Physical activity pattern and its relation to glucose metabolism in

Transcription

Physical activity pattern and its relation to
glucose metabolism in Greenland
- a country in transition
PhD thesis
Inger Katrine Dahl-Petersen
Centre for Health Research in Greenland
December 2013
PhD thesis
National Institute of Public health, Faculty of Health Science
University of Southern Denmark
Physical activity pattern and its relation to
glucose metabolism in Greenland
- a country in transition
Inger Katrine Dahl-Petersen
Centre for Health Research in Greenland
December, 2013
Physical activity pattern and its relation to glucose metabolism in Greenland – a country in transition
Preface
Preface
This thesis is based on data from the Inuit Health in Transition study in Greenland, which aimed to
contribute to a better understanding of the health effects of the transition from a traditional lifestyle to a
modern, industrialized life. The study was conducted as a collaboration between the Centre for Health
Research in Greenland at the National Institute of Public Health, University of Southern Denmark, Steno
Diabetes Center and the Department of Health in Greenland. Karen Elise Jensen’s Foundation was the main
source of funding for the study. The PhD was funded by Karen Elise Jensen’s Foundation and the University
of Southern Denmark. The data collection was carried out between 2005 and 2010. Both as a research
assistant and during the PhD study at the Centre of Health Research in Greenland, I was lucky to be a part
of the data collection team. Participation in the data collection allowed me to travel all over Greenland and
gave me valuable insights in the study methods and procedures. It also left me with many fond and
valuable memories and stories from the participants and of the peaceful, but at times challenging, nature. I
wish to thank my supervisors Peter Bjerregaard, Marit Eika Jørgensen and Søren Brage for sharing their
great insight and experience in the world of epidemiology and population health, and for the inspiring and
stimulating discussions and talks about health among the population in Greenland, which have kept me
motivated since my first stay in Greenland in 2004 and throughout the progress of my PhD. Thanks should
also be given to the employees at Steno Diabetes Center, especially Anne-Louise Schmidt Hansen for
providing me many Thursdays in a friendly and academically inspiring atmosphere, to Stefanie Mayle and
Kate Westgate from the PA Tech team at the MRC Epidemiology Unit in Cambridge for expert assistance in
processing combined sensor data, and Andreas W. Hansen for his valuable contribution as a co-author of
one of the papers in this thesis. I have had the privilege of working with great colleagues and friends at the
National Institute of Public Health, and I am grateful for the support they have given me. Especially, I want
to thank my colleagues and friends at the Centre for Health Research in Greenland: Cecilia, Christina,
Susanne, Ingelise, Charlotte, Anni and Nina for creating a supportive atmosphere with room for inspiring
discussions and fun times also when things were busy, and especially to Susanne and Cecilia for help with
the graphical layout and Vibeke and Majken for help with the English grammar. Above all, I owe my thanks
to every single man and woman who participated in the Inuit Health Transition study in Greenland. Finally, I
wish to thank my good friends and family for their patience and encouragement and especially Peter and
Frida for dragging me out into the (wild) nature.
Inger Katrine Dahl-Petersen, December 2012.
"Tab for alt ikke lysten til at gå.
Jeg går mig hver dag det daglige velbefindende til og går fra enhver sygdom;
Jeg har gået mig mine bedste tanker til
og jeg kender ingen tanke så tung, at man jo ikke kan gå fra den."
Søren Kierkegaard
Papers
This thesis is based on the following papers:
Paper l: Dahl-Petersen IK, Hansen AW, Bjerregaard P, Jørgensen ME, Brage S.
Validation of the long International Physical Activity Questionnaire in the Arctic - measures of physical
activity in Greenland; Medicine and Science in Sports and Exercise, 2013: 45(4): 728-736A.
Paper ll: Dahl-Petersen IK, Jørgensen ME, Bjerregaard P.
Physical activity patterns in Greenland: A country in transition; Scandinavian Journal of Public Health, 2011;
39: 678–686.
Paper lll: Dahl-Petersen IK, Bjerregaard P, Brage S, Jørgensen ME.
Physical activity energy expenditure is associated with 2-h insulin independently of obesity among Inuit in
Greenland; Diabetes Research and Clinical Practice, 2013; article in press.
/ŶĐůƵĚĞĚŝŶƚŚŝƐƚŚĞƐŝƐŝŶƐƵďŵŝƚƚĞĚĨŽƌŵ͘
Academic supervisors:
Professor Peter Bjerregaard MD. National Institute of Public Health, Centre for Health Research in
Greenland, University of Southern Denmark.
Marit Eika Jørgensen, PhD, MD. Steno Diabetes Centre, Gentofte Denmark.
Søren Brage, PhD. MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge, United Kingdom.
Abbreviations
IHT 2005–2010
The Inuit Health in Transition study 2005–2010
PA
Physical Activity
PAEE
Physical Activity Energy Expenditure
MET
Metabolic Equivalence
HR
Heart Rate
Acc
Accelerometer
Acc and HR
Combined Accelerometry and Heart Rate monitoring
IPAQ-L
International Physical Activity Questionnaire, long version
BMI
Body Mass Index
WC
Waist Circumference
MVPA
Moderate and Vigorous intensity Physical Activity
LPA
Light intensity Physical Activity
OGTT
Oral Glucose Tolerance Test
IGT
Impaired Glucose Tolerance
IFG
Impaired Fasting Glucose
CI
Confidence Interval
OR
Odds Ratio
SE
Standard Error
SD
Standard Deviation
Abbreviations
Table of contents
Table of contents
Introduction ....................................................................................................................................................... 6
Aims of the thesis .......................................................................................................................................... 6
Background ........................................................................................................................................................ 8
The concept of physical activity and key definitions ..................................................................................... 8
The impact of physical activity on public health ........................................................................................... 9
The physical activity transition in populations undergoing rapid transition – what is the problem?........... 9
Differences in living conditions as a marker of the PA transition and type 2 diabetes............................... 11
Measurements of population-based physical activity in a non-Western context ...................................... 11
Material and methods ..................................................................................................................................... 14
The Inuit Health in Transition Study ............................................................................................................ 14
Places of data collection and procedures.................................................................................................... 14
Population sample ....................................................................................................................................... 15
Ethical considerations.................................................................................................................................. 18
Outcome measures and exposures ................................................................................................................. 20
Measures of physical activity ...................................................................................................................... 20
Measures of glucose metabolism ................................................................................................................ 22
Anthropometric measures .......................................................................................................................... 23
Sociodemographic variables ........................................................................................................................ 23
Social transition ........................................................................................................................................... 23
Confounders ................................................................................................................................................ 24
Data analysis ................................................................................................................................................ 24
Summary of main results................................................................................................................................. 26
Is the International Physical Activity Questionnaire valid to use in an arctic population? ......................... 26
Is there an association between physical activity patterns and social transition in Greenland? ............... 28
Is physical activity energy expenditure associated with glucose metabolism in Greenland? .................... 29
Intensities of daily life PA – an overview ..................................................................................................... 30
Discussion ........................................................................................................................................................ 32
4
Table of contents
Main findings ............................................................................................................................................... 32
Is measuring physical activity by questionnaire a feasible method in Greenland? .................................... 33
What characterizes the physical activity transition in Greenland? ............................................................. 36
Physical activity and glucose metabolism in Greenland - what is the evidence? ....................................... 38
Methodological considerations ................................................................................................................... 39
Conclusion ....................................................................................................................................................... 44
Implications for public health in Greenland ................................................................................................ 45
Implication for future research ................................................................................................................... 45
Summary.......................................................................................................................................................... 48
Dansk resuméͬ'ƌƆŶůĂŶĚƐŬƌĞƐƵŵĠ................................................................................................................. 50
Reference list ................................................................................................................................................... 5ϰ
Appendix .......................................................................................................................................................... 6ϰ
Publications ..................................................................................................................................................... 8ϰ
5
Introduction
Introduction
Since the 1940s, indigenous populations in the Arctic, including Inuit in Greenland have undergone rapid
cultural, economic and social changes accompanied by a modernization of lifestyle. Parallel to these
changes, Greenland has experienced a major health transition with substantial increases in chronic
diseases, such as type 2 diabetes (1;2). Recent studies in Greenland have shown a prevalence of type 2
diabetes at 7-10% among adult Inuit (2). Among the main causes, changes in habitual physical activity (PA)
have been suggested (2). This PhD thesis focuses on PA in relation to the social transition within the
country of Greenland and the impact on glucose metabolism. The traditional lifestyle in Greenland was
characterized by physically demanding activities, such as subsistence hunting and fishing, berry picking,
kayaking, dog sledging and transportation of water to the household. Today, these activities are still
widespread all over Greenland, but more often as a leisure activity in the larger towns. Sedentary
occupations have become more prevalent and mechanization of equipment, such as motorized boats, cars,
snow mobiles, washing machines and computers, has resulted in a less physically challenging daily life.
Sedentary leisure pursuits (e.g. TV and computer use) have become increasingly available, but also
common modern leisure-time activities, such as football, skiing, biking and fitness training, have become
popular all over Greenland, although there are large regional differences in the availability of such facilities.
A population-based survey among adult Inuit in Greenland in 1993-1994 showed that a high level of PA
during leisure time was more common in villages than in towns. The proportion of physically inactive
individuals increased by age, and men were more physically active compared with women. In total, 22% of
the population was physically inactive in leisure time during summer and winter (3). Information on PA
across the Arctic is limited, and the use of different measures and measurement tools for PA complicates
comparisons of PA patterns. Overall, the PA level is found to be lower among women than men, and to
decrease by age, but the knowledge of how living conditions, income and educational level affect PA
patterns is lacking (4-9). Only one prospective study has investigated how rapid cultural, economic and
social changes have influenced PA patterns among indigenous populations in the Arctic (10). Since the level
of adaptation to a westernized lifestyle still varies markedly within the population of Greenland, the
opportunity to study the PA transition in relation to the social transition is obvious. Such studies can
provide information to public health interventions with the aim of improving health in populations going
through a similar process of social transition. This thesis is based on investigations carried out to contribute
with novel information on PA patterns in an arctic population, and the central research question was: What
characterizes the PA transition within the adult Inuit population in Greenland and how is PA associated with
glucose metabolism.
Aims of the thesis
The overall aim of this PhD thesis was to evaluate PA patterns in an arctic population undergoing rapid
social transition and to add to the epidemiological evidence of how PA relates to glucose metabolism in an
6
Introduction
Inuit population in Greenland. Moreover, the thesis aimed to study the feasibility of a questionnaire-based
measure of PA at a population-based level in Greenland by comparing questionnaire-based information
with objectively measured PA. Three investigations formed the basis of the thesis. The aim of paper I was to
validate the long International Physical Activity Questionnaire (IPAQ-L) against accelerometry and heart
rate monitoring (Acc and HR) in the Inuit population of Greenland. The aim of paper II was to study the PA
transition among Inuit in Greenland by examining differences in PA patterns in relation to the social
transition. The aim of paper III was to analyze the association between objectively measured PA and
glucose metabolism in Inuit in Greenland.
7
Background
Background
The concept of physical activity and key definitions
Physical activity (PA) may be defined as: “any bodily movement produced by skeletal muscles that result in
energy expenditure” (11). As an interpretational framework it is important to distinguish between exercise
and PA. Exercise is a subcategory of PA and includes sports activities that are planned, structured and
repetitive in order to maintain or improve physical fitness, overall health and well-being, and often these
activities are performed at vigorous intensity. Besides exercise, PA comprises activities of daily life involving
any bodily movement as well as activities such as active transportation (walking and biking), household
activities, and occupational PA. These activities are normally unstructured activities, performed at varying
levels of intensity. The focus of this PhD thesis is on the habitual PA during daily life. PA can be categorized
into domains of daily life: leisure time, household and gardening, occupation and transportation, and
comprises four different subdimensions: frequency, duration, intensity and type (12). The thesis will focus
on all four dimensions and domains of PA. Frequency relates to how often or how many bouts of PA are
performed e.g. daily, weekly or monthly. Duration refers to time spent on PA, most often described in
minutes or hours of PA. Intensity refers to how much effort is required to perform the specific activity, e.g.
rate of energy expenditure per unit of time. The intensity can be expressed relatively in percent of maximal
oxygen uptake (VO2max), as resting metabolic rate (RMR) or as absolute intensity, most often expressed in
MET (Metabolic Equivalent Task), with 1 MET corresponding to a standard value for the resting metabolic
rate; 3.5 mL O2·kg-1·min-1 (13;14). PA can be divided into light, moderate and vigorous intensity according
to the rate of energy expenditure. Moreover, energy expenditure of specific types of PA can be quantified.
Widely accepted is the use of the Compendium of physical activities now presenting 821 MET values for
specific activities of daily life (13-15). Figure 1 illustrates the continuum of PA and the corresponding rate of
energy expenditure in MET. The type of PA refers to the specific PA behaviour, such as running or walking,
or the classification of an activity into aerobic or anaerobic. Duration and frequency can be multiplied
providing total amount of time spent on PA. These durations can be multiplied with the intensity of each
activity type or category (activityi) and added up across activities (∑Durationi x frequencyi x intensityi); this
sum is referred to as the total PA energy expenditure, expressed for example in MET-hrs per week, kcal per
week, or kJ/kg/day (16).
8
0.9-1.0 MET
Sleep
1.0 - 1.5 MET
1.5 - 3.0 MET
Sedentary
Light intensity
3.0 - 6.0 MET
>6 MET
Moderate
intensity
Vigorous
intensity
Background
Figure 1. Overview of the continuum of PA and the energy expenditure corresponding to the specific activities (14).
The impact of physical activity on public health
Worldwide, it has been estimated that physical inactivity causes 6-7% of the burden of disease from major
non-communicable diseases like coronary heart disease and type 2 diabetes, and 9% of premature
mortality (17). Inactivity is concluded to be a risk factor similar to that of smoking or obesity (17). Since the
study of London transport workers in 1953 showed that the physically active conductors on double-deckers
were at lower risk of coronary heart disease compared with the sedentary drivers (18), the number of
studies on PA and health has substantially increased, and the evidence has been summarised in several
consensus documents (19-21). Evidence from a recently published prospective cohort study showed that
individuals who performed as little as 15 minutes a day or 90 minutes a week of moderate-intensity
exercise had a 14% reduced risk of all-cause mortality and a 3-year longer life expectancy compared with
inactive individuals in all age groups, for both men and women and for those with cardiovascular disease
risks (22). A positive dose response effect of total PA on health has also been documented (23;24). A metaanalysis concluded that some PA is better than none, and that additional health benefits occur with more
PA (25). Studies have identified a great potential in increasing the level of PA, particularly among the most
inactive individuals (22); however, whether there is a specific threshold for the effect of PA energy
expenditure (PAEE) on health and how specific intensities influence health still need further clarification.
This information can be used in recommendations for PA. In Greenland, the current National
recommendation is one hour of daily PA for both adults and children, however, without specifying the
recommended level of intensity (26).
The physical activity transition in populations undergoing rapid transition – what is the
problem?
In Greenland, social changes started to evolve rapidly at the beginning of the 20th century when cod fishing
was replacing the traditional hunting of sea mammals as the main livelihood of the Inuit, and cod were sold
for cash (27). After World War 2, Greenland went through rapid cultural, economic and social changes
characterized by population movement from small villages to larger towns, changes in living conditions and
increased availability of formal education (1;27;28). These changes were accompanied by a more
modernized lifestyle, especially in the larger towns. Parallel to these changes, Greenland experienced a
major health transition with a gradual reduction in the prevalence of tuberculosis and acute infectious
diseases, paralleled by a substantial increase in chronic lifestyle diseases, such as type 2 diabetes and
obesity, and increasing prevalence of mental health problems, such as youth suicides, and alcohol problems
(1;2;29). In this regard, the history of Greenland shares similar traits with the epidemiological transition
9
Background
among populations undergoing rapid social, cultural and economic changes in other parts of the world (3032). In the framework of the epidemiological transition, the physical transition explains how rapid changes
in PA have occurred in parallel with the increasing prevalence of obesity and other chronic diseases (33).
According to Katzmarzyk and colleagues “the physical activity transition seeks to explain the potential
effects of changes in PA on health and life expectancy in countries experiencing rapid economic
development” (33). The consequences of the PA transition are relevant for all populations, but may be
more marked in populations experiencing rapid social change, such as Inuit in Greenland. Little is known
about how PA has changed in relation to the agrarian, industrial and technological breakthrough. Among
indigenous populations in the Arctic, a cross-sectional comparison among the Yakut of Siberia found that
total energy expenditure (TEE) adjusted for body mass was correlated with participation in subsistence
activities, such as hunting and fishing. Individuals with a traditional lifestyle reflecting participation in
subsistence tasks had higher energy expenditure than individuals with a more modern lifestyle, indicating a
decrease in PA with modernization (34). Contrary, a study among Hadza hunter-gathers in Tanzania
presented similar daily energy expenditure as their Western counterparts and suggested the energy
expenditure to be independent of cultural differences (35). The impact of the transition from a traditional
hunter-gatherer lifestyle to a more Western lifestyle on physical fitness levels has only been exemplified in
one prospective study a 20-year study in an Inuit community in the Northwest Territories, which showed a
temporally decreasing level of fitness along with rapid acculturation and an increasing sedentary lifestyle
(10;36).
Physical activity and glucose metabolism among indigenous populations in the Arctic
Worldwide, the prevalence of diabetes is expected to increase from 4.0 to 5.4% (35% increase) between
1995 and 2025, with a proportionally greater increase in developing countries and a considerable excess of
diabetes in urban areas by 2025 (37). Physical inactivity is a strong and well-known risk factor for type 2
diabetes (38-43). The potential of PA in the treatment of type 2 diabetes is also well established in several
larger intervention studies (44-47). Most of the evidence on PA and metabolic risk is based on studies in
Western populations, and only a few studies have investigated the association among indigenous
populations in the Arctic. One study showed a positive effect of PA on fasting insulin concentrations in a
subarctic native Canadian population (48). Another study demonstrated an association of PA with the
prevalence of Impaired Glucose Tolerance (IGT) and diabetes in Greenland (2), and a study among Yup’ik
Eskimos and Athabaskan Indians in Alaska showed that a moderate and high level of PA were associated
with a lower prevalence of glucose intolerance compared to a reference group with a low level of PA (49).
In Greenland, physical inactivity due to a decrease in subsistence hunting and fishing activities was
suggested to explain, that Westernization was found to increase the metabolic risk for men only (50).
Contrary, another study in Greenland found a higher prevalence of type 2 diabetes and glucose intolerance
in rural areas compared with towns despite a higher level of PA in rural areas (66). These studies are all
based on self-reported PA, and only a few studies (51-56), all conducted in non-Arctic populations, have
reported on objectively measured free-living PA and its association with glucose metabolism. Information
10
Background
from such studies is important as a part of the prevention of further increases in type 2 diabetes in this
population.
Differences in living conditions as a marker of the PA transition and type 2 diabetes
To examine differences in PA patterns in relation to epidemiological transition changes or modernization,
most research has used an urban-rural dichotomy as a model, mainly since longitudinal PA data are almost
non-existent. A review of studies in developing countries summarized that both men and women living in
urban areas were more likely to be inactive compared with those living in rural areas (57). Studies in
populations undergoing transition have found that urban living is associated with lower PA and higher
prevalence of pre-stages of- and type 2 diabetes (52;58-63) and research in developing countries and
countries undergoing rapid transition has shown that the risk of impaired glucose metabolism increases
with urbanization (64;65). This is contrary to Greenland, where a population-based study showed higher
prevalence of type 2 diabetes and glucose intolerance in rural areas compared with towns (66). In addition
to current residence, a study found that both lifetime exposure to an urban environment and recent
migration history influenced the association between obesity and diabetes (67). Overall, there are
considerable differences in the definition and measurements of urbanity, modernization and social change,
and in a review investigating how urbanization has been measured, it was emphasized that measures of
urbanization as a process are needed to obtain more detailed information on changes in urbanicity and
impact on health (68).
Detailed information about the disease patterns of indigenous peoples in the North has only during the last
generation become available, and, therefore, the health impact of social change can only be studied at the
ecological level. In a recent study among Inuit in Greenland, we examined the secular differences in the
health outcomes between two population-based surveys among adult Inuit in Greenland in 1993-1994 (N
=1,580) and 2005-2009 (N=2,834). Furthermore, we defined and ranked six subgroups; from participants at
a presumed early stage of social transition (more traditional) to those at a later stage (more modern),
defined from current and childhood residence in a village or town, family job type, and education. We
compared the distribution of socioeconomic, behavioral, and clinical/biochemical risk factors for
cardiovascular disease among these groups, using data from the Inuit Health in Transition Greenland Survey
2005-2009, with the secular trends found from the two surveys. We found that in the absence of
longitudinal data, cross-sectional data could be used, although with caution, to mirror social change for
selected analyses of cardiovascular risk (69). This grouping was used in paper II as a proxy for changes in PA
patterns along with the social transition.
Measurements of population-based physical activity in a non-Western context
Using adequate measures of PA is fundamental in the assessment of PA, whether the purpose is to
measure time trends, associations with health outcomes or to evaluate interventions to promote PA. When
interpreting the results on PA one must take the quality of the measurement tools into consideration (70).
Properties, such as validity, reliability and responsiveness, are not always assessed or they have been
11
Background
studied differently, which makes it difficult to rate one questionnaire better than the other (70;71). Further,
the diversity of questionnaires available are substantial due to the fact that different questionnaires have
been developed for different purposes (e.g. surveillance, activity group categorization, etiology). The
population-based studies of PA worldwide are still mainly based on self-reported information obtained
through interviews or self-administered questionnaires. However, most questionnaires have been
developed for use in non-indigenous populations, and the reliability and validity are far from always
examined in the specific cultural setting where the questionnaire is used. The International PA
Questionnaire was developed to measure PA in different cultural settings and has been widely used
worldwide (72). The questionnaire exists in a short (IPAQ-S) and a long (IPAQ-L) form and as interviewer- or
self-administered. The short form is recommended for national monitoring (7 items), whereas the long
version is more comprehensive (27 items) and assesses time spent at different intensities of PA within four
domains of daily life: transportation, work, leisure time, and domestic activities (73). The IPAQ has been
translated, adapted, used and validated in several populations, including populations undergoing transition
(74-77). Craig and colleagues found in a 12-country evaluation that the IPAQ was as reliable and valid as
other questionnaires. However, the questionnaire showed different validity used in different populations
(78), which underlines the need to assess the measurement properties in the specific target population.
The gold standard for measuring PA energy expenditure (PAEE) is a combination of doubly labelled water
(DLW) and measurement of resting metabolic rate; however, this method is not feasible to use in
population-based surveys and does not provide information on duration, domains and intensity (79).The
advancing technological possibilities have resulted in increasing possibilities to use device-based measures
on a large study population (80). Most common is the use of different kinds of accelerometers, pedometers
and heart rate monitors. Accelerometers (uni- or tri-axial) provide measures of biomechanical intensity,
duration and frequency, and has been shown to provide greater precision when compared to self-reporting
of the total amount of PA and energy expenditure spent on specific activities (81) but also present
limitations regarding information on certain activities, such as upper-arm activities, kayaking, weightlifting
and cycling as well as high-intensity PA, and provide no information about the domain in which the activity
is performed (82;83). Heart rate monitoring can be used as an objective measure of energy expenditure
based on the premise that heart rate and oxygen consumption are linearly related (84;85). However, heart
rate is easily influenced by factors such as medicine, temperature and fitness, and is most suitable for
measuring activities at high intensity(84). The combination of accelerometry and heart rate monitoring has
been shown, in most cases, to provide a more precise and accurate estimation of the energy expenditure
for PA among both adults and children, compared with each of the methods used alone, and this method
has been validated in a non-Western context (86-88). However, the method also presents methodological
challenges, such as wear-time issues and cost. Figure 2 presents an overview of different methods and the
inverse relationship between validity and feasibility (Søren Brage, personal communication).
12
Background
Measuring Physical Activity – Levels of Sophistication
Room
Calorimetry
Precision
Doubly Labeled Water
Indirect Calorimetry
Heart Rate
Movement sensors
Self-report
Ease of Assessment
Figure 2. Overview of different methods and the inverse relationship between validity and feasibility. Søren Brage,
personal communication.
13
Material and methods
The Inuit Health in Transition Study
The present PhD thesis is based on data from the Inuit Health in Transition Greenland Survey 2005-2010,
which is a part of an international collaboration across the Arctic including Inuit in Greenland, Canada
(Nunavik and Nunavut) and Alaska. The Inuit Health in Transition Study (IHIT) was designed as a longitudinal
cohort study with the purpose of studying the interaction between the environment and genetic factors on
the health and disease pattern of the Inuit in the regions of Greenland, Canada and Alaska. Thus, the
project in Greenland is part of an international study with data collection in several villages and towns in all
three countries. The project aimed to contribute to a better understanding of the health effects of the
transition from a traditional lifestyle to a modern, industrialized life, which takes place in most present-day
developing countries. One of the specific aims of the study was to assess risk factors for cardiovascular
disease and diabetes. This PhD thesis is based on the initial cross-sectional data. A follow-up has been
scheduled in 2014 for both Nunavik and Greenland.
Places of data collection and procedures
Figure 3. Map of Greenland with study communities, names of the towns included. Inuit Health in Transition
Greenland Survey 2005-2010.
Greenland – or Kalaallit Nunaat in Greenlandic - is the world’s largest island and a country in the Arctic with
a population of about 57,000, of whom 90% are ethnic Greenlanders (Inuit). Genetically, Greenlanders are
14
Inuit (Eskimos) with a mixture of mainly Danish genes, and are genetically and culturally closely related to
the Inuit/Iñupiat in Canada and Alaska and, somewhat more distantly, to the Yupiit of Alaska and Siberia.
Only 18% of the total area of the island is free of ice. Greenland’s 80 communities are all located on the
coast and are divided into towns (with population ranging between 469 and 15,469) and villages (with
population ranging from less than 10 to about 550) with no connecting roads.
The data collection took place from 2005 to 2010, both during summer and winter time. For logistical
reasons it was not possible to distribute the data collection in the specific communities throughout the
year. With the exception of Upernavik, Tasiilaq and Qaanaaq, the towns were visited by public transport
(flight) and the villages were visited on three expeditions by a chartered boat (m/s Kisaq) (figure 3). Data
was collected by a team of local persons responsible for the recruitment of participants, a supervisor, one
or two laboratory technicians, 2-4 interviewers, and two clinical assistants. The participants were informed
about the investigation by a personal letter, and after the arrival of the team they were contacted by the
person responsible for recruitment. The participants were asked to show up fasting (i.e. at least 8 hours
without eating or drinking), they were informed about the investigation and signed an informed consent.
The participants went through a 2-hour oral glucose tolerance test, interview, filled in a questionnaire,
went through various clinical tests and were provided with the Actiheart device (combined accelerometer
and heart rate monitor). The interviews were conducted in both Greenlandic and Danish according to the
choice of the participant. After 2 hours, another blood sample was drawn. At the end of the session,
participants were informed about the results of the investigation. When the Actiheart device was returned,
a compensation of DKK 200 was paid to each participant.
