Protocol for the modeling the epidemiologic transition study: a longitudinal observational study of energy balance and change in body weight, diabetes and cardiovascular disease risk.
ABSTRACT The prevalence of obesity has increased in societies of all socio-cultural backgrounds. To date, guidelines set forward to prevent obesity have universally emphasized optimal levels of physical activity. However there are few empirical data to support the assertion that low levels of energy expenditure in activity is a causal factor in the current obesity epidemic are very limited.
The Modeling the Epidemiologic Transition Study (METS) is a cohort study designed to assess the association between physical activity levels and relative weight, weight gain and diabetes and cardiovascular disease risk in five population-based samples at different stages of economic development. Twenty-five hundred young adults, ages 25-45, were enrolled in the study; 500 from sites in Ghana, South Africa, Seychelles, Jamaica and the United States. At baseline, physical activity levels were assessed using accelerometry and a questionnaire in all participants and by doubly labeled water in a subsample of 75 per site. We assessed dietary intake using two separate 24-hour recalls, body composition using bioelectrical impedance analysis, and health history, social and economic indicators by questionnaire. Blood pressure was measured and blood samples collected for measurement of lipids, glucose, insulin and adipokines. Full examination including physical activity using accelerometry, anthropometric data and fasting glucose will take place at 12 and 24 months. The distribution of the main variables and the associations between physical activity, independent of energy intake, glucose metabolism and anthropometric measures will be assessed using cross-section and longitudinal analysis within and between sites.
METS will provide insight on the relative contribution of physical activity and diet to excess weight, age-related weight gain and incident glucose impairment in five populations' samples of young adults at different stages of economic development. These data should be useful for the development of empirically-based public health policy aimed at the prevention of obesity and associated chronic diseases.
- [show abstract] [hide abstract]
ABSTRACT: Obesity could be considered a new global health epidemic above all others, especially when it is characterized by central fat distribution. This is illustrated by dramatic provisional data, indicating a continuous increase in the trend of overweight and obese individuals in several countries, including the USA and countries in Europe. Several epidemiological, pathophysiological and clinical studies clearly indicate that two of the major independent risk factors for cardiovascular disease or events are being overweight, and obesity. Accordingly, weight loss and prevention of weight gain has to be considered one of the most important strategies to reduce the incidence of cardiovascular disease.Expert Review of Cardiovascular Therapy 04/2004; 2(2):203-12.
- [show abstract] [hide abstract]
ABSTRACT: This paper explores the major changes in diet and physical activity patterns around the world and focuses on shifts in obesity. Review of results focusing on large-scale surveys and nationally representative studies of diet, activity, and obesity among adults and children. Youth and adults from a range of countries around the world. The International Obesity Task Force guidelines for defining overweight and obesity are used for youth and the body mass index > or =25 kg/m(2) and 30 cutoffs are used, respectively, for adults. The nutrition transition patterns are examined from the time period termed the receding famine pattern to one dominated by nutrition-related noncommunicable diseases (NR-NCDs). The speed of dietary and activity pattern shifts is great, particularly in the developing world, resulting in major shifts in obesity on a worldwide basis. Data limitations force us to examine data on obesity trends in adults to provide a broader sense of changes in obesity over time, and then to examine the relatively fewer studies on youth. Specifically, this work provides a sense of change both in the United States, Europe, and the lower- and middle-income countries of Asia, Africa, the Middle East, and Latin America. The paper shows that changes are occurring at great speed and at earlier stages of the economic and social development of each country. The burden of obesity is shifting towards the poor.International Journal of Obesity 11/2004; 28 Suppl 3:S2-9. · 5.22 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: There is an epidemic of obesity affecting adolescents worldwide. Both in developing and in developed countries, adolescents are increasingly becoming more obese. The number of adolescents exceeding previously identified cut-points as well as the weight and fatness of the most obese individuals is increasing at a progressive rate. Despite their benign appearance, epidemics of noncommunicable disease (or their risk factors) are no less devastating to the health of populations. The two key responses to any epidemic are to discover the causes of the epidemic disease and to characterize the epidemic. The latter needs to occur in relation to prevalence, distribution across the populations (are some population groups more likely to be affected than others?), and secular trends. This chapter reviews what is currently known about the epidemiology of overweight and obesity among adolescents throughout the world. To clarify terms of epidemiology, this chapter first identifies what are considered the most appropriate measures of adiposity and defines how much fat is too much fat.Adolescent medicine (Philadelphia, Pa.) 03/2003; 14(1):1-9.
STUDY PROTOCOL Open Access
Protocol for the modeling the epidemiologic
transition study: a longitudinal observational
study of energy balance and change in body
weight, diabetes and cardiovascular disease risk
Amy Luke1*†, Pascal Bovet2, Terrence E Forrester3, Estelle V Lambert4, Jacob Plange-Rhule5, Dale A Schoeller6,
Lara R Dugas1†, Ramon A Durazo-Arvizu1†, David Shoham1, Richard S Cooper1, Soren Brage7, Ulf Ekelund7and
Nelia P Steyn8
Background: The prevalence of obesity has increased in societies of all socio-cultural backgrounds. To date,
guidelines set forward to prevent obesity have universally emphasized optimal levels of physical activity. However
there are few empirical data to support the assertion that low levels of energy expenditure in activity is a causal
factor in the current obesity epidemic are very limited.
Methods/Design: The Modeling the Epidemiologic Transition Study (METS) is a cohort study designed to assess
the association between physical activity levels and relative weight, weight gain and diabetes and cardiovascular
disease risk in five population-based samples at different stages of economic development. Twenty-five hundred
young adults, ages 25-45, were enrolled in the study; 500 from sites in Ghana, South Africa, Seychelles, Jamaica and
the United States. At baseline, physical activity levels were assessed using accelerometry and a questionnaire in all
participants and by doubly labeled water in a subsample of 75 per site. We assessed dietary intake using two
separate 24-hour recalls, body composition using bioelectrical impedance analysis, and health history, social and
economic indicators by questionnaire. Blood pressure was measured and blood samples collected for
measurement of lipids, glucose, insulin and adipokines. Full examination including physical activity using
accelerometry, anthropometric data and fasting glucose will take place at 12 and 24 months. The distribution of
the main variables and the associations between physical activity, independent of energy intake, glucose
metabolism and anthropometric measures will be assessed using cross-section and longitudinal analysis within and
Discussion: METS will provide insight on the relative contribution of physical activity and diet to excess weight,
age-related weight gain and incident glucose impairment in five populations’ samples of young adults at different
stages of economic development. These data should be useful for the development of empirically-based public
health policy aimed at the prevention of obesity and associated chronic diseases.
