each component of daily energy
expenditure (Ravussin, Fontvieille, Swinburn, and Bogardus 1993). Body
composition, age, gender, genetics, hormones (such as free T3) and
sympathetic nervous system activity are thought to determine resting
metabolic rate (Ravussin 1995). By far, however, the major determinant of
resting metabolic rate is fat-free mass (Zurlo, Larson, Bogardus, and Ravussin
1990, Elia 1992). Adipose tissue is relatively metabolically inactive and so
contributes minimally to the resting metabolic rate in lean individuals but can
increase in its importance in obese individuals (as the relative fat mass
increases) (Felber and Golay 1995). With respect to the thermic effect of food,
the amount and composition of food consumed during an eating episode, as
well as the hormonal response (insulin) and in particular the sympathetic
32


nervous system response, can influence the extent of the post-prandial rise in
metabolic rate. The physical activity component of daily energy expenditure is
mainly dependent on the quantity of activity performed (intensity and duration).
What determines the amount of activity a person undertakes has not only
genetic and physiological determinants (body weight and the sympathetic
nervous system), but also behavioral components (spontaneous physical
activity, purposeful physical activity).
Although all of the major components and determinants of daily energy
expenditure are important, this project will focus on the associations of body
composition, gender, and ethnicity with resting metabolic rate. Resting
metabolic rate represents the largest percentage of daily energy
expenditure, and thus variations in resting metabolic rate could potentially
have a significant impact on long-term body weight and adiposity changes
compared to other components of daily energy expenditure. It is
recognized, however, that the energy expended in physical activity can also
be an important contributor to daily energy expenditure but this was not a
parameter on which data were collected for the current dataset used for this
thesis. This will be considered when addressing the limitations of the data
set in the discussion.
33


2.10 Relationship of Gender, Ethnicity, Body Weight
and Body Composition with Resting Metabolic
Rate
2.10.1 Gender Differences in Energy Expenditure
Previous investigations have reported that there are differences in energy
expenditure by gender. Using a whole-room calorimeter, Ferraro, et al.
(1992) found 24-hour energy expenditure (adjusted for body composition,
age and physical activity) higher in male versus female adults by
approximately 100 kcal/day. Although non-significant, both resting
metabolic rate and sleeping metabolic rate tended to be higher in male
subjects. Arciero, Goran, and Poehlman (1993) had similar, though
significant, findings using indirect calorimetry (via ventilated hood) studies to
measure resting metabolic rate in 522 adults (328 males, 194 females).
Unadjusted resting metabolic rate was 23% higher in the men. After
adjustment for fat-free mass, fat mass, and aerobic fitness, resting metabolic
rate was still 3% lower in female versus male subjects (p<0.01).
The hormonal status, in terms of the sex-steroids, during different phases of
the menstrual cycle can impact the measured resting metabolic rate values
of female subjects, therefore, this should be considered if gender
comparisons in resting metabolic rate are a primary target of an
investigation. The studies mentioned above (Ferraro, et al. 1992, Arciero, et
34


al. 1993) accounted for the menstrual cycle phase and menopausal status of
subjects in their studies by relying on self-reports. The majority of studies
show that women studied in the follicular phase of their menstrual cycle
have a lower metabolic rate compared to when they are studied in the luteal
phase, whereas post-menopausal women have a lower resting metabolic
rate compared to pre-menopausal women (Poehlman 2002). In the present
sample, the maximum age at the time of the follow-up exam (Year Five) was
about 45 years, so most of the female participants had not yet gone through
natural menopause, although there were some with surgical menopause.
We expect that we had a random distribution of menstrual cycle phase at
the time of the baseline and follow-up exams so that any menstrual cycle
phase effects on resting metabolic rate would be predominantly
representative of the average value across the follicular and luteal phases of
the menstrual cycle. Additionally, menstrual cycle phase and menopausal
status were not included in the dataset used for the current project, so any
analyses regarding the effects of menstrual cycle phase and menopausal
status on resting metabolic rate cannot be addressed in the current study.
Gender differences per se in resting energy expenditure will not form part of
the data analysis for this thesis. Gender will be considered, however, in the
context of ethnicity to establish whether or not differences in resting
35


metabolic rate contribute to any gender by ethnic differences in body weight
gain, and/or body fat increases, over time.
2.10.2 Ethnic Differences in Energy Expenditure
Several reports of a low resting metabolic rate in African American women
have raised the question of whether a low rate of energy expenditure is a
cause of the high prevalence of obesity in African American versus
Caucasian women. In general, African American women have been
observed to have a 6-11 % lower resting metabolic rate, when resting
metabolic rate is adjusted for body weight and fat-free mass, compared to
Caucasian women (Foster, Wadden, and Vogt 1997, Jakicic and Wing 1998,
Kushner, Racette, Neil, and Schoeller 1995, Albu, et al. 1997, Carpenter, et
al. 1998, Sharp, et al. 2002).
Nicklas, et al. (1999) addressed racial differences in metabolic predictors of
obesity, and concluded that there are differences between African American
and Caucasian postmenopausal obese women with respect to the predictive
capacity of resting metabolic rate, fat oxidation, and maximal oxygen
consumption (VC^max) on obesity. All three variables were higher in
Caucasian women, thus potentially predisposing the African American
36


women to gain more weight after menopause than their Caucasian
counterparts.
Data comparing resting metabolic rate between African American and
Caucasian men are more limited. A lower adjusted resting metabolic rate
(Carpenter, et al. 1998) and sleeping metabolic rate (Weyer, Snitker,
Bogardus, and Ravussin 1999) have been reported in African American men
compared to Caucasian men.
All of these previous studies that have compared ethnic differences in
metabolic rate do suggest that differences in metabolic rate could be a
contributory factor to ethnic differences in the etiology of obesity. However,
data from these studies have all been cross-sectional in nature.
Longitudinal data would allow us to confirm whether or not there is an
association between metabolic rate and body weight changes, and if there
are ethnic differences in these associations. Longitudinal data are not
presently available in African Americans compared to Caucasians to
address the relationship of change in metabolic rate to change in body
weight and adiposity over a prolonged time period.
37