Population sample
Participants for the Inuit Health in Transition Greenland Survey were selected as a stratified random sample
of adults aged 18 years and older and born in Greenland or Denmark. Greenland was divided into strata
based on geography (Southwest coast; Central West coast; Northwest coast; East Greenland; North
Greenland) and community size (towns with ≥ 2000 inhabitants; towns with < 2000 inhabitants; and
villages). From each of these strata one or more towns and 2-3 villages were selected for the study as being
representative of the stratum with regard to living conditions. A random sample was drawn from the
central population register to obtain around 300 participants from each town; this number represents the
practical limit for a research team during a 4-6 week visit. Villages were chosen at random in the strata, and
in the selected villages all adults were invited to participate. We collected data in 9 towns and 13 villages in
Greenland. At the study location, the invited participants were contacted by telephone, person-to-person
or contacted by asking their neighbor of their whereabouts. The final sample was revised to exclude
participants no longer living in the community, pregnant women and deceased persons. Ethnicity as
Greenlander or Dane was determined at enrolment, based on the primary language of the participant and
self-identification. The current PhD thesis focuses on Greenlanders only. According to community size, the
participation was 61.4% in Nuuk (the capital), 65.1% in other large towns, 69.9% in small towns and 68.5%
15
in the villages (p<0.001). Participation rates also varied by age and sex. Women were more likely than men
to participate, and particularly young men were under-represented. The reasons for non-participation can
be seen in the flow chart (figure 4).
16
The Inuit Health in Transition – Greenland Survey 2005-2010
N=6,016
Stratified random sample of adult inhabitants in Greenland (18+) born in Greenland
(‘Greenland Inuit’) or Denmark (‘Danes’). Drawn from the central population register.
N=1,005
Reduction of initial sample:
Moved (746); Not in study base* (144);
Pregnant (60); Deceased (55)
*) e.g. unknown in the community, in
prison, irrelevant ethnicity
N=356
N=5,011
Revised sample
Exclusion of ethnic Danes
N=4,655
Revised sample ‘Greenland Inuit’
N=1,553
Non-participation:
Doesn´t want to participate (803); Illness
or disability (115); Hunting, fishing or in
mine camp (97); Out of town for other
reasons (24); No contact* (514)
*) e.g. Interviewers gave up after 2-3
attempts to get in touch
N=3,102
Participants participating in the health survey. The study surveyed 9.2% of the adult,
Greenland born population.
N=6
Missing information
N=79
Mistakes in values or items
N=2,053 (66.2%)
Participants with ACC AND
HR monitoring
N=166
Participants with step test.
Included in sample
N=58
Participants excluded before
data analyses due to poor
quality of both ACC and HR
N=1,995 (64.3%)
Participants with information
on ACC AND HR - revised
N=2,079**
Participants in paper II.
N=95
Flagged observations,* Included
in the analysis
N=97
Exclusion of participants if ACC
AND HR<24 hours
N=1,898 (61.2%)
Participants with
information on ACC AND HR
- revised
N=353
Exclusion of participants if ACC
AND HR<48 hours
N=1,545 (49.8%)
Participants with
information on ACC AND
HR revised
N=3,017**
Participants available for analyses
of physical activity.
Data processing are described in
appendix III and paper I
N=47
Flagged observations,*
Included in the analysis
* Recordings were flagged for the reason of a poor
HR or acceleration signal, absence of data during
sleep (required for estimation of SHR, which is used
in activity intensity calculations), heart disease or
calibration errors. Another branch model or heart
flex model were suggested for some of the flagged
records. Flagged observations are included in the
analyses (See Appendix III for further details).
** In Paper II, the participants with more than 6,720
minutes of weekly reported time spent on total PA
were excluded, according to IPAQ guidelines. In
Paper I we have used a scaling of these
observations instead.
Figure 4. Overview of the study sample and drop-outs.
17
Ethical considerations
The study was approved by the Ethical Review Committee for Greenland. Participants were informed about
the study objectives and the data collection procedures orally and in writing, and accordingly gave their
written informed consent.
18
19
Outcome measures and exposures
Measures of physical activity
The International Physical activity Questionnaire – long version
In this PhD thesis information on PA was collected using a modified version of the intervieweradministrated seven-day International Physical Activity Questionnaire (IPAQ) (long version) (IPAQ-L).
Participants were asked to report time spent on PA in the previous seven days: how often (number of days
per week) and for how long (average duration per day). Questions were asked separately for vigorous
intensity, moderate intensity and walking in the four domains: work, transportation, domestic and leisure
time. Participants were also asked to report number of days and time spent sitting during the week and in
the weekend. The original English version of the PA questionnaire was translated into Greenlandic and
back-translated by two interpreters bilingual in Danish and Greenlandic and familiar with Greenlandic living
conditions. The questions were adjusted to arctic living conditions by replacing some of the activity
examples with culturally relevant examples based on a pilot study comparing IPAQ-L and a short
questionnaire with combined Acc and HR in Greenland. In the domestic domain we combined the two
questions concerning moderate intensity (outdoor and indoor activity) into one; gardening is non-existent
in arctic living conditions, and common activities such as getting fishing equipment ready take place both
inside and outside the house. Data were initially scored according to guidelines from the IPAQ group (89).
Some exceptions were made, as described in detail in paper I. An overview is also provided in appendix IV.
PA energy expenditure was calculated by multiplying time reported (minutes/week) by the net metabolic
cost of each activity, which was expressed in metabolic equivalents (METs). The net metabolic cost of each
activity was assigned according to the PA Compendium’s gross MET values (13), subtracted by 1 MET to
account for resting metabolic rate (RMR). An estimate of total daily sedentary time was calculated from
time spent sitting, such as TV and computer use and reading. In paper I we added 8 hours as presumed
time spent sleeping (sleep information not included in IPAQ-L) (Appendix 1, the Greenlandic version of
IPAQ-L).
20
70.
During the last 7 days, on how many days did you do vigorous physical
activities in your home? (for instance heavy lifting, shovelling snow, digging,
fetching water)
BBBBBGD\VSHUZHHN
'LGQRWGRYLJRURXVSK\VLFDODFWLYLW\DWKRPH →JRWRTXHVW
71.
How much time did you usually spend on one of those days doing vigorous
physical activities in your home?
BBBBBKRXUVSHUGD\
BBBBBPLQXWHVSHUGD\
Figure
5. Example of question in IPAQ-L. Vigorous PA in the domestic domain. PA examples adapted to Greenlandic
living conditions.
Combined accelerometry and heart rate monitoring
As a part of the clinical health examination we issued a combined accelerometer and heart rate monitor
(ACC and HR) (Actiheart®, CamNtech Ltd, Cambridge, UK) to a subgroup of the participants all over
Greenland (n=2053). The monitor was set up to measure acceleration and heart rate at 30-second intervals
and attached to the participant’s chest by two standard ECG electrodes (MXC55 MediMax UK)(figure 6).
The participants were told to leave the monitor on for 24 hours a day, also for sleep and showering. A
subgroup of participants conducted an individual calibration test (8-minute step test). Step tests were used
to define a population-specific calibration equation of the heart rate-activity energy expenditure
relationship. Due to study logistics (travel distances, weather conditions and the data collection time
schedule), only limited time was available at each study location, especially for data collection in villages.
Together with a finite stock of monitors, this explains why not all participants were given a monitor, and
why the length of recordings from some participants was of shorter duration. A detailed description of data
processing and sample is available in paper I and appendix III. Caloric intensity of PA was estimated by
combining the acceleration-based estimate of intensity (90) with the heart rate-based estimate from the
population-specific equation in a branched equation modelling framework (91). Briefly, this method
predominantly uses the accelerometer estimate during low levels of heart rate and movement, and the
heart rate estimate when both heart rate and acceleration levels are high, with equal weighting for other
conditions (appendix III for details on branched equation). Resulting time series of activity intensity (in
J/min/kg) were summarised into total PAEE (in kJ/kg/day) and time spent at different intensity levels
(sedentary as <1.5MET, moderate as 3-6MET, and vigorous as >6MET). We included individuals with >48 hrs
of monitor wear data.
21
Figure 6. The placement of a combined accelerometer and heart rate monitor.
We validated the interviewer-administered long form of the International Physical Activity Questionnaire
(IPAQ-L) modified and adapted to arctic living conditions (paper l) and utilized it for analyses of patterns of
PA in relation to social transition in paper II. The combined accelerometer and heart rate monitor was
applied in a subsample of the participants. In paper I this method has been used as a criterion measure for
validation, and in an addition to paper II as a descriptive outcome measure across transition categories, and
in paper III as exposure in an etiological analysis of the association with precursors of type 2 diabetes.
Interview with main interviewers
We conducted individually based interviews with the main interviewers about their experiences of
interviewing about PA and how the questions were interpreted by the participants. Moreover, preliminary
results were presented and possible explanations were highlighted by the interviewers. The answers from
the interviewers were used in paper I and as background knowledge on how the concept of PA, including
intensities and domains, is being interpreted in Greenland.
Measures of glucose metabolism
After a minimum of 8 hours of fasting, participants underwent a standardized 2-hours oral glucose
tolerance test (75 g), except for those with known type 2 diabetes at the time of health examination.
Fasting and 2hr blood samples were taken. Plasma glucose was measured fasting, plasma was separated
and frozen at –20°C and transported to one central laboratory for measurement of plasma glucose. Serum
insulin was analyzed with a flouroimmunoassay technique. The inter assay precision CV was 6%. Non
fasting participants (self-reported) or participants with known type 2 diabetes were not included in the
further analysis involving glucose and insulin parameters. Glucose tolerance; impaired fating glucose (IFG),
impaired glucose tolerance (IGT) and type 2 diabetes were classified according to WHO criteria (table 1).
22
Table 1. WHO diagnostic criteria for diagnosis of diabetes mellitus and intermediate hyperglycemia (World Health
Organization 1999).
WHO Diagnostic Criteria
IFG
Fasting plasma glucose from 6.1 to 6.9 mmol/l and 2h
plasma glucose <7.8 mmol/l
IGT
Fasting plasma glucose <7.0 mmol/l and 2h plasma
glucose ≥7.8 mmol/l and <11.1 mmol/l
Fasting plasma glucose ≥7.0 mmol/l or 2h plasma glucose
≥11.1 mmol/l
Diabetes
Anthropometric measures
Height (nearest 0.1 cm) and weight (nearest 0.1 kg) were measured with the participants wearing
underwear. BMI was calculated as weight/height2 (kg/m2). Waist circumference was measured midway
between the rib cage and the iliac crest, hip circumference at its maximum on the standing participant.
Weight was measured on a standard electronic clinical scale. Bioimpedance and calculation of fat
percentage were performed on a leg-to-leg Tanita TBF-300MA. Based on a single reading, fat percentage
was calculated by the internal algoritm of the device, which is based on height, weight, sex, impedance and
age; body type was set to “standard”.
Sociodemographic variables
From the interviewer-administrated questionnaire, residence at age 10 was obtained and recoded into
residence in village or town. Job type was determined from questions about job title of participant and
spouse. Formal education was determined from questions about highest school education attained and
further vocational or academic education and recoded as primary school/high school only, short vocational
education (less than three years), and longer vocational/academic education. Place of residence was
divided into the capital of Nuuk, villages and towns.
Social transition
Individuals were divided into six groups defined from occupation type, education, and place of residence
(present and 10 years old). This variable is used in paper II as a proxy for secular changes:
A. hunters and fishermen in villages;
B. other inhabitants of villages;
C. blue-collar migrants (inhabitants of towns, with no vocational education, having lived in villages at
age 10);
D. other blue-collar participants (inhabitants of towns, with no vocational education, having lived in
towns at age 10);
E. intermediate (inhabitants of towns, with short vocational education);
F. professionals (inhabitants of towns, with longer vocational or academic education).
23
In order not to misclassify participants who had not yet finished their education and to minimize the
proportion of participants outside the workforce, analyses only included those aged 25-64 years.
Confounders
In papers I and II, we stratified the analyses by sex, age group and place of residence. In analyses of the
association between PAEE and glucose metabolism outcomes (paper III), information on smoking habits,
family history of diabetes, sex, age, BMI, waist circumference (WC) and fat percentage was included in
regression analyses to address the issue of confounding and mediating by these variables.
Data analysis
The analyses were performed in STATA 10-12 and SPSS 18.
Paper I
The association between questionnaire- and monitor-based PA estimates was examined by the nonparametric Spearman rank correlation coefficient (ρ). Level of agreement was examined by modified BlandAltman plots (Bland and Altman ) (92). Bland and Altman recommend graphical presentations (plots) for
method comparison, so that the error structure can be explored throughout the range of the variable of
interest. We used a modification of the classic Bland-Altman plot by plotting the difference between the
measurements (IPAQ-L minus Acc and HR) against the objective estimate; with lines indicating the median
difference (median bias) and 95% limits of agreement (2.5 and 97.5 centiles of the difference). Median
instead of mean and centiles of the difference instead of 1.96SD of the difference were used due to the
non-normal distribution of data (non-parametric). Moreover, we chose to plot the difference against the
absolute measure of PA by accelerometry and heart rate monitoring, because we considered this monitorbased measurement as a more accurate and precise representation of the true underlying exposure,
compared with the questionnaire data. The differences of the medians were analysed by a Wilcoxon
signed-rank test. Sensitivity analyses were performed including only participants with ≥72 hours Acc and HR
of valid monitoring data.
Paper II
Time spent on PA was presented in median hours per day with interquartile ranges for each domain of PA
as well as for total PA. Differences in time spent on PA across social transition groups were tested using a
multiple linear regression model with time spent on PA as dependent variable. A square root
transformation of time spent on PA was applied in order to approximate a normal distribution of the
variable. The analyses were stratified on sex and adjusted by age. A test for linear trend in PA across the six
transition groups was applied adjusted for age (Likelihood-ratio; STATA version 10). Moreover the
proportion of participants that did not report any time spent on PA in the specific domains of PA was
presented. Time spent on moderate and vigorous intensity PA was analyzed. No transformation was
applied for time spent on sedentary activity.
24
Paper III
Associations between PAEE and glucose outcomes were analyzed in multiple linear regression models.
Potential confounders and mediators were chosen a priori: age, sex, WC, family history of diabetes and
smoking were included stepwise. Three models were presented; model A: no adjustments; model B: age
and sex adjustments and model C: further adjustments by WC. Interaction terms were included, and
models with and without interaction were compared using a log likelihood ratio test. The variable PAEE2
was included to test and adjust for nonlinearity. The distribution of outcome variables was graphically
viewed before analysis (qqplot), and a model control was performed to test if the variance of the residuals
was normally distributed. Accordingly, fasting and 2-hour insulin concentrations were logarithm
transformed before analysis and back-transformed and reported as percentage decrease or increase.
Impaired fasting glucose (IFG), impaired glucose tolerance (IGT) and type 2 diabetes were classified
(dichotomized). Logistic regression models were utilized to compare individuals with diabetes versus
individuals with Normal Glucose Tolerance (NGT), individuals with IGT versus NGT+IFG and individuals with
IFG versus NGT.
25
Summary of main results
Is the International Physical Activity Questionnaire valid to use in an arctic population?
This validation study is based on PA data from IPAQ-L and Acc and HR monitoring (n=1508 adult Inuit).
Questionnaire-based PAEE was moderately correlated with accelerometry and heart rate monitored PAEE
(r=0.20–0.35, p<0.01). The agreement analysis showed that the median difference for the level of PAEE
measured by the two methods was small and indicated a moderate agreement between the two methods;
however, 95% limits of agreement were wide. This pattern was similar for subgroup analyses of sex, place
of residence (Nuuk, town, village) and age groups. A weak correlation was found for questionnaire-based
time spent at different intensities of PA (moderate and vigorous) and sedentary time versus Acc and HR
monitoring (r=0.11–0.31). The agreement plots showed that time spent at moderate intensity PA was
substantially over-reported by IPAQ-L when walking was included as a moderate intensity activity
(>1.5hrs/day, p<0.001); however, the agreement was substantially better when excluding walking (figure
7).
Table 2. PA characteristics. Self-reported (IPAQ-L) and objectively measured PA (Acc and HR) presented as daily
physical activity energy expenditure (PAEE). Inuit in Greenland, n=1508.
Total PAEE
P value
(kJ/day/kg)
Sex
Men n=659
Self-report
51.7
23.6-97.0
0.2
Objective measure
56.6
40.3-75.5
Women n=849
Self-report
47.3
24.9-76.9
0.002
Objective measure
45.7
34.2-60.1
Place of residence
Nuuk n=323
Self-report
45.9
24.2-80.9
0.3
Objective measure
50.9
36.6-64.1
Towns n=906
Self-report
48.4
23.6-85.0
0.06
Objective measure
49.6
36.2-67.8
Villages n=279
Self-report
50.0
30.0-86.3
0.02
Objective measure
49.4
36.8-66.2
26
Total PAEE
(kJ/day/kg)
Table 2 (continue)
Age groups
18–44 years n=829
Self-report
P value
57.7
33.2-92.3
57.8
44.6-75.2
Self-report
47.5
23.3-85.7
Objective measure
47.2
34.3-60.1
Self-report
31.6
15.3-57.7
Objective measure
34.3
22.5-47.5
Objective measure
0.04
45–54 years n=349
0.02
55+ years n=330
Women (moderate intensity, walking not included)
-10
-10
IPAQ-(Acc+HR)
-5
0
5
IPAQ-(Acc+HR)
-5
0
5
10
10
Men (moderate intensity, walking not included)
0.6
0
2
4
6
Hours per day
8
0
Men (moderate intensity, walking included)
2
4
Hours per day
6
8
-5
-5
IPAQ-(Acc+HR)
0
5
10
IPAQ-(Acc+HR)
0
5
10
15
15
Women (moderate intensity, walking included)
0
2
4
6
Hours per day
8
0
2
4
Hours per day
6
8
Figure 7. Median difference between self-reported and objectively measured time spent at moderate intensity PA
(IPAQ-Acc and HR) plotted against (Acc and HR) stratified on sex (presented with and without walking included). The
lines represent median and 2.5 and 97.5 centiles. Inuit in Greenland, n=1508.
27
Is there an association between physical activity patterns and social transition in
Greenland?
Based on data from IPAQ-L (n=2079 adult Inuit), we found that total age-adjusted hours spent on PA were
significantly higher among hunters and fishermen living in villages compared with wage earners with long
vocational or academic education living in towns (men p<0.001; women p=0.002); however, no significant
linear trend in relation to social transition was shown. For men and women in the latest stage of social
transition, men spent significantly less time on occupational PA and women significantly less time on
domestic PA compared with men and women with the most traditional lifestyle and a linear negative trend
was found in the level of PA by stage of social transition for men (p=0.01) and for women (p=0.06).
Significantly less time was spent on PA during transportation for men and women in the latest stage of
social transition compared with the earliest stage (men p=0.02, women p=0.01). No significant differences
were found for time spent on leisure time PA in relation to social transition. The average time used on
sedentary behavior increased along the stages of social transition (p<0.001). In preliminary unpublished
analyses, we examined the distribution of time spent at different intensities of PA and total PAEE in relation
to the six transition groups based on data from Acc and HR monitoring. The results are presented in table 3
and table 4. Overall the results indicated a linear trend for decreasing PAEE with stages of social transition
for men, but not for women. No significant linear trend was identified for time spent at different intensities
of PA, although a borderline significant trend was found for decreasing time spent at vigorous intensity PA
for men only (p=0.08).
Table 3. Total PAEE and daily hours spent at different intensities of PA across transition groups. Results based on Acc
and HR monitoring (men). Unpublished and preliminary analyses. Inuit Health in Transition study in Greenland.
Men N=512
Total daily PAEE
MVPA
Moderate PA
Vigorous PA
Light PA
kJ/kg/day
(>3METs)
(3–6METs)
(>6 METs)
(1.5–3 METs)
Median hours
Median hours
Median hours
Median hours
Median (IQR)
(IQR)
(IQR)
(IQR)
(IQR)
A Hunters/fishermen
69.1 (42.8-77.9)
2.4 (1.3-3.0)
2.1 (1.1-2.5)
0.1 (0.04-0.4)
7.3 (6.1-8.8)
B Other villagers
56.6 (47.5-75.8)
2.2 (1.2-3.2)
1.9 (1.2-2.7)
0.1 (0.01-0.3)
6.8 (5.6-8.1)
C Blue-collar migrants
56.1 (39.9-67.4)
1.9 (1.1-3.0)
1.7 (1.1-2.8)
0.09 (0.01-0.3)
6.7 (4.9-7.5)
D Other blue-collar
58.0 (45.1-75.4)
1.95 (1.3-3.1)
1.8 (1.2-2.8)
0.1 (0.04-0.3)
7.8 (5.5-8.2)
E Intermediate
56.0 (39.8-71.3)
1.9 (1.1-3.1)
1.9 (1.0-2.7)
0.07 (0.01-0.2)
7.0 (5.5-8.0)
F Professionals
52.5 (38.8-62.1)
1.6 (1.2-2.5)
1.5 (1.2-2.5)
0.1 (0.02-0.28)
6.4 (5.0-7.8)
P=0.046
P=0.3
P=0.3
P=0.08
P=0.2
Trend age-adjusted
28
Table 4. Total PAEE and daily hours spent at different intensities of PA across transition groups. Results based on Acc
and HR monitoring (women). Unpublished and preliminary analyses. Inuit Health in Transition study in Greenland.
Women N=690
Total daily PAEE
MVPA
Moderate PA
Vigorous PA
Light PA
kJ/kg/day
(>3METs)
(3–6METs)
(>6 METs)
(1.5–3 METs)
Median hours
Median hours
Median hours
Median hours
Median (IQR)
(IQR)
(IQR)
(IQR)
(IQR)
A Hunters/fishermen
46.9 (38.1-55.5)
1.4 (1.0-2.1)
1.4 (0.98-1.99)
0.04 (0.007-0.1)
6.8 (6.0-8.3)
B Other villagers
46.4 (34.9-62.9)
1.7 (0.8-2.7)
1.5 (0.8-2.6)
0.06 (0-0.2)
6.7 (5.3-7.9)
C Blue-collar migrants
46.4 (34.9-56.5)
1.4 (0.8-2.2)
1.3 (0.7-2.0)
0.04 (0-0.1)
6.6 (5.5-7.8)
D Other blue-collar
43.3 (32.4-59.1)
1.4 (0.8-2.3)
1.3 (0.8-2.1)
0.04 (0.003-0.1)
6.3 (5.1-7.8)
E Intermediate
47.1 (37.2-59.6)
1.6 (1.1-2.4)
1.5 (1.1-2.3)
0.05 (0.01-0.16)
6.9 (5.5-8.1)
F Professionals
43.0 (34.1-56-6)
1.6 (0.98-2.4)
1.5 (0.95-2.2)
0.06 (0.007-0.2)
6.0 (5.0-7.6)
P=0.6
P=0.5
P=0.5
P=0.4
P=0.3
Trend age-adjusted
Is physical activity energy expenditure associated with glucose metabolism in
Greenland?
This is the first study to report on associations between objectively measured PAEE and glucose
metabolism among Inuit. Associations between PAEE and fasting insulin, 2-hour insulin, fasting glucose, fat
mass, BMI and waist circumference (WC) were found for 1,545 adult Inuit presenting valid data from Acc
+HR monitoring (≥48 hours of wear-time). After adjustments for age and sex, only the association with
fasting and 2-hour insulin remained significant. Further adjustment for waist circumference revealed that
only the association between PAEE and 2-hour insulin was independent of WC. An increase in PAEE, in
particular for those participants with the lowest level of PAEE (<35 kJ/kg/day), was associated with a lower
2-hour insulin concentration, indicating a dose-response relation of the amount of PAEE as seen in figure 8;
on average, fasting and 2-hour insulin levels were 3% and 9% lower for every 10kJ/kg/day difference in
PAEE. This difference could be achieved with an extra hour of gentle walking each day.
29
Fasting Insulin
160
140
120
100
80
60
40
20
0
2-h insulin
Pmol/l
6,4
2-h glucose
6,2
6
5,8
5,6
5,4
5,2
kJ/kg/day
28
Fasting glucose
mmol/l
BMI
kg/m2
27
26
25
24
23
kJ/kg/day
kJ/kg/day
98
96
94
92
90
88
86
84
82
WC
CM
kJ/kg/day
Figure 8. Age- and sex-adjusted means of fasting glucose, 2-hour glucose, fasting insulin and 2-hour insulin, BMI and
waist circumference across deciles of PAEE. Inuit in Greenland (n=1545).
Intensities of daily life PA – an overview
Table 5 provides an overview of the Acc and HR measurements and demonstrates that light intensity PA
(<3METs) contributed to a substantial part of the daily life PA. In contrast, a very limited amount of time
was spent at vigorous intensity PA, but the relative contribution to total PAEE was substantial.
Table 5. Absolute median number of hours spent at different intensities of PA (24 hours) and the contribution of
different intensities of PA for total PAEE. Preliminary and unpublished analysis. Inuit Health in Transition study in
Greenland.
Hr/day
IQR
% of total PAEE
IQR
Light intensity (1.5–3 METs)
6.8
5.3-8.1
52.6
42.6–61.1
Moderate intensity (3–6METs)
1.8
1.1-2.7
33.9
26.3–41.9
Vigorous intensity (>6 METs)
0.1
0.02-0.3
4.3
0.8–11.4
14.9
13.0-17.1
4.0
2.1–6.4
Light intensity (1.5–3 METs)
6.5
5.3-7.9
55.8
47.7–64.3
Moderate intensity (3–6METs)
1.5
0.9-2.2
31.4
23.1–39.2
Vigorous intensity (>6 METs)
0.1
0.007-0.2
2.5
0.2–6.8
Sedentary time (<1.5 METs)
15.7
13.7-17.4
5.1
2.7–7.9
Men (n=568)
Sedentary time (<1.5METs)
Women (n=770)
Values include all epochs (day and night). Intensity is defined as multiples of RMR, estimated by age, sex, height, and
weight (93).
30
31
Discussion
Discussion
The overall aim of this PhD thesis was to study PA patterns in an arctic population undergoing rapid social
transition and to add to the epidemiological evidence of how PA relates to glucose metabolism in an Inuit
population in Greenland. Moreover, the thesis aimed to study the feasibility of a questionnaire-based
measure of PA at a population-based level in Greenland.
Main findings
•
The IPAQ-L modified to arctic living conditions and interpreted with truncation of extreme outliers
is a moderately valid measure for overall physical activity at population level, but not valid to
measure different intensities of PA and sedentary activity when compared with accelerometry and
heart rate monitoring (Acc and HR). In particular, moderate intensity is substantially over-reported
if walking is included in the measure. However, the questionnaire provides important
complementary information on domain-specific PA, which is specifically of interest in populations
undergoing rapid social changes, such as Greenland.