* Correspondence: email@example.com
† Contributed equally
1Stritch School of Medicine, Loyola University Chicago, Maywood, IL, USA
Full list of author information is available at the end of the article
Luke et al. BMC Public Health 2011, 11:927
© 2011 Luke et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
Populations all over the world are experiencing rapid
increases in the prevalence of obesity and diabetes [1-5].
To date, the public health response to the emerging epi-
demics has been largely ineffective. As a first line of
response, professional bodies and government organiza-
tions have issued prevention guidelines, all of which
include recommendations on levels of physical activity
(PA) required to prevent weight gain [6-10]. There is
limited direct evidence, however, that can be brought to
bear on the question of whether the obesity epidemic
has resulted primarily or even partially from population-
wide declines in habitual PA . Thus, even if fully
implemented, it is not clear that the current recommen-
dations on PA would impact the trend in age-related
Energy balance is defined by the direct relationship
between energy intake and expenditure. Fat mass accu-
mulation and subsequently obesity can only result from
an excess of calories consumed over calories expended.
It is often assumed, therefore, that a similarly straight-
forward, simple relationship exists between variation in
PA within the range of normal for a population and the
risk of weight gain. The second hypothesis which fol-
lows from this line of reasoning is that population-level
weight gains taking place have resulted, at least in part,
from declining PA, and that small increases in PA for
individuals will prevent weight gain. In fact, neither of
these lines of thoughts is justified by theory alone or
supported by evidence. Volitional increases in physical
expenditure are typically accompanied by increases in
energy intake [12,13], while excess intake can be stimu-
lated independently by high caloric density of food,
changes in food availability, and eating patterns. As
pointed out repeatedly by investigators in this field
[14-17], the regulation of energy stores and body com-
position must be seen as a complex, dynamic process
influenced by the interplay between factors that modify
both intake and expenditure and these proximal factors
are admittedly importantly influenced by distal environ-
mental factors at the society and global levels . An
individual ultimately responds to biological stimuli, e.g.,
the hormones that control satiety; and environmental
cues, e.g. food aromas or congested walkways, to modu-
late energy balance [18-21]. Whether variation in pat-
terns of activity observed in modern, free-living
populations plays a key role in this process is thus lar-
Lifestyle trends are clearly taking place in many socie-
ties; however, it is difficult to identify clear temporal
associations with increases risk of obesity. As demon-
strated in the trends from the CDC’s Behavioral Risk
Factor Surveillance System and the National Health and
Nutrition Examination Surveys conducted since the
1960’s, the US experienced an upward deviation of the
trend for body mass index (BMI) in the mid-1980s,
resulting in the current “obesity epidemic” . How-
ever, there was no obvious temporally related break in
activity patterns or eating habits [11,22]. In the absence
of direct evidence most observers have defaulted to a
“common sense” explanation that people are “eating
more and exercising less”. Unfortunately, this formula-
tion is both an oversimplified concept and unsubstan-
tiated causally  and there have been relatively few
longitudinal studies investigating this premise. Thus the
central purpose of our project is to test whether change
in PA can be identified as a contributory mechanism to
the population-wide weight gain, and, if so, to quantify
It is repeatedly stated that the environment of modern
society is “obesogenic” in part because of the greatly
reduced need for physical exertion and increased supply
of foods universally [23-25]. While it seems self-evident
that industrialized societies require less PA across a
range of domains, the empirical data on this question
lead to the opposite conclusion. As summarized by
Ferro-Luzzi and Martino , “The seemingly obvious
conclusion that energy expenditure is systematically
higher in Third World countries is not supported by the
evidence... (Based on a review of the data) we conclude,
therefore, that there are no systematic differences in the
level of habitual activity between developed and develop-
ing countries” . Additionally, the results of a recent
meta-analysis of total energy expenditure and PA levels
in adult samples found no difference in either, after
adjustment for body size, between developing and more
industrialized countries . While PA has many indis-
putable health benefits, its role in the regulation of body
weight requires careful additional study particularly in
populations in early stages of the epidemiologic transi-
tion. We recognize that surveillance for PA will be cru-
cial to unravel the causes of the obesity epidemic, and it
must be conducted using objective measures.
Modeling the Epidemiologic Transition Study Hypotheses
The Modeling the Epidemiologic Transition Study
(METS) was designed to test four hypotheses associated
with the relationship between PA, body weight and dia-
betes and cardiovascular disease (CVD) risk. We
hypothesized that: 1) population mean levels of PA are
negatively related to population mean levels of obesity
and relative weight, 2) PA is negatively related to per-
cent body fat in the study populations at baseline, inde-
pendent of dietary intake, 3) PA is negatively related to
change in body weight during follow-up, independent of
dietary intake, and as an exploratory hypothesis, 4) PA
modifies the association between adipocytokines (e.g.,
Luke et al. BMC Public Health 2011, 11:927
Page 2 of 10
adiponectin and leptin) and hormones (e.g., ghrelin) and
weight regulation and insulin sensitivity.
In order to test these hypotheses, the investigators of
METS have enrolled 500 participants, ages 25-45 years,
from each of five African-origin populations, i.e., Ghana,
South Africa, Seychelles, Jamaica and the United States.