2.10.3 Ethnic and Gender Differences in Resting
Metabolic Rate
In a more recent study, Sharp and colleagues (2002) assessed resting
metabolic rate in 395 adults aged 28-40 years (100 African American men,
95 Caucasian men, 94 African American women and 106 Caucasian
women), recruited from participants in the CARDIA Birmingham, Alabama
and Oakland, California field centers. Body composition was measured
using Dual Energy X-ray Absorptiometry (DEXA) and resting metabolic rate
by indirect calorimetry.
The results of this study confirmed prior reports that African American
women have a lower absolute resting metabolic rate than Caucasian women
(Foster, et al. 1997, Jakicic and Wing 1998, Kushner, et al. 1995, Albu, et al.
1997). In addition, results indicate that African American men also have a
lower resting metabolic rate than Caucasian men, and that the magnitude of
the difference between ethnic groups was similar for men and women.
When adjusted for commonly accepted determinants of metabolic rate (fat-
free mass, fat mass, visceral fat, and age), resting metabolic rate remained
significantly lower in African Americans (1585 11 kcal/day) than in
Caucasians (1665 11 kcal/day) by an average of 80 16 kcal/day
(p<0.0001). The results of this cross-sectional study are consistent with the
38


notion that a lower resting metabolic rate could contribute to the higher
prevalence of obesity in African American versus Caucasian women.
However, the observation that African American men in this study also have
a lower resting metabolic rate than Caucasian men but had similar amounts
of body fat, suggests that a lower resting metabolic rate may not be a major
determinant of obesity in men. Thus, this study suggests that there may be
a gender difference in the causes of obesity development.
Data suggest that a low resting metabolic rate may predispose subjects to
obesity, which results predominantly in an increase in fat mass, but may
also result in small increases in fat-free mass. It has been hypothesized that
this accumulation of total body mass may eventually normalize resting
metabolic rate and produce weight stability (Ravussin, et al. 1993). Sharp
and colleagues (2002) observed that the difference in resting metabolic rate
between African American and Caucasian subjects was similar across a
wide range of body weights. That is, African American subjects had a lower
resting metabolic rate compared to Caucasian subjects whether or not they
were lean or obese. Body fat mass contributed similarly to resting metabolic
rate in both ethnic groups. Thus, there was no evidence that becoming
obese in African American subjects normalized their resting metabolic rate.
Nevertheless, this could reflect a lack of weight stability in this sample of
39


African Americans. Increased body weight is a result of increased fat mass
as well as small increases in fat-free mass. This suggests that a higher
energy expenditure with higher amounts of fat-free mass would be similar in
both groups, a situation consistent with the results of other studies (Sharp,
Reed, Sun, Abumrad, and Hill 1992). However, as mentioned above, even
when resting metabolic rate was adjusted for commonly accepted
determinants of energy expenditure, resting metabolic rate was still lower in
the African American subjects as compared to the Caucasian subjects.
Some studies (Schutte, et al. 1984, Ortiz, et al. 1992) have noted
differences in bone mass or density due to gender or ethnicity, and such
differences were also noted in the baseline VIM study (Sharp, et al. 2002).
The metabolic activity of bone, however, contributes minimally to energy
expenditure at rest. Nevertheless, Sharp and colleagues (2002) conducted
parallel analyses with and without bone mass as a component of the non-fat
tissue (ie., fat-free mass vs. lean mass, respectively) and found essentially
the same variability in resting metabolic rate between the ethnic and gender
groups. This indicates that differences in bone mass do not contribute to the
differences seen in resting metabolic rate between African American and
Caucasian subjects in the VIM baseline study (Sharp, et al. 2002).
40


2.10.4 Body Weight, Body Composition and Energy
Expenditure
Assessment of body composition is an important component of energy
balance research. Fat-free mass is a major determinant of energy
expenditure, and accounts for the greatest source of variation in resting
metabolic rate (Poehlman, et al. 1992, Ravussin, Lillioja, Anderson, Christin,
and Bogardus 1986). Although not as important, fat mass and body weight
have also been shown to be contributory factors to metabolic rate (Bray,
Schwartz, Rozin, and Lister 1970, Ravussin, Zurlo, Ferraro, and Bogardus
1990). In the context of the present study, fat mass could play an important
part in assessing differences in resting metabolic rate adjusted for body
composition parameters if African American subjects have more fat mass
than Caucasian subjects.
2.11 Impact of Respiratory Quotient on Obesity
Development, and Factors Affecting
Respiratory Quotient
Ravussin and Gautier (1999) identified metabolic predictors for weight gain. In
addition to a low metabolic rate and a low physical activity level, low rates of
fat oxidation, insulin insensitivity, and low concentrations of plasma leptin were
additional determinants of obesity.
41


Respiratory quotient, defined as the rate of carbon dioxide production divided
by the rate of oxygen consumption, has been widely used as an indicator of
whole-body fuel oxidation. A respiratory quotient of 0.70 indicates primarily
fat oxidation whereas a respiratory quotient of 1.0 indicates primarily
carbohydrate oxidation. Individuals with low rates of fat oxidation (ie., a higher
respiratory quotient) may be predisposed to weight gain due to their
preferential use of carbohydrate as a fuel source. This preference for
carbohydrate oxidation, may result in higher levels of fat storage, and
therefore, weight gain if fat intake exceeds fat oxidation, a situation
predominantly associated with positive energy balance. Data to support this
concept was initially described by Zurlo, et al. (1990), who assessed 24-hour
respiratory quotient in a group of 111 Pima Indians, and found a significant
positive correlation between elevated baseline respiratory quotient and
subsequent weight gain over a two-year period. Weyer and colleagues (1999)
also found that African American men had higher 24-hour respiratory quotient
values than Caucasian men, indicating lower 24-hour fat oxidation rates, which
could predispose the African Americans to greater increases in body weight
and adiposity.
42


2.11.1 Factors Affecting Resting Metabolic Rate and
Respiratory Quotient
A non-fasting (post-prandial) state affects the accuracy of resting metabolic
rate assessments due to the "interference" of diet-induced thermogenesis
(thermic effect of food). Likewise, a non-fasting state would also result in
erroneously high respiratory quotient measures if the test is meant to assess
fasting respiratory quotient. This is because ingested carbohydrate is a
readily available fuel source that is preferentially oxidized by the body
(Berne and Levy, 1988). Defining the conditions under which respiratory
quotient is measured is thus important for appropriate interpretation of the
data. Besides protocol-related factors, Ravussin and Gautier (1999) report
the following influences on respiratory quotient:
1. Energy balance: More recent negative energy balance, or
long-term negative energy balance like that seen in weight
loss, could result in lower respiratory quotient values
indicating greater fat oxidation.
2. Adiposity: A higher fat mass is associated with a higher fat
oxidation.
43


3. Possible genetic determinants: Some research has
indicated a possible familial correlation for respiratory
quotient. However, much of this research is related to
studies in Pima Indians, and may not be generalizable to
other populations.
In the current dataset, all measures were made in the overnight fasted state
to avoid the variable impact of time and amount of food eaten on resting
metabolic rate and respiratory quotient. Participants also abstained from
exercise during the preceding 24 hours. State of energy balance, however,
could not be controlled for and this will be discussed in the limitations
section. Adiposity, however, was a variable that was measured and is
addressed in the Current analyses. Finally, genetics cannot be accounted
for other than in the analyses of gender and ethnicity.
2.12 Summary of Literature Review
The aforementioned studies provide a summary of the literature addressing
the issue of the ethnic and gender differences seen in the prevalence of
obesity and obesity-related diseases. Information was also provided that
corresponds to the suggestion that differences in energy and substrate
balance and body composition could be contributors to the current problem
44


of obesity in the United States. The information provided in the literature
supports the justification for the analyses suggested in this thesis. These
analyses are intended to assess the relationship of longitudinal measures of
resting metabolic rate, respiratory quotient, body weight, and adiposity, and
to determine if there are ethnic and gender differences in these results that
might be suggestive of reasons for the ethnic and gender differences seen
in the development of obesity and subsequent disease risk. The following
chapter, Participants and Methodology, provides information on the parent
study (CARDIA), the ancillary study (VIM), a description of the protocol
participants, and a description of the data collection and statistical analyses
methodology used to address the hypotheses presented in Chapter 1.
45