•
When using residence, education and occupational status to rank the population into six subgroups
as a proxy for different stages of social transition, we found that PA patterns differed between
transition groups, and we were to some extent able to identify changes in PA patterns in relation to
the social transition. Less time was spent on occupational, domestic (women only) and
transportation-related PA and more time was spent on sedentary activity among the group of
participants in towns, with longer vocational or academic education (more modern lifestyle)
compared with hunters and fishermen in villages (more traditional lifestyle). No difference was
found for time spent on leisure time PA across transition groups. The overall time spent on PA did
not decrease linearly. However, preliminary analyses based on Acc and HR monitoring show that
physical activity energy expenditure (PAEE) decreased across the transition groups for men, but not
for women. The transition groups only work as a proxy for longitudinal information; hence, changes
must be interpreted with caution.
•
A strong association was found between objectively measured PAEE and BMI and waist
circumference. PAEE and 2-hour insulin only was shown to be associated independently of
abdominal obesity. Age, sex and weight were confounding factors for the association between
PAEE and fasting glucose, 2-hour glucose and fasting insulin. The results indicate a positive doseresponse relation and it is suggested that increasing PAEE, in particular for those participants with
the lowest level of PAEE (<35 kJ/kg/day), is associated with a lower 2-hour insulin concentration.
Our results suggest that both obesity and low levels of PAEE may be important contributing risk
factors for the increasing prevalence of type 2 diabetes among Inuit in Greenland. Nevertheless,
the study also points out the importance of examining factors other than lifestyle, i.e. genetic or
early-life factors, which could play a role in the development of impaired glucose metabolism.
32
Discussion
Is measuring physical activity by questionnaire a feasible method in Greenland?
A valid measure of PA is of great importance for the future monitoring of PA in Greenland. The finding of a
moderate agreement between the modified version of IPAQ-L and Acc and HR monitoring in the measure
of total PAEE is contrary to several other studies where overestimation of PA by IPAQ has been shown,
especially at high levels of PA, a bias for which social desirability has been suggested as a plausible
explanation (94-97). The attention from the media on the positive health impact of PA may have been less
marked in Greenland compared with more westernized countries, and thus the risk of social desirability
bias may be somewhat lower in our study. Moreover we found the IPAQ-L to be valid for use in both towns
and villages in Greenland. This is contrary to previous studies demonstrating IPAQ to be less valid in rural
areas (57;75;78). According to the interviewers in our study the use of face-to-face interviews undertaken
by Greenlandic interviewers most likely have diminished potential interpretational differences between
towns and villages in relation to the wide differences in living conditions, climatic differences and dialects
across Greenland. Likewise, a status report on the assessment of PA found the use of interviewers to
increase the validity of the responses compared with self-reporting (12). Furthermore the adapted
examples of PA might have made the reporting of PA easier to report adequately.
Asking for domains of PA makes the IPAQ-L a rather long and time-consuming questionnaire compared with
other PA questionnaires. However, our results show that the domain-specific information of PA was highly
valuable in Greenland to identify domain-specific differences in PA patterns along with the social transition.
Moreover, substantial information from occupational and domestic PA would have been missing if we had
measured PA during leisure time only and would have resulted in a substantial lower overall level of PA.
Similarly to most other studies, the IPAQ-L substantially overestimated moderate intensity PA in our study.
Time spent walking was included in all four domains in the questionnaire, which might have increased the
risk of reporting the same walking activity twice. Ekelund and colleagues found that walking is difficult to
accurately quantify (98), and studies have shown large errors when assessing simple activities such as
walking (13;99). When excluding walking from the analyses we found a substantially better agreement
between IPAQ-L and Acc and HR monitoring. According to guidelines from the IPAQ group (89), walking is
set to moderate intensity and is assigned the MET value of 3.3 METs. In the compendium of PA by
Ainsworth et al, various intensities of walking corresponding to different MET values are listed (2.3 to
3.6METs)(14). One could argue that a slow pace of walking corresponds to light intensity and not moderate
intensity. Moreover, qualitative information from the interviewers in our study revealed that occupational
activities, such as teaching or working in a shop, were sometimes misinterpreted as walking activity instead
of light intensity PA. Walking is very common in Greenland due to an infrastructure with a limited number
of roads and cars, as well as small residential areas; therefore, walking might contribute to a substantial
misclassification of overall moderate intensity. In addition, the IPAQ, as well as most other questionnaires,
does not include questions about light intensity activities which may result in participants classifying light
intensity PA as moderate intensity PA. Acc and HR monitoring revealed that light intensity PA contributes to
a substantial part of daily life PA. Such error will therefore clearly result in substantial overestimation of
moderate intensity PA.
33
Discussion
Previous research has shown that in particular activities of light and moderate PA are difficult to recall
because these activities often include unstructured activities in contrast to vigorous intensity PA, which
includes more structured activities like sports (12;70;99). However, in Greenland we found that
questionnaire-based vigorous intensity PA also showed low validity. According to the interviewers, one
possible explanation could be that vigorous intensity was sometimes interpreted as psychological
demanding instead of physiological demanding despite of the adapted PA examples. Likewise, a study
based on cognitive interviews demonstrated that in some cultural settings intensity level was frequently
interpreted as emotional or psychological intensity rather than the level of physical effort (100).
The IPAQ-L did not measure light intensity PA and showed low validity for the measure of sedentary
behavior in our study. Along with the increasing availability of sedentary and light intensity pursuits, the
relative importance of specific levels of intensity on health has been the subject of much current research.
Both light intensity PA and sedentary behavior have been related to decreased metabolic health, but the
evidence is still limited and conflicting (53;54;56;101-105). Moreover, it is emphasized that differential
measurement uncertainty between intensities of PA challenges the interpretation of their relative
importance (106).
The IPAQ-L only allows reporting of PA for a minimum duration of 10 minutes. However, as populations are
getting more sedentary it could be argued that questionnaires also must take into account short bouts of
PA (<10 min) and low intensity PA to avoid the risk of “floor effects”. In statistics, it means that data cannot
present a value lower than some particular number, which could hinder the ability to differentiate between
low levels of PA (107). This effect might be present in our study. The evidence for a minimum duration of
activity to induce health benefits is limited, as is the effect of accumulated versus continuous bouts of
exercise. A study by Eriksen et al showed that 3x10 minutes had a greater impact on glycemic control than
one bout of 30 minutes (108), whereas a review of empirical studies was not able to make firm conclusions
on the effect of continuous versus accumulated exercise on health (109).
In the IPAQ-L, PA is reported from the previous 7 days. The climate in Greenland includes substantial
seasonal differences, which could potentially influence the PA level reported and provide differences in PA
results relating to when data is collected. In our study, we collected data both in winter and summer time,
although villages were only visited during summer time. Further information on seasonal variation would
be valuable to include in future measurements of PA in an arctic population. Although, a study of seasonal
differences in the level of fitness in an Inuit population found that despite substantial seasonal differences
in hunting patterns, fitness remained at a high level throughout the year, with no indication of differences
between summer and winter. Similar results were found for Inuit living more permanently in settlements
(110).
Overall, we find the modified version of the IPAQ-L to provide important domain-specific information and
valid to use in an arctic population to provide an estimate for overall PAEE at population level, but not to
distinguish between intensities of PA and sedentary behavior. Furthermore, IPAQ-L does not provide a
measure of light intensity PA. In relation to feasibility, we find the questionnaire-based method to have a
low participant burden and to minimize reactivity (an individually changing behaviour due to being
34
Discussion
measured). However, in principle self-reported PA is perceived as a relatively low cost method and
therefore often used in population-based studies, but in an arctic context like Greenland, the use of trained
bilingual interviewers and a need for a Danish and Greenlandic version of the questionnaire culturally
adapted and modified to arctic living conditions, made the questionnaire method rather costly. Moreover,
IPAQ-L is very comprehensive and therefore a time-consuming questionnaire, in particular if used for
surveillance purpose.
Larger-scale studies have started to combine self-reported measures with device-based measures. In
Greenland, the use of combined heart rate and movement monitoring provided us with a unique
opportunity to obtain supplemental information on objectively measured PA and intensity level. Although
we did not validate the method in Greenland, we obtained comprehensive practical experience of the use
of a relatively new method under arctic living conditions and the subsequent data processing (appendix III).
We found that both methods contributed with important knowledge on the different dimensions of PA in
Greenland. Table 6 presents an overview of what is found to be the main advantages and disadvantages
using the two methods in Greenland.
Table 6. An overview of main advantages and disadvantages of using the interviewer-administered IPAQ-L and Acc and
HR monitoring in Greenland.
IPAQ-L
Advantages
››Domain and activity-specific information
››Information on sedentary time and intensities
››Limited reactivity
››Relative high feasibility e.g. logistic and low participant burden
››Culturally adaptable to arctic living conditions
››Valid measure for overall PAEE (both villages and towns)
Disadvantages
››Risk of systematic and non-systematic bias (recall Bias, social desirability bias)
››Limited validity for measuring intensity and sedentary behavior
››Data processing issues, outliers
››Context and cultural-dependent
››Costly (interviewer-administered, language) and time-consuming
››No measure of light intensity PA
35
Discussion
Acc and HR monitoring
Advantages
››Information of total amount, frequency, duration and individual patterns of physical activity
››Avoiding bias seen by self-report, such as recall bias
››More accurate and precise estimation of individually based PAEE than self-report
››Not prone to recall bias, social desirability bias
Disadvantages
››High participant burden, risk of reactivity
››Resource demanding due to logistic and cost
››Complex data managing, and processing of sensor noise for HR
››Limited information of context and type of PA
››Practical issues (wear time, administration and placement)
What characterizes the physical activity transition in Greenland?
Most of the research on changes in chronic diseases and risk factors in populations undergoing rapid
transition has focused on urbanicity (nature of urban environments) measured by a simple dichotomized
measure (urban versus rural). The use of an urban-rural dichotomy has been criticized for ignoring the
heterogeneity of environments within urban and rural areas and for inability to detect changes over time
because rural areas themselves are being modernized (111;112). Modernization in Greenland has resulted
in increased mechanization of hunting and fishing activities both in villages and towns. Likewise sedentary
service-oriented occupations and sedentary pursuits, such as computer use and TV viewing during leisure
time, have not only increased in availability in towns but also in the most remote villages, and walking
activity is still very common in both towns and villages. The modernization of both urban and rural areas
results in a less clear distinction between urban and rural (Champion and Hugo, 2004), and important
differences in the process of urbanization or modernization might be overlooked using this simple variable.
In Greenland, various definitions of Westernization have been used. One study defined the degree of
Westernization by language and current place of residence (50). Another study used parents’ place of birth
and occupation, residence during childhood, knowledge of Greenlandic and Danish and school education to
divide the population into a group of Greenlanders with a predominantly traditional childhood and a group
with a more Westernized childhood (3). In order to obtain more detailed information on the ongoing
modernization process in Greenland, we used the participant’s current place of residence and childhood
residence combined with formal education and family job type as a proxy for secular changes in PA
patterns (69). The various definitions of urbanization and modernization also complicate the comparison of
the physical activity transition and its consequences between populations and within populations over
time. However, most studies worldwide agree that occupational PA has decreased with modernization
(52;57;59;60;113), similarly to what we found in our study in Greenland. We showed that for men this
decrease was mostly explained by the difference in occupational PA between hunters and fishermen in
villages and participants with longer vocational or academic education living in a town. This is most likely
explained by more sedentary occupational activities available when higher educational status. In our study,
36
Discussion
the decrease in transportation-related PA found both for women and men, is most likely explained by the
limited possibility for mechanized transport in villages compared with larger towns in Greenland.
Furthermore, we found a substantial difference in household PA for women along with the social transition,
which for some part might be explained by increasing mechanization of household chores and water
facilities, and less time available at home especially among the group of professionals in towns compared
with villages. Men spent relatively less time on household PA compared with women across all groups of
the transition variable, which might be an expression of social norms rather than a consequence of social
transition. In accordance with other populations undergoing transition we found no differences in
participation in leisure-time PA in relation to social transition (59;60). This finding is contrary to the upward
tendencies found in time-trend studies in Western populations, although only a few exist (113;114). One
explanation could be a greater focus on health-enhancing PA, such as weight control and well-being, more
time eligible for leisure-time pursuits in Western populations or the question of availability of activities
suitable for leisure-time pursuits. Another explanation could be that traditional activities, such as hunting
and fishing are being misclassified as occupational activities even though they have increasingly gained
status as a leisure activity in Greenland because of the potential overlap of these activities in this cultural
context.
The supplemental information from Acc and HR monitoring in our study revealed additional important
information on gender differences in relation to social transition. The energy expenditure spent on PA
decreased linearly by stages of social transition for men. This decrease seems to be partly explained by
decreasing time spent on moderate-to vigorous intensity PA. For women, the overall PAEE and time spent
on different intensities of PA was not significantly different. The results indicate that PA patterns have
changed as a result of the social transition both for men and women, but had an impact on total PAEE for
men only. This finding is in line with a previous study in Greenland that showed an association between
Westernization and metabolic risk for men only (50). Knowledge of differences in PA patterns in relation to
the modernization process in Greenland can help to differentiate and target the promotion of PA.
We also did the analyses of PA patterns comparing villages and towns as an expression of the traditional
urban-rural distinction and found only small differences in PA patterns, which might indicate that more
detailed information is obtained using the social transition variable. The transition variable was developed
specifically for Greenland and should not be applied to other populations in the Arctic without further
examination. Moreover, data are cross-sectional and, therefore, changes in PA can only be seen as a proxy
for longitudinal changes.
Research has been done to develop more detailed measures for urbanization. One study used residence
and occupation to measure urbanization (115). Another study developed an urbanization index score on
the basis on ten measures both at individual level and area level: population size, population density,
access to markets, transportation, communication possibilities, economic factors, housing quality,
education, sanitation and health (116). Dahly et Adair constructed a multivariable scale of urbanicity using
community level data: population size, population density, communication possibilities, transportation,
educational facilities, health services and markets. The scale was shown to be able to detect differences in
37
Discussion
urbanicity between communities and across time (111). It would be valuable to examine if some of these
variables, such as ownership of different assets, economic factors as level of income and access to
communication and transportation facilities could provide more insight in the process of modernization in
Greenland.
Physical activity and glucose metabolism in Greenland - what is the evidence?
The inverse association between PAEE and 2-hour insulin independent of abdominal obesity found in our
study is in line with findings from other populations undergoing rapid social transition (48;52;117). The
indicated dose-response relation between the volume of PAEE and 2-hour insulin, in particular for those
participants with the lowest level of PAEE, corresponds well with research suggesting a positive doseresponse effect of total PA on health (23;24) and that even small increases in PA among the most physically
inactive is shown to be beneficial for health (22). However, the results must be confirmed in future
prospective studies. Contrary to most previous research, we were not able to identify an association
between PAEE and 2-hour plasma glucose and IGT when measurements of abdominal fat were included in
the analysis (48;51;118;119). However, not all studies have included body composition measures as
potential confounding or mediating factors. Research examining the patho-physiology and aetiology of
impaired glucose tolerance (IGT) showed that IGT was predominantly related to physical inactivity,
unhealthy diet and short stature (120). Although, we adjusted our analysis for factors known to be related
to both PA and glucose metabolism outcomes, such as age, sex, smoking and family history of type 2
diabetes, residual confounding might be present. For example, diet, early life factors or genetic disposition
not captured by family history, which we were not able to adjust for, could play a significant role in our
population. A study found fasting glucose to be a marker of beta-cell dysfunction and hepatic glucose
production rather than peripheral insulin resistance, and predominantly related to genetic factors, smoking
and male sex which could be a plausible explanation for our findings for PAEE and fasting glucose (120).
Insufficient physical activity may contribute to impaired glucose tolerance through a pathway including
alterations in obesity and fat distribution. Our regression analysis showed that abdominal fat was
significantly associated with glucose and insulin concentrations and that PAEE was inversely associated with
BMI, waist circumference and fat percentage. It is suggested that overweight or obesity have a significant
role in explaining differences in 2-hour insulin and fasting insulin in our study population. A study of Rana et
al. demonstrated that obesity and physical inactivity contributed to the development of type 2 diabetes
independently; however, the magnitude of risk contributed by obesity was much greater than the lack of
PA (121). The evidence of the relative influence of obesity and physical inactivity on the risk of developing
diabetes is however still sparse and conflicting. A Finish study found that increasing PA was associated with
a significantly reduced risk of type 2 diabetes, especially among obese patients (122). Contrary, Weinstein
and colleagues concluded that PA had relatively small effects on diabetes in overweight and obese patients
(123). Our results suggest that both obesity and low levels of PAEE may be important contributing risk
factors for the increasing prevalence of type 2 diabetes among Inuit in Greenland. Nevertheless, the study
38
Discussion
also points out the importance of examining other factors, than just those related to current lifestyle, such
as genetic or early life factors, which could play a role in the development of impaired glucose metabolism
in this indigenous population. Findings in this thesis can be incorporated in public health strategies in the
prevention of type 2 diabetes in Greenland. However, the cross-sectional design does not allow us to draw
conclusions about the direction of associations or any strong inference on causality.
Methodological considerations
Strengths
Some strength of the three studies should be put forward. The validation study is based on a large study
sample (n=1508) encompassing all fractions of the population, which is contrary to most validation studies
based on a small number of participants. Such a sampling strategy for validation increases the probability
that all possible interpretations and lifestyles are included, and forces one to think about how to deal with
outliers without introducing too much selection bias (paper I). The social transition variable provided us
with the possibility to obtain a proxy for the physical activity transition when longitudinal data were not
available (paper II). The association between PAEE and glucose metabolism has mostly been studied by the
use of self-reported measures. We used an objective measure for PAEE. Furthermore, the measures of
glucose tolerance and insulin concentrations were based on blood samples instead of self-reported
measures of type 2 diabetes (paper III). However, the three studies also present several potential
limitations, of which the main are discussed in the following.
Selection Bias
The Inuit Health in Greenland study demonstrated a participation rate of 66.7%, which is high compared
with population-based surveys in general (124), and especially for this setting where data collection is
challenged by infrastructure and weather conditions. Moreover, the study surveyed a large proportion of
the adult, Greenland born population (9.2%). However, we have very limited information on the non‐
participants, and therefore the risk of selection bias cannot be ruled out; some potential differences
between participants and non-participants should be emphasized in relation to the existence of possible
selection bias. The variation in participation rates across the country and between villages, towns and the
capital and the stratification procedure of the random sample means that the study sample includes
proportionately more participants from some regions despite their small percentage of the total
population. Because of logistic challenges it would be almost impossible and costly to base this population
survey on a non-stratified sample. The non-random distribution of non-participants could introduce bias for
the precision of countrywide estimates.
There are several potential explanations for non-participating in the study. The rather long duration of the
health examination, including both clinical and questionnaire measurements, could be an explanation for
the higher proportion of non-participants from the larger towns compared with villages due to a more busy
daily life. This scenario is underlined by a higher proportion of participants in Nuuk indicating lack of time as
the reason for not wanting to participate (17% of the non-participants compared with 2% in the rest of the
39
Discussion
communities). We suspect that socially exposed persons, alcohol abusers and persons who frequently go in
and out of jobs and the unemployed likewise are over-represented among the non-participants. It was the
impression of the interviewers that there was a distinct downward social trend from the beginning to the
end of data collection in a town. In some towns it could be demonstrated that during the first week of the
study in a particular community 10% of those who had made an appointment did not show up, while during
the last week of the study as many as 26% failed to show up. Analyses of register-based income showed
that the personal income of participants was higher than among non-participants, which confirms this
possible social selection. However, we find it unlikely that this selection would have affected the
association between PA and glucose metabolism. There is a potential risk that hunters and fishermen are
underrepresented in this study due to longer periods of time being away from home and therefore not
being present at the health examination. It is difficult to evaluate the effect of such selection; however, as
shown in paper II this group is most physically active, and therefore such a selection could bias the
population estimate of PA.
We know that persons with serious illness or disability are overrepresented among the non-participants;
this bias might play a role for the ability to generalize the estimates of median PA to the population as a
whole (106). However, we find it unlikely that this bias would have affected the validity of the association
between physical activity and glucose metabolism within this population. The participants differed from
non-participants by age and sex. Women more often participated than men, and particularly young men
were under-represented. A lower proportion of young participants will most likely be accompanied by a
higher prevalence of type 2 diabetes but also of a lower level of PA. Since our sample did include some
young participants, it is unlikely that this bias will alter the age- and sex-adjusted association between PA
and glucose metabolism in this thesis.
Papers I and III are based on a reduced subsample (n=1545) due to a limited number of participants with
accelerometry and heart rate monitoring. The subsample was selected from all over Greenland, and only
small differences were found between the subsample and the total study sample (further details in
appendix III). The odds for being monitored by ACC and HR did not differ significantly between sexes, was
slightly lower only for age groups above 70 years and 40-44 years old and for participants living in a town,
but was higher for participants living in Nuuk. Overall, we have no reason to believe that the association
between PAEE and diabetes should be prone to substantial selection bias.
Information bias and validity of the physical activity measurements
Misclassification in relation to the level and dimension of physical activity is likely when using self-reported
methods. Social desirability is a plausible explanation for misclassification due to over-reporting of PA by
self-report (12). The moderate agreement between the two methods for overall PAEE may imply that the
risk of non-differential misclassification due to social desirability bias may be somewhat lower in our study.
Physical activity is a multidimensional behaviour and therefore most likely prone to recall bias. Structured
40
Discussion
activities, such as type of exercise, have been shown to be easier to recall compared with activities like
walking and other moderate-intensity activities, which are seldom structured (12). This could explain the
highly overestimated level of moderate intensity and walking in paper I (non-differential misclassification).
The IPAQ asks about PA in the previous 7 days. Studies have shown that this time limit is easier to
remember compared with monthly averages (70), and it therefore lowers the risk of recall bias in our study.
Individual characteristics, such as level of fitness, could possibly play a role in the perception of intensity,
e.g. individuals with a higher level of fitness may perceive moderate (3-6 METs) and vigorous activity (>6
METs) differently compared to those who are more sedentary, providing a potential differential
misclassification of self-reported PA.
The use of device-based measures of PA in paper III avoids issues of recall bias (88;125).
However, contrary to the estimates from the IPAQ-L calculated as the average of the previous 7 days (no
information on sleeping hours, but 8 hours were estimated for sleep), the estimation of physical activity
energy expenditure from ACC and HR monitoring was based on individual recordings from 48 hours to 5 full
days mostly representing both week and weekend days. Rennie et al. estimated that 3 days of recording
yielded a validity coefficient of 0.85 for the assessment of energy expenditure in a European sample (126).
In our study, only 858 of the participants had more than 3 days of wear data, but our sensitivity analyses
showed similar results when applying this stricter inclusion criterion. Ideally, more days of objective
recording would have been preferable to capture variations in PA during the week, but logistics made this
unfeasible.
As a consequence of the administration of the two instruments they did not refer to the same time period.
The monitor was given to the participants on the day they were interviewed about their PA in the
preceding 7 days. However, the short interval between the periods is unlikely to have introduced
substantial bias in the results, and one may even consider the present results to reflect more truly the
convergent validity of these instruments to assess habitual physical activity. In the interpretation of validity
it is crucial which reference method is chosen as criterion measure. The ACC and HR monitoring in this
study has shown itself to be valid compared with DLW in a non-Western context (86)). Another crucial
factor is whether the two methods measure different aspects of PA. The IPAQ asks for PA of moderate or
vigorous intensity for a minimum of ten minutes, and no information on sleep is available, whereas the
device-based method provides estimates of PA for 24 hours including all PA intensities. This provides a
potential bias in the estimates of especially different intensities of PA during an average day estimated by
IPAQ-L.
41
Discussion
Confounding
The multiple linear regressions between PAEE and glucose metabolism in paper III were adjusted for
potential confounders/mediators identified a priori based on identification from previous research of risk
factors for type 2 diabetes and biological plausibility: sex, age, smoking, family history of diabetes and
waist circumference. Still, there is a risk of residual confounding, or confounding due to unmeasured
covariates, which could skew the association between PAEE and glucose metabolism in our population,
such as diet, early life factors or genetic disposition not captured by family history, which we were not able
to adjust for. Contrarily, also a potential risk of over-adjustment is present if considering waist
circumference as a confounder and not a mediator for the association between PAEE and glucose
metabolism. In paper II differences in time spent on physical activity across social transition were analyzed
in linear regression model with time spent on PA as dependent variable. The analyses were age adjusted
and stratified by sex.
Causal relationship
The cross-sectional design of the study did not allow us to make conclusions about the direction of
associations or any strong inference on causality. There is a potential risk of reversed causality if
participants with type 2 diabetes have a lower level of PA due to type 2 diabetes-specific complications. We
tried to minimize this by excluding participants with known diabetes from the analyses.
42
43
Discussion
Conclusion
Conclusion
The beneficial effects of physical activity (PA) on glucose metabolism are well-established in Western
populations. The knowledge is, however, much more limited when it comes to populations undergoing
rapid social transition, such as Inuit in Greenland. It has been suggested that decreasing PA is an important
contributing risk factor for the increasing prevalence of type 2 diabetes in this population.
The work presented in this thesis shows that physical activity energy expenditure (PAEE) is associated with
2-hour postload insulin independent of abdominal obesity. A dose-response relation indicates a statistically
significant beneficial effect of increasing PAEE, in particular for those participants with the lowest levels of
PAEE. However, age, sex and waist circumference were confounding factors for the association between
PAEE and fasting glucose, 2-hour glucose and fasting insulin. PAEE was strongly associated with BMI and
abdominal obesity. Our results suggest that both obesity and low levels of PAEE may be important
contributing risk factors for the increasing prevalence of type 2 diabetes among Inuit in Greenland.
Nevertheless, the study also points out that other factors, such as genetic predisposition and early lifefactors, must play a role for the high prevalence of type 2 diabetes in Greenland. Due to the cross-sectional
data in this thesis, causality cannot be established and the association should be further investigated in
prospective studies.
The physical activity patterns in Greenland have changed markedly along with the social transition. By
grouping the population into stages of social transition we were able to identify information of PA patterns
along with the modernization process. A lower level of occupational, domestic and transportation-related
PA was found among professionals in towns (most modern lifestyle) compared with hunters and fishermen
in villages (most traditional lifestyle). Nonetheless, no difference in leisure time PA was found as a result of
the social transition. Leisure time PA could be an important domain for the promotion of PA in order to
prevent decreasing levels of overall PA along with the ongoing social transition. Despite the difference in PA
patterns, the overall PAEE decreased by stages of social transition for men only, this is most likely explained
by decreasing time spent at moderate to vigorous intensity PA. However, due to the cross-sectional design
of the study, changes can only be seen as a proxy for longitudinal changes.