In all participants at baseline PA was measured using
accelerometry and dietary intake by 24-hour dietary
recalls, fasting glucose/insulin, and adipocytokines were
measured by standard laboratory methods, and social
and environmental factors, PA patterns and alcohol,
smoking and medication histories were assessed by
questionnaire. In a random subset of 75 participants
from each site at baseline, total energy expenditure
(TEE) was measured using the doubly labeled water
(DLW) method and resting energy expenditure (REE) by
indirect calorimetry. Follow-up measurements of weight,
waist circumference and blood pressure will be made at
12-months and all measures except dietary intake, TEE
and REE will be repeated at 24-months. The statistical
analyses will examine the relationship between PA and
body composition within and among populations at
baseline and between PA and weight change at follow-
up, and assess inter-relations of PA, diet, glucose, insulin
Methods and Design
Design and Settings
Twenty-five hundred adults, ages 25-45, were enrolled
in METS between January 2010 and September 2011
and have had energy expenditure, dietary intake, body
weight and composition, and biomarkers of obesity and
diabetes measured at baseline. The participants will all
be followed over the subsequent 24 months to assess
change in body weight, composition, and diabetes and
CVD risk. Five hundred participants, 50% of whom are
female, were enrolled in each of five study sites: rural
Ghana, urban South Africa, Seychelles, urban Jamaica
and suburban United States (Chicago area). The popula-
tions sampled are of African descent and provide a
range of body sizes, i.e., the mean BMIs of adults from
the study sites vary from a low of about 24 kg/m2in
rural Nkwantakese (Ghana) to a high of 31 kg/m2in
suburban Maywood (USA). The study sites also repre-
sent a range of social and economic development as
defined by the UN Human Development Index (HDI)
2010: Ghana is defined as a low HDI country, South
Africa as middle HDI, Jamaica and Seychelles as high
HDI and the US as a very high HDI country .
The in-country study sites include the town of
Nkwantakese in the Afigya-Kwabre District of the
Ashanti Region of Ghana and its surrounding villages.
The town is situated to the southwest of Agona Ashanti
the District Capital, and is about 20km from Kumasi
with a population of approximately 17,000. Khayelitsha
is the 3rdlargest township in South Africa and is adja-
cent to the city of Cape Town. The population is about
500,000 with 80 percent of the residents living in tem-
porary housing and 40 percent unemployed. The study
site in Jamaica is in Kingston, the capital and largest city
with a population of 651,880. In Seychelles, individuals
have been recruited from the main island of the archipe-
lago, Mahé, which includes approximately 75,000 inhabi-
tants for a surface of 155 km2. Mahé can be qualified as
semi-urban and its economy is mainly driven by tour-
ism, industrial fishing and services. Seychelles is located
approximately 1,600 km east of Kenya in the Indian
Ocean, and approximately 2,000 km north of the island
of Mauritius, and has a total population of about 87,000.
Maywood, in the US, is an African-American working
class community adjacent to the western border of Chi-
cago, Illinois, with a population of approximately 24,903
people. The research clinic is on the campus of Stritch
School of Medicine, Loyola University, within walking
distance of most neighborhoods and is well known to
local residents. The coordinating center for METS is
also located in Maywood, IL at Loyola University.
We excluded individuals with obvious infectious dis-
eases (including active malaria), pregnant or lactating
women, and HIV positive individuals. Any individual
that has a condition preventing them from engaging in
normal physical activities, such as severe osteo’- or
rheumatoid arthritis, lower extremity disability, was also
excluded from METS. Population-based surveys have
been previously carried out in each of the sites, thus
investigators from each site decided upon the best
means of recruiting a representative sample from their
respective communities. In Nkwantakese, Ghana, a sim-
ple random sample was generated for the age-range of
the study from the population census for Nkwantakese
and its surrounding villages. In both Seychelles and
South Africa sex- and age-stratified random samples
were generated from their respective national censuses.
In Kingston, Jamaica, districts were randomly sampled;
beginning from a fixed point in each district (e.g., the
north-west corner), and door-to-door recruitment took
place. Similarly, in Maywood, IL, USA, all city blocks in
the community were randomized and door-to-door
recruitment was conducted.
The protocol for METS was approved by the Institu-
tional Review Board of Loyola University Chicago, IL,
USA; the Committee on Human Research Publication
and Ethics of Kwame Nkrumah University of Science
and Technology, Kumasi, Ghana; the Research Ethics
Committee of the University of Cape Town, South
Africa; the Board for Ethics and Clinical Research of the
Luke et al. BMC Public Health 2011, 11:927
Page 3 of 10
University of Lausanne, Switzerland; the Ethics Commit-
tee of the University of the West Indies, Kingston,
Jamaica; and the Health Sciences Institutional Review
Board of the University of Wisconsin, Madison, WI,
USA. Written informed consent was obtained from all
Baseline, 12-month and 24-month follow-up examina-
tions will be conducted for the METS study. The project
coordinators for each field site were jointly trained and
certified in all measurement protocols by coordinating
center staff; the measurements included in METS are
presented in Table 1. All measurements were underta-
ken at outpatient clinics located in each of the METS
Physical Activity (Accelerometer)
Physical activity was assessed using the Actical acceler-
ometer (Phillips Respironics, Bend, OR, USA) in all
2500 METS participants. Previous studies have shown
that accelerometer-based activity monitors can discrimi-
nate differing intensities of activity [29-34], making it
possible to adequately characterize each of the study
communities with regard to overall and subcomponents
of PA. The activity monitor records the intensity, dura-
tion and frequency of physical motion through the use
of an accelerometer which produces a variable electrical
current based on the combination of the amplitude and
frequency of motion. Accelerometers are omnidirec-
tional motion sensors that count the vertical and
horizontal acceleration of the user. This information is
stored within the instrument as activity counts per
epoch, or specified subunit of time, e.g., per minute. As
the intensity of the activity increases, so does the num-
ber of activity counts per epoch.
The monitor was worn at the waist, positioned just
behind the left hip. Each participant was asked to wear
the activity monitor at all times over 8 days, including
during sleep; the only time the monitor should be
removed was while bathing, showering, or swimming.
We chose to monitor all participants for six complete
days (i.e., a total of 8 days after taking the two partial
days into consideration); based on preliminary work,
this will provide a good level of reliability at 0.83-0.92%.
Total Energy Expenditure (Doubly Labeled Water Method)
A subset of 75 participants per site were enrolled in the
DLW protocol (N = 375 total). For these participants,
PA energy expenditure was calculated as the difference
between TEE as measured using DLW and REE (PA =
TEE - REE - TEF; where TEF, thermic effect of food,
will be estimated as 10% of TEE) .