3.
Participants and Methodology
3.1 CARDIA Background and Methods
The Coronary Artery Risk Development in Young Adults Study (CARDIA)
began in 1984 as a multi-center trial investigating the ethnic and gender
differences associated with lifestyle and other risk factors for the
development of coronary heart disease. African American and Caucasian
men and women (n=5116) between the ages of 18 to 30 years were
recruited via contact letters and telephone calls from four urban areas for
participation in this longitudinal trial (Oakland, CA, Birmingham, AL,
Chicago, IL, and Minneapolis, MN). A major recruitment focus was the
desirability of sampling approximately equal numbers by age, ethnic group,
gender, and education level as compared with sampling numbers
representative of the population base within these urban centers. The
baseline CARDIA examination included: informed consent acquisition,
blood pressure measures, sociodemographic questionnaire, phlebotomy,
medical and psychosocial questionnaires, Type A/B personality interview,
pulmonary function, diet interview assessing dietary intake during the
previous month, anthropometry (height, weight, skinfold measures,
46


circumference measures, elbow breadth), and a symptom-limited treadmill
test.
Data analysis of baseline information indicated that African Americans had
higher mean systolic and diastolic blood pressures than Caucasians, and
that males had higher mean systolic and diastolic blood pressures than
women. It was also determined that African Americans had higher mean
total cholesterol than Caucasians, but cholesterol did not differ by gender.
Body weights were significantly higher in African American compared to
Caucasian participants, and in male compared to female participants. Mean
body mass index (BMI) was significantly higher in African American and
older participants, but did not differ by gender or educational level.
Additional information regarding CARDIA recruitment, and baseline analysis
and results has been previously published (Friedman, et al. 1988, Hughes,
et al. 1987, Cutter, etal. 1991).
47


3.2 Background on VIM
The Visceral Fat, Metabolic Rate, and CHD Risk in Young Adults (VIM)
Study, was initiated in 1995, during the time of the 10-year follow-up of the
CARDIA exams, to increase our understanding of the ethnic and gender
differences in adiposity, and subsequent risk for coronary heart disease.
The primary objectives of the VIM study were to: a) determine the ethnic and
gender differences in visceral adipose tissue, and determine the relationship
of visceral adipose tissue with coronary heart disease risk in the ethnic and
gender groups, and b) assess ethnic and gender differences in resting
metabolic rate, and to relate potential differences to the differences seen in
adiposity within the ethnic and gender groups. The VIM protocol was
approved by the Institutional Review Boards of the University of Alabama-
Birmingham and the Kaiser Permanente Division of Research in Oakland,
CA.
The five-year follow-up of the VIM study (CARDIA year 15 exam) has
recently been completed. In addition to the initial objectives stated above,
the follow-up data allow us to address the following objectives: a) to
determine ethnic and gender relationships in changes in visceral and total
adiposity over a five-year period, and relate these to coronary heart disease
risk, and b) to determine ethnic and gender relationships between change in
48


resting metabolic rate, respiratory quotient, body weight, and adiposity over
a five-year period. This latter objective is the focus of the data analyses for
this thesis.
3.3 VIM Study Methods
Two of the original four CARDIA field centers were used as sites for the VIM
study; Kaiser Permanente Medical Care Program in Oakland, CA, and the
University of Alabama-Birmingham, Birmingham, AL. Participation in the
VIM study involved two additional measurements; the determination of
resting metabolic rate (via indirect calorimetry) and the assessment of
visceral adipose tissue (via computed tomography (CT) scan). Baseline
assessments for VIM were conducted in 1995-96, and the five-year follow-
up period (Year Five) occurred in 2000-01. These correspond to the year 10
and year 15 CARDIA examination periods. The same protocols and
equipment were used for both the VIM baseline and VIM Year Five follow-up
exams, and are described in detail below.
49


3.4 Participants
Participants for this project included African American and Caucasian adults
who participated in both the Baseline and Year Five follow-up of the VIM
study (1995-96, 2000-01) (Sidney, et al. 1999). There were 395 participants
in the baseline VIM assessment period (100 African American males, 95
Caucasian males, 94 African American females, and 106 Caucasian
females). The intention was to recruit equal numbers from each ethnic and
gender group. BMI was distributed similarly (both above and below the
median ethnic- and gender-specific BMI values from the previous (Year
Seven) exam in each group (Sidney, et al. 1999). However, over the five-
year period, there was some attrition. Therefore, participants are excluded
from the current analyses if they did not participate in both the Baseline and
Year Five follow-up VIM exams. Of the initial 327 individuals who
participated in both VIM exam periods, three participants are excluded from
the analysis because of abnormal values for thyroid stimulating hormone at
their baseline VIM assessment, six participants are excluded because of
missing DEXA data from the five-year follow-up exam, and 21 are excluded
because of missing or abnormal (due to equipment malfunction) calorimetry
data. Therefore, 297 individuals who have complete assessment data from
both exam periods are included in this thesis project; 72 African American
males, 79 Caucasian males, 72 African American females, and 74
50


Caucasian females. This sample size provides adequate power (.80) to
detect expected differences at a significance level of 0.05 for the outcome
variables of interest.
3.5 Resting Metabolic Rate
Resting metabolic rate was measured by indirect calorimetry using a
Sensormedics 2900 Metabolic Cart (Yorba Linda, CA). Indirect calorimetry
assesses whole-body oxygen consumption and carbon dioxide production
by monitoring the flow of reference air through an open circuit system, and
minute-by-minute measures of oxygen and carbon dioxide concentrations of
the expired air. Oxygen consumption and carbon dioxide production can
then be calculated from the ventilation rate and the differentials in gas
concentrations between the reference air and the expired air using the
calculations of Weir (1949). Energy expenditure, or metabolic rate, was
then calculated from rates of oxygen consumption and carbon dioxide
production (McLean and Tobin 1987, Jequier, Acheson and Schutz 1987).
The respiratory quotient was calculated as the ratio of carbon dioxide
production to oxygen consumption. It should be noted that no correction
was made to the respiratory quotient value for the contribution of protein to
whole-body substrate oxidation. Measurements of 24-hour urinary nitrogen
are needed to calculate whole-body protein oxidation, but were not acquired
51