Surveillance of changes in PA is of great importance due to the increasing metabolic disorders reported in
Greenland. We find the modified interviewer-administered IPAQ-L as a valid method to measure overall
PAEE but it cannot be used to differentiate between intensities of PA. Furthermore, the method is feasible
to use in Greenland but the interviewer-administered version, which seemed to be important to prevent
cultural barriers in the interpretation of the questions made the questionnaire-based measurement of PA
relatively costly and time consuming. Furthermore the lack of information on light intensity PA, shown to
be contributing to a large part of daily life PA, must be considered if this measurement tool is to be used. In
Greenland, the use of combined heart rate and movement monitoring provided a unique opportunity to
obtain supplemental information on objectively measured PA and intensity level and is feasible to use in an
arctic setting, although the method is still costly, and logistically as well as technically demanding.
44
Conclusion
Implications for public health in Greenland
The promotion of a physically active lifestyle in Greenland played a central role in the public health
program Inuuneritta 2007-2012. Our findings underline the need to continuously include PA in prevention
and promotion strategies, such as the Inuuneritta 2013-2016, in order to maintain and promote a physically
active lifestyle in relation to the ongoing process of modernization. The suggested dose-response relation
of physical activity energy expenditure (PAEE) on a precursor for type 2 diabetes, with benefits for those
with the lowest level of PAEE in particular, is an important public health message in the future prevention
of type 2 diabetes.
The difference in PA patterns in relation to the process of modernization indicates the necessity of
targeting the promotion of PA to specific population groups, age-groups and gender. Our results suggest
that inhabitants in towns with a longer vocational or academic education are one important group. The
domain-specific information on PA in our study points at leisure time as an important domain to promote
PA in order to maintain or increase PA. Furthermore our findings suggest a need to focus on reducing time
spent on sedentary behaviors. Evidence for negative health consequences of prolonged sitting, such as
increased metabolic risk has increased (104;105;127) and even small breaks in the sedentary time have
shown beneficial effects on metabolic risk (128). However, controversies still exist whether to include
quantitative recommendations or just advise against reducing sedentary behaviors in national
recommendations for PA.
Systematic surveillance of PA and sedentary behavior should be carried out in order to monitor time trends
and changes in PA patterns along with the social, cultural and economic changes in Greenland. This
surveillance should be based on valid methods and comparable measurements. Our findings, illustrate that
the combination of a self-reported and a device-based method provides several advantages, such as
measurements of domain-specific PA with great importance for providing a valid measurement of the
overall PAEE and time spent at different intensities of PA.
Implication for future research
The association between objectively measured PA and glucose metabolism has not been investigated
before in an Inuit population and only rarely in other populations. The association must be examined in a
prospective design to explore whether the dose-response relation can be confirmed. Furthermore the
association between social transition, PA patterns and metabolic risk should be further investigated with
the use of objective measures for PA. The contribution of insufficient PA to impaired glucose metabolism
through a pathway including alterations in obesity and fat distribution should also be investigated further
using data from the Inuit Health in Transition study in Greenland. Data from the Inuit Health in Transition
study in Greenland 2005-10 provides the opportunity to study this association in future studies and
knowledge about these associations is of great importance for evidence-based-guidelines for PA.
The influence of various levels of intensity (light to vigorous intensity) and domains of PA on glucose
metabolism was not investigated in this study. In a future perspective the domain-specific information can
45
Conclusion
be used to study the association between leisure time PA and metabolic health. E.g., a study in a Danish
population, found opposing effects of occupational and leisure-time physical activity on global health (129).
It has been suggested that a large part of the effect of PA in decreasing insulin resistance is short lived and
therefore the effect may last only a few days (130;131). In order to learn more about the underlying
mechanism for the association between PA and glucose metabolism in Greenland, the consistency of an
individual´s PA could be further assessed by measuring PA more frequently (132).
In order to reduce misclassification of PA by self-report, future studies with a qualitative approach should
be carried out examining the concepts of physical activity and different intensities of PA. Such an approach
would contribute to further knowledge of the interpretation of physical activity in an arctic setting and also
contribute to the further development of appropriate activity examples in questionnaires.
Further research is still needed to increase knowledge on suitable methods to measure PA. The rapid
technological development in device-based measurements means that device-based methods are
becoming better to integrate context-specific information on PA by using methods, such as Global
Positioning System (GPS), Geographical Information System (GIS) and integrated cameras in the monitoring
programs (133). These methods might provide valuable information on patterns of PA, and should be
investigated for future use in populations-based studies in the Arctic.
Overall there is a lack of knowledge on what motivates or hinders populations in the Arctic to be physically
active (134). There is a need for more intervention studies to form the basis for successful PA promotion
strategies. Environmental and policy interventions are based on ecological models of behavior and have
shown to have a potential to affect the entire population. Cross-sectional data indicate that environmental
and policy variables are associated with physical activity behaviors of young people and adults (135). Sallis
and colleagues concluded that PA in the different domains of daily life, such as occupation and transport
are associated with different environmental factors (136). The influence of environmental and policy
factors on PA in Greenland is a subject that deserves much greater exploration and considerations.
Furthermore, there is a lack of large-scale studies with comparable data on PA to study the physical activity
transition and implications for type 2 diabetes among indigenous populations in the Arctic. Collaborative
work should be established in order to develop comparable and standardized measurements as well as
survey procedures for cross-country comparisons of PA among indigenous populations in the Arctic.
Comparisons of prospective data across populations would provide knowledge on successful intervention
and prevention strategies. Already existing data on combined accelerometry and heart rate monitoring
among Alaska natives are promising in order to increase the knowledge about PA in the Arctic. Overall,
measuring PA in different contexts can help us to clarify how economic and social conditions, as well as the
environmental and cultural context within the specific country and across countries play a role for the
physical activity transition.
46
47
Conclusion
Summary
Summary
Since the 1940s indigenous populations in the Arctic, including Greenland, have undergone rapid cultural,
economic and social changes characterized by a shift from a traditional lifestyle to a more westernized
lifestyle, especially in the larger towns. This process has resulted in a less physically demanding lifestyle
with changes from subsistence hunting and fishing to sedentary occupational activity and increased
mechanization of society. Parallel to these changes a decrease in infectious diseases and an increase in
lifestyle-related chronic diseases, such as type 2 diabetes has been observed. Changes in physical activity
patterns are suggested to be an important contributor to the rise in chronic lifestyle diseases. However,
little is known about physical activity in arctic populations and how physical activity is related to social and
cultural changes in society, the so-called physical activity transition.
The main objective of this thesis was to examine the physical activity transition and the relation to glucose
metabolism in an arctic population undergoing rapid social transition. Moreover the aim was to study the
feasibility of a questionnaire-based measurement of PA at a population-based level in Greenland. The
overall objective was divided into three specific research objectives:
 to validate a modified version of the long International Physical Activity Questionnaire against
accelerometry and heart rate monitoring in an arctic population (Paper I).
 to study physical activity pattern in relation to the social transition among Inuit in Greenland (Paper
II).
 to analyze the objectively measured association between physical activity energy expenditure and
glucose metabolism in Greenland (Paper III).
This thesis is based on data from the Inuit Health in Transition Study (IHT) in Greenland collected in 20052010. Data are collected from 9 towns and 13 villages in different parts of Greenland and comprise clinical
examinations, and an interviewer- and self-administered questionnaire. The overall participation rate was
64.9%. In total 3102 adult Inuit, aged 18 years and above, were interviewed. The International physical
activity questionnaire (IPAQ-long version) was used to obtain data on physical activity (PA) and a subgroup
of participants was monitored by combined accelerometry and heart rate monitoring (n=1995).
In the first paper, we found that the IPAQ-L adapted to arctic living conditions in Greenland showed a
moderate level of agreement with combined accelerometry and heart rate monitoring for total Physical
Activity Energy Expenditure (PAEE) at population level, but was less valid to measure different intensities of
PA and sedentary behavior. Validity did not differ markedly between rural and urban communities.
In the second paper, we identified changes in physical activity patterns in relation to the social transition
evaluated as differences between groups of social change defined by residence, occupation and education.
Men in the latest stage of the social transition spent less time on occupational PA and women less time on
48
Summary
domestic PA, compared with men and women in the earliest stage of the social transition. A similar pattern
was found for physical activity in the transportation domain and sitting time for men and women. No
differences were found for leisure time PA. The overall time spent on PA was not found to decrease;
however physical activity energy expenditure (PAEE) decreased for men only. Due to cross-sectional data,
changes must be interpreted with caution.
In the third paper, we revealed that only the association between objectively measured PAEE and 2-h
insulin was independent of obesity. Age, sex and waist circumference were confounding factors for the
association between PAEE and fasting glucose, 2 hour glucose and fasting insulin. The results underline a
need to examine additional potential risk factors in the prevention of type 2 diabetes in Greenland.
This thesis underlines the importance of a continuous monitoring of changes in physical activity in relation
to the economic, cultural, and social changes in Greenland. The use of combined heart rate and movement
monitoring provides a unique opportunity to obtain supplemental information on objectively measured PA
and intensity level and is feasible to use in an arctic setting. From a public health perspective it is important
to promote PA during leisure time and reduce sedentary behavior to maintain a physically active lifestyle
thereby reducing the development of type 2 diabetes in Greenland.
49
Dansk resumé
Dansk resumé
Siden 1940´erne har oprindelige folk i Arktis, herunder Grønland, gennemgået en hurtig kulturel,
økonomisk og social forandring karakteriseret ved et skift fra en traditionel livstil til en mere vestlig
moderne livstil. Forandringen har især fundet sted i de større byer. Ændringen fra et selvforsørgende fanger
og fisker samfund til et samfund, der i højere grad er præget af stillesiddende aktiviteter har medført en
mindre fysisk krævende livstil. Sideløbende med disse forandringer er der fundet en høj forekomst af
livstilsrelaterede kroniske sygdomme som type 2 diabetes. En ændring i det fysiske aktivitets mønster anses
for at bidrage væsentligt til stigningen i de kroniske livstilssygdomme. Der eksisterer kun sparsom viden om
fysisk aktivitet og om hvordan fysisk aktivitet relaterer sig til de sociale, økonomiske og kulturelle
forandringer, også kaldet den fysiske aktivitets transition.
Det overordnede formål med afhandlingen var at undersøge den fysisk aktivitets transition og
sammenhængen mellem fysisk aktivitet og glukose metabolisme blandt et repræsentativt udsnit af inuit i
Grønland. Endvidere var formålet at undersøge anvendeligheden af et spørgeskema til at måle fysisk
aktivitet i en arktisk kontekst. Afhandlingen havde tre delformål:
 At validere the long International Physical Activity Questionnaire tilpasset arktiske levevilkår
sammenholdt med kombineret accelerometri og hjerterytme måling (artikel I).
 At undersøge det fysiske aktivitetsmønster i relation til den sociale transition i Grønland (artikel II).
 At analysere associationen mellem fysisk aktivitet og glucose metabolisme i Grønland målt ved en
objektiv metode (artikel III).
Afhandlingen er baseret på data fra Befolkningsundersøgelsen i Grønland (Inuit Health in Transition Study)
indsamlet i perioden 2005-2010. Data er indsamlet i 9 byer og 13 bygder i forskellige dele af Grønland og
omfattede kliniske undersøgelser, et interviewerbaseret spørgeskema samt et selvudfyldt spørgeskema. I
alt blev 3102 voksne Inuit (>18 år) interviewet. Den overordnede deltagerprocent var 64.9%. En modificeret
udgave af The International Physical Activity Questionnaire (IPAQ-L) blev brugt til at indsamle data omkring
fysisk aktivitet. Endvidere indgik målinger af kombineret accelerometri og hjerterytme (Actiheart®) fra en
subgruppe af deltagerne (n=1995).
Afhandlingen viser, at IPAQ-L er anvendeligt til at måle det totale energiforbrug brugt på fysisk aktivitet på
befolkningsniveau både blandt by- og bygdebefolkningen i Grønland, men ikke til at differentiere mellem
tid brugt på forskellige intensiteter af fysisk aktivitet og stillesiddende adfærd.
Spørgeskemaet giver vigtig information omkring det fysiske aktivitetsmønster og afhandlingen viser, at
mønstret har ændret sig i takt med den sociale transition i Grønland målt ud fra en gruppering af bopæl,
erhverv og uddannelse. Mænd med længerevarende uddannelse og bosiddende i en by var mindst fysisk
50
Dansk resumé
aktive i deres arbejde og kvinder tilsvarende mindst fysisk aktive i hjemmet sammenlignet med mænd og
kvinder med en mere traditionel livsstil med bopæl i en bygd. Et tilsvarende mønster blev fundet for den
transportrelaterede fysiske aktivitet og for stillesiddende aktivitet både for mænd og kvinder. Der var ingen
forskel at finde for fysisk aktivitet i fritiden. Den overordnede tid brugt på fysisk aktivitet ændrede sig ikke i
takt med den sociale transition, hvorimod foreløbige analyser viste, at det samlede energiforbrug brugt på
fysisk aktivitet faldt for mænd. Studiets tværsnitsdesign betyder at ændringer i fysisk aktivitetsmønster må
fortolkes med forsigtighed.
Afhandlingen viser endvidere en sammenhæng mellem energiforbruget forbrugt på fysisk aktivitet og
insulin koncentrationen to timer efter efter en oral glukose tolerance test uafhængig af abdominal fedme.
Denne sammenhæng kunne ikke genfindes for faste glukose, 2 timers glucose, faste insulin og type 2
diabetes. Resultaterner tyder på, at fysisk aktivetet har en betydning for type 2 diabetes, men at der er
behov for at undersøge yderligere risikofaktorer der kan have betydning for udviklingen af type 2 diabetes i
Grønland. Denne afhandling understreger betydningen af fortsat at måle udviklingen i fysisk aktivitet i
Grønland i takt med den økonomiske, kulturelle og sociale transition. Kombinationen af spørgeskemadata
og data fra en kombineret accelerometer og hjerterytme monitor gav værdifuld information om forskellige
dimensioner af fysisk aktivitet og var anvendelig i en arktisk kontekst. I et folkesundhedsperspektiv er det
væsentligt at fremme fysisk aktivitet i fritiden og reducere stillesiddende adfærd i forhold til at bibeholde
en fysisk aktiv livsstil og for at bidrage til forebyggelsen af type 2 diabetes i Grønland.
51
Grønlandsk resumé
Grønlandsk resumé
Kalaallisut eqikkaaneq
Issittumi nunap inoqqaavi, tassungalu ilanngullugu Kalaallit Nunaat, 1940-kunnili kulturikkutaningaasaqarnikkut- inuiaqatigiinnilu inooriaatsimikkut atugaat, sukkasuumik allanngoriartorsimapput,
tamannalu ilisarnaatigisimavaa inooriaatsikkut mutiusumik, nunani kippasissuni assingusumik
inooriaaseqalersimaneq. Illoqarfinni annerusuni inooriaatsikkut allanngorneq annertuneq pisimavoq.
Inuiaqatigiit piniartutut aalisartutullu imminut napatissimasuniit, ullumikkut issianerulluni
suliaqartalernerup nassatarisimava, timimik minnerusumik atuinermik inooriaaseqalersimaneq.
Allanngornerit taakku saniatigut anigorsinnaanngisanik nappaatinik, soorlu inooriaatsimut
attuumassutilimmik sukkorneq annertuumik nassaarfiusimavoq. Timimik atuisarnerup allanngorsimanera,
inooriaatsimut attuumassutilinnut nappaatit amerlisimanerannut peqqutaaqataasorineqarpoq. Timip
atortarnerannut, qanorlu inuiaqatigiinni inooriaatsikkut- aningaasaqarnikkut- kulturikkullu imminnut
attuumassuteqarnersut ilisimasagut annikipput, tamanna aamma timip atortarneranut atatillugu
allanngoriartorneranik (transition) taaneqartartoq.
Ilisimatuutut allaatigisap matuma anguniagaa pingaarneq tassaasimavoq, Kalaallit Nunaani kalaallit
akornanni, timip atortarnerata allanngorsimanera, saniatigullu timip atortarnera, timimi sukku
suliarineqartarneranut atatillugu qanoq sunniuteqarnersoq. Taassumalu saniatigut immersugassaq
tunngavigalugu apeqqutit issittumi naleqqunnersut paasiniarlugit. Ilisimatuutut allaatigisaq pingasunik
siunertaqarpoq.
 Timip atortarnerannut immersugassaq tunngavigalugu apeqqutit, the long Physikal Activity
Questionaire, issittumi inooriaatsimut atorsinnaaneri, timillu atortarnerannut uummatillu
tillernerannut uuttuut ataqatigiitsillugit naliliiffiginssaat (allakkiaq I)
 Kalaallit Nunaanni timip atortarnera inuiaqatigiinni inooriaatsimut atatillugu allanngorsimaneranut
misissuineq (allakkiaq ll)
 Kalaallit Nunaanni timip atortarnera, timimi sukku suliarineqartarnerannut qanoq
sunniuteqarnersoq (allakkiaq lll)
Ilisimatuuttut allaaserisaq 2005 – 2010-mi Kalaallit Nunaanni innuttaasunik misissuinermit paasissutissanit
tunngaveqarpoq (Inuit Health in Transition Study). Paasissutissat illoqarfinni 9-ni nunaqarfinnilu 13-ni
tunngaveqarput, tassanilu timikkut misissuinerit, apersuinermi immersugassat, namminerlu apeqqutit
akisassat, katillugulu kalaallinit inersimasunit (>18 ukiullit) 3102-t peqataaffigineqarsimalluni. Nuna
tamakkerlugu peqataasut 64,9%-iupput. Apersuinermik immersugassaq naleqqussagaq, The International
Physical Activity Questionaire (IPAQ-L) timip atortarnerannut paasissutissanik katersuinermi atorneqarpoq.
52
Grønlandsk resumé
Saniatigut misissuinermi peqataasut ilaannut (n= 1995) timip atortarnerannut uummatillu tillernerannut
uuttuut atorneqarsimalluni.
Ilisimatuutut allaatigisap innutaasunut illoqarfinni nunaqarfinnilu najugalinnut IPAQ-L tamakkiisumik, timip
atortarnerannut nukinik atuinermi uuttuutissatut atorneqarsinnaasoq takutippaa, kisiannili timi
atorneranut- issiallunilu suliaqarnermut piffissamik atukkamik immikkoortitsinermi atorneqarsinnaanani.
Apersuinermi immersugassaq pingaarutilinnik timimik atuisarnermut takussutissanut paasissutissiivoq,
ilisimatuutullu allaatigisap Kalaallit Nunaanni sumiissuseq- inuutissarsiut ilinniagaqarnerlu aallaavigalugit
inuiaqatigiinni inooriaaseq allanngorsimanera ilutigalugu timimik atueriaaseq allanngorsimasoq takutippaa.
Angutit sivisunerusumik ilinniagallit illoqarfimmilu najugallit, suliffimmini timimik atuinnginnerupput, arnat
taamatulli angerlarsimaffimmini timimik atuinnginnerullutik, arnanut angutillu nunaqarfinni najugalinnut
ileqqutoqqat malinnerullugit inooriaasilinnut sanilliullugit. Tamanna aamma angallassinermut atatillugu
timimik atuisarnermut issiaanernerusumillu suliaqarnermut arnat angutillu akornanni takussaavoq. Suliffiup
avataatigut timimik atuisarnermut atatillugu immikkooruteqanngilaq. Tamakkiisumik timip atortarneranut
piffissaq atugaq, inuiaqatigiinni inooriaatsip allanngorsimanera aallaavigalugu allanngoriartunngilaq,
illuatungaanili misissuinerit siulliit timimik atuinermi nukinik atuineq angutit akornanni apparsimasoq
takutippaat. Misissuinerup ilusilersornera pissutigalugu timimik atuisarneq
allanngorsimaneranut naliliinissaq mianersortumik pissaaq.
Saniatigut ilisimatuuttut allaaserisap timip atornerannut atatillugu, nukinik atuineq insulin aammi
kimittussusaa, sukkornermut misissuinermut atatillugu sukkutorsimanerup nalunaaqutaq marluk
kingunerini, naakkut orsoqassusermut atuumassuteqanngitsoq. Tamannalu assigisaanik sukkornermut
misissuinermut atatillugu sukkutorsimanerup kingorna akusiuinikkut ((faste glucose, 2 timers glucose, faste
insulin type 2 diabetes-ilu) takussaasimanani.
Misissuinerup inerneri timip atortarnera inooriaatsimut tunngasumik sukkornermut
attuumassuteqarsinnaasoq takutippaa, kisiannili sukkornermut pilersitsisartut allat aamma Kalaallit
Nunaanni misissorneqarnissaat pisariaqarpoq. Ilisimatuut allaaserisap matuma timip atortarnera, qanoq
ineriartornera Kalaallit Nunaanni aningaasaqarnerup- kulturikkut- inuiaqatigiinnilu inooriaatsip
allanngoriartornera ilutigalugu, uuttortarneqarnissaa pingaaruteqarnera naqissuserpaa. Apersuinermi
immersugassat, timip atortarnerinut assigiinngitsunut, uummatillu tillernerannut uuttuut pingaarutilinnik
paasissutissanik pissarsiffiuvoq, issittumilu atorneqarsinnaallutik. Innuttaasut peqqissuunissaannut
atatillugu, sunngiffimmi timip atortarnissaannut kaammattuinissaq pingaaruteqarpoq, issiaannarlunilu
suliaqartarneq annikillisarnissaa, taamaalilluni inooriaatsimut atatillugu sukkortarnermut pitsaaliuinissamut
tapertaasinnaalluni.
53
Reference list
Reference list
(1) Bjerregaard P, Young TK, Dewailly E, Ebbesson SO. Indigenous health in the Arctic: an overview of
the circumpolar Inuit population. Scand J Public Health 2004;32(5):390-5.
(2) Jørgensen ME, Bjerregaard P, Borch-Johnsen K. Diabetes and impaired glucose tolerance among the
inuit population of greenland. Diabetes Care 2002;25(10):1766-71.
(3) Bjerregaard P, Mulvad G, Pedersen HS. Cardiovascular risk factors in Inuit of Greenland. Int J
Epidemiol 1997;1;26(6):1182-90.
(4) Chateau-Degat ML, Dewailly E, Louchini R, Counil E, Noel M, Ferland A, et al. Cardiovascular burden
and related risk factors among Nunavik (Quebec) Inuit: insights from baseline findings in the
circumpolar Inuit health in transition cohort study. Can J Cardiol 2010;26(6):190-6.
(5) Hopping BN, Erber E, Beck L, De Roose E, Sharma S. Inuvialuit adults in the Canadian Arctic have a
high body mass index and self-reported physical activity. J Hum Nutr Diet 2010;23 Suppl 1:115-9.
(6) Hopping BN, Erber E, Mead E, Roache C, Sharma S. High levels of physical activity and obesity coexist amongst Inuit adults in Arctic Canada. J Hum Nutr Diet 2010;23(suppl. 1):110-4.
(7) Redwood D, Schumacher MC, Lanier AP, Ferucci ED, Asay E, Helzer LJ, et al. Physical activity
patterns of American Indian and Alaskan Native people living in Alaska and the Southwestern
United States. Am J Health Promot 2009;23(6):388-95.
(8) Young TK, Katzmarzyk PT. [Physical activity among aboriginals in Canada]. Appl Physiol Nutr Metab
2007;32 Suppl 2F:S165-S178.
(9) Findlay LC. Physical activity among First Nations people off reserve, Metis and Inuit. Health Rep
2011;22(1):47-54.
(10) Rode A, Shephard RJ. Physiological consequences of acculturation: a 20-year study of fitness in an
Inuit community. Eur J Appl Physiol Occup Physiol 1994;69(6):516-24.
(11) Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness:
definitions and distinctions for health-related research. Public Health Rep 1985;100(2):126-31.
(12) Sallis JF, Saelens BE. Assessment of physical activity by self-report: status, limitations, and future
directions. Res Q Exerc Sport 2000;71(2 Suppl):S1-14.
(13) Ainsworth BE, Haskell WL, Leon AS, Jacobs DR, Jr., Montoye HJ, Sallis JF, et al. Compendium of
physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc
1993;25(1):71-80.
(14) Ainsworth BE, Haskell WL, Whitt MC, Irwin ML, Swartz AM, Strath SJ, et al. Compendium of physical
activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 2000;32(9
Suppl):S498-S504.
5ϰ
Reference list
(15) Ainsworth BE, Haskell WL, Herrmann SD, Meckes N, Bassett DR, Tudor-Locke C, et al. 2011
Compendium of Physical Activities: A Second Update of Codes and MET Values. Med Sci Sports
Exerc 2011;43(8):1575-81.
(16) Besson H, Brage S, Jakes RW, Ekelund U, Wareham NJ. Estimating physical activity energy
expenditure, sedentary time, and physical activity intensity by self-report in adults. Am J Clin Nutr
2010;91(1):106-14.
(17) Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT. Effect of physical inactivity on major
non-communicable diseases worldwide: an analysis of burden of disease and life expectancy.
Lancet 2012;380(9838):219-29.
(18) Morris JN, Heady JA, Raffle PA, Roberts CG, Parks JW. Coronary heart-disease and physical activity
of work. Lancet 1953;265(6795):1053-7.
(19) Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al. Physical activity and public
health: updated recommendation for adults from the American College of Sports Medicine and the
American Heart Association. Med Sci Sports Exerc 2007;39(8):1423-34.
(20) Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, 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):402-7.
(21) Garber CE, Blissmer B, Deschenes MR, Franklin BA, Lamonte MJ, Lee IM, et al. American College of
Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining
cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance
for prescribing exercise. Med Sci Sports Exerc 2011;43(7):1334-59.
(22) Wen CP, Wai JP, Tsai MK, Yang YC, Cheng TY, Lee MC, et al. Minimum amount of physical activity
for reduced mortality and extended life expectancy: a prospective cohort study. Lancet
2011;378(9798):1244-53.
(23) Kohl HW, III. Physical activity and cardiovascular disease: evidence for a dose response. Med Sci
Sports Exerc 2001;33(6 Suppl):S472-S483.
(24) Lee IM, Skerrett PJ. Physical activity and all-cause mortality: what is the dose-response relation?
Med Sci Sports Exerc 2001;33(6 Suppl):S459-S471.
(25) Sattelmair J, Pertman J, Ding EL, Kohl HW, III, Haskell W, Lee IM. Dose response between physical
activity and risk of coronary heart disease: a meta-analysis. Circulation 2011;124(7):789-95.
(26) De ti kostråd. Ernæringsrådet i Grønland . 17-12-2012. Available from:
http://www.paarisa.gl/home/suliniutit/nerisat_aalanerlu/nerisassat.aspx
(27) Bjerregaard P. Rapid socio-cultural change and health in the Arctic. Int J Circumpolar Health
2001;60(2):102-11.
(28) Young TK, Bjerregaard P. Health transitions in Arctic populations. Toronto: University of Toronto
Press; 2008;23-39.
5ϱ
Reference list
(29) Bjerregaard P, Curtis T. Cultural change and mental health in Greenland: the association of
childhood conditions, language, and urbanization with mental health and suicidal thoughts among
the Inuit of Greenland. Soc Sci Med 2002;54(1):33-48.
(30) Gracey M, King M. Indigenous health part 1: determinants and disease patterns. Lancet 2009
4;374(9683):65-75.