Detailed descriptions of the DLW method and labora-
tory analysis have previously been published [36-39]. In
brief, TEE was measured over a 7-day period. On the
morning of the initial clinic examination, prior to the
measurement of REE, a spot urine collection was made
and the participant given an oral dose of DLW contain-
ing approximately 1.8 g of 10% H218O and 0.12 g 99.9%
2H2O per kg body water. Spot urine samples were then
collected at approximately 1, 3 and 4 hours after isotope
Table 1 Study Measures
Objectively Measured Energy Expenditure
Physical activity (Actical)
Total energy expenditure (DLW;subset N = 375)
Resting energy expenditure (indirect calorimetry; subset N = 375)
Physical Activity (GPAQ)
Dietary Intake (24-hr Recall)
Medication & supplement use
Smoking status & alcohol consumption
Household SES, education level, industry & occupation
Bioelectrical impedance analysis
Isotope Dilution (subset N = 375)
Weight, height, waist & hip circumferences, blood pressure & pulse
Hba1c, total cholesterol, HDL and LDL
cholesterol, triglyceride, glucose, insulin,
adiponectin, leptin, ghrelin, urinary albumin &
creatinine, and T4, T3& TSH in subset (N = 375)
Luke et al. BMC Public Health 2011, 11:927
Page 4 of 10
administration. Participants collected an early morning
urine void on day 7 and returned to the clinic later the
same day (± 1 day) to provide a final urine sample. All
urine samples were aliquotted in duplicate and stored in
5-mL o-ring-sealed cryovials at -20°C. Mass spectro-
metric analyses of the DLW urine samples were carried
out at the Stable Isotope Core Laboratory at University
of Wisconsin, Madison, WI, USA. The CO2production
was calculated using equation 6.6 of the IAEA technical
 and energy expenditure using the modified Weir
equation . The country specific average RER from
dietary records was used for the latter.
Resting Energy Expenditure
Resting EE was measured in the DLW subsample using
indirect calorimetry. In the US, Ghana and Seychelles
sites, REE was measured using the MaxIIa indirect
calorimeter (AEI Technologies, Aurora, IL, USA); in
Jamaica the Columbia Instruments Oxymax 4.0.
(Columbus, Ohio, USA) was used; and in South Africa,
the VMax indirect calorimeter by SensorMedics (Viasys
Health Care, Waukegan, IL, USA) was used. Cross-vali-
dation of instruments was carried out through external
calibrations. The investigators have had extensive
experience in the measurement of REE across multiple
sites, with over 2500 measurements made in the US and
The detailed description of measurement of REE using
indirect calorimetry has been previously published
[42-45]. Participants were asked to fast from 10 pm the
evening prior to the initial examination and were rested
for at least 15 minutes prior to the REE measurement.
Respiratory gases were collected for 30 minutes, the first
10 minutes of data were discarded and last 20 minutes
used to estimate REE. Oxygen and carbon dioxide were
continuously sampled during the procedure and minute-
by-minute consumption and production values were cal-
culated; EE was calculated according to the modified
Weir equation .
Physical Activity by Questionnaire
All participants had PA also assessed using the Global
Physical Activity Questionnaire (GPAQ, version 2) .
The GPAQ was developed by the World Health Organi-
zation (WHO) as part of the WHO STEPwise approach
to chronic disease risk-factor surveillance  to pro-
duce reliable and valid estimates of PA for use in devel-
oping countries. The main outcome variables are: a
categorical variable of total PA (high, moderate and low)
and a continuous variable of PA within the domains of
work, transport and leisure.
Each participant has completed two 24-hour recalls
using the multiple pass method [48-50], one at the
initial baseline examination and the second when the
activity monitor is collected, typically 6-9 days later. At
each site, the 24-hour recalls were collected by centrally
trained interviewers using pen and paper, scanned and
sent via secure FTP  to the Coordinating Center at
Loyola University Chicago where data entry and analysis
using the Nutrient Data System for Research (NDSR;
University of Minneapolis, MN, USA) [48-50] took
place. Food ingredient identification and portion size
estimation were augmented using photographs and
usual portions of local foods assembled at each site
prior to study initiation; this methodology is based on
the Dietary Assessment Education Kit developed by one
of the study consultants for the Medical Research Coun-
cil South Africa . Primary endpoints of interest are
total energy intake and macronutrient composition (i.e.,
% kcals from fat, carbohydrate and protein), as well as
indicators of intake of processed foods (e.g., # of swee-
tened beverages, pre-packaged foods and restaurant fast
foods per day) and of intake of fruits, vegetables, as well
as site-specific commonly eaten foods. While we recog-
nize the importance of micronutrients for overall health
and well-being of individuals, many nutrient databases
lack sufficient data on local food micronutrient content.
Smoking Status & Alcohol Consumption
Participants were classified as non-smoker, occasional or
smoker based on questions on cigarettes, cigar or pipe
smoking and chewing tobacco use. In addition, there
were multiple questions concerning alcohol use in an
effort to describe type, volume and frequency of
Health History, Medication & Supplement Use
Basic health history information, with a focus on cardio-
vascular conditions and diabetes, was collected including
age of first diagnosis where applicable. Participants were
asked about medication and dietary supplement use,
with emphasis on vitamin D and calcium supplements.
Household SES, Education and Occupation
Fifty-four questions were included which covered gen-
eral household characteristics, participant and significant
other’s occupation, parental education and household
assets and amenities. These questions were based on the
Core Welfare Indicators Questionnaire from the World
Bank, designed to monitor social indicators in Africa
Body Composition (Bioelectrical Impedance Analysis)
Body composition was assessed in all participants using
bioelectrical impedance analysis (BIA); BIA measures
the impedance to the flow of an applied mild alternating
current by body tissues. The measured impedance of
body tissues can be used to estimate total body water,
from which fat-free mass and fat mass can be calculated
. With participants in the supine position with limbs
abducted, current-supplying electrodes were placed on
the dorsal surfaces of the right hand and foot at the
metacarpals and metatarsals, respectively, and detection
Luke et al. BMC Public Health 2011, 11:927
Page 5 of 10
electrodes were placed at the pisiform prominence of
the right wrist and the anterior surface of the true ankle
joint . A single-frequency instrument (BIA Quan-
tum, RJL Systems, Clinton Township, MI) was attached
to the electrodes and generated an excitation current of
800 μA at 50 kHz; resistance and reactance measures
were recorded. Unless data from the isotope dilution
analyses, described below, indicate a need for the devel-
opment of separate equations for our sample, we pro-
pose to use pre-existing validated BIA equations  for
estimation of total body water. Impedance measure-
ments will be taken at baseline and at the 24-month
examination to estimate change in body composition.