as part of the VIM study protocol. The potential impact of not accounting for
protein oxidation differences related to the ethnic and gender groups, with
respect to the interpretation of respiratory quotient, is discussed in the
section dealing with the limitations of the current dataset.
Both testing locations had identical systems, and all technicians were
trained by the same individual (T. Sharp). The same testing protocol was
used at the baseline and Year 5 VIM exams. Gas analyzer and flowmeter
calibrations were performed prior to each subject measurement.
Participants were instructed to report to the laboratory in the morning after
an overnight fast and to refrain from exercise for 24 hours prior to the test.
Smokers refrained from smoking the morning of the test. Participants rested
in a supine position for 30 minutes, after which a plexiglass hood was placed
over the head. Oxygen consumption and carbon dioxide production were
measured for 15 continuous minutes as long as the expired fraction of CO2
and O2 (%EC02 and %E02, respectively) were within 0.1 % of the preceding
minute-by-minute %EC02 and %E02 values acquired during the test. The
respiratory quotient values had to be within 0.1 of the other minute-by-
minute respiratory quotient values acquired during the test. These cutoff
values were used to indicate that the participant had achieved a steady-
state level of gas exchange for use in the calculation of resting metabolic
52


rate and resting respiratory quotient. Any minutes that contained data that
did not meet these criteria were excluded from the test, and the gas
exchange measurements were continued until 15 consecutive minutes of
appropriate data were collected. Data from all resting metabolic rate test
reports were evaluated by one investigator (T. Sharp).
3.6 Body Weight, Height, and Body Mass Index
(BMI)
Standing height in centimeters (cm) was measured without shoes using a
stadiometer (Perspective Enterprises, Inc., Portage, Ml). Body weight in
kilograms (kg) was measured after voiding and with minimal clothing using a
Detecto balance beam scale (Cardinal Scale Mfr. Co., Webb City, MO).
Body mass index (BMI) is a measure of body weight in relation to height,
and is calculated as weight in kilograms divided by the square of height in
meters (Khosla and Lowe 1967). A BMI of 19 to 24.9 kg/m2 is considered
normal, from 25 to 29.9 kg/m2 is considered overweight, and a BMI over 30
kg/m2 is considered obese (NIH Clinical Guidelines 1998). Individuals with
BMI above 25 kg/m2 are assumed to be at risk for the development of
cardiovascular co-morbidities (Poirier and Eckel 2000), so this value is used
in the present study to reflect overweight status, and below 25 kg/m2 is
used to reflect non-overweight status.
53


3.7 Body Composition
Body composition was measured using Dual Energy X-ray Absorptiometry
(DEXA) in the enhanced total-body-array scanning mode with a Hologic
Model QDR-2000 scanner (Waltham, MA). Daily phantom scanning was
performed to check the accuracy of the detector throughout the course of
the study. Dual energy x-ray for use in measuring body composition and
bone density in human subjects is based on the measurement of the
attenuation of an x-ray beam of two distinct energy levels (70 and 140
kiloVolts). Each of the three major components of the body (bone mass, fat
mass, fat-free mass) illustrates different attenuation properties. The system
notes the different attenuations of the x-ray beam, and calculates the
amount of each of these masses using specialized software. The
individuals body weight is entered at the beginning of the scan, and the
system uses this weight as the basis for calculating kilogram weights of
each of the components of the individuals body.
The motor of the scanner operates a pair of source and receiving electrodes
in a series of transverse scans that pass across the subjects body in one-
centimeter intervals. These scans move in the direction from the head to
toes of the individual. Previous trials have indicated that DEXA provides a
precise method of assessing body composition and bone mineral content
54


with low exposure to radiation (Nord and Payne 1995, Heymsfield, Smith,
and Aulet 1990, Mazess, Barden, Bisek, and Hanson 1990).
Earlier studies confirm that there is a greater amount of bone density in
African American compared to Caucasian subjects. Trotter, Brosnan, and
Peterson (1960) reported 8-12% higher bone density in Black male
skeletons compared to White male skeletons. Black women had 5-8%
higher bone mineral content than age- and body size-matched White women
(Cohn, et al. 1977).
Data also suggest that there is a greater amount of lean mass in African
American compared to Caucasians (Schutte, et al. 1984, Ortiz, et al. 1992,
Cote and Adams 1993). Methods such as underwater weighing and whole-
body potassium counting that were derived from studies of white subjects to
assess percent body fat, underestimate body fat in black subjects (Schutte ,
et al. 1984, Ortiz, et al. 1992). This underestimation is related to the inability
of these methods to differentiate the bone mass and fat-free mass
components of lean tissue. Errors such as these can have both clinical and
epidemiological significance. The ability of the DEXA method to differentiate
between bone mass, fat mass, and fat-free mass makes it the preferred
method for body composition assessment in this study.
55


3.8 Statistics
Statistical analysis was performed using SPSS Version 11.0 statistical
software. Graphical analysis was used to examine the data for outliers and
for non-normality. Gender and ethnic interactions related to changes in
resting metabolic rate, respiratory quotient, body weight, and adiposity over
the five-year period are of primary interest in this study. Therefore, all
analyses, results, and discussion will reflect the total sample, the two ethnic
groups (African American and Caucasian), and the four ethnic and gender
groups (African American men, Caucasian men, African American women,
and Caucasian women).
3.8.1 Methods of Data Expression
Due to the body size-dependent nature of daily energy expenditure, it has
been recommended that energy expenditure be adjusted for components of
body habitus (Toth 2002). Body weight has been used for the adjustment of
energy expenditure data, but fat-free mass is considered by many to be the
primary determinant of energy expenditure in humans (Weinsier, Schutz and
Bracco 1992, Zurlo, et al. 1990, DeLany and Lovejoy 1996, Ravussin, et al.
1986). In the present study, both absolute resting metabolic rate (in
kcal/day), and resting metabolic rate adjusted for body weight and fat-free
mass are used in the analyses. The reasoning for including both absolute
56


and adjusted values is in an attempt to allow for more generalizability of the
findings for populations similar to those used here. In some clinical and
epidemiologic trials, access to techniques such as dual-energy X-ray
absorptiometry (DEXA) for body composition assessment may not be
available, and therefore, adjustment for fat-free mass may be more difficult
or less accurate if field-based methods are employed. It is therefore,
worthwhile to include analyses using absolute measures of resting metabolic
rate to provide information for investigations encompassing large field-based
datasets. It is also important to include measures of resting metabolic rate
adjusted for body habitus to enable better comparison of participants and
groups of differing body size and/or body composition.
3.8.2 Descriptive Data
Descriptive statistics of the total sample and of each ethnic and gender
group for resting metabolic rate, respiratory quotient, body weight, BMI, fat-
free mass, fat mass, and percent body fat at each time point, and change
from baseline are expressed as mean standard deviation.
3.8.3 Analyses of Baseline and Year Five Data
Analysis of variance was used to study subject characteristics at each time
point (Baseline and Year Five follow-up) and resting metabolic rate
unadjusted for body weight, body composition, and age. Scheffe test of
57