(31) Omran AR. The epidemiologic transition: a theory of the epidemiology of population change. 1971.
Milbank Q 2005;83(4):731-57.
(32) Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: general
considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation
2001;104(22):2746-53.
(33) Katzmarzyk PT, Mason C. The physical activity transition. J Phys Act Health 2009;6(3):269-80.
(34) Snodgrass JJ, Leonard WR, Tarskaia LA, Schoeller DA. Total energy expenditure in the Yakut (Sakha)
of Siberia as measured by the doubly labeled water method. Am J Clin Nutr 2006;84(4):798-806.
(35) Pontzer H, Raichlen DA, Wood BM, Mabulla AZ, Racette SB, Marlowe FW. Hunter-gatherer
energetics and human obesity. PLoS One 2012;7(7):e40503.
(36) Rode A, Shephard RJ. Ten years of "civilization": fitness of Canadian Inuit. J Appl Physiol
1984;56(6):1472-7.
(37) King H, Aubert RE, Herman WH. Global burden of diabetes, 1995-2025: prevalence, numerical
estimates, and projections. Diabetes Care 1998;21(9):1414-31.
(38) Helmrich SP, Ragland DR, Leung RW, Paffenbarger RS, Jr. Physical activity and reduced occurrence
of non-insulin-dependent diabetes mellitus. N Engl J Med 1991;325(3):147-52.
(39) Helmrich SP, Ragland DR, Paffenbarger RS, Jr. Prevention of non-insulin-dependent diabetes
mellitus with physical activity. Med Sci Sports Exerc 1994;26(7):824-30.
(40) Laaksonen DE, Lindstrom J, Lakka TA, Eriksson JG, Niskanen L, Wikstrom K, et al. Physical activity in
the prevention of type 2 diabetes: the Finnish diabetes prevention study. Diabetes 2005;54(1):15865.
(41) Lynch J, Helmrich SP, Lakka TA, Kaplan GA, Cohen RD, Salonen R, et al. Moderately intense physical
activities and high levels of cardiorespiratory fitness reduce the risk of non-insulin-dependent
diabetes mellitus in middle-aged men. Arch Intern Med 1996;156(12):1307-14.
(42) Manson JE, Nathan DM, Krolewski AS, Stampfer MJ, Willett WC, Hennekens CH. A prospective study
of exercise and incidence of diabetes among US male physicians. Jama 1992;268(1):63-7.
(43) Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P, et al. Prevention of
type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N
Engl J Med 2001;344(18):1343-50.
5ϲ
Reference list
(44) Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, et al. Reduction in
the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med
2002;346(6):393-403.
(45) Kriska AM, Edelstein SL, Hamman RF, Otto A, Bray GA, Mayer-Davis EJ, et al. Physical activity in
individuals at risk for diabetes: Diabetes Prevention Program. Med Sci Sports Exerc 2006;38(5):82632.
(46) Lindstrom J, Louheranta A, Mannelin M, Rastas M, Salminen V, Eriksson J, et al. The Finnish
Diabetes Prevention Study (DPS): Lifestyle intervention and 3-year results on diet and physical
activity. Diabetes Care 2003;26(12):3230-6.
(47) Pi-Sunyer X, Blackburn G, Brancati FL, Bray GA, Bright R, Clark JM, et al. Reduction in weight and
cardiovascular disease risk factors in individuals with type 2 diabetes: one-year results of the look
AHEAD trial. Diabetes Care 2007;30(6):1374-83.
(48) Kriska AM, Hanley AJ, Harris SB, Zinman B. Physical activity, physical fitness, and insulin and glucose
concentrations in an isolated Native Canadian population experiencing rapid lifestyle change.
Diabetes Care 2001;24(10):1787-92.
(49) Adler AI, Boyko EJ, Schraer CD, Murphy NJ. The negative association between traditional physical
activities and the prevalence of glucose intolerance in Alaska Natives. Diabet Med 1996;13(6):55560.
(50) Jørgensen ME, Moustgaard H, Bjerregaard P, Borch-Johnsen K. Gender differences in the
association between westernization and metabolic risk among Greenland Inuit. Eur J Epidemiol
2006;21(10):741-8.
(51) Assah FK, Ekelund U, Brage S, Mbanya JC, Wareham NJ. Free-Living Physical Activity Energy
Expenditure Is Strongly Related to Glucose Intolerance in Cameroonian Adults Independently of
Obesity. Diabetes Care 2009;32(2):367-9.
(52) Assah FK, Ekelund U, Brage S, Mbanya JC, Wareham NJ. Urbanization, physical activity, and
metabolic health in sub-Saharan Africa. Diabetes Care 2011;34(2):491-6.
(53) Healy GN, Wijndaele K, Dunstan DW, Shaw JE, Salmon J, Zimmet PZ, et al. Objectively measured
sedentary time, physical activity, and metabolic risk: the Australian Diabetes, Obesity and Lifestyle
Study (AusDiab). Diabetes Care 2008;31(2):369-71.
(54) Balkau B, Mhamdi L, Oppert JM, Nolan J, Golay A, Porcellati F, et al. Physical activity and insulin
sensitivity: the RISC study. Diabetes 2008;57(10):2613-8.
(55) Ekelund U, Franks PW, Sharp S, Brage S, Wareham NJ. Increase in physical activity energy
expenditure is associated with reduced metabolic risk independent of change in fatness and fitness.
Diabetes Care 2007;30(8):2101-6.
(56) Assah FK, Brage S, Ekelund U, Wareham NJ. The association of intensity and overall level of physical
activity energy expenditure with a marker of insulin resistance. Diabetologia 2008;51(8):1399-407.
5ϳ
Reference list
(57) Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide variability in physical inactivity a 51country survey. Am J Prev Med 2008;34(6):486-94.
(58) Mbalilaki JA, Masesa Z, Stromme SB, Hostmark AT, Sundquist J, Wandell P, et al. Daily energy
expenditure and cardiovascular risk in Masai, rural and urban Bantu Tanzanians. Br J Sports Med
2010;44:121-126.
(59) Monda KL, Gordon-Larsen P, Stevens J, Popkin BM. China's transition: the effect of rapid
urbanization on adult occupational physical activity. Soc Sci Med 2007;64(4):858-70.
(60) Sobngwi E, Mbanya JC, Unwin NC, Kengne AP, Fezeu L, Minkoulou EM, et al. Physical activity and its
relationship with obesity, hypertension and diabetes in urban and rural Cameroon. Int J Obes Relat
Metab Disord 2002;26(7):1009-16.
(61) Yamauchi T, Umezaki M, Ohtsuka R. Influence of urbanisation on physical activity and dietary
changes in Huli-speaking population: a comparative study of village dwellers and migrants in urban
settlements. Br J Nutr 2001;85(1):65-73.
(62) Christensen DL, Friis H, Mwaniki DL, Kilonzo B, Tetens I, Boit MK, et al. Prevalence of glucose
intolerance and associated risk factors in rural and urban populations of different ethnic groups in
Kenya. Diabetes Res Clin Pract 2009;84(3):303-310.
(63) Levine JA, McCrady SK, Boyne S, Smith J, Cargill K, Forrester T. Non-exercise physical activity in
agricultural and urban people. Urban Stud 2011;48(11):2417-27.
(64) Ebrahim S, Kinra S, Bowen L, Andersen E, Ben-Shlomo Y, Lyngdoh T, et al. The effect of rural-tourban migration on obesity and diabetes in India: a cross-sectional study. PLoS Med
2010;7(4):e1000268.
(65) Mohan V, Deepa M, Anjana RM, Lanthorn H, Deepa R. Incidence of diabetes and pre-diabetes in a
selected urban south Indian population (CUPS-19). J Assoc Physicians India 2008;56:152-7.
(66) Jorgensen ME, Borch-Johnsen K, Witte DR, Bjerregaard P. Diabetes in Greenland and its
relationship with urbanization. Diabet Med 2012;29(6):755-60.
(67) Sobngwi E, Mbanya JC, Unwin NC, Porcher R, Kengne AP, Fezeu L, et al. Exposure over the life
course to an urban environment and its relation with obesity, diabetes, and hypertension in rural
and urban Cameroon. Int J Epidemiol 2004;33(4):769-76.
(68) Allender S, Foster C, Hutchinson L, Arambepola C. Quantification of urbanization in relation to
chronic diseases in developing countries: a systematic review. J Urban Health 2008;85(6):938-51.
(69) Bjerregaard P, Dahl-Petersen IK. How well does social variation mirror secular change in prevalence
of cardiovascular risk factors in a country in transition? Am J Hum Biol 2011;23(6):774-9.
(70) van Poppel MNM, Chinapaw MJM, Mokkink LB, van Mechelen W, Terwee CB. Physical Activity
Questionnaires for Adults A Systematic Review of Measurement Properties. Sports Medicine
2010;40(7):565-600.
5ϴ
Reference list
(71) Helmerhorst HJ, Brage S, Warren J, Besson H, Ekelund U. A systematic review of reliability and
objective criterion-related validity of physical activity questionnaires. Int J Behav Nutr Phys Act
2012;9:103.
(72) Bauman A, Bull F, Chey T, Craig CL, Ainsworth BE, Sallis JF, et al. The International Prevalence Study
on Physical Activity: results from 20 countries. Int J Behav Nutr Phys Act 2009;6(1):21.
(73) IPAQ-group. IPAQ. IPAQ-group 2007. Available from: URL: www.ipaq.ki.se
(74) Hallal PC, Simoes E, Reichert FF, Azevedo MR, Ramos LR, Pratt M, et al. Validity and Reliability of
the Telephone-Administered International Physical Activity Questionnaire in Brazil. Journal of
Physical Activity & Health 2010;7(3):402-9.
(75) Lachat CK, Verstraeten R, Khanh lN, Hagstromer M, Khan NC, Van Ndo A, et al. Validity of two
physical activity questionnaires (IPAQ and PAQA) for Vietnamese adolescents in rural and urban
areas. Int J Behav Nutr Phys Act 2008;5:37.
(76) Macfarlane D, Chan A, Cerin E. Examining the validity and reliability of the Chinese version of the
International Physical Activity Questionnaire, long form (IPAQ-LC). Public Health Nutr 2010 13;1-8.
(77) Nang EEK, Ngunjiri SAG, Wu Y, Salim A, Tai ES, Lee J, et al. Validity of the international physical
activity questionnaire and the Singapore prospective study program physical activity questionnaire
in a multiethnic urban Asian population. Bmc Medical Research Methodology 2011 13;11.
(78) Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, et al. International
physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc
2003;35(8):1381-95.
(79) Westerterp KR, Plasqui G. Physical activity and human energy expenditure. Curr Opin Clin Nutr
Metab Care 2004;7(6):607-13.
(80) Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United
States measured by accelerometer. Med Sci Sports Exerc 2008;40(1):181-8.
(81) John D, Freedson P. ActiGraph and Actical physical activity monitors: a peek under the hood. Med
Sci Sports Exerc 2012;44(1 Suppl 1):S86-S89.
(82) Freedson PS, Lyden K, Kozey-Keadle S, Staudenmayer J. Evaluation of artificial neural network
algorithms for predicting METs and activity type from accelerometer data: validation on an
independent sample. J Appl Physiol 2011;111(6):1804-12.
(83) Corder K, Brage S, Ekelund U. Accelerometers and pedometers: methodology and clinical
application. Curr Opin Clin Nutr Metab Care 2007;10(5):597-603.
(84) Ceesay SM, Prentice AM, Day KC, Murgatroyd PR, Goldberg GR, Scott W, et al. The use of heart rate
monitoring in the estimation of energy expenditure: a validation study using indirect whole-body
calorimetry. Br J Nutr 1989;61(2):175-86.
(85) Livingstone MB. Heart-rate monitoring: the answer for assessing energy expenditure and physical
activity in population studies? Br J Nutr 1997;78(6):869-71.
5ϵ
Reference list
(86) Assah FK, Ekelund U, Brage S, Wright A, Mbanya JC, Wareham NJ. Accuracy and validity of a
combined heart rate and motion sensor for the measurement of free-living physical activity energy
expenditure in adults in Cameroon. Int J Epidemiol 2011;40(1):112-20.
(87) Brage S, Brage N, Franks P, Ekelund U, Wareham N. Reliability and validity of the ActiHeart®: A
single-piece instrument that measures acceleration, heart rate, and heart rate variability. 2003.
(88) Crouter SE, Churilla JR, Bassett DRJ. Accuracy of the Actiheart for the assessment of energy
expenditure in adults. Eur J Clin Nutr 2008;62(6):704-11.
(89) IPAQ-group. IPAQ analyse guidelines . IPAQ-group 2011. Available from: URL:
http://www.ipaq.ki.se/scoring.pdf
(90) Brage S, Ekelund U, Brage N, Hennings MA, Froberg K, Franks PW, et al. Hierarchy of individual
calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol
2007;103(2):682-92.
(91) Brage S, Brage N, Franks PW, Ekelund U, Wong MY, Andersen LB, et al. Branched equation
modeling of simultaneous accelerometry and heart rate monitoring improves estimate of directly
measured physical activity energy expenditure. J Appl Physiol 2004;96(1):343-51.
(92) Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical
measurement. Lancet 1986;1(8476):307-10.
(93) Henry CJK. Basal metabolic rate studies in humans: measurement and development of new
equations. Public Health Nutrition 2005;8(7A):1133-52.
(94) Boon RM, Hamlin MJ, Steel GD, Ross JJ. Validation of the New Zealand Physical Activity
Questionnaire (NZPAQ-LF) and the International Physical Activity Questionnaire (IPAQ-LF) with
Accelerometry. Br J Sports Med 2010;44(10):741-6.
(95) Hagstromer M, Oja P, Sjostrom M. The International Physical Activity Questionnaire (IPAQ): a study
of concurrent and construct validity. Public Health Nutr 2006;9(6):755-62.
(96) Sebastiao E, Gobbi S, Chodzko-Zajko W, Schwingel A, Papini CB, Nakamura PM, et al. The
International Physical Activity Questionnaire-long form overestimates self-reported physical activity
of Brazilian adults. Public Health 2012;126(11):967-75.
(97) Hagstromer M, Ainsworth BE, Oja P, Sjostrom M. Comparison of a subjective and an objective
measure of physical activity in a population sample. J Phys Act Health 2010;7(4):541-50.
(98) Ekelund U, Sepp H, Brage S, Becker W, Jakes R, Hennings M, et al. Criterion-related validity of the
last 7-day, short form of the International Physical Activity Questionnaire in Swedish adults. Public
Health Nutr 2006;9(2):258-65.
(99) Klesges RC, Eck LH, Mellon MW, Fulliton W, Somes GW, Hanson CL. The accuracy of self-reports of
physical activity. Med Sci Sports Exerc 1990;22(5):690-7.
(100) Altschuler A, Picchi T, Nelson M, Rogers JD, Hart J, Sternfeld B. Physical activity questionnaire
comprehension: lessons from cognitive interviews. Med Sci Sports Exerc 2009;41(2):336-43.
ϲϬ
Reference list
(101) Ekelund U, Brage S, Griffin SJ, Wareham NJ. Objectively measured moderate- and vigorous-intensity
physical activity but not sedentary time predicts insulin resistance in high-risk individuals. Diabetes
Care 2009;32(6):1081-6.
(102) Healy GN, Dunstan DW, Salmon J, Cerin E, Shaw JE, Zimmet PZ, et al. Objectively measured lightintensity physical activity is independently associated with 2-h plasma glucose. Diabetes Care
2007;30(6):1384-9.
(103) Dunstan DW, Salmon J, Owen N, Armstrong T, Zimmet PZ, Welborn TA, et al. Associations of TV
viewing and physical activity with the metabolic syndrome in Australian adults. Diabetologia
2005;48(11):2254-61.
(104) Hamilton MT, Healy GN, Dunstan DW, Zderic TW, Owen N. Too Little Exercise and Too Much Sitting:
Inactivity Physiology and the Need for New Recommendations on Sedentary Behavior. Curr
Cardiovasc Risk Rep 2008;2(4):292-8.
(105) Owen N, Bauman A, Brown W. Too much sitting: a novel and important predictor of chronic disease
risk? Br J Sports Med 2009;43(2):81-3.
(106) Wareham NJ, Wong MY, Day NE. Glucose intolerance and physical inactivity: the relative
importance of low habitual energy expenditure and cardiorespiratory fitness. Am J Epidemiol
2000;152(2):132-9.
(107) Shephard RJ, Aoyagi Y. Measurement of human energy expenditure, with particular reference to
field studies: an historical perspective. Eur J Appl Physiol 2012;112(8):2785-815.
(108) Eriksen L, Dahl-Petersen I, Haugaard SB, Dela F. Comparison of the effect of multiple short-duration
with single long-duration exercise sessions on glucose homeostasis in type 2 diabetes mellitus.
Diabetologia 2007;50(11):2245-53.
(109) Murphy MH, Blair SN, Murtagh EM. Accumulated versus Continuous Exercise for Health Benefit A
Review of Empirical Studies. Sports Medicine 2009;39(1):29-43.
(110) Rode A, Shephard RJ. Fitness of the Canadian Eskimo--the influence of season. Med Sci Sports
1973;5(3):170-3.
(111) Dahly DL, Adair LS. Quantifying the urban environment: a scale measure of urbanicity outperforms
the urban-rural dichotomy. Soc Sci Med 2007;64(7):1407-19.
(112) McDade TW, Adair LS. Defining the "urban" in urbanization and health: a factor analysis approach.
Soc Sci Med 2001;53(1):55-70.
(113) Stamatakis E, Ekelund U, Wareham NJ. Temporal trends in physical activity in England: the Health
Survey for England 1991 to 2004. Prev Med 2007;45(6):416-23.
(114) Petersen CB, Thygesen LC, Helge JW, Gronbaek M, Tolstrup JS. Time trends in physical activity in
leisure time in the Danish population from 1987 to 2005. Scand J Public Health 2010;38(2):121-8.
ϲϭ
Reference list
(115) Vorster HH, Venter CS, Wissing MP, Margetts BM. The nutrition and health transition in the North
West Province of South Africa: a review of the THUSA (Transition and Health during Urbanisation of
South Africans) study. Public Health Nutr 2005;8(5):480-90.
(116) Mendez MA, Monteiro CA, Popkin BM. Overweight exceeds underweight among women in most
developing countries. Am J Clin Nutr 2005;81(3):714-21.
(117) Kriska AM, Pereira MA, Hanson RL, de Court, Zimmet PZ, Alberti KG, et al. Association of physical
activity and serum insulin concentrations in two populations at high risk for type 2 diabetes but
differing by BMI. Diabetes Care 2001;24(7):1175-80.
(118) Kriska AM, LaPorte RE, Pettitt DJ, Charles MA, Nelson RG, Kuller LH, et al. The association of
physical activity with obesity, fat distribution and glucose intolerance in Pima Indians. Diabetologia
1993;36(9):863-9.
(119) Pereira MA, Kriska AM, Joswiak ML, Dowse GK, Collins VR, Zimmet PZ, et al. Physical inactivity and
glucose intolerance in the multiethnic island of Mauritius. Med Sci Sports Exerc 1995;27(12):162634.
(120) Faerch K, Borch-Johnsen K, Holst JJ, Vaag A. Pathophysiology and aetiology of impaired fasting
glycaemia and impaired glucose tolerance: does it matter for prevention and treatment of type 2
diabetes? Diabetologia 2009;52(9):1714-23.
(121) Rana JS, Li TY, Manson JE, Hu FB. Adiposity compared with physical inactivity and risk of type 2
diabetes in women. Diabetes Care 2007;30(1):53-8.
(122) Hu G, Lindstrom J, Valle TT, Eriksson JG, Jousilahti P, Silventoinen K, et al. Physical activity, body
mass index, and risk of type 2 diabetes in patients with normal or impaired glucose regulation. Arch
Intern Med 2004;164(8):892-6.
(123) Weinstein AR, Sesso HD, Lee IM, Cook NR, Manson JE, Buring JE, et al. Relationship of physical
activity vs body mass index with type 2 diabetes in women. Jama 2004;292(10):1188-94.
(124) Morton LM, Cahill J, Hartge P. Reporting participation in epidemiologic studies: a survey of practice.
Am J Epidemiol 2006;163(3):197-203.
(125) Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart
rate and movement sensor Actiheart. Eur J Clin Nutr 2005;59(4):561-70.
(126) Rennie KL, Wareham NJ. The validation of physical activity instruments for measuring energy
expenditure: problems and pitfalls. Public Health Nutr 1998;1(4):265-71.
(127) Bauman A, Ainsworth BE, Sallis JF, Hagstromer M, Craig CL, Bull FC, et al. The Descriptive
Epidemiology of Sitting A 20-Country Comparison Using the International Physical Activity
Questionnaire (IPAQ). American Journal of Preventive Medicine 2011;41(2):228-35.
(128) Healy GN, Dunstan DW, Salmon J, Cerin E, Shaw JE, Zimmet PZ, et al. Breaks in sedentary time:
beneficial associations with metabolic risk. Diabetes Care 2008;31(4):661-6.
6Ϯ
Reference list
(129) Holtermann A, Hansen JV, Burr H, Sogaard K, Sjogaard G. The health paradox of occupational and
leisure-time physical activity. Br J Sports Med 2012;46(4):291-5.
(130) Heath GW, Gavin JR, III, Hinderliter JM, Hagberg JM, Bloomfield SA, Holloszy JO. Effects of exercise
and lack of exercise on glucose tolerance and insulin sensitivity. J Appl Physiol 1983;55(2):512-7.
(131) Koivisto VA, Yki-Jarvinen H, DeFronzo RA. Physical training and insulin sensitivity. Diabetes Metab
Rev 1986;1(4):445-81.
(132) Kriska A. Can a physically active lifestyle prevent type 2 diabetes? Exerc Sport Sci Rev
2003;31(3):132-7.
(133) Intille SS, Lester J, Sallis JF, Duncan G. New horizons in sensor development. Med Sci Sports Exerc
2012;44(1 Suppl 1):S24-S31.
(134) Coble JD, Rhodes RE. Physical Activity and Native Americans: A Review. American Journal of
Preventive Medicine 2006;31(1):36-46.
(135) Sallis JF, Bauman A, Pratt M. Environmental and policy interventions to promote physical activity.
Am J Prev Med 1998;15(4):379-97.
(136) Sallis JF, Cervero RB, Ascher W, Henderson KA, Kraft MK, Kerr J. An ecological approach to creating
active living communities. Annu Rev Public Health 2006;27:297-322.
(137) Strath SJ, Swartz AM, Bassett DR, Jr., O'Brien WL, King GA, Ainsworth BE. Evaluation of heart rate as
a method for assessing moderate intensity physical activity. Med Sci Sports Exerc 2000;32(9
Suppl):S465-S470.
(138) Strath SJ, Brage S, Ekelund U. Integration of physiological and accelerometer data to improve
physical activity assessment. Med Sci Sports Exerc 2005;37(11 Suppl):S563-S571.
(139) Stegle O, Fallert SV, Mackay DJC, Brage S. Gaussian process robust regression for noisy heart rate
data. Ieee Transactions on Biomedical Engineering 2008;55(9):2143-51.
6ϯ
Appendix I
Appendix I – danish questionnaire
/EQU 0 UNDWIUDEORGSU¡YH
VHGGHONO EHVLQGKHU
,QWHUYLHZHUBBBBBBBBBBBBBBBBBBBBBBBBBB
'DWR
6WDUWSnLQWHUYLHZBBBBBBBBBBB
6OXWSnLQWHUYLHZBBBBBBBBBBB
6ϰ
De næste spørgsmål drejer sig om hvor lang tid du har været fysisk aktiv
de sidste 7 dage. Den første del handler om dit arbejde, herunder jagt og
fiskeri som erhverv, men ikke husarbejde
57.
Har du for øjeblikket arbejde uden for hjemmet?
58.
I løbet af de sidste 7 dage, hvor mange dage har du udført hård fysisk
aktivitet på dit arbejde? Tænk kun på aktiviteter som du udfører mindst 10
minutter ad gangen? (Hård fysisk aktivitet er aktivitet, som er meget fysisk
anstrengende, og hvor du øger din vejrtrækning meget; f.eks. tunge løft,
gravearbejde, tungt byggearbejde, trappegang)
59.
Hvor lang tid brugte du i gennemsnit om dagen på hård fysisk aktivitet?
BBBBBWLPHURPGDJHQ
BBBBBPLQXWWHURPGDJHQ
60.
I løbet af de sidste 7 dage, hvor mange dage har du udført moderat fysisk
aktivitet som på dit arbejde? (Moderat aktivitet er mindre anstrengende og øger
vejrtrækningen i nogen grad; f.eks. mindre løft)
BBBBBGDJHRPXJHQ
KDULNNHKnUGWI\VLVNDUEHMGH →JnWLOVSP
MD 
QHM →JnWLOVSP
BBBBBGDJHRPXJHQ
KDULNNHPRGHUDWI\VLVNDUEHMGH →JnWLOVSP
61.
Hvor lang tid brugte du i gennemsnit om dagen på moderat fysisk aktivitet?
BBBBBWLPHURPGDJHQ
62.
I løbet af de sidste 7 dage, hvor mange dage har du gået mindst 10 min. ad
gangen på dit arbejde? Medregn ikke gang til og fra arbejde
63.
Hvor lang tid brugte du i gennemsnit om dagen på at gå på dit arbejde?
BBBBBWLPHURPGDJHQ
BBBBBGDJHRPXJHQ
KDULNNHJnHQGHDUEHMGH →JnWLOVSP
De næste spørgsmål handler om transport fra sted til sted
64.
I løbet af de sidste 7 dage, hvor mange dage har du kørt i bil, bus eller
snescooter?
65.
Hvor lang tid brugte du i gennemsnit om dagen på at transportere dig med
bil, bus eller snescooter?
BBBBBWLPHURPGDJHQ
66.
BBBBBGDJHRPXJHQ
KDULNNHN¡UWPHGELOEXVHOOHUVQHVFRRWHU →JnWLOVSP
I løbet af de sidste 7 dage, hvor mange dage har du cyklet mindst 10 min. ad
gangen for at komme fra sted til sted?
BBBBBGDJHRPXJHQ
KDULNNHF\NOHW →JnWLOVSP
67.
Hvor lang tid brugte du i gennemsnit om dagen på at cykle fra sted til sted?
BBBBBWLPHURPGDJHQ
68.
gangen fra sted til sted?
69.
Hvor lang tid brugte du i gennemsnit om dagen på at gå fra sted til sted?
BBBBBWLPHURPGDJHQ
BBBBBGDJHRPXJHQ
KDULNNHJnHWIUDVWHGWLOVWHG →JnWLOVSP
De næste spørgsmål handler om arbejde i og omkring hjemmet f.eks.
husarbejde, reparationer, vedligeholdelse og pasning af børn og familie.