Body Composition (Isotope Dilution)
Total body water was measured using isotope dilution in
the DLW subset of participants at baseline and again at
the 24-month examination. The basis of this measure-
ment is the dilution principle: total body water was cal-
culated using the measurement of the abundance of
either isotope (deuterium or 18-oxygen) from the DLW
procedure after complete equilibration with body water
 and correction for non-aqueous exchange of 1.042
and 1.007, respectively . Fat-free mass was calculated
using a hydration constant (0.732 ) from total body
water, and fat was calculated as the difference between
body weight and fat-free mass . The 375 isotope
dilution measurements at both baseline and follow-up
will be used to calibrate BIA measurements for change
in individual body composition.
Height, Weight & Circumferences
At the initial clinic examination, height, weight, and
waist and hip circumferences were measured. Weight
was measured without shoes with the participant
dressed in light clothing to the nearest 0.1 kg using the
same model standard calibrated balance at all 5 sites
(Seca 770, Hamburg, Germany). Height was measured
to the nearest 0.1 cm using a stadiometer (e.g. Invicta
Stadiometer, Invicta, London, UK) without shoes and
with the participant’s head held in the Frankfort plane.
Waist circumference was measured to the nearest 0.1
cm at the umbilicus using a flexible metal tape measure
(Gulick, Creative Engineering, Michigan, USA). Hip cir-
cumference was measured to the nearest 0.1 cm at the
point of maximum extension of the buttocks. For both
circumferences, repeat measures will be taken and if the
two measurements differ by more than 0.5 cm, a third
measurement was taken.
Blood pressure was measured using the protocol and
training procedures developed for our ongoing interna-
tional hypertension studies [44,59-62]. Systolic and dia-
stolic blood pressure and pulse were measured using the
Omron Automatic Digital Blood Pressure Monitor
(model HEM-747Ic, Omron Healthcare, Bannockburn,
IL, USA). With the antecubital fossa at heart level, three
measurements were made at each of two time points
separated by approximately 60 minutes.
Participants were asked to fast from the evening prior to
the baseline clinic examination. Fasting blood samples
were drawn for analysis of adipose-related hormones
and adipocytokines, glucose, insulin, lipids, albumin. A
random spot urine sample was collected for assessment
of urinary creatinine and albumin. The blood samples
were processed and plasma or serum separated within
two hours of collection and stored at -80°C in the
laboratory at each study site. Fasting plasma glucose was
measured using the glucose oxidase method at each site
at the time of collection. Insulin, total ghrelin, leptin
and adiponectin from all sites were measured using
radioimmunoassay kits at the departmental laboratory at
Loyola University Chicago (Linco Research, Inc., St.
Charles, MO). All remaining assays were conducted at
the Zentrum fϋr Lambormedizin, Leiter Klinische Che-
mie und Hämatologie, St. Gallen, Switzerland.
All protocols for METS were approved by the Institu-
tional Review Board or Ethics Committee of all partici-
pating institutions. Written informed consent was
obtained from all participants in all study sites; all parti-
cipants received a copy of the signed consent containing
site-specific contact information in case of questions or
Participants were provided results of plasma glucose
levels and blood pressure at the time of their baseline or
follow-up clinic examinations. In cases where elevated
plasma glucose or blood pressure was identified, accord-
ing to clinical cut-points, participants were referred to
appropriate clinics or their own physicians.
Data management is centralized at the coordinating cen-
ter at Loyola University Chicago. All data forms, ques-
tionnaires and dietary recall instruments are scanned
and, along with electronic Actical data files, are sent via
secure FTP (Bitvise Tunnelier ) to the data manager
at the coordinating center. All scanned forms are coded
by experienced, trained personnel and double data entry
is carried out. A series of logic checks are then per-
formed and, when outliers are encountered, discrepan-
cies are followed up with staff at the appropriate field
In the analysis phase of METS descriptive characteristics
from each study site (e.g., mean levels and distributions
Luke et al. BMC Public Health 2011, 11:927
Page 6 of 10
of PA, dietary factors, body size and composition, adipo-
cytokines and hormones, and prevalence of hyperglyce-
mia and high blood pressure) will be explored and the
univariate correlation structure for the continuous vari-
ables will be described. Cross-sectional, multivariable
regression models will then be constructed to assess the
role of PA as a predictor of weight and risk factor sta-
tus, independent of diet. Similar analyses will be con-
ducted to assess the relationship between PA and
change in weight and risk factors. An important focus
of these analyses will be looking for potential heteroge-
neity of risk relationships across sites. Three main
aspects are of interest, namely whether the PA-risk fac-
tor (e.g. insulin level) relationship is present, or of the
same magnitude, in each sample and whether there
exists co-variation with the mean level of PA; testing for
interaction will be undertaken where appropriate.
The magnitude of association between PA and risk
factors will be assessed using both regression coefficients
and standardized regression coefficients, allowing within
and between site comparisons. For situations in which
the standard deviation of the variables differ across sites
standardized regression coefficients will be obtained
using pooled estimates of the mean and standard devia-
tion. Multilevel models will then be applied, in which
the site specific variable, mean PA, will be used to esti-
mate the effect of the average PA on the regression
coefficient. Multiplicative interaction terms will be
included in regression models to assess how much site
modifies the association between PA and the risk factor.
There will be three weight determinations over time:
one at baseline, a second after 12 months, and a final
assessment 2 years after entering the study. Mixed effect
models will be used to assess the association between
PA and weight change. This approach will allow for the
use of all the available information thus improving the
statistical power , and account for the within-subject
correlation of the weight determinations, as well as
adjust for subject-specific covariates that may confound
the associations. In addition, the dependence of weight
change on PA values can be modeled by the use of ran-
dom coefficients. Potential confounders will include
demographic variables and energy intake variables (e.g.
total intake, percent calories from fat, etc). Similar mul-
tilevel analysis will be used to assess the associations
between PA and change in risk factors such as blood
pressure and fasting glucose concentration.