multiple comparisons was used to compare mean Baseline and Year Five
values of adjusted resting metabolic rate (adjusted for fat-free mass and
body weight) between ethnic and gender groups to determine if associations
seen in the total sample were being driven by one of the ethnic and gender
groups. When Scheffe tests are used, data are presented using mean and
standard error of the differences between the means of each group. T-test
of independent samples was used to compare differences between ethnic
and gender groups for Baseline and Year Five data.
3.8.4 Analyses of Change Data
The evaluation of change over time in targeted variables was conducted
using a one-sample t-test with the change values. Bivariate correlations
were used to test the relationship of change in resting metabolic rate with
change in fat mass and body weight, and to test the relationship of change
in respiratory quotient with change in body weight and fat mass.
3.8.5 Ability of Data to Predict Five-year Change
Pearson bivariate correlations were used to assess the relationships
between baseline variables (resting metabolic rate, body weight, fat-free
mass, fat mass, and age) and five-year change in body weight, fat mass and
percent body fat. T-tests were used to assess the strength of association of
baseline age with five-year change in resting metabolic rate, respiratory
58


quotient, body weight, fat mass, and fat-free mass. Analysis of variance
with comparison of means was used to compare baseline weight status
(overweight and non-overweight) with five-year change in fat mass and body
weight for the total sample and for each ethnic and gender group. A
stepwise multiple regression procedure was used to predict five-year
change outcomes (for resting metabolic rate, body weight, and fat mass)
from baseline measures of resting metabolic rate, respiratory quotient, body
weight, fat mass, fat-free mass, percent body fat, and age) for the total
sample, for each ethnic group, and for each ethnic and gender group.
59


4.
Results
The following results reflect analyses intended to address each of the five
hypotheses presented in Chapter 1. The results related to each of the
hypotheses will be presented separately. Descriptive data for baseline and
five-year time-points will be presented first, followed by results for each
hypothesis. Discussion of change in variables from baseline, predictive
ability of baseline measures for five-year results, and associations of change
in variables will be addressed with the appropriate hypothesis.
4.1 . Descriptive Data for VIM Baseline and VIM
Year Five Time Points
The descriptive data presented are compared cross-sectionally at each time
point between each ethnic group and between each ethnic and gender
group. Changes in each variable will be presented in more detail later in this
section. Participant characteristics at the two study time points (Baseline
VIM and VIM Year Five) for resting metabolic rate, respiratory quotient, body
weight measured in kilograms, body mass index calculated as weight in
kilograms divided by height in meters squared, fat-free mass measured in
kilograms, fat mass measured in kilograms, and percent body fat are
presented in Tables 4.1 and 4.2. There was no significant difference in the
60


mean age between any of the four ethnic and gender groups at the time of
the VIM baseline exam, which was intended in the study design.
61


Table 4.1. Participant Characteristics: Baseline Age, and Baseline
and Year Five Body Weight (Wt kg), Body Mass Index (BMI), and
Body Composition (FFM kg, FM kg, and % fat) Mean (+ SD)
All Subjects All AA All C AAM CM AAW CW
n 297 144 153 72 79 72 74
Baseline 35.2 34.6 35.8 34.2 35.2 34.9 36.5
Age (3-5) (3.6) (3.4) (3.5) (3.4) (3.7) (3.2)
Baseline 79.6 83.4 76.0 85.3 84.4 81.6 67.0
Wt kg (15.6) (15.1)*** (15.3) (14.2) (12.0) (15.8)*** (13.3)
Year 5 83.5 87.9 79.4 89.7 87.2 86.1 70.9
Wtkg (17.0) (16.9)*** (16.0) (16.4) (13.8) (17.4)*** (13.8)
Baseline 26.9 28.5 25.3 26.8 26.2 30.2 24.3
BMI (5.1) (5.3)*** (4.3) (4.2) (3.4) (5.8)*** (4.9)
Year 5 28.4 30.1 26.8 28.3 27.7 32.0 25.9
BMI (5.7) (5.9)*** (5.0) (4.9) (4.5) (6.4)*** (5.4)
Baseline 54.1 56.1 52.2 65.8 62.0 46.4 41.8
FFM kg (12.3) (12.2)** (12.0) (8.2)** (7.5) (6.5)*** (5.2)
Year 5 54.4 56.6 52.3 66.3 62.0 46.9 41.8
FFM kg (12.5) (12.5)** (12.1) (8.6)** (7.9) (7.1)*** (4.9)
Baseline 24.2 26.0 22.6 18.1 21.2 34.0 24.1
FM kg (11.1) (12.5)** (9.4) (8.3)* (6.8) (11.0)*** (11.5)
Year 5 27.8 30.2 25.6 22.2 23.4 38.1 27.9
FM kg (12.7) (14.2)** (10.7) (10.3) (8.6) (13.1)*** (12.2)
62


Table 4.1 (Cont.)
All Subjects All AA All C AAM CM AAW CW
Baseline 30.4 31.0 29.9 20.8 25.1 41.4 35.1
% fat (11.0) (12.6) (9.3) (7.2)*** (5.7) (7.3)*** (9.6)
Year 5 33.2 33.9 32.5 24.2 26.9 43.6 38.4
% fat (11.4) (12.7) (10.0) (7.8)* (6-3) (8.7)** (9.9)
Significant difference by ethnicity: p<0.05, **p<0.01, ***p<0.001
All AA=all African American subjects, All C=all Caucasian subjects, AAM=African American
men, CM=Caucasian men, AAW=African American women, CW=Caucasian women, Wt
kg=body weight in kilograms at Baseline VIM, and VIM year 5,
BMI=Body Mass Index (wt kg/ht m2 ) at Baseline VIM and VIM year 5,
FFM kg= fat-free mass in kilograms by DEXA at Baseline VIM, and VIM year 5,
FM kg=fat mass in kilograms by DEXA at Baseline VIM, and VIM year 5,
% fat=percent body fat by DEXA at Baseline VIM, and VIM year 5.
4.1.1 BMI and Weight
Body weight alone does not indicate whether or not an individual is
overweight. Body mass index (BMI, wt kg/ht in m2) is a measure of weight in
relation to height, and is typically used as an indicator of weight status in the
general population. Baseline and Year Five BMI and body weight data are
presented in Table 4.1.
At the time of the baseline VIM exam, 58.6% of the total sample was
overweight based upon a BMI greater than 25 kg/m2, which is considered
overweight according to CDC (2001) and NIH (1998) criteria. Five years
63


later, 69.9% of the total sample was overweight based upon the same
criteria, representing a significant increase in overweight status (p=0.035).
Males were heavier at both time points than females. African Americans
were significantly heavier (p<0.001) and had significantly greater BMI values
(p<0.001) at both time points than Caucasians. At baseline, 70.1% of the
African American subjects were overweight (62.5% males, 77.8% females),
and 47.7% of the Caucasian subjects (63.3% males, 31.1% females) were
overweight. Five years later, these percentages increased to 79.9% of the
African Americans (73.6% males, 86.1% females), and 60.8% of the
Caucasians (72.2% males, 48.6% females) being overweight. Although
overweight prevalence increased over the five-year period in both ethnic
groups, this change over time was not different between African Americans
and Caucasians (p=0.658).
Figure 4.1 shows the mean BMI values at Baseline and the Year Five time
points for the total sample. The mean BMI at Baseline was 26.85 ( 5.05)
and by Year Five, the mean BMI for the total sample had increased to 28.42
(5.74).
64