Tænk kun på fysisk aktivitet som du udfører mindst 10 minutter ad
gangen
70.
aktivitet i hjemmet? (f.eks. tunge løft, skovle sne, gravearbejde, hente vand)
BBBBBGDJHRPXJHQ
KDULNNHXGI¡UWWXQJWI\VLVNDUEHMGHLKMHPPHW →JnWLOVSP
71.
Hvor lang tid brugte du i gennemsnit om dagen på at udføre hårdt fysisk
arbejde i hjemmet?
BBBBBWLPHURPGDJHQ
72.
Hvor mange dage har du udført moderat fysisk aktivitet i hjemmet? (f.eks.
reparationer i hjemmet, reparation af udstyr, rengøring og tøjvask, pleje af børn
eller gamle)
73.
Hvor lang tid brugte du i gennemsnit om dagen på moderat aktivitet i
hjemmet?
BBBBBWLPHURPGDJHQ
BBBBBGDJHRPXJHQ
KDULNNHXGI¡UWPRGHUDWI\VLVNDNWLYLWHW →JnWLOVSP
De næste spørgsmål handler om motion, sport og anden fysisk aktivitet i
fritiden. Medregn ikke aktiviteter, som du allerede har beskrevet i de
foregående afsnit
74.
gangen i fritiden?
75.
Hvor lang tid brugte du i gennemsnit om dagen på at gå i fritiden?
BBBBBWLPHURPGDJHQ
BBBBBGDJHRPXJHQ
KDULNNHJnHWLIULWLGHQ →JnWLOVSP
76.
aktivitet i fritiden? (f.eks. aerobics, løb, kampsport, fodbold, skiløb)
BBBBBGDJHRPXJHQ
LQJHQKnUGI\VLVNDNWLYLWHWLIULWLGHQ →JnWLOVSP
77.
Hvor lang tid brugte du i gennemsnit om dagen på hård fysisk aktivitet i
fritiden?
BBBBBWLPHURPGDJHQ
78.
Hvor mange dage har du udført moderat fysisk aktivitet i fritiden? (f.eks.
cykling i lavt tempo, svømning i lavt tempo, vandreture)
79.
Hvor lang tid brugte du i gennemsnit om dagen på moderat fysisk aktivitet i
fritiden?
BBBBBWLPHURPGDJHQ
BBBBBGDJHRPXJHQ
LQJHQPRGHUDWDNWLYLWHWLIULWLGHQ →JnWLOVSP
De sidste spørgsmål handler om den tid, du sidder stille på arbejdet og i
fritiden (f.eks. sidde ved et skrivebord, besøge venner, læse, computer og TV)
Medregn ikke bilkørsel o.l.
80.
I løbet af de sidste 7 dage, hvor meget tid har du brugt på stillesiddende
aktiviteter på hverdage?
BBBBBWLPHURPGDJHQ
81.
I løbet af de sidste 7 dage, hvor meget tid har du brugt på stillesiddende
aktiviteter om dagen i weekenden?
BBBBBWLPHURPGDJHQ
Appendix II
Appendix II – greenlandic questionnaire
/EQU 0 UNDWIUDEORGSU¡YH
VHGGHONO EHVLQGKHU
,QWHUYLHZHUBBBBBBBBBBBBBBBBBBBBBBBBBB
'DWR
6WDUWSnLQWHUYLHZBBBBBBBBBBB
6OXWSnLQWHUYLHZBBBBBBBBBBB
ϳϬ
Apeqqutit tulliit ulluni kingullerni piffissami qanoq sivisutigisumik timinnik
atuisimanernut tunngassuteqarput. Immikkoortoq siulleq sulinernut
tunngassuteqarpoq, tassani inuussutissarsiutigalugu aallaaniarneq aalisarnelu
ilanngullugit, angerlarsimaffimmili suliat ilanngunnagit
57.
Massakkut angerlarsimaffiup avataani suliffeqarpit?
58.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit suliffinni
oqimaatsunik suliaqartarpit? Suliat 10 minuttinit sivisunerusumik
suliarineqartut kisiisa eqqarsaatigikkit? (Suliat oqimaatsut annertuumik
nukissorfiusut, anerterinerulernernillu kinguneqartut; soorlu oqimaatsunik
kivitsinerit, assaanerit, sanaartornermi oqimaatsunik suliaqarneq,
tummeqqatigoornerit)
VDSDDWLSDNXQQHUDQXWXOOXWBBBBB
RTLPDDWVXQLNVXOLDTDQQJLODQJD →DSHTTQXXJLW
59.
Ulluni taakkunani piffissaq qanoq sivisutigisoq agguaqatigiissillugu
oqimaatsunik suliaqarlutit atorsimaviuk?
XOORUPXWQDOXQDDTXWWDSDNXQQHULBBBBB
XOORUPXWPLQXWWLWBBBBB
DDS 
QDDPLN →DSHTTQXXJLW
60.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit
oqimaakannersunik suliffinni suliaqartarsimavit? (Oqimaakannersunik sulineq
ilungersunannginnerusuuvoq, anerterinerunermillu annikinnerusumik
kinguneqartarluni; soorlu kivitsinerit annikinnerusut)
RTLPDDNDQQHUVXQLNVXOLDTDUWDQQJLODQJD →DSHTTQXXJLW
61.
Ulluni taakkunani suliffinni agguaqatigiissillugu qanoq sivisutigisumik
oqimaakannersunik suliaqartarpit?
62.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit sulinerni
minnerpaamik minuttini 10-ni pisuttarpit? Suliartornermi soraarnermilu
pisuttarnerit ilanngunneqassanngillat
63.
Ulluni taakkunani sulinernut atatillugu agguaqatigiissillugu qanoq
sivisutigisumik pisuttarpit?
VXOLQLQQLSLVXWWDULDTDUWDQQJLODQJD →DSHTTQXXJLW
Apeqqutit tulliuttut piffimmit piffimmut angallanermut tunngapput
64.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit biilit busit
snescooterilluunniit atorlugit ingerlasimavit?
65.
ELLOLWEXVLWVQHVFRRWHULOOXXQQLWDWRUOXJLWLQJHUODQQJLODQJD →DSHTTQXXJLW
Ulluni taakkunani qanoq agguaqatigiissillugu sivisutigisumik biilit, busit
snescooterilluunniit atorlugit angallanneqartarpit?
66.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit piffimmit
piffimmut ingerlanernut atatillugu minnerpaamik minuttsini 10-ni
sivisussusilimmi cykelertarpit?
67.
Ulluni taakkunani piffimmit piffimmut cykilernerni agguaqatigiissillugu
piffissaq qanoq sivisutigisoq atortarpiuk?
68.
F\NHOLQQJLODQJD →DSHTTQXXJLW
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit piffimmit
piffimmut minnerpaamik minuttini 10-ni pisuttarsimavit?
SLIILPPLWSLIILPPXWSLVXWWDULDTDUWDUVLPDQQJLODQJD →DSHTTQXXJLW
69.
Ulluni taakkunani piffimmit piffimmut pisunnerni agguaqatigiissillugu
piffissaq qanoq sivisutigisoq atortarpiuk?
Apeqqutit tulliuttut angerlarsimaffimmi eqqaanilu soorlu
angerlarsimaffimmut tunngasunik suliaqarnernut, iluarsaassinernut,
aserfallatsaaliuinernut meeqqanillu ilaquttanillu paarsinernut
tunngassuteqarput. Suliat 10 minuttinit sivisunerusumik suliarineqartut
kisiisa eqqarsaatigikkit
70.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit angerlarsimaffinni
oqimaatsunik suliaqartarpit? (soorlu oqimaatsunik kivitsinerit, apummik
nivannerit, assaanerit, imertartornerit)
71.
Ulluni taakkunani angerlarsimaffimmi agguaqatigiissillugu piffissaq qanoq
sivisutigisoq oqimaatsunik suliaqarlutit atortarpiuk?
72.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit angerlarsimaffinni
oqimaakannersunik suliaqartarsimavit? (soorlu angerlarsimaffimmi
iluarsaassinernut, atortunik iluarsaassinerni, eqqiaanerni atisanik errorsinerni,
meeqqanik utoqqarnillu paaqqutarinninermi)
RTLPDDNDQQHUVXQLNVXOLDTDUWDQQJLODQJD →DSHTTQXXJLW
73.
DQJHUODUVLPDIILPPLRTLPDDWVXQLNVXOLDTDQQJLODQJD →DSHTTQXXJLW
Ulluni taakkunani angerlarsimaffinni oqimaakernersunik suliaqarlutit
agguaqatigiissillugu piffissaq qanoq sivisutigisoq atortarpiuk?
Apeqqutit tulliuttut sunngiffimmi timigissarnermut, timersornermut
timiluunniit atorlugu sammisaqarnermut tunngassuteqarput.
Immikkoortuni siuliini suliat ilannguteriikkatit matumani
ilanngutissanngilatit
74.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit sunngiffinni
minnerpaamik minuttini 10-ni pisuttarsimavit?
75.
Ulluni taakkunani sunngiffinni agguaqatigiissillugu piffissaq qanoq
sivisutigisoq pisuttarsimavit?
76.
Ullut kinguliit arfineq marluk ingerlaneranni ullut qassit sunngiffinni
oqimaatsunik sammisaqartarsimavit? (soorlu aerobics,arpanneq, imminut
illersorluni timersuutinik sammisaqarneq, arsarneq, sisorarneq)
VXQQJLIILQQLSLVXWWDQQJLODQJD →DSHTTQXXJLW
VXQQJLIILQQLRTLPDDWVXQLNVXOLDTDUWDQQJLODQJD →DSHTTQXXJLW
77.
Ulluni taakkunani sunngiffinni oqimaatsunik sammisaqarlutit
78.
Ulluni qanoq amerlatigisuni sunngiffinni oqimaakannersunik
sammisaqartarsimvit? (soorlu sukkavallaanngitsumik cykilerneq, kingaatsumik
nalunneq, pisuttuarneq)
79.
VXQQJLIILQQLRTLPDDNDQQHUVXQLNVXOLDTDUWDQQJLODQJD →DSHTTQXXJLW
Ulluni taakkunani sunngiffinni oqimaakannersunik sammisaqarlutit
Apeqqutit tulliuttut sulinerni sunngiffinnilu issiasarnernut piffissamut
atortakkannut tunngassuteqarput (Soorlu allaffimmi issianerit, ikinngutinut
pulaarnerit, atuarneq, qarasaasiarsoqneq TV-lu).
Biilernerit assigisaallu ilanngunneqassanngillat
80.
Ullut kinguliit arfineq marluk ingerlaneranni piffissaq qanoq sivisutigsoq
ingillutit ulluinnarni suliaqartarsimavit?
81.
Ulluni arfineq marlunni kingullerni sapaatit akunnerisa naanerini ullormut
piffissami qanoq sivisutigisumi ingillutit suliaqartarsimavit?
Appendix III
Appendix III – Overview of Accelerometry and heart rate monitoring
For a description of procedures in the data collection is referred to paper I.
A limited number of monitors was available for the study and therefore randomly assigned to a subgroup of
participants at each study location. No monitors were given to participants in villages in Avanersuaq
because of the limited time available at the study location (1-2 days). Moreover, we did not include
recordings from the pilot study because the participants were not randomly assigned and were not
included in the master sample.
Data from the monitors were manually trimmed to indicate the end of each participant’s recording. The
value of the first night of sleeping heart rate (SHR) was excluded. Based on the lowest HR, a 10% deviation
was accepted as variation in SHR for the following nights. A mean SHR was calculated for every participant.
The quality of the recordings based on the interpretation by the researcher was divided into 5 levels and
reported on a log sheet. The ID number, sex, age, height and weight were validated with the master
database and revised manually in the database. Moreover, it was noted if the participant suffered from
heart disease or received medication that could influence the heart rate (beta-blockers). It was noted if the
monitor was not given to a participant. An example could be that the participant was leaving the town for
several days. On the log sheet it was noted if the calibration factor of the monitor was outside of the
prescribed ranges. Data were cleaned according to written guidelines and with support from the MRC
Epidemiology Unit in Cambridge, UK. A revised log sheet was provided by MRC including further changes to
data, i.e. change of accelerometer calibration factors, which can get corrupted, but can be recovered by
cross-referencing to other records obtained with the same monitor.
Step test data were available from 166 participants from two towns (Aasiiaat and Qasianguit). To keep the
duration of the health examination reasonable for the participant (maximum 2.5 hours) it was decided to
skip the test in the following health examinations. The step test data were used to calculate a group
calibration model specific for this population. Individual calibration means available information on heart
rate response to a known workload at the individual level. HR can be influenced by several factors, such as
age, sex, training state, stroke volume and mental stress among others (137;138). It has been shown that
some of these limitations can be overcome by individual calibration. However, in a paper by Brage et al it
was suggested, that a group calibration model was reasonable to use at population level, although it would
have a larger random error (90). The limitations of Accelerometry are mainly biomechanical, i.e. the
accelerometry-PAI relation is different for different activities.
Heart rate data were pre-processed using robust Gaussian Process Regression for inference of latent heart
rate trace, as described elsewhere (139). The combination of prolonged time periods of large heart rate
uncertainty accompanied by no acceleration was used to classify all measured time points as wear or nonwear.
Caloric intensity of PA was estimated by combining the acceleration-based estimate of intensity (90) with
the heart rate-based estimate from the population-specific equation (see above) in a branched equation
7ϴ
Appendix III
modelling framework (90;91). If available, step-calibrated HR was used instead of the group-calibrated HR
estimate. Briefly, the branched equation modelling method predominantly uses the accelerometer
estimate during low levels of heart rate and movement estimate and the heart rate estimate when both
heart rate and acceleration levels are high, with equal weighting for other conditions. Resulting time series
of activity intensity (in J/min/kg) were summarised into total PAEE (in kJ/kg/day) and time spent on
different intensity levels (sedentary as <1.5MET, moderate as 3-6MET, and vigorous as >6MET), whilst
minimizing diurnal bias from potentially unbalanced data accumulated over the day. This weighting
technique ensures equal representation of all the hours of the day and minimizes the impact of records
containing for example 3 nights and 2 days’ worth of data. Intensity categories were defined using
multiples of RMR as derived using the Oxford equations using age, sex, height, and weight (93). Branched
equation modelling of simultaneous accelerometry and heart rate monitoring has been shown to improve
estimates of directly measured PAEE. Brage et al, 2004 suggest that individual calibration may be less
necessary when branched modelling is employed (91).
After data cleaning, 2,053 recordings from Inuit were available for analysis, corresponding to 63.5% of the
total study population. Data were merged with the master database. Four recordings figured only in the AH
database and were deleted from further analysis due to the following: One recording was stated with an ID
number not identified in the master database and it was not possible to identify the correct ID number
from the information on weight, height and CPR number, one recording was from a test person and should
not be included in the analysis, two participants were examined twice in two different places and the first
record of each participant was deleted.
One recording was recommended by the MRC Epidemiology Unit in Cambridge to be deleted due to a very
poor signal. Furthermore, 57 recordings were considered missing because no, or almost no, data were
available from the monitor, and therefore data processing was not possible. This might be explained by the
participant removing the monitor shortly after it was handed over, poor acceleration signal or noisy HR
data. It could be that the monitors were susceptible to interferences from electrical appliances or other
sources of static current or that the electrodes were in poor contact with the skin. One recording had
missing information on weight.
Recordings were flagged by the MRC Epidemiology Unit in Cambridge for the reason of a poor HR or
acceleration signal, SHR estimation problems (no valid data during nights), heart disease or calibration
errors. Comments were made to use another estimation model for some of the flagged records. The
analyses include flagged recordings unless otherwise stated (see flowchart for cleaning process, table l). If
for some reason the combined Acc and HR estimate was deemed invalid, for example due to excessive
amounts of noise on the HR channel, the single-measure estimates of PAEE were used to impute such
missing values. These were scaled to minimize bias, and scaling factors were derived based on the sample
with valid data for both channels. Similarly, a flex HR estimate (84;90) was used if only heart rate data was
valid (corrupt and unrecoverable acceleration signal).
7ϵ
Appendix III
A wear time of less than twenty-four hours was presented for 4.8% of the valid AH data. A wear time of
more than 48 hours was presented for about 77% of the valid data. In villages 11% did not wear the
monitor for more than 24 hours compared with 5% in Nuuk and 2% in towns.
Table l. Proportional differences and odds ratios for participants with and without Acc and HR monitoring > 48 hours.
Characteristics
Sample without ACC and
Sample with ACC and HR
OR(CI95%)
HR
recordings>48 hours
(n=1049)
(n=1546)
Sex
Men
44.4
43.5
1
Women
55.6
56.5
1.04(0.89;1.2)
Nuuk
5.7
21.4
1
Town
76.1
59.8
0.2(0.20.3)
Village
18.2
18.8
0.3(0.2;0.4)
18-24
9.9
10.5
1
25-29
7.8
7.8
0.9(0.6;1.4)
30-34
7.7
7.1
0.9(0.6;1.3)
35-39
9.3
10.7
1.1(0.8:1.5)
40-44
11.6
18.9
1.5(1.1;2.1)
45-49
13.2
12.8
0.9(0.7;1.3)
50-54
10.5
10.4
0.9(0.7;1.3)
55-59
7.7
7.6
0.9(0.6;1.3)
60-64
6.8
5.7
0.8(0.5;1.2)
65-69
5.3
4.3
0.8(0.5;1.2)
10.1
4.3
0.4(0.3;0.6)
Place of residence
Age (years)
>70
ϴϬ
Appendix III
Table II. Overview of start day of the week presented for total ACC and HR sample and ACC and HR>48 hours.
Frequency
Percent
Frequency
Total sample
Percent
>48 hrs of PA data
Start day of the week
Monday
309
15.5
246
15.9
Tuesday
346
17.3
277
17.9
Wednesday
354
17.7
260
16.8
Thursday
292
14.6
218
14.1
Friday
287
14.4
232
15.0
Saturday
243
12.2
193
12.5
Sunday
164
8.2
119
7.7
1995
1945
Table III. Overview of wear time combined week day and weekend day presented for total ACC and HR sample and
ACC and HR>48 hours.
Wear time
N
%
Wear time weekend and week – total ACC and HR sample
1172
58.7
Wear time weekend and week day ≥48 hours
1032
66.8
Table IV. Overview of wear time presented for total ACC and HR sample and ACC and HR>48 hours.
Mean
Median
Range
IQR
Wear time overall, hours (n=1995)
66
70.6
(2-136)
IQR (49-75.9)
Wear time overall, hours (n=1945)
75
72.7
(48-136)
IQR (48-90.0)
ϴϭ
Appendix IV
Appendix IV – Overview of IPAQ data processing
Domain
Variabel
Limit
Iht57
Work
Transport
Iht58
Iht59
0-7
Iht60
Iht61
0-7
Iht62
Iht63
0-7
IHT64
IHT65
0-7
Iht66
Iht67
0-7
Iht68
IHT69
0-7
Iht70
0-7
Iht71
0-7
Iht72
Iht73
0-7
Iht74
Iht75
0-7
Iht76
Iht77
0-7
0-59
Iht78
Iht79
0-7
Iht80
Sitting-time
Iht81
Mets
(minus
BMR)
Yes/no
Domestic
Leisure-time
”Value”
Days
Minutes and
hours daily
Days
Minutes and
hours daily
Days
Minutes and
hours daily
Days
Minutes and
hours daily
Days
Minutes and
hours daily
Days
Minutes and
hours daily
Days
Minutes and
hours daily daily
Days
Minutes and
hours a day
Days
Minutes and
hours a day
Days
Minutes and
Hours a day
Days
Minutes and
hours a day
Minutes and
hours a day
Minutes and
hours a day
Activity
Comments
Occupational activity
outside home
.
8.0 (7.0)
Vigorous
4.0 (3.0)
Moderate
3.3 (2.3)
Walking
0.0
Sitting/standing
6.0 (5.0)
Biking
3.3 (2.3)
Walking
Hunting and fishing activities not
mentioned in activity examples.
Adapted by adding snowmobile.
Adapted by adding fetching water
and snow shoveling
5.5 (4.5)
Vigorous
4.0 (3.0)
Moderate
3.3 (2.3)
Walking
8.0 (7.0)
Vigorous
4.0 (3.0)
Moderate
0.0
Sitting week and
weekend day
Moderate intensity outside and
inside is combined into one;
Moderate garden activity
excluded. Gardening is nonexistent and common activities
such as getting fishing equipment
ready are done both inside and
outside the house. Activity
examples differ from the Danish
version by including care taking
and reparation of equipment.
More activity examples added in
the Greenlandic version, such as
skiing
Activity such as hiking is added
We have used the scoring protocol by IPAQ: http://www.ipaq.ki.se/scoring.htm. Specific rules of truncation
and scaling have been added to deal with outliers.
8Ϯ
8ϯ
Appendix IV
Publications
8ϰ
Publications
Paper I
PAPER I
Validation of the long International Physical Activity Questionnaire in the
Arctic - measures of physical activity in Greenland
Dahl-Petersen IK
Hansen AW
Bjerregaard P
Jørgensen ME
Brage S
Medicine and Science in Sports and Exercise, 2013: 45(4): 728-736A
Validity of the International Physical Activity
Questionnaire in the Arctic
INGER KATRINE DAHL-PETERSEN1, ANDREAS WOLFF HANSEN1, PETER BJERREGAARD1,
MARIT EIKA JKRGENSEN2, and SKREN BRAGE3
1
National Institute of Public Health, University of Southern Denmark, Copenhagen, DENMARK; 2Steno Diabetes Center,
Gentofte, DENMARK; and 3MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge, UNITED KINGDOM
ABSTRACT
EPIDEMIOLOGY
DAHL-PETERSEN, I. K., A. W. HANSEN, P. BJERREGAARD, M. E. JKRGENSEN, and S. BRAGE. Validity of the International
Physical Activity Questionnaire in the Arctic. Med. Sci. Sports Exerc., Vol. 45, No. 4, pp. 728–736, 2013. Purpose: Information about
physical activity (PA) in Greenland is limited, partly because of a lack of validated instruments in countries with non-Western living
conditions. We modified the long form of the International Physical Activity Questionnaire (IPAQ-L) to arctic living conditions. The aim
of the study was to compare IPAQ-L estimates with combined accelerometry and heart rate monitoring (ACC + HR) in a populationbased study of adult Inuit in Greenland. Methods: Cross-sectional data were collected by face-to-face interview and ACC + HR monitoring
among Inuit (18 yr and above) in Greenland during 2005–2010 (n = 1508). PA energy expenditure (PAEE) and time spent sedentary and on
PA at moderate and vigorous intensity were derived from IPAQ-L and ACC + HR. Estimates were compared using Bland–Altman
agreement analysis and Spearman correlations stratified by sex, place of residence (capital, towns, and villages), and age groups. Results:
Questionnaire-based PAEE was moderately correlated with objectively measured PAEE (r = 0.20–0.35, P G 0.01). Self-reported time spent at
moderate- and vigorous-intensity PA and time spent sedentary were weakly correlated with the objective measure (r = 0.11–0.31).
Agreement analyses showed relatively small median differences for all measures of PA; however, time spent at moderate-intensity PA was
substantially overreported by IPAQ-L when including walking (91.5 hIdj1, P G 0.001) but not when excluding walking. Conclusions: The
IPAQ-L adapted to arctic living conditions in Greenland had a moderate level of agreement with combined accelerometry and heart rate
monitoring for total PAEE at population level, but it was less valid to measure different intensities of PA and sedentary activity. Validity did
not differ markedly between rural and urban communities. Key Words: PHYSICAL ACTIVITY ASSESSMENT, SELF-REPORT,
ACCELEROMETRY, HEART RATE, INUIT, INDIGENOUS POPULATIONS
I
nuit in Greenland have experienced a substantial increase in chronic lifestyle diseases such as type 2 diabetes along with the rapid cultural and social transition
over the last 50 yr (5,22). Differences are found in physical
activity (PA) patterns, suggesting a less physically active
lifestyle in relation to the social change (13). Knowledge
about the level of PA still remains limited mainly because of
lack of validated instruments to assess everyday life PA in
countries with non-Western living conditions as in Greenland.
A need for further research in different cultural settings has
been suggested (11). Measurement of PA by questionnaire
is still the most commonly used method at population level
because it is inexpensive, is feasible to use in large populations, and can provide information on PA patterns. The
International Physical Activity Questionnaire (IPAQ) is a
questionnaire developed for measuring PA in different cultural settings and is the most frequently used (34). It exists in
a short (IPAQ-S) and a long (IPAQ-L) form. The short form
is recommended for national monitoring (seven items),
whereas the long version is more comprehensive (27 items)
and assesses time spent at different intensities of PA within
four domains of daily life: transportation, work, leisure time,
and domestic activities (16). Both forms have been tested
for reliability and validity in adult populations in various
countries against accelerometers (7,11,17–19,24) and have
shown fair to moderate validity, although lower in rural
areas (23). To our knowledge, only the short form has been
used in the Arctic and found to be significantly correlated to
body fat and waist circumference among Liyiyiuch, a Cree
community in Canada (14). We planned to modify IPAQ-L
to arctic living conditions in Greenland and to compare
its main PA variables with objective estimates from combined accelerometry and heart rate monitoring (ACC + HR).
Estimates of PA energy expenditure (PAEE) from ACC +
HR compare favorably to doubly labeled water (DLW)
estimates of PAEE in free-living adults in urban and rural
population samples (2) and also provide valid estimates of
PA intensity (12,32,33). The aim of the study was to assess
the validity of IPAQ-L against combined accelerometry and
Address for correspondence: Inger Katrine Dahl-Petersen, MSPH National
Institute of Public Health, University of Southern Denmark, Kster Farimagsgade
5A,2, DK 1353 Copenhagen K, Denmark; E-mail: [email protected].
Submitted for publication August 2012.
Accepted for publication October 2012.
0195-9131/13/4504-0728/0
MEDICINE & SCIENCE IN SPORTS & EXERCISEÒ
Copyright Ó 2013 by the American College of Sports Medicine
DOI: 10.1249/MSS.0b013e31827a6b40
728
Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
heart rate monitoring in both rural and urban communities
in a country undergoing rapid social transition.
METHODS
MEASURE OF PHYSICAL ACTIVITY IN GREENLAND
Medicine & Science in Sports & Exercised
729
EPIDEMIOLOGY
Study population. Data for this population-based, crosssectional study were collected in Greenland during 2005–
2010. The total population of Greenland is 57,000, of which
90% are Inuit. Twenty-two communities, the capital (Nuuk,
pop = 16,181), eight smaller towns (pop = 469–5571), and
13 villages (pop = 7–513) (25% of all communities) were
selected as study areas being representative of each region in
Greenland. Nuuk represents the most Westernized living conditions compared with smaller towns and villages. From capital and towns, random population samples were drawn from
the central population register. From villages, all adults were
invited to participate. Pregnant women, individuals not born in
Greenland or Denmark, and individuals who had moved out
of the study area at the time of the study were excluded from
the population sample. We confined the study to Inuit as defined by the participant and the interviewer based on language
and self-perceived ethnicity at the time of enrolment in the
health examination. In total, 2874 adult Inuit age 18 yr and
older participated in a clinical examination and were
interviewed. A detailed description of the methods is available
elsewhere (4). The study was approved by the ethical review
committee for Greenland. Written informed consent was
obtained from all participants.