Sample Size Estimation
A statistical justification for the selected sample size is
provided for each aim of the study. Cross-sectional asso-
ciation of PA to obesity/relative weight across sites:
Through preliminary studies, we have estimated that the
within-site correlation between PA and adiposity of indivi-
dual data could be as large as 0.6, in addition we anticipate
that the correlation between mean PA levels and mean
body composition will exceed 0.85. A one-sided Fisher’s
Z-transformation 5%-significance test will have over 55%
power to detect this correlation with 5 values correspond-
ing to the 5 sites of the study. However, after dividing our
samples by gender there will be 10 observations (2 × 5
countries), resulting in over 95% statistical power. Assum-
ing an effective sample size of 7, to account for within-
country correlations, the statistical power is 80%. Note
that additional power will be gained when weighted linear
regression methods are applied to estimate the correlation
between PA and adiposity using meta-regression.
Cross-sectional association of PA to adiposity and
CVD risk factors within site: The correlation between
PA as measured using DLW and body composition is
between -0.6 and -0.5. A correlation of -0.33 or smaller
will be detectable with 83% power and maximum prob-
ability of type I error alpha = 5% using a one-sided Fish-
er’s Z test and 75 subjects per site .
Longitudinal association of PA to change in body weight
within site: PA, measured using accelerometry, is asso-
ciated with changes in body composition. Two PA groups
will be formed by using the median total PA and moder-
ate-vigorous PA as cut points. Mean changes in weight
between groups will be compared in a longitudinal analy-
sis, adjusting for baseline values . In this method the
Figure 1 Study sites and affiliated institutions for the Modeling the Epidemiologic Transition Study (METS).
Luke et al. BMC Public Health 2011, 11:927
Page 7 of 10
following assumptions are made: the correlation between
any 2 weight determinations is ~ 0.60 and the minimum
acceptable difference between the means is 0.25 standard
deviations (i.e., effect size equals 0.25) . There will be
one baseline measure of weight and two follow-up mea-
surements. An analysis of covariance approach will be
used to adjust for baseline values. To detect the pre-speci-
fied minimum difference with 90% statistical power we
needed to accrue 460 subjects per PA group (230 in the
low and 230 in the high group).
Although the “obesity epidemic” has received enormous
attention in recent years, the manner and mechanisms
through which the energy budget is being perturbed are
still not well understood. In particular, the role of
declining PA as a causal mechanism in population-wide
increases in relative weight deserves much greater atten-
tion and new methods employed in epidemiologic
research are required before we can base causal infer-
ences or public health recommendations on something
more than assumptions. As with other diseases, interna-
tional comparisons may be particularly informative.
METS will provide information on the relative contribu-
tions of physical activity and energy intake to excess
relative weight, weight gain and diabetes risk in young
adults from five African-origin populations at different
stages of economic development.
The tests of our study hypotheses will be “two-sided”.
That is to say, the presence or absence of an association
can be given a clear interpretation. Of course, given the
body of evidence that increased PA has health advan-
tages, absence of an association would not be taken to
suggest that PA is of no value. Rather, the correct inter-
pretation would be that declining PA is not likely to be
the driving force behind the rise in obesity, nor will
modest increases in PA be adequate to reverse the
trends. These data will be critically important for the
development of meaningful public health policy targeted
at the prevention of obesity, particularly for populations
at early stages of the epidemiologic transition.
List of abbreviations
PA: Physical activity; CDC: Centers for Disease Control and Prevention; BMI:
Body Mass Index; METS: Modeling the Epidemiologic Transition Study; CVD:
Cardiovascular disease; TEE: Total energy expenditure; DLW: Doubly labeled
water; REE: Resting energy expenditure; USA: United States of America; UN:
United Nations; HDI: Human development index; TEF: Thermic effect of food;
RER: Respiratory exchange ratio; GPAQ: Global physical activity questionnaire;
WHO: World Health Organization; FTP: File transfer program; NDSR: Nutrient
data system for research; BIA: Bioelectrical impedance analysis.
Acknowledgement and Funding
The authors would like to acknowledge the site-specific clinic staff members
as well as the 2,500 participants.
METS is funded in part by the National Institutes of Health (1R01DK80763).
1Stritch School of Medicine, Loyola University Chicago, Maywood, IL, USA.
2Institute of Social & Preventive Medicine, Lausanne University Hospital,
Lausanne, Switzerland & Ministry of Health, Republic of Seychelles.3Tropical
Medicine Research Institute, University of the West Indies, Mona, Kingston,
Jamaica.4Research Unit for Exercise Science and Sports Medicine, University
of Cape Town, Cape Town, South Africa.5Kwame Nkrumah University of
Science and Technology, Kumasi, Ghana.6University of Wisconsin, Madison,
WI, USA.7MRC Epidemiology Unit, Addenbrooke’s Hospital, Cambridge, UK.
8Health Sciences Research Council, Cape Town, South Africa.
AL is the overall principal investigator for METS. PB, TEF, EVL, and JPR are the
principal investigators at each of the four study sites outside of the United
States. DAS is responsible for the doubly labeled water analyses and data
interpretation. LRD is the project coordinator. RADA is the study statistician.
DS is the social epidemiologist and RSC is the cardiovascular epidemiologist.
SB and UE are consultants aiding in the analysis and interpretation of
accelerometry data and NPS is a consultant assisting with the dietary data.
AL, LRD and RADA drafted the manuscript. All authors approved the final
The authors declare that they have no competing interests.
Received: 7 October 2011 Accepted: 14 December 2011
Published: 14 December 2011
1.Zimmet P: The burden of type 2 diabetes: are we doing enough?
Diabetes Metab 2003, 29(4 Pt 2):6S9-18.
2.Scaglione R, Argano C, Di Chiara T, Licata G: Obesity and cardiovascular
risk: the new public health problem of worldwide proportions. Expert Rev
Cardiovasc Ther 2004, 2(2):203-212.
3. Popkin BM, Gordon-Larsen P: The nutrition transition: worldwide obesity
dynamics and their determinants. Int J Obes Relat Metab Disord 2004,
4.Kohn M, Booth M: The worldwide epidemic of obesity in adolescents.
Adolesc Med 2003, 14(1):1-9.