Figure 4.1. Mean ( sd) Body
Mass Index (BMI) for the total
sample at Baseline and Year 5.
*
Time Period
*= significant change from Baseline, p=0.035
Figure 4.2 shows the mean BMI values at each of the two time points for
each of the four ethnic and gender groups. The results of paired samples T-
tests indicate that the BMI of all four groups increased significantly (p<0.001)
over the five-year period. African American women had the highest mean
BMI values at both time points, and had the greatest mean change in BMI
(1.81 3.16). The mean BMI increased in African American men by 1.43
2.04 units, in Caucasian men by 1.46 3.46, and in Caucasian women by
1.48 2.29.
65


Figure 4.2. Mean BMI (+ sd) at
Baseline and Year 5 for each
ethnic and gender group.
Ethnic and Gender Groups
*=significant change from Baseline, p<0.0001
4.1.2 Body Composition
Baseline and Year Five body composition data are presented in Table 4.1.
Males had higher fat-free mass, but lesser fat mass and mean percent body
fat than females at both time points. African Americans as a group had
greater fat-free mass (p<0.01), and fat mass (p<0.01) at both time points
than Caucasians. African American men had more fat-free mass (p<0.01),
but less fat mass (p<0.05) and percent body fat (p<0.001 at baseline,
p<0.05 at Year Five), at both time points than Caucasian men. African
American women were heavier (p<0.001) and had more fat-free mass
66


(p<0.001) than Caucasian women. Unlike African American men, African
American women also had higher fat mass (p<0.001) and percent body fat
(p<0.001 at baseline, p<0.01 at Year Five) than Caucasian women.
4.1.3 Resting Metabolic Rate and Respiratory
Quotient
Baseline and Year Five resting metabolic rate and respiratory quotient data
are presented in Table 4.2. Males had greater mean resting metabolic rate
values at both time points than females. There was no difference in
respiratory quotient between genders at either time point.
Mean resting metabolic rate and mean respiratory quotient were
similar at both time points in African American and Caucasian
subjects. Although at both time points Caucasian men had slightly
higher mean resting metabolic rate values than African American
men, and African American women had slightly higher mean resting
metabolic rate values than Caucasian women, these differences in
absolute resting metabolic rate by ethnicity were not significantly
different.
67


Table 4.2. Participant Characteristics: Baseline and Year Five
Resting Metabolic Rate (RMR, kcal/day) and Respiratory Quotient
(RQ). Mean (+ SD)
All Subjects All AA All C AAM CM AAW CW
n 297 144 153 72 79 72 74
Baseline 1654 1655 1652 1785 1821 1526 1471
RMR (283.9) (282.4) (286.2) (268.4) (255.4) (233.9) (191.1)
Year5 1690 1699 1682 1826 1860 1571 1492
RMR (295.9) (277.1) (313.3) (249.4) (292.0) (244.1) (205.8)
Baseline .77 .77 .77 .78 .78 .76 .76
RQ (.06) (.06) (.05) (.06) (.06) (.06) (.05)
Year 5 .77 .77 .77 .77 .78 .77 .77
RQ (.05) (.05) (.05) (.05) (.05) (.05) (.05)
All AA=all African American subjects, All C=all Caucasian subjects, AAM=African American
men, CM=Caucasian men, AAW=African American women, CW=Caucasian women.
Baseline RMR=resting metabolic rate expressed as kcal/day at Baseline VIM exam,
Year 5 RMR= resting metabolic rate expressed as kcal/day at VIM Year 5 exam
RQ=respiratory quotient at Baseline VIM, and VIM Year 5.
The data in Table 4.2 show that there is no difference in either VIM baseline
or Year Five absolute resting metabolic rate between African American and
Caucasian men. There is also no difference in baseline or Year Five resting
metabolic rate between women of the two ethnic groups.
68


In this group of participants, despite having increases in weight and
adiposity over the five-year period, there is no significant difference between
African American and Caucasian adults in either Baseline or Year Five
respiratory quotient (Table 4.2).
4.1.4 Resting Metabolic Rate Expressed per
Kilogram of Fat-free Mass, and per Kilogram
of Body Weight
It is not uncommon to express resting metabolic rate in terms of body weight
or per unit of fat-free mass, since both are significant contributors to resting
metabolic rate. Tables 4.3 and 4.4 show differences between African
American and Caucasian men and women in Baseline and Year Five values
of resting metabolic rate expressed per kilogram of body weight (Table 4.3)
and per kilogram of fat-free mass (Table 4.4). Using a Scheffe test of
multiple comparisons, when baseline resting metabolic rate is expressed per
kilogram fat-free mass, African American men and women have significantly
lower resting metabolic rate than their Caucasian counterparts (p<0.001).
(Scheffe tests provide the mean difference and the standard error of the
differences between the means of each group.) When resting metabolic
rate is expressed per kilogram of body weight, there is a difference seen
only between women of the two groups, with African American women
69


having significantly lower baseline resting metabolic rate per kilogram of
body weight than Caucasian women (p<0.001).
These relationships stayed the same at the Year Five time point, with resting
metabolic rate per kilogram of fat-free mass being lower in African
Americans compared to Caucasians, and resting metabolic rate per
kilogram of body weight being significantly lower (p<0.001) in African
American compared to Caucasian women.
70


Table 4.3. Scheffe Multiple Comparisons Test of Differences in Mean
Values of Resting Metabolic Rate per Kilogram of Body Weight
(Kcal/wtkg) at VIM Baseline and VIM Year Five Between African
Americans and Caucasians.
Baselin e Kcal/wt kg Year 5 Kcal/wtkg Baseline Kcal/wtkg Year5 Kcal/wtkg
Mean SD Mean SD Mean SE p value Mean SE p value
AAM 21.15 2.75 20.72 3.18 AA vs C Men -.539 .471 .727 -.718 .476 .518
CM 21.69 2.09 21.44 2.20
AAW 19.09 3.08 18.63 2.86 AA vs C Women -3.34 .478 p<.001 -2.92 .485 p<.001
CW 22.44 3.50 21.55 3.36
Mean differences are calculated as the mean value for African American (AA) subjects
minus the mean value for Caucasian subjects (C) in kcals per kilogram of body weight
(wtkg) at Baseline and Year 5.
71


Table 4.4. Scheffe Multiple Comparisons Test of Differences in Mean
Values of Resting Metabolic Rate per Kilogram of Fat-free Mass
(Kcal/FFM) at VIM Baseline and VIM Year Five Between African
Americans and Caucasians.
Baseline Kcal/FFM Year 5 Kcal/FFM Baseline Kcal/FFM Year 5 Kcal/FFM
Mean SD Mean SD Mean SE p value Mean SE p value
AAM 27.19 3.02 27.70 3.16 AA vs C Men -2.24 .538 p=.001 -2.33 .509 p<.001
CM 29.44 2.70 30.03 2.99
AAW 33.13 4.11 33.64 2.97 AA vs C Women -2.18 .547 p=.001 -2.08 .518 p=.001
CW 35.31 3.27 35.72 3.37
Mean differences are calculated as the mean value for African American (AA) subjects
minus the mean value for Caucasian subjects (C) in kcals per kilogram of fat-free mass
(FFM) at Baseline and Year 5.
72


4.2 Results for Hypothesis #1
Hypothesis #1: Baseline age and baseline measures of body weight,
and adiposity will be positively associated with, but baseline resting
metabolic rate inversely associated with, changes in body weight and
adiposity at the end of five-years follow up in both African American
and Caucasian subjects.
This hypothesis addresses the relationship between baseline age and
baseline measures of resting metabolic rate, body weight, fat mass, and
percent body fat to changes in body weight, fat mass, and percent body fat
over the five-year period. A significant relationship between the
independent (baseline) variables and the dependent (change) variables
could reflect the ability of baseline measures to predict body weight and
adiposity changes over time.
73


4.2.1 Relationship of Baseline Measures of
Resting Metabolic Rate, Body Weight, and
Adiposity to Five-year Change in Weight and
Adiposity
Table 4.5 shows the Pearson correlation coefficients and significance values
for each of the comparisons of baseline variables (age, resting metabolic
rate, body weight, fat-free mass, fat mass, and percent body fat) to change
over time in body weight, fat mass, and percent body fat. The results
indicate that for the most part, baseline resting metabolic rate was not
related to the changes in body weight and body composition in the total
sample, nor within each of the four ethnic and gender groups. The
exceptions to this were for African American men, in whom baseline resting
metabolic rate was positively correlated to change in body weight (.236,
p<0.05) and for Caucasians as a whole, in whom baseline resting metabolic
rate was negatively correlated to the change in percent body fat (-0.211, p <
0.01). There was also a positive relationship between baseline resting
metabolic rate and change in fat-free mass in African American women
(0.245, p<0.05).
The majority of the significant correlations were observed in relation to the
baseline measures of adiposity, and/or body weight, and percent body fat
change. In the total sample, baseline body weight, fat mass, and percent
74


body fat were inversely related to percent body fat change. When
evaluating the relationships by ethnic group, baseline fat mass was inversely
related to percent body fat change in both African Americans and
Caucasians. In addition, baseline measures of body weight and fat-free
mass were significantly inversely related to percent body fat change in
Caucasians. In looking at the data by gender and ethnicity, baseline percent
body fat was inversely related to percent body fat change in African
American men only.
None of the baseline measures of body weight or body composition were
consistently correlated with body weight or fat mass change when looking at
the total sample or by ethnic and gender group. The exception was in
Caucasian women, in whom baseline fat-free mass and percent body fat
were negatively and positively correlated, respectively, with body weight
change, and in whom baseline fat-free mass was negatively correlated with
fat mass change. Therefore, these data suggest that fat-free mass at
baseline is the best predictor of body weight and fat mass change in
Caucasian women.
For the total sample, body weight and adiposity were positively related to
five-year change in fat-free mass. Both baseline fat mass and % body fat
75


were positively related to change in fat-free mass in all but Caucasian
women. Higher levels of adiposity at baseline were related to greater gains
in fat-free mass over the five-year period.
76


Table 4.5. Relationship of Baseline Measures of Age, Resting
Metabolic Rate (RMR), Body Weight (WT), Fat-free Mass (FFM), Fat
Mass (FM), or Percent Body Fat (% fat) to the Change in Weight or
Adiposity over Five Years.
Body Wt Change Fat-free Mass Change Fat mass Change % Fat Change
Pearson P= Pearson P= Pearson P= Pearson P=
Total Sample
BL Age -.142 .015 -.149** .010 -.107 .067 -.062 .291
BL RMR .020 .736 .105 .071 -.013 .821 -.085 .146
BL Wt -.012 .843 .149** .010 -.023 .697 -.172** .003
BL FFM -.027 .646 -.009 .883 -.032 .584 -.055 .345
BL FM .068 .246 .185** .001 .003 .953 -.182** .002
BL %fat .094 .106 .160** .006 .034 .555 -.122* .035
All African Americans
BL Age -.144 .086 -.219** .008 -.092 .273 -.041 .624
BL RMR .086 .308 .157 .061 .076 .369 .027 .747
BL Wt .014 .871 .181* .030 .041 .630 -.099 .239
BL FFM -.002 .977 -.015 .858 .038 .654 .065 .442
BL FM .047 .576 .175* .037 -.006 .940 -.194* .020
BL %fat .061 .471 .132 .115 -.004 .960 -.172* .040
All Caucasians
BL Age -.110 .178 -.016 .843 -.092 .256 -.077 .342
BL RMR -.052 .527 .040 .622 -.123 .129 -.211** .009
BL Wt -.086 .294 .066 .420 -.155 .056 -.282** .0001
BL FFM -.085 .298 -.038 .640 -.156 .053 -.205* .011
BL FM .068 .406 .175 .030 -.020 .811 * o 00 r .026
BL %fat .135 .097 .199 .013 .087 .286 -.051 .532
African Amer. Men
BL Age .120 .315 -.205 .084 .165 .165 .194 .102
BL RMR .236* .046 .126 .293 .189 .113 .047 .694
BL Wt .144 .229 .191 .107 .137 .251 -.080 .503
BL FFM .164 .168 -.043 .717 .167 .160 .041 .734
BL FM . .171 .150 .250* .034 .048 .689 -.180 .130
BL %fat .128 .284 .234* .048 -.008 .945 -.210 .076
Caucasian Men
BL Age -.174 .124 .001 .993 -.173 .128 -.169 .136
BL RMR .061 .592 .146 .200 .016 .886 -.108 .342
BL Wt .002 .984 .185 .102 -.035 .763 -.202 .074
BL FFM .070 .541 .063 .580 .032 .777 -.067 .559
BL FM .044 .700 .266* .018 -.051 .653 -.242 * .031
BL %fat -.025 .829 .257* .022 -.118 .299 -.282 * .012
African Amer.Women
BL Age -.335** .004 -.237* .045 -.276* .020 -.222 .063
BL RMR -.012 .923 .245* .038 .002 .987 -.095 .431
BL Wt -.071 .552 .177 .137 -.022 .857 -.140, .245
BL FFM -.141 .236 .020 .868 -.031 .795 -.109 .365
BL FM -.005 .964 .238* .045 -.039 .749 -.159 .185
BL %fat .084 .484 .274* .021 -.005 .967 -.103 .395
77