5.James PT: Obesity: the worldwide epidemic. Clin Dermatol 2004,
6.Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, Buchner D,
Ettinger W, Heath GW, King AC, et al: Physical activity and public health.
A recommendation from the Centers for Disease Control and Prevention
and the American College of Sports Medicine. JAMA 1995, 273(5):402-407.
7. Fletcher GF, Balady G, Blair SN, Blumenthal J, Caspersen C, Chaitman B,
Epstein S, Sivarajan Froelicher ES, Froelicher VF, Pina IL, et al: Statement on
exercise: benefits and recommendations for physical activity programs
for all Americans. A statement for health professionals by the
Committee on Exercise and Cardiac Rehabilitation of the Council on
Clinical Cardiology, American Heart Association. Circulation 1996,
Physical activity and cardiovascular health. NIH Consensus Development
Panel on Physical Activity and Cardiovascular Health. JAMA 1996,
9. US Department of Health and Human Services: Physical activity and health:
a report of the Surgeon General Atlanta: US Department of Health and
Human Services, Centers for Disease Control and Prevention, National
Center for Chronic Disease Prevention and Health Promotion; 1996.
10.Institute of Medicine: Dietary reference intakes: energy, carbohydrate, fiber, fat,
fatty acids, cholesterol, protein, and amino acids Washington, DC: National
Academy Press; 2002 [http://www.nap.edu/openbook.php?
11.Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML,
Gortmaker SL: The global obesity pandemic: shaped by global drivers
and local environments. Lancet 378(9793):804-814.
12.Cook CM, Schoeller DA: Physical activity and weight control: conflicting
findings. Curr Opin Clin Nutr Metab Care 14(5):419-424.
13.Blundell JE, Stubbs RJ, Hughes DA, Whybrow S, King NA: Cross talk
between physical activity and appetite control: does physical activity
stimulate appetite? Proc Nutr Soc 2003, 62(3):651-661.
Luke et al. BMC Public Health 2011, 11:927
Page 8 of 10
14.Tremblay A, Perusse L, Bouchard C: Energy balance and body-weight
stability: impact of gene-environment interactions. Br J Nutr 2004,
Ravussin E, Bogardus C: Energy balance and weight regulation: genetics
versus environment. Br J Nutr 2000, 83(Suppl 1):S17-20.
Lakka TA, Bouchard C: Physical activity, obesity and cardiovascular
diseases. Handb Exp Pharmacol 2005, , 170: 137-163.
Bray GA: Etiology and pathogenesis of obesity. Clin Cornerstone 1999,
Wu JT, Kral JG: Ghrelin: integrative neuroendocrine peptide in health and
disease. Ann Surg 2004, 239(4):464-474.
Meier U, Gressner AM: Endocrine regulation of energy metabolism:
review of pathobiochemical and clinical chemical aspects of leptin,
ghrelin, adiponectin, and resistin. Clin Chem 2004, 50(9):1511-1525.
Margetic S, Gazzola C, Pegg GG, Hill RA: Leptin: a review of its peripheral
actions and interactions. Int J Obes Relat Metab Disord 2002,
Harris RB: Leptin–much more than a satiety signal. Annu Rev Nutr 2000,
Behavioral Risk Factor Surveillance System Survey Data. [http://www.cdc.
Blair SN, Nichaman MZ: The public health problem of increasing
prevalence rates of obesity and what should be done about it. Mayo Clin
Proc 2002, 77(2):109-113.
Hill JO: Understanding and addressing the epidemic of obesity: an
energy balance perspective. Endocr Rev 2006, 27(7):750-761.
Lake A, Townshend T: Obesogenic environments: exploring the built and
food environments. J R Soc Health 2006, 126(6):262-267.
Ferro-Luzzi A, Martino L: Obesity and Physical Activity. In Ciba Foundation
Symposium 201: The Origins and Consequences of Obesity. Edited by:
Chadwick DJ, Cardew G. New York: John Wiley 1996:207-227.
Dugas LR, Harders R, Merrill S, Ebersole K, Shoham DA, Rush EC, Assah FK,
Forrester T, Durazo-Arvizu RA, Luke A: Energy expenditure in adults living
in developing compared with industrialized countries: a meta-analysis of
doubly labeled water studies. Am J Clin Nutr 93(2):427-441.
Barro RJ, Lee JW: A New Data Set of Educational Attainment in the
World, 1950-2010. Volume NBER Working Paper No. 15902 The National
Bureau of Economic Research; 2011.
Welk GJ, Schaben JA, Morrow JR Jr: Reliability of accelerometry-based
activity monitors: a generalizability study. Med Sci Sports Exerc 2004,
Trost SG, McIver KL, Pate RR: Conducting accelerometer-based activity
assessments in field-based research. Med Sci Sports Exerc 2005, 37(11
Puyau MR, Adolph AL, Vohra FA, Zakeri I, Butte NF: Prediction of activity
energy expenditure using accelerometers in children. Med Sci Sports Exerc
Johnson RK, Russ J, Goran MI: Physical activity related energy expenditure
in children by doubly labeled water as compared with the Caltrac
accelerometer. Int J Obes Relat Metab Disord 1998, 22(11):1046-1052.
Ekelund U, Sjostrom M, Yngve A, Poortvliet E, Nilsson A, Froberg K,
Wedderkopp N, Westerterp K: Physical activity assessed by activity
monitor and doubly labeled water in children. Med Sci Sports Exerc 2001,
Brage S, Wedderkopp N, Ekelund U, Franks PW, Wareham NJ, Andersen LB,
Froberg K: Objectively measured physical activity correlates with indices
of insulin resistance in Danish children. The European Youth Heart
Study (EYHS). Int J Obes Relat Metab Disord 2004, 28(11):1503-1508.
Reed GW, Hill JO: Measuring the thermic effect of food. Am J Clin Nutr
Luke A, Dugas LR, Ebersole K, Durazo-Arvizu RA, Cao G, Schoeller DA,
Adeyemo A, Brieger WR, Cooper RS: Energy expenditure does not predict
weight change in either Nigerian or African American women. Am J Clin
Nutr 2009, 89(1):169-176.
Luke A, Durazo-Arvizu RA, Rotimi CN, Iams H, Schoeller DA, Adeyemo AA,
Forrester TE, Wilks R, Cooper RS: Activity energy expenditure and
adiposity among black adults in Nigeria and the United States. Am J Clin
Nutr 2002, 75(6):1045-1050.