Table 4.5 (Cont.)
Body Wt Change Fat-free Mass Change Fat mass Change % Fat Change
Pearson P= Pearson P= Pearson P= Pearson P=
Caucasian Women
BL Age -.061 .610 -.042 .722 -.053 .657 -.033 .782
BL RMR -.142 .232 -.127 .283 -.157 .183 -.185 .114
BL Wt -.130 .273 -.050 .672 -.161 .171 -.266 * .022
BL FFM -.286* .014 -.345** .003 -.266* .022 -.204 .081
BL FM .073 .537 .130 .269 -.041 .732 -.217 .063
BL %fat .233* .047 .256 .027 .113 .337 -.113 .339
* Correlation is significant at the 0.05 level (2-tailed).
** Correlation is significant at the 0.01 level (2-tailed).
BL age= age measured at the baseline exam, BL RMR = resting metabolic rate in kcal/day
measured at baseline, BL Wt = body weight in kg measured at baseline, BL FFM= fat-free
mass in kg measured at baseline, BL FM= fat mass in kg measured at baseline, BL %Fat=
percent body fat measured at baseline.
Total Sample= all participants (n=297), All African American= all African American subjects
(n=144), All Caucasians= all Caucasian subjects (n=153), African Amer. Men=African
American men (n=72), Caucasian Men=Caucasian men (n=79), African Amer.
Women=African American women (n=72), Caucasian Women=Caucasian women (n=74).
4.2.2 Baseline Age as a Predictor of Five-year
Change in Body Weight, Adiposity, Resting
Metabolic Rate, and Respiratory Quotient
Although by study design all subjects aged five years over the course of the
VIM study, it is still of interest to determine if age at baseline was associated
with five-year change in body weight, adiposity, metabolic rate, respiratory
quotient and fat-free mass. Table 4.5 shows that in the total sample
baseline age was negatively associated with change in body weight
(p=0.015) and negatively associated with change in fat-free mass (p=0.010),
but was not associated with adiposity changes. Baseline age was
significantly and negatively associated with change in fat-free mass
78


(p=0.008) in the African American group as a whole, however this
association was mainly driven by the African American women. However,
when looking at each of the ethnic and gender groups separately, age was
negatively correlated with body weight change (p=0.004), fat-free mass
change (p=0.045), and fat mass change (p=0.020) only in African American
women. There were no other associations between baseline age and five-
year change in body weight or body composition within any of the other
ethnic and gender groups.
T-tests were used to assess the strength of the relationship between age at
baseline, and five-year change in resting metabolic rate, respiratory
quotient, and fat-free mass within the total sample and within each group
separately. Table 4.6 shows these relationships. Baseline age was
inversely associated with fat-free mass change in the total sample, in African
Americans, but particularly in African American women. Thus, the greater
the initial age in African Americans, the smaller the increase in fat-free mass
over time. Baseline age was only predictive of a change in resting metabolic
rate for African American men; higher baseline age was inversely
associated with resting metabolic rate. Thus, in African American men, the
older an individual was at baseline, the lesser the increase in resting
metabolic rate. Baseline age was unrelated to the change in respiratory
79


quotient over the five years, except in Caucasian women where there was a
significant inverse relationship (p < 0.012). Overall, although the change in
the individual variables were somewhat different within each ethnic and
gender group, the effect of age was to decrease the amount of change in
the measured variables over the five-year period.
80


Table 4.6. Association of Age at VIM Baseline with Change in
Resting Metabolic Rate (RMR), Respiratory Quotient (RQ), and
Fat-free Mass (FFM).
RMR change RQ change FFM change
Group and n Pearson Coefficient Pearson Coefficient Pearson Coefficient
Total Sample (n=297) -0.082 -0.044 -0.149*
All African Americans (n=144) -0.112 0.026 -0.219**
All Caucasians (n=153) -0.034 -0.131 -0.016
African American Men (n=72) -0.242 * 0.046 -0.205
Caucasian Men (n=79) 0.010 -0.007 0.001
African American Women (n=72) 0.041 -0.016 -0.237 *
Caucasian Women (n=74) -0.078 -0.290 * -0.042
*=p<0.05, **=p<0.01
RMR change = change over five-year period in resting metabolic rate in kcal/day
RQ change = change over five-year period in respiratory quotient
FFM change = change over five-year period in fat-free mass in kilograms
81


To further analyze the effect of age on change over time, the groups were
subdivided into two age groups by division at the median of the baseline age
range. Analyses were performed stratifying age to determine if there was a
difference in the amount of change in weight and adiposity change between
the participants who were younger at baseline (28-34 years) and those
who were older at baseline (35 years and older). Table 4.7 shows the
relationship between stratified age within each of the ethnic and gender
groups and change in these variables. All groups within both age ranges
had significant increases in body weight, fat mass, and percent body fat
from baseline except African American women who were 35 or older at
baseline. Only African American men and women in the younger age group
at baseline had significant increases in fat-free mass, and only younger
African American men had a significant increase in resting metabolic rate
from baseline. There were no significant changes in respiratory quotient
over the five-year period within either of the age groups.
82


Table 4.7. Association of Stratified Age (younger and older) at VIM
Baseline with Change in Resting Metabolic Rate (RMR), Respiratory
Quotient (RQ), Weight, and Adiposity.
RMR change RQ change Wt kg change FFM kg change FM kg change % Fat change
Group Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD
BL 28-34 yrs
AAM (n=42) 86 194 ** -0.00 0.08 4.05 6.55 *** 0.95 2.63* 3.46 5.55 *** 2.64 4.41 ***
CM (n=32) 52 230 0.00 0.05 4.00 7.78** 0.12 2.09 3.37 6.40** 2.64 + 4.88 **
AAW (n=31) 46 158 0.01 0.07 7.44 7.09 **** 1.16 2.81 * 6.38 6.76 **** 3.58 5.03 ***
CW (n=21) 37 151 0.02 0.06 4.09 7.76* 0.07 1.50 4.02 5.28** 3.39 3.90 ***
35+ yrs at BL
AAM (n=30) -21 213 -0.01 0.08 4.97 0 <| **** -0.16 3.15 5.15 5.96 **** 4.39 5.37 ****
CM (n=47) 29 142 0.01 0.08 2.02 6.03* -0.02 2.26 1.53 5.04* 1.28 4.64
AAW (n=41) 44 166 0.01 0.07 2.35 8.99 0.05 2.62 2.53 8.42 1.33 5.83
CW (n=53) 15 126 -0.00 0.06 3.47 5.47 **** 0.02 1.77 3.64 4.48 **** 3.30 3.68 ****
*=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001
BL= baseline, ns= non-significant, AAM= African American men, CM= Caucasian men,
AAW= African American women, CW= Caucasian women, RMR=resting metabolic rate in
kcal/day, RQ=respiratory quotient, Wtkg=body weight in kilograms, FFM=fat-free mass in
kilograms, FM=fat mass in kilograms, %Fat=percent body fat.
83