Schoeller DA, Ravussin E, Schutz Y, Acheson KJ, Baertschi P, Jequier E:
Energy expenditure by doubly labeled water: validation in humans and
proposed calculation. Am J Physiol 1986, 250(5 Pt 2):R823-830.
39.Schoeller DA, van Santen E: Measurement of energy expenditure in
humans by doubly labeled water method. J Appl Physiol 1982,
Assessment of Body Composition and Total Energy Expenditure in Humans
Using Stable Isotope Techniques. IAEA Human Health Series 3 Vienna, Austria;
Weir JB: New methods for calculating metabolic rate with special
reference to protein metabolism. J Physiol 1949, 109(1-2):1-9.
Luke A, Rotimi CN, Adeyemo AA, Durazo-Arvizu RA, Prewitt TE, Moragne-
Kayser L, Harders R, Cooper RS: Comparability of resting energy
expenditure in Nigerians and U.S. blacks. Obes Res 2000, 8(5):351-359.
Luke A, Durazo-Arvizu R, Adeyemo A, al E: Resting metabolic rate is
positively associated with weight change in a lean population. Obes Res
Luke A, Adeyemo A, Kramer H, Forrester T, Cooper RS: Association
between blood pressure and resting energy expenditure independent
of body size. Hypertension 2004, 43(3):555-560.
Luke A, Durazo-Arvizu R, Cao G, Adeyemo A, Tayo B, Cooper R: Positive
association between resting energy expenditure and weight gain in a
lean adult population. Am J Clin Nutr 2006, 83(5):1076-1081.
Armstrong T, Bull F: Development of the World Health Organization
Global Physical Activity Questionnaire (GPAQ). Journal of Public Health
World Health Organization: WHO STEPS Surveillance Manual: The WHO
STEPwise approach to chronic disease risk factor surveillance Geneva: World
Health Organization; 2005.
Carriquiry AL: Estimation of usual intake distributions of nutrients and
foods. J Nutr 2003, 133(2):601S-608S.
Guenther PM, Kott PS, Carriquiry AL: Development of an approach for
estimating usual nutrient intake distributions at the population level.
J Nutr 1997, 127(6):1106-1112.
Carriquiry AL, Fuller WA, Goyeneche JJ, Dodd KW: Estimation of the Usual
Intake Distributions of Ratios of Dietary Components. Dietary Assessment
Research Series Report 5, Staff Report 95-SR Ames, Iowa: Center for
Agricultural and Rural Development, Iowa State University, Ames, Iowa
Bitvise Tunnelier. [http://www.bitvise.com/tunnelier].
Steyn NP, Nel JH, Parker WA, Ayah R, Mbithe D: Dietary, social, and
environmental determinants of obesity in Kenyan women. Scand J Public
African Region InfoBriefs. Ghana: Poverty Monitoring with the Core
Welfare Indicators Questionnaire. The World Bank Group 1999, 39.
Lukaski HC, Bolonchuk WW, Hall CB, Siders WA: Validation of tetrapolar
bioelectrical impedance method to assess human body composition.
J Appl Physiol 1986, 60(4):1327-1332.
Luke A, Durazo-Arvizu R, Rotimi C, Prewitt TE, Forrester T, Wilks R,
Ogunbiyi OJ, Schoeller DA, McGee D, Cooper RS: Relation between body
mass index and body fat in black population samples from Nigeria,
Jamaica, and the United States. Am J Epidemiol 1997, 145(7):620-628.
Schoeller DA, van Santen E, Peterson DW, Dietz W, Jaspan J, Klein PD: Total
body water measurement in humans with 18O and 2H labeled water.
Am J Clin Nutr 1980, 33(12):2686-2693.
Wang Z, Deurenberg P, Wang W, Pietrobelli A, Baumgartner RN,
Heymsfield SB: Hydration of fat-free body mass: new physiological
modeling approach. Am J Physiol 1999, 276(6 Pt 1):E995-E1003.
Moore FD, Olesen KH, McMurrey JD, et al: The Body Cell Mass and Its
Surrounding Environment Philadelphia: Saunders; 1963.
Cooper R, Rotimi C, Ataman S, McGee D, Osotimehin B, Kadiri S, Muna W,
Kingue S, Fraser H, Forrester T, et al: The prevalence of hypertension in seven
populations of west African origin. Am J Public Health 1997, 87(2):160-168.
Cooper RS, Guo X, Rotimi CN, Luke A, Ward R, Adeyemo A, Danilov SM:
Heritability of angiotensin-converting enzyme and angiotensinogen: A
comparison of US blacks and Nigerians. Hypertension 2000,
Cooper RS, Wolf-Maier K, Luke A, Adeyemo A, Banegas JR, Forrester T,
Giampaoli S, Joffres M, Kastarinen M, Primatesta P, et al: An international
comparative study of blood pressure in populations of European vs.
African descent. BMC Med 2005, 3:2.
Luke A, Kramer H, Adeyemo A, Forrester T, Wilks R, Schoeller D, Leman C,
Cooper RS: Relationship between blood pressure and physical activity
assessed with stable isotopes. J Hum Hypertens 2005, 19(2):127-132.
Luke et al. BMC Public Health 2011, 11:927
Page 9 of 10
63. Royston P, Altman DG, Sauerbrei W: Dichotomizing continuous predictors
in multiple regression: a bad idea. Stat Med 2006, 25(1):127-141.
Rosner B: Fundamentals of Biostatistics Pacific Grove, CA: Duxbury; 2000.
Frison L, Pocock SJ: Repeated measures in clinical trials: analysis using
mean summary statistics and its implications for design. Stat Med 1992,
Cohen J: Applied Power Analysis for the Behavioral Sciences New York:
Lawrence Erlbaum Associates, Inc; 1988.
The pre-publication history for this paper can be accessed here:
Cite this article as: Luke et al.: Protocol for the modeling the
epidemiologic transition study: a longitudinal observational study of
energy balance and change in body weight, diabetes and
cardiovascular disease risk. BMC Public Health 2011 11:927.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
Luke et al. BMC Public Health 2011, 11:927
Page 10 of 10