Procedures of health examination. At the day of the
health examination, face-to-face interviews were conducted
by trained native Greenlandic-speaking interviewers in the
language chosen by the participant (Greenlandic or Danish).
Information on sociodemographic factors and lifestyle, including PA in the form of IPAQ-L, was obtained during the
interview. Height (nearest 0.1 cm) and weight (nearest 0.1 kg)
were measured.
Self-reported PA. Information on PA was collected using a modification of the interviewer-administered IPAQ-L.
Participants indicated time spent on PA in the previous 7 d:
how often (the number of days per week) and for how long
(the average duration per day) separately for vigorous intensity, moderate intensity, and walking in the four domains
(work, transportation, domestic, and leisure time). The original English version of the PA questionnaire was translated into
Greenlandic and back-translated by two translators bilingual
in Danish and Greenlandic and familiar with Greenlandic
living conditions. The questions were adjusted to arctic living
conditions by replacing some of the activity examples by
culturally relevant examples. In the domestic domain, we
combined the two questions concerning moderate intensity
(outside and inside activity) into one; gardening is nonexistent
in arctic living conditions, and common activities such as
getting fishing equipment ready are done both inside and
outside the house. We also did a brief interview with five of
the main interviewers about their experience with interpretation of the questions.
Combined accelerometry and heart rate monitoring. A combined accelerometer and heart rate monitor
(ACC + HR) (ActiheartÒ; CamNtech Ltd, Cambridge, UK),
described in technical detail elsewhere (8), was provided to a
subgroup of the participants (n = 2055). The monitor was
set up to measure acceleration and heart rate in 30-s intervals
and attached to the participant_s chest by two standard
ECG electrodes (MXC55; MediMax, Edison, NJ). The participant was instructed to wear the monitor for 24 hIdj1 for at least
2 d and preferably 4 d from the day of the health examination.
Because of study logistics, only a limited time was available at each study location, especially for data collection in
villages. Together with a finite stock of monitors, this explains why not all participants were given a monitor and why
the length of recordings from some participants was shorter.
A subgroup of participants (n = 135) conducted an individual calibration test (8-min step test) as described previously
(10). The step tests were used to define a population-specific
calibration equation of the heart rate–activity energy expenditure relationship.
IPAQ-L data processing. Data on PA from the modified IPAQ-L were initially scored according to guidelines
from the IPAQ group (16). All participants who reported
days (frequency) but not time (duration) of PA or vice versa
were treated as missing. Total daily PA of more than 960 min
(16 h) was scaled linearly (sedentary time not included). In
each domain, minutes spent on PA at moderate and vigorous
intensity and walking for more than 180 minIdj1 (at each activity) were truncated to 180 min. Reported time spent on PA
between 0 and 10 min was accepted even though the questionnaire was restricted to activity of at least 10-min duration.
In a review of Murphy et al. (27), most of the studies did not
find alterations between accumulated and continuous patterns
of exercise, and it was concluded that further research was
required to determine whether even shorter bouts (G10 min)
could provide a health benefit. Time spent at moderate- and
vigorous-intensity PA and walking and total time spent in
each domain were calculated. PAEE on each activity was
calculated by multiplying time reported (minIwkj1) by the
net metabolic cost of each activity, which was expressed in
METs. Net metabolic cost of each activity was assigned
according to the physical activity compendium_s gross MET
values (1), subtracted by 1 MET to account for resting metabolic rate (RMR). An estimate of total daily sedentary time
was calculated from time spent sitting (including activities,
such as TV and computer use and reading), to which we
added 8 h as presumed time spent sleeping (not included in
IPAQ-L). Time spent at moderate-intensity activity was analyzed with and without the inclusion of walking.
Accelerometry and heart rate data processing. Data
from ACC + HR monitoring were manually trimmed to indicate the end of each participant_s recording, after which heart
rate data were preprocessed using robust Gaussian process regression for inference of latent heart rate trace as described
elsewhere (31). The combination of prolonged periods of large
heart rate uncertainty (disturbances in heart rate signal mostly
TABLE 1. Characteristics of the study population, Inuit in Greenland, n = 1508.
Age (yr)
Body mass index (kgImj2)
Place of residence (%)
Nuuk,a n = 323
Smaller towns, n = 906
Villages, n = 279
Age groups (%)
18–44 yr, n = 829
45–54 yr, n = 349
55 yr and older, n = 330
Job status (%)
White collar, n = 158
Skilled, n = 279
Blue collar, n = 444
Hunters and fishermen, n = 76
Students, n = 75
Unemployed and others, n = 285
Men (n = 659)
Women (n = 849)
P Value
44.8 T 14.2
25.8 T 4.6
43.1 T 13.8
26.5 T 5.5
0.02
0.02
19.9
63.4
16.7
22.6
57.5
19.9
0.06
50.7
24.4
24.9
58.3
22.1
19.6
G0.01
12.0
20.6
31.6
11.3
4.8
19.9
12.0
21.7
35.3
1.6
6.4
23.0
G0.01
Values are mean SD unless otherwise noted.
P values for sex comparison by chi-square test for categorical data and t-test for continuous data.
The capital of Greenland.
EPIDEMIOLOGY
a
related to the wear of the monitor and eventually interference by cloth or electrode detachment) accompanied by no
acceleration was used to classify all measured time points
as wear or nonwear. We included individuals with 948 h of
monitor wear data. Caloric intensity of PA was estimated
by combining the acceleration-based estimate of intensity
(10) with the heart rate-based estimate from the populationspecific equation (see previous discussions) in a branched
equation modeling framework (9). Briefly, this method predominantly uses the accelerometer estimate during low levels
of heart rate and movement and the heart rate estimate when
both heart rate and acceleration levels are high, with equal
weighting for other conditions. Resulting time series of activity intensity (JIminj1Ikgj1) were summarized into total
PAEE (kJIkgj1Idj1) and time spent at different intensity levels
(sedentary as G1.5 MET, moderate as 3–6 MET, and vigorous
as 96 MET) while minimizing diurnal bias from potentially
unbalanced data accumulated over the day. Intensity categories were defined using multiples of 1 MET as derived
using the Oxford equations for resting metabolic rate (21).
Statistical methods. Descriptive characteristics of the
study sample are presented as means with SDs for normally
distributed continuous variables and medians with interquartile ranges for nonnormal distributed variables. Results
are stratified by sex, age groups, and place of residence.
Differences between men and women were tested by t-test
for continuous data and chi-square test for categorical data.
The linear association between self-reported and objective
activity estimates was examined by the nonparametric
Spearman rank correlation coefficient (Q). Level of agreement was examined by modified Bland–Altman plots (6).
Median difference between the measurements (IPAQ-L minus ACC + HR) was plotted against the objective estimate,
with lines indicating the median difference (median bias)
and 95% limits of agreement (2.5 and 97.5 centiles) (nonparametric data). The differences of the medians were analyzed by Wilcoxon signed-rank test. Sensitivity analyses
730
Official Journal of the American College of Sports Medicine
were performed including only participants with ACC + HR
monitoring of Q72 h. Gaussian process regression of heart
rate was performed in JAVA using a MySQL database, and
all other analyses were carried out in STATA version 11.
RESULTS
The study population consisted of 2874 Inuit adults. After
data processing, valid data from the IPAQ-L were obtained
from 2798 participants (97.4%), of whom 1999 (71.4%) had
worn an ACC + HR monitor. After excluding recordings with
insufficient valid PA data (G48 h), 1508 participants with
complete data from both IPAQ-L and ACC + HR were
available for analysis. The proportion of men and women that
was monitored by ACC + HR for more than 48 h did not differ
from those not monitored (men, 44% vs 45%; women, 56% vs
55%; P = 0.6); however, a smaller proportion of participants
from villages and towns was monitored by ACC + HR
(Nuuk, 21% vs 8%; towns, 61% vs 70%; villages, 18% vs
23%; P G 0.001). Moreover, a smaller proportion of participants age 70 yr or older was monitored by ACC + HR,
compared with those not monitored (4% vs 11%, P G 0.001).
The characteristics of the study population are displayed in
Table 1. Men (range, 18–84 yr old) were slightly older
than women (range, 18–85 yr old), and women had higher
body mass index (26.5 kgImj2) than men (25.8 kgImj2)
(P = 0.02). One fifth of the participants lived in the capital,
Nuuk, and more than half in other towns.
The median PAEE estimated by the IPAQ-L was almost
similar compared with the median estimate from ACC + HR
for all subgroups, although significant differences were found
for women, for the participants in villages, and in the age group
of 45–54 yr (Table 2). Self-reported PAEE (kJIdj1Ikgj1) was
moderately, although significantly, correlated with objectively measured PAEE in analyses stratified by sex, age
groups, and place of residence (r = 0.20–0.35, P G 0.001)
http://www.acsm-msse.org
24.2–80.9
36.6–64.1
23.6–85.0
36.2–67.8
30.0–86.3
36.8–66.2
33.2–92.3
44.6–75.2
23.3–85.7
34.3–60.1
45.9
50.9
48.4
49.6
50.0
49.4
57.7
57.8
47.5
47.2
15.3–57.7
22.5–47.5
24.9–76.9
34.2–60.1
47.3
45.7
31.6
34.3
23.6–97.0
40.3–75.5
51.7
56.6
Total PAEE
(kJIdj1Ikgj1)
0.6
0.02
0.04
0.02
0.06
0.3
0.002
0.2
P Value
0.9
1.0
1.1
1.5
1.7
1.9
2.0
1.6
1.1
1.6
1.2
1.6
1.5
1.4
1.1
1.8
0.1–2.3
0.4–1.6
0.4–2.7
0.9–2.2
0.6–3.1
1.2–2.9
0.6–3.1
0.9–2.5
0.3–3.0
0.9–2.5
0.5–2.6
1.0–2.4
0.5–3.0
0.9–2.2
0.3–3.0
1.1–2.7
Moderate Intensity
(3–6 METs)
(hIdj1)
0.1
0.2
0.01
0.002
0.01
0.009
0.002
G0.001
P Value
2.8
1.0
3.8
1.5
4.2
1.9
3.7
1.6
3.9
1.6
3.8
1.6
3.9
1.4
3.7
1.8
1.4–4.8
0.4–1.6
1.7–6.1
0.9–2.2
2.4–6.6
1.2–2.9
2.0–6.1
0.9–2.5
1.9–6.1
0.9–2.5
1.9–6.1
1.0–2.4
2.0–6.2
0.9–2.2
1.8–6.0
1.1–2.7
Moderate Intensity
(walking included,
3–6 METs) (hIdj1)
0.0
4.7
0.0
3.6
4.3
4.4
2.9
7.5
0.0
3.0
G0.001
G0.001
G0.001
G0.001
G0.001
0.0
0.3
0.0
2.9
G0.001
G0.001
7.1
6.2
G0.001
P Value
0.0–0.0
0.0–2.9
0.0–34.3
0.2–9.4
0.0–51.4
2.1–19.3
0.0–60.0
0.3–12.7
0.0–34.3
0.4–12.9
0.0–30.0
0.8–14.6
0.0–12.9
0.2–9.9
0.0–68.6
0.9–18.9
Vigorous Intensity
(96 METs)
(minIdj1)
0.007
0.002
G0.001
G0.001
0.2
0.3
0.05
G0.001
P Value
11.0
17.4
11.9
15.7
11.6
14.6
10.5
15.3
11.4
15.5
12.9
15.6
11.3
15.8
11.9
14.9
9.8–12.9
15.7–19.4
10.3–13.9
13.9–17.6
10–13.9
12.6–16.4
9.7–11.9
13.3–17.3
10.0–13.5
13.3–17.4
11.2–15.0
13.6–17.4
10–13.2
13.8–17.5
10.0–14.0
13.0–17.2
Sedentary Activity
(G1.5 METs) (hIdj1)
G0.001
G0.001
G0.001
G0.001
G0.001
G0.001
G0.001
G0.001
P Value
Self-reported (IPAQ-L) and objectively measured (ACC + HR) PA presented as daily PAEE and time spent on PA at moderate-intensity (with and without walking included), vigorous-intensity, and sedentary activity stratified by sex, place of
residence, and age groups. Inuit in Greenland, n = 1508.
All values are presented in median and interquartile range. P value: Wilcoxon rank test for difference between medians of the two methods.
a
Sex
Men, n = 659
Self-report
Objective measure
Women, n = 849
Self-report
Objective measure
Place of residence
Nuuk,a n = 323
Self-report
Objective measure
Towns, n = 906
Self-report
Objective measure
Villages, n = 279
Self-report
Objective measure
Age groups
18–44 yr, n = 829
Self-report
Objective measure
45–54 yr, n = 349
Self-report
Objective measure
Self-report
Objective measure
TABLE 2. PA characteristics.
EPIDEMIOLOGY
731
TABLE 3. Spearman correlation coefficients (Q) between self-reported PA (IPAQ-L) and objectively measured PA (ACC + HR).
Total PAEE
(kJIdj1Ikgj1)
Sex
Men, n = 659
Women, n = 849
Place of residence
Nuuk,a n = 323
Smaller towns = 906
Village, n = 279
Age groups
18–44 yr, n = 829
45–54 yr, n = 349
Moderate Intensity
(3–6 METs) (hIdj1)
Moderate Intensity
(walking included,
3–6 METs) (hIdj1)
Vigorous Intensity
(96 METs) (minIdj1)
Sedentary Activity
(G1.5 METs) (hIdj1)
Q
P Value
Q
P Value
Q
P Value
Q
P Value
Q
P Value
0.33
0.28
G0.001
G0.001
0.25
0.19
G0.001
G0.001
0.26
0.24
G0.001
G0.001
0.27
0.17
G0.001
G0.001
j0.01
0.06
0.8
0.1
0.35
0.30
0.29
G0.001
G0.001
G0.001
0.15
0.20
0.26
0.01
G0.001
G0.001
0.25
0.22
0.26
G0.001
G0.001
G0.001
0.27
0.23
0.26
G0.001
G0.001
G0.001
j0.1
0.001
0.13
0.3
0.97
0.03
0.23
0.20
0.33
G0.001
0.0002
G0.001
0.13
0.16
0.23
0.002
0.003
G0.001
0.15
0.17
0.31
G0.001
0.002
G0.001
0.19
0.11
0.21
G0.001
0.04
0.001
0.05
0.02
0.09
0.2
0.7
0.1
Results are stratified by sex, place of residence, and age groups. Inuit in Greenland, n = 1508.
a
tions across strata of sex, age groups, and residence, however,
were similar, with a tendency toward higher correlations when
walking was included (Table 3). Figure 2 illustrates MPA including walking. The median bias was large, and the 95%
limits of agreement indicated high individual variability for
both men and women in the different subgroups. The asymmetry of the 95% limits of agreement around the median
bias highlights the substantial overestimation of MPA by selfreport. The same pattern was found for all subgroups (data not
shown). In contrast, time spent at moderate-intensity PA
without walking included was only slightly overestimated by
EPIDEMIOLOGY
(Table 3). The level of agreement between the two methods
for measuring PAEE stratified by sex and residence is illustrated in Figure 1. The median bias for PAEE was small
and indicated a fair agreement between the two methods at
population level; however, large individual differences in
PAEE were found (Table 2). The same tendencies were found
for all subgroups (data not shown).
Self-reported time spent on PA at moderate intensity (MPA)
was significantly lower than objectively assessed MPA when
walking was not included in the estimation but was significantly higher when walking was included (Table 2); correla-
FIGURE 1—Self-reported and objectively measured PAEE by sex and residence. Median difference between self-reported PAEE and objectively
measured PAEE (IPAQ-L–ACC + HR) plotted against (ACC + HR) stratified on sex and place of residence. The lines represent median and 2.5 and
97.5 centiles. Inuit in Greenland, n = 1508.
732
FIGURE 2—Self-reported and objectively measured moderate-intensity PA by sex and residence. Median difference between self-reported and
objectively measured time spent at moderate intensity PA (IPAQ-L–ACC + HR) plotted against (ACC + HR) stratified on sex and place of residence
(walking included). The lines represent median and 2.5 and 97.5 centiles. Inuit in Greenland, n = 1508.
We repeated the analysis for the subgroup of participants
who wore the ACC + HR monitor for Q72 h and found
similar results (data not shown).
DISCUSSION
We found moderate validity for questionnaire-based
overall PAEE and weak to moderate validity for different
intensities of PA and sedentary time compared with ACC +
HR monitoring stratified by sex, age groups, and residence.
The Bland–Altman plots showed relatively small median
differences for all variables of PA; however, the individual
variability in PA measures was high.
Studies testing the validity of the IPAQ-L questionnaire
against different criterion measures have shown different
levels of correlation. Craig et al. (11) validated IPAQ-L in
12 different cultural settings with accelerometry and found
a rank correlation of around 0.30 for overall PA, although
this varied greatly between study sites, from j0.27 to
0.61. Similarly, a study examining the Chinese version of
IPAQ-L reported a correlation of 0.35 for overall PA
against accelerometry (25). Only one study has validated
the long form of IPAQ-L using DLW, which is considered
as the gold standard for total energy expenditure during
free living and found a correlation of 0.38 (26). Taken together, these previous validation results are more or less
in line with our correlation coefficients for overall PA
(0.20–0.35).
733
EPIDEMIOLOGY
IPAQ-L for women and for participants in villages and slightly
underestimated for the rest of the subgroups. For all subgroups, a significant difference between median time spent
on moderate-intensity PA measured by IPAQ-L and ACC +
HR was found except for participants age 45 yr and older
(Table 2). Light-intensity PA is not included in IPAQ-L,
so estimates areonly available from ACC + HR data. More
than 50% of the population spent 6 h daily or more on lightintensity PA (1.5–3 METs), which made up 54.4% (interquartile range, 44.9–62.8) of total PAEE (data not shown).
The median duration of vigorous-intensity PA measured
by IPAQ-L and ACC + HR monitoring differed significantly
for all subgroups except for women and for participants
living in Nuuk and in towns (Table 2). Men reported more
time spent at vigorous-intensity PA and women reported less
vigorous PA as compared with ACC + HR measurements.
Time spent on PA at vigorous intensity estimated from
IPAQ-L was significantly correlated with the ACC + HR
measurement in all subgroups, but the correlation was
generally weak (r = 0.11–0.27, P G 0.05) (Table 3). Both
IPAQ-L and ACC + HR showed that more than 50% of the
population spent less than 10 min daily at vigorous-intensity
PA. A poor and nonsignificant correlation was found for
sedentary activity measured by self-report and ACC + HR
for all subgroups, and the median differed significantly for
all subgroups (Tables 2 and 3). Sedentary behavior was
highly underestimated by the questions in IPAQ-L, even
after adding 8 h of presumed time spent sleeping.
EPIDEMIOLOGY
We found that self-reported PA at different intensities
was more weakly correlated with ACC + HR estimates
than overall PA. The findings from other studies are not
clear (7,18); however, two studies among non-European
populations (19,28) have shown comparably low correlations
as in our study for moderate and vigorous-intensity PA. It is
suggested that cultural differences may affect the interpretation of the intensity of the activity. Median self-reported time
spent at vigorous-intensity PA was found to be substantially
higher (more than fourfold) compared with objective estimates among participants living in villages. Accordingly, we
stratified the analysis by job status and found that hunters and
fishermen substantially overreported vigorous-intensity PA
(median: IPAQ-L, 34 minIdj1, vs ACC + HR, 8 minIdj1).
Traditional activities, such as hunting and fishing, may be
more difficult to recall and classify into moderate or vigorous
intensity because these activities do not have a regular time
schedule and vary in intensity. According to the interviewers
in our study, traditional activities such as hunting and fishing
could be misinterpreted as a vigorous-intensity activity because of its psychological exhausting and time-consuming
character and because of demanding climate conditions.
Going hunting is generally considered to be physically demanding; however, hunting often includes periods with waiting time.
Despite the substantial overreporting of time spent at
moderate-intensity PA (walking included) by IPAQ-L, the
Bland–Altman plots showed only small median bias for
overall PAEE measured by self-report and ACC + HR
monitoring, which suggests that IPAQ-L is a valid measure
for overall PAEE at the group level; however, it is less valid
when measuring different intensities of PA. Light-intensity
PA was only measured by ACC + HR monitoring but contributed substantially to daily PA. Our findings are contrary
to other studies that have demonstrated that PAEE on the
group level is overestimated by IPAQ-L, a bias for which
social desirability has been suggested as a plausible explanation (7,18). The attention from the media on the positive health effect of PA might have been less marked in
Greenland compared with more Westernized countries,
and thus, the risk of social desirability bias may be somewhat lower.
Moderate intensity with and without walking. We
found that IPAQ-L substantially overestimated moderateintensity PA when walking was included as a moderateintensity PA. Ekelund et al. (15) found in a study of IPAQ-S
that walking might be difficult to accurately quantify. In
IPAQ-L, walking activity is asked for in all domains of the
questionnaire, which might increase the risk of reporting the
same walking activity twice. According to guidelines from
the IPAQ group, walking is defined as moderate-intensity
PA and assigned the MET value of 3.3 METs; however, in
the compendium of Ainsworth et al. is listed various intensities of walking corresponding to different MET values
(2.0 to 8.0 METs) (1). One could argue that a slow pace
of walking corresponds to light intensity and not moderate-
734
intensity PA. Moreover, the IPAQ-L does not ask for lightintensity activities, which may result in participants classifying
light-intensity PA as moderate-intensity PA. A qualitative
interview with the interviewers in our study pointed out
the difficulties in estimating total time spent on moderateintensity PA and that standing activity, such as teaching or
working in a shop, was sometimes misinterpreted as walking
activity; this may add to explain the high amount of moderateintensity activity reported in our study. Accordingly, we did the
analyses of moderate-intensity activity with and without
walking and walking considered as a light-intensity activity
(data not shown) and found a substantially higher level of
agreement between the two methods when walking was not
included as MPA.
Sedentary time. Knowledge about the health risks of
sedentary behavior is increasing (20,35). The question about
sitting in IPAQ-L has demonstrated acceptable validity and
reliability and has been used to compare the prevalence of
sitting time in an international study in 20 countries (3,11).
We added 8 h of presumed sleeping time to sedentary time
estimated from IPAQ-L and compared it with time spent on
activity of less than 1.5 MET from the day-and-night ACC +
HR recordings. We found a substantial underestimation of
sedentary behavior by IPAQ-L, which could be explained by
the fact that frequent activities such as standing and lying
(and sleep) are estimated by ACC + HR monitoring as sedentary activity (G1.5 METs), whereas even though the question of sitting time in the IPAQ-L includes some aspects of
lying, it may not capture all sedentary activities in daily life.
Moreover, we estimated time spent sleeping to 8 h in the
IPAQ-L processing, and individual variations in sleeping
time could be another explanation.
Study population. The present study is a populationbased study including a representative sample of Inuit in
Greenland. For logistic reasons, not all participants were given
a monitor or fulfilled the criteria of ACC + HR monitoring for
48 h or more (65.4%). Nevertheless, we found only small
differences in age, sex, and residence between our smaller
study sample and the entire study population, which imply
that the results of this study are applicaple to the population of
Greenland. In the present study, we did not find significant
differences in validity between rural and urban communities,
defined by living in a village, a town, or in Nuuk. However,
traditional rural activities such as hunting and fishing might
influence the interpretation of the different intensities of PA.
Strengths and limitations. The gold standard for
measuring PAEE in free-living individuals is the DLW
method, combined with a measure of RMR. However, this
method is expensive and cannot provide information about
the intensity, frequency, and patterns of PA. Studies in nonWestern countries have shown that it is particularly important to monitor both HR and movement in the estimation of
PAEE in rural populations because of a higher number
of activities that cannot be fully measured by a classic uniaxial accelerometer, such as digging and heavy lifting,
which are activities comparable with traditional activities in
Rennie and Wareham (29) estimated that 3 d of recording
yielded a validity coefficient at 0.85 for the assessment of
energy expenditure. In our study, that was the case for 810
(56.3%) of the participants. Ideally, 7 d of objective recording would have been preferable to capture variations
in PA during the week, but this was not feasible. However,
our sensitivity analyses that included only participants with
more than 72 h of recording showed very similiar results
as the primary analyses, which suggests that the law of
diminishing returns may govern these behavioral data.
Although objectively measured PA is considered a more
valid measure of PA, the IPAQ-L has the advantage of
measuring domains of daily life PA, which are important,
particularly in non-European countries where being physically active at work and at home is more common compared
with leisure time PA. The domains in the IPAQ-L provide
the opportunity to track changes in PA patterns along with
social changes, e.g., whether decreasing occupational PA is
compensated by an increase in leisure time PA, not necessarily changing the total amount of PA; therefore, objective
and subjective measures complement each other.
CONCLUSIONS
The long version of IPAQ modified to arctic living conditions is a valid measure for overall PAEE among adult
Inuit in Greenland at population level; it is, however, less
valid to measure different intensities of PA and sedentary
activities. Time spent at moderate-intensity PA was substantially overreported by IPAQ-L when walking was included in this category. The validity did not differ significantly
between rural and urban communities. Using IPAQ-L at the
individual level will be subject to a high degree of uncertainty.
Studies on how culture, social norms, and language affect
the interpretation of PA questions are important to improve
the validity of the IPAQ-L in non-Western countries.
This study was funded by the Karen Elise Jensen Foundation,
Denmark. The authors are grateful to the participants and the participating communities. The authors would also like to thank Kate
Westgate and Stefanie Mayle at the MRC Epidemiology Unit,
Cambridge, United Kingdom, for assistance in data processing.
The authors declare that there are no conflicts of interest.
The results of the present study do not constitute endorsement by
the American College of Sports Medicine.
REFERENCES
1. Ainsworth BE, Haskell WL, Herrmann SD, et al. 2011 Compendium of Physical Activities: a second update of codes and MET
values. Med Sci Sports Exerc. 2011;43(8):1575–81.
2. Assah FK, Ekelund U, Brage S, Wright A, Mbanya JC, Wareham
NJ. Accuracy and validity of a combined heart rate and motion sensor for the measurement of free-living physical activity energy expenditure in adults in Cameroon. Int J Epidemiol. 2011;40(1):112–20.
3. Bauman A, Ainsworth BE, Sallis JF, et al. The descriptive epidemiology of sitting a 20-country comparison using the International
Physical Activity Questionnaire (IPAQ). Am J Prev Med. 2011;
41(2):228–35.
4. Bjerregaard P. Inuit health in transition. Greenland survey 2005–2009.
Population sample and survey methods. In: National Institute of
Public Health. 2009 (cited 19 October 2012):1–13 Available from:
http://www.si-folkesundhed.dk/upload/inuit_health_in_transition_
greenland_methods_5_002.pdf.
5. Bjerregaard P, Young TK, Dewailly E, Ebbesson SO. Indigenous
health in the Arctic: an overview of the circumpolar Inuit population. Scand J Public Health. 2004;32(5):390–5.
6. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;
327(8476):307–10.
735
EPIDEMIOLOGY
Greenland. We therefore consider the use of combined
accelerometry and heart rate monitoring for estimating
PAEE in this study as a strength (2,8,12), although the lack
of dynamic individual calibration in everybody is a potential
weakness (10). A status report on the assessment of PA by
self-report finds the use of an interviewer-administrated
questionnaire to increase the validity of the responses (30).
According to the interviewers in our study, some participants found the interpretation of moderate and vigorousintensity PA difficult. Furthermore, because of the wide
differences in living conditions, climatic differences, and
dialects across the country, the interviewers had to pay
particular attention to the choice of words and the examples
given of different activities. Therefore, the use of face-toface interviews undertaken by Greenlandic interviewers bilingual in Danish and Greenlandic is a strength in this study.
An important observation in the translation of the questionnaire into Greenlandic was that no word exists for PA. PA is
translated to ‘‘use of the body’’ and that may, in a higher
degree, refer to sports activities instead of activities of daily
living. However, we did not find this misclassification likely
because we would have assumed a substantial degree of
underreporting.
Our study has some potential limitations. First, we did
not conduct any repeated administrations of IPAQ-L because of logistical reasons in this comprehensive data collection process. Knowledge about reliability is an important
metric of any instrument, and it is recommended that future
research on measurements of PA in the Arctic include a
test–retest element.
Second, the administration of the two instruments meant
that they did not refer to the same period. The monitor was
given to the participants on the day they were interviewed
about their PA in the preceding 7 d. However, the short
interval between the periods is unlikely to have introduced
substantial bias in the results, and one may even consider the
present results to reflect more truly the convergent validity
of these instruments to assess habitual PA. Third, the estimates from the IPAQ-L were calculated as the average of the
previous 7 d (no information on sleeping hours but 8 h were
estimated for sleep), whereas estimates from the ACC + HR
monitor were calculated as an average of 2–5 d including
nights. ACC + HR monitoring from both week and weekend
day were obtained from 965 (67%) of the participants.
EPIDEMIOLOGY
7. Boon RM, Hamlin MJ, Steel GD, Ross JJ. Validation of the New
Zealand Physical Activity Questionnaire (NZPAQ-LF) and the
International Physical Activity Questionnaire (IPAQ-LF) with
Accelerometry. Br J Sports Med. 2010;44(10):741–6.
8. Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart rate and movement
sensor Actiheart. Eur J Clin Nutr. 2005;59:561–70.
9. Brage S, Brage N, Franks PW, et al. Branched equation modeling
of simultaneous accelerometry and heart rate monitoring improves
estimate of directly measured physical activity energy expenditure.
J Appl Physiol. 2004;96:343–51.
10. Brage S, Ekelund U, Brage N, et al. Hierarchy of individual calibration levels for heart rate and accelerometry to measure physical
activity. J Appl Physiol. 2007;103:682–92.
11. Craig CL, Marshall AL, Sjostrom M, et al. International Physical
Activity Questionnaire: 12-country reliability and validity. Med Sci
Sports Exerc. 2003;35(8):1381–95.
12. Crouter SE, Churilla JR, Bassett DRJ. Accuracy of the Actiheart
for the assessment of energy expenditure in adults. Eur J Clin Nutr.
2008;62(6):704–11.
13. Dahl-Petersen IK, Joergensen ME, Bjerregaard P. Physical activity
patterns in Greenland: a country in transition. Scand J Public
Health. 2011;39:678–86.
14. Egeland GM, Denomme D, Lejeune P, Pereg D. Concurrent validity of the International Physical Activity Questionnaire (IPAQ)
in an liyiyiu Aschii (Cree) community. Can J Public Health.
2008;99(4):307–10.
15. Ekelund U, Sepp H, Brage S, et al. Criterion-related validity of the
last 7-day, short form of the International Physical Activity Questionnaire in Swedish adults. Public Health Nutr. 2006;9(2):258–65.
16. Guidelines for data processing and analysis of IPAQ—short and
long forms. Web site (Internet); (cited 19 October 2012). Available
from: http://learnonline.canberra.edu.au/pluginfile.php/511212/mod_
resource/content/0/IPAQ_scoring_long.pdf.
17. Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide
variability in physical inactivity a 51-country survey. Am J Prev
Med. 2008;34(6):486–94.
18. Hagstromer M, Oja P, Sjostrom M. The International Physical
Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006;9(6):755–62.
19. Hallal PC, Simoes E, Reichert FF, et al. Validity and reliability of
the Telephone-Administered International Physical Activity
Questionnaire in Brazil. J Phys Act Health. 2010;7(3):402–9.
20. Healy GN, Wijndaele K, Dunstan DW, et al. Objectively measured
sedentary time, physical activity, and metabolic risk: the Australian
Diabetes, Obesity and Lifestyle Study (AusDiab). Diabetes Care.
2008;31:369–71.
21. Henry CJK. Basal metabolic rate studies in humans: measurement
and development of new equations. Public Health Nutr. 2005;
8(7A):1133–52.
736
22. JLrgensen ME, Bjerregaard P, Borch-Johnsen K. Diabetes and impaired glucose tolerance among the Inuit population of Greenland.
Diabetes Care. 2002;25(10):1766–71.
23. Lachat CK, Verstraeten R, Khanh IN, et al. Validity of two physical activity questionnaires (IPAQ and PAQA) for Vietnamese
adolescents in rural and urban areas. Int J Behav Nutr Phys Act.
2008;5:37.
24. Lee PH, Macfarlane DJ, Lam TH, Stewart SM. Validity of
the International Physical Activity Questionnaire short form
(IPAQ-SF): a systematic review. Int J Behav Nutr Phys Act. 2011;
8:115.
25. Macfarlane D, Chan A, Cerin E. Examining the validity and reliability of the Chinese version of the International Physical Activity
Questionnaire, long form (IPAQ-LC). Public Health Nutr. 2010;
13:1–8.
26. Maddison R, Ni MC, Jiang Y, et al. International Physical Activity
Questionnaire (IPAQ) and New Zealand Physical Activity Questionnaire (NZPAQ): a doubly labelled water validation. Int J Behav
Nutr Phys Act. 2007;4:62.
27. Murphy MH, Blair SN, Murtagh EM. Accumulated versus continuous exercise for health benefit a review of empirical studies.
Sports Med. 2009;39(1):29–43.
28. Nang EEK, Ngunjiri SAG, Wu Y, et al. Validity of the International Physical Activity Questionnaire and the Singapore prospective study program physical activity questionnaire in a
multiethnic urban Asian population. BMC Med Res Methodol.
2011;11:241.
29. Rennie KL, Wareham NJ. The validation of physical activity
instruments for measuring energy expenditure: problems and pitfalls. Public Health Nutr. 1998;1(4):265–71.
30. Sallis JF, Saelens BE. Assessment of physical activity by selfreport: status, limitations, and future directions. Res Q Exerc
Sport. 2000;71(2):1–14.
31. Stegle O, Fallert SV, MacKay DJC, Brage S. Gaussian process
robust regression for noisy heart rate data. IEEE Trans Biomed
Eng. 2008;55(9):2143–51.
32. Strath SJ, Brage S, Ekelund U. Integration of physiological and
accelerometer data to improve physical activity assessment. Med
Sci. Sports Exerc. 2005;37(11 Suppl):S563–71.
33. Thompson D, Batterham AM, Bock S, Robson C, Stokes K. Assessment of low-to-moderate intensity physical activity thermogenesis
in young adults using synchronized heart rate and accelerometry with
branched-equation modeling. J Nutr. 2006;136:1037–42.
34. van Poppel MN, Chinapaw MJ, Mokkink LB, Van Mechelen W,
Terwee CB. Physical activity questionnaires for adults. A systematic review of measurement properties. Sports Med. 2010;
40(7):565–600.
35. Wijndaele K, Brage S, Besson H. Television viewing time independently predicts all-cause and cardiovascular mortality: the
EPIC Norfolk Study. Int J Epidemiol. 2011;40:150–9.
Paper II
PAPER II
Physical activity patterns in Greenland: A country in transition
Dahl-Petersen IK
Jørgensen ME
Bjerregaard P
Scandinavian Journal of Public Health, 2011; 39: 678–6
Scandinavian http://sjp.sagepub.com/
Journal of Public Health
Physical activity patterns in Greenland : A country in transition
Inger K Dahl-Petersen, Marit E Jørgensen and Peter Bjerregaard
Scand J Public Health published online 22 September 2011
DOI: 10.1177/1403494811420486
The online version of this article can be found at:
http://sjp.sagepub.com/content/early/2011/09/21/1403494811420486
Published by:
http://www.sagepublications.com
Additional services and information for Scandinavian Journal of Public Health can be found at:
Email Alerts: http://sjp.sagepub.com/cgi/alerts
Subscriptions: http://sjp.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Sep 22, 2011
What is This?
Downloaded from sjp.sagepub.com at Syddansk Universitetsbibliotek on October 12, 2011
XML Template (2011)
K:/SJP/SJP 420486.3d
[6.9.2011–1:26pm]
(SJP)
[1–9]
[PREPRINTER stage]
Scandinavian Journal of Public Health, 2011; 0: 1–9
ORIGINAL ARTICLE
INGER K. DAHL-PETERSEN1, MARIT E. JØRGENSEN2 & PETER BJERREGAARD1
1
Centre for Health Research in Greenland, National Institute of Public Health, University of Southern Denmark,
Denmark, and 2 Steno Diabetes Center, Gentofte, Denmark
Abstract
Aims: To examine differences in physical activity patterns among Inuit in Greenland in relation to social transition. The Inuit
in Greenland are an indigenous population in the circumpolar north who are experiencing rapid social transition. Methods:
Physical activity patterns were assessed by the International Physical Activity Questionnaire (long version). The population
was divided into six groups according to different stages of social change, measured on the basis of education, current
residence and occupation. Data were collected in a country-wide cross-sectional population survey among adult Inuit in
Greenland from 2005 to 2009. Results: Men with long vocational or academic education living in towns (latest stage of social
change) spent significantly less time on occupational physical activity (p ¼ 0.001) compared with hunters and fishermen in
villages (earliest stage of social change) (trend test p ¼ 0.01). Women in the latest stage of change spent significantly less time
on domestic physical activity (p < 0.001) (trend test p ¼ 0.06) compared with women in the earliest stage of social change.
This was also found for physical activity during transportation (p ¼ 0.02 and p ¼ 0.01 for men and women, respectively). No
significant difference was found for leisuretime physical activity. Men and women in the latest stage of social change spent
more time on sedentary activity (p < 0.001). Conclusions: Differences in physical activity patterns among Inuit in
Greenland included decreasing time spent on domestic and occupational physical activity and increasing time
spent on sedentary activities along with social change. Knowledge of changes in physical activity patterns in
relation to social transition is important in prevention of obesity, type 2 diabetes and lifestyle diseases.
Key Words: Cardiovascular risk, gender, physical activity, social change, social transition
Background
Indigenous populations in the circumpolar north
have experienced a rapid cultural and social transition that has changed many aspects of everyday life.
Furthermore, a decrease in infectious diseases and an
increase in lifestyle-related chronic diseases, such as
type 2 diabetes, has been observed [1,2] similar to the
experience of developing populations throughout the
world [3]. Changes in physical activity patterns could
be an important contributor to this rise in chronic
lifestyle diseases. Changes in physical activity patterns in relation to social transition have been coined
as the physical activity transition [4]. The consequences of the physical activity transition are relevant
for all populations in Europe and North America but
might be more marked in populations experiencing
rapid social changes, such as indigenous populations
in the circumpolar north. A study among the Yakut of
Siberia showed that individuals with more traditional
lifestyles had higher energy expenditure than individuals with more modern lifestyles [5]. A temporally
decreasing level of fitness along with an increasing
sedentary lifestyle was found among Inuit in Canada
[6]. Although a reduction over time in fitness was
found among populations in the circumpolar north,
very few studies have examined the physical activity
transition in terms of differences in physical activity
patterns.
In order to examine the physical activity transition
in a population going through rapid transition we
used a population-based study among Inuit in
Greenland. In Greenland profound and rapid
Correspondence: Inger K. Dahl-Petersen, Centre for Health Research in Greenland, National Institute of Public Health, University of Southern Denmark,
Denmark.
E-mail: [email protected]
(Accepted 25 July 2011)
ß 2011 the Nordic Societies of Public Health
DOI: 10.1177/1403494811420486
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
2
[1–9]
[PREPRINTER stage]
I. K. Dahl-Petersen et al.
changes in the social structure have occurred over the
last 50 years, resulting in a shift from a traditional
lifestyle to a more Westernized lifestyle, especially in
the large towns. This process has included economic,
social and cultural changes, such as changes from
subsistence hunting and fishing to wage earning,
population movement from small villages to larger
towns, and increased availability of formal education
accompanied by changes in lifestyle [7]. These
changes are similar to those observed in other
indigenous populations in the circumpolar north [7].
Within the population of Greenland there are still
large differences in socioeconomic status and living
conditions between the small villages and the larger
towns. Inuit in Greenland are therefore considered
an ideal population for investigating the physical
activity transition.
In order to examine the differences in physical
activity patterns, most research worldwide has used
an urban-rural dichotomy based on residence.
Comparing 51 countries, mainly developing countries, Guthold et al. found that both men and women
living in urban areas were more likely to be inactive
compared with those living in rural areas [8]. Other
studies have shown that urban people were characterized by a reduced intensity of occupational activity
as a result of modernization [9,10]. The use of an
urban/rural dichotomy has been criticized for ignoring the heterogeneity of environments within urban
and rural areas [11] and an inability to detect changes
in urbanicity over time [12].
To examine the physical activity patterns in relation to the physical activity transition we used
residence, education and occupation status to rank
the Inuit population in our cross-sectional study into
six sub-groups as a proxy for different stages of social
change.
The aim of the study was to study the physical
activity transition among Inuit in Greenland as
differences in physical activity patterns in relation to
level of social change. Information about the physical
activity transition in Greenland could be useful for
indigenous populations in the circumpolar north.
Methods
Study population
The data for this population-based cross-sectional
study was collected in Greenland from 2005 to 2009.
The total population of Greenland is 57,000 of which
90% are Inuit. During 2005 and 2009 a total of 2,834
adult Inuit aged 18 years and above were interviewed.
Eight towns (population ranging from 1,150–
14,700) and 10 villages (population ranging from
100–425) in Greenland (25% of all communities)
were selected as study areas to represent different
community sizes and geographical locations. From
these 18 communities a random population sample
was drawn from the central population register.
Pregnant women, individuals not born in
Greenland or Denmark, and individuals who had
moved out of the study area, were excluded from the
population sample. We furthermore confined
the study to Inuit as defined by the participant and
the interviewer at the time of enrolment.
A detailed description of the methods is available
elsewhere [13].
Interview and clinical measurements
An interviewer administered questionnaire was
developed in the Danish language, translated into
Greenlandic and translated back into Danish.
Information on socio-demographic factors and lifestyle, including physical activity, was obtained from
the questionnaire. Participants were interviewed face
to face by trained native Greenlandic speaking interviewers and the interviews were conducted in the
language of choice of the participant.
Physical activity
Information on physical activity was collected using a
modified version of the long version of the seven-day
International Physical Activity Questionnaire (IPAQ)
[14]. IPAQ assesses four domains of physical activity
in daily life (work, transportation, domestic and
leisure time) and time spent sitting down.
Participants were asked to indicate time spent on
physical activity in the last seven days; how often (the
number of days per week) and for how long (the
average duration in minutes) divided into vigorous
intensity, moderate intensity and walking for the four
domains separately. The original English version of
the questionnaire was translated into Greenlandic
and back-translated in accordance with the guidelines from the IPAQ group [15]. The questionnaire
was adjusted to Greenlandic living conditions by
replacing the activity examples with culturally appropriate examples in accordance with guidelines given
by the IPAQ group. Furthermore, one question about
gardening was excluded because this question was
not relevant under Arctic conditions. In a small study
(n ¼ 79) in one Greenlandic town, the adapted
version of IPAQ was validated against a combined
accelerometer and heart rate monitor (Actiheart)
showing validity comparable with other population
studies. Spearman correlation coefficient for activity
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
[1–9]
[PREPRINTER stage]
energy expenditure (AEE) was 0.50 (data not
published).
Transition variable
Six levels of social transition were defined from family
occupation type, education, and place of residence
including A) hunters and fishermen in villages; B)
inhabitants of villages who were not hunters or
fishermen; C) wage earners with no vocational education, currently living in towns but having lived in
villages at age 10; D) wage earners with no vocational
education, currently living in towns and having also
lived in towns at age 10; E) wage earners with short
vocational education, living in towns irrespective of
childhood residence; and F) wage earners with long
vocational or academic education, living in towns. In
a Greenland context the largest community in each of
the 17 districts is by definition a town while the rest
are villages. The population of towns ranges from 485
to 15,000 and the population of villages from less
than 50 to around 450. In the towns are located the
school, health centre, church, main shops and
administration of the district. Because full educational achievement cannot be expected before the
mid-20s and the proportion of participants with an
active working life is rapidly reduced after the mid60s, we chose to analyze only those aged 25–64 years.
The term ‘‘social change’’– from an earlier to a later
stage, was chosen to refer to the process studied and
included elements of urbanization, modernization
and westernization.
3
Data analysis
The analyses were performed separately for men and
women. Data on physical activity are presented in
hours per day for each domain and only for participants reporting physical activity. A square root
transformation was applied to time spent on physical
activity in the different domains in order to approximate a normal distribution. Time used on physical
activity is presented as median hours/day with
interquartile ranges for each domain of physical
activity as well as for total physical activity. The
results for sedentary behaviour are shown for men
and women together and a 2.5% to 97.5% range is
presented. The proportion reporting no physical
activity is presented separately. Comparison of physical activity patterns of individuals in an earlier to a
later stage of social change were carried out.
Differences in physical activity patterns in each
group of the transition variable were tested using a
multiple linear regression model with the transformed physical activity variable as the dependent
variable adjusted for age. p values are presented for
differences in physical activity patterns between the
individuals/group in the latest stage of social change
(F) compared with the group/individuals in the
earliest stage of change (A).
A test for linear trend of differences in physical
activity patterns in relation to the groups of transition
(Likelihood-ratio; STATA version 10) was performed, adjusted for age. The study was ethically
approved by the ethical review committee for
Greenland. Written informed consent was obtained
from all participants.
Data processing
Data on physical activity were collected from 2,831
Inuit participants. All participants who reported days
(frequency) but not time (duration) of physical
activity or vice versa were treated as missing, n ¼ 65
(2.3%). According to the IPAQ scoring protocol
[15], 132 (4.8%) participants reporting more than
960 minutes (16 hours) of physical activity a day were
excluded from the final analysis. In each domain
minutes spent on physical activity at moderate and
vigorous intensity activity and walking for more than
180 minutes a day at each activity was truncated to
180 minutes. If the participant reported time spent
on physical activity between 0 and 10 minutes the
answers were not truncated to 0. In total, 555
(21.0%) participants were excluded because they
were younger than 25 or older than 64 years of age,
resulting in a total of 2,079 participants. Thirty-three
participants (1.6%) could not be placed in any of the
transition groups due to missing information on
education, residency or occupational status.
Results
Sample characteristics
The study provides information on physical activity
patterns from 2,079 Inuit participants aged 25–64
years old. The participant rate was 68% for Inuit.
Table I shows the characteristics of the study population separately for men and women.
No differences were found between the proportion
of men and women categorized in the different social
transition groups. Mean age for the women was
found to be slightly lower compared with men.
Physical activity in different domains of daily life in
relation to social change
Table II presents time used on physical activity in the
four physical activity domains stratified by the six
transition groups. The level of total physical activity
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
4
[1–9]
[PREPRINTER stage]
Table I. Characteristics of the study sample. Age 25–64 years. Greenland, 2005–2009.
Men
Age (years) mean (SD)
Place of residence
Nuuk
Town
Village
Transition groups
A Hunters/fishermen
B Other villagers
C Blue collar migrants
D Other blue collar
E Intermediate
F Professionals
No physical activity reported
Women
n
%
n
%
p value*
887
44.6 (10.4)
1,192
43.6 (10.0)
0.03
0.2
132
531
224
14.9
59.9
25.3
214
697
281
18.0
58.5
23.6
65
159
74
198
286
93
13
7.4
18.2
8.5
22.6
32.7
10.6
1.5
81
200
132
283
336
139
8
6.9
17.1
11.3
24.2
28.7
11.9
0.7
0.14
% was presented unless otherwise noted. Differences in gender were tested by Chi-square test for categorical data, t-test for
normally distributed data.
*indicates difference between men and women.
decreased along the social transition categories, but
not significantly (trend test: men p ¼ 0.075; women
p ¼ 0.2). Total hours spent on physical activity were
significantly higher among hunting and fishing families living in a village (group A) compared with wage
earners with long vocational or academic education,
living in towns (group F), adjusted for age (men
p < 0.001; women p ¼ 0.002).
A significant test for trend was found for decreasing occupational physical activity by social change
among men (p ¼ 0.01), but is most likely explained
by the significant difference in median hours spent on
physical activity between group A and F.
Women in the latest stage of social change were
found to spend significantly less time on domestic
physical activity (p < 0.001) compared with those in
the earliest stage of social change. For the transportation domain no significant trend was found.
However a significantly less amount of physical
activity was found for men and women in the latest
stage of social change compared with the earliest
stage (men p ¼ 0.02; women p ¼ 0.01)
Contrary to what was expected, no significant
difference was found for leisuretime physical activity
between the most modern group (group F) compared
with the most traditional (group A). And no significant trend was seen across the social transition
groups for men and women.
Table II also shows the proportion of participants
reporting no physical activity, separately by domain
and stratified by groups of social transition.
For physical activity at work there was a clear
pattern that a smaller proportion of hunters and
fishermen (group A) reported no physical activity
compared with professionals (group F). For leisuretime activity the pattern was contrary, with a greater
proportion of individuals in the most traditional
group reporting no physical activity compared with
the most Westernized group.
The questionnaire provided information on time
used on both moderate and vigorous intensity physical activity. Moderate intensity physical activity
includes activities such as child care and cleaning,
while shovelling snow and heavy lifting are activities
of vigorous intensity. A consistent pattern was found
that the group of wage earners who had a long
vocational or academic education and who lived in
towns (group F) used less time on moderately
intensive activities compared with hunters and fishermen in villages (group A) (men and women
p < 0.001), however no significant trend was
found across the whole spectrum of social change
(Figure 1). This pattern was also found for vigorous
physical activity (men and women p ¼ 0.02). For all
groups less time was used on vigorous intensity
activity compared with moderate intensity activity.
The results suggest that less time spent on physical
activity at work or at home in the group with long
vocational or academic education and living in a town
is associated with more sedentary behaviour and not
compensated by a higher level of physical activity in
the other domains. The time spent on sedentary
activity was markedly higher among wage earners
with long vocational or academic education, living in
towns (group F) compared with hunting and fishing
families living in a village (group A) for both men and
women adjusted for age (p < 0.001) (data for men
and women shown together) (Figure 2). Half of the
3.0 (0.4–3.2)
2.5 (1.3–3.4)
3.2 (2.1–5.0)
3.0 (1.7–4.4)
2.9 (0.6–3.3)
1.6 (0.7–3.0)
0.33
<0.001
4.5 (2.8–6.7)
4.4 (2.9–6.0)
4.3 (2.2–7.0)
4.0 (2.2–6.6)
4.3 (2.3–6.6)
3.4 (1.9–4.9)
0.002
0.2
0.02
(24.5)
(17.0)
(8.9)
(13.8)
(20.9)
(25.6)
No PA
reported
n ¼ 154
(%)
<0.001
(5.6)
(8.6)
(9.3)
(12.9)
(14.6)
(35.6)
No PA
reported n ¼ 92
(%)
0.73
1.0 (0.6–1.5)
0.8 (0.4–1.4)
0.9 (0.4–1.4)
0.9 (0.3–1.6)
0.7 (0.3–1.6)
0.6 (0.4–1.3)
0.07
Daily PA
n ¼ 898
Median
hours (IQR)
0.002
(41.5)
(25.2)
(24.3)
(19.2)
(19.6)
(17.2)
No PA
reported
n ¼ 195
(%)
<0.001
(35.8)
(28.5)
(31.1)
(20.8)
(18.5)
(18.0)
No PA
reported
n ¼ 273 (%)
Leisuretime domain
0.81
1.1 (0.5–1.9)
1.1 (0.4–2.0)
0.7 (0.4–1.9)
0.7 (0.3–2.0)
1.0 (0.4–1.7)
0.9 (0.4–1.7)
0.22
Daily PA
n ¼ 680
Median hours
(IQR)
0.06
2.0 (1.1–3.0)
2.0 (1.0–3.0)
1.1 (0.6–2.0)
1.0 (0.5–2.0)
1.0 (0.5–2.0)
0.7 (0.3–1.3)
<0.001
Daily PA
n ¼ 1,050
Median
hours (IQR)
Domestic domain
0.26
0.8 (0.3–1.7)
1.0 (0.4–2.0)
0.5 (0.2–1.2)
0.5 (0.2–1.0)
0.5 (0.3–1.0)
0.5 (0.2–1.1)
0.07
Daily PA
n ¼ 679
Median hours
(IQR)
0.002
(2.5)
(8.0)
(8.3)
(16.6)
(9.5)
(9.4)
No PA
reported
n ¼ 121
(%)
0.6
(24.6)
(17.6)
(24.3)
(23.7)
(24.1)
(19.4)
No PA
reported
n ¼ 196
(%)
Domestic domain
0.27
1.0 (0.3–1.0)
1.0 (0.4–2)
0.6 (0.3–1.0)
0.5 (0.3–1.0)
0.5 (0.3–1.0)
0.5 (0.3–1.0)
0.01
Daily PA
n ¼ 1041
Median
hours (IQR)
0.08
(18.5)
(14.5)
(7.6)
(11.0)
(9.2)
(10.1)
No PA
reported
n ¼ 130
(%)
0.4
(20.0)
(19.5)
(12.2)
(12.6)
(17.5)
(14.0)
No PA
reported
n ¼ 141
(%)
Transportation domain
0.16
1.0 (0.2–2.0)
0.9 (0.3–2.0)
0.5 (0.3–1.0)
0.6 (0.3–1.5)
0.5 (0.3–1.0)
0.5 (0.3–1.0)
0.017
Daily PA
n ¼ 734
Median hours
(IQR)
Transportation domain
Time presented in median hours/day with interquartile range. p values for group F relative to group A (hunters and fishermen) derived from linear regression on transformed data. Analysis
adjusted for age. ** Likelihood-ratio test for trend adjusted for age. ***Only participants reporting occupational activity outside the home. **** Chi-square test. A: Hunters and fishermen in
villages; B: Inhabitants of villages who were not hunters or fishermen; C: Wage earners with no vocational education, currently living in towns but having lived in villages at age 10; D: Wage
earners with no vocational education, currently living in towns and having lived in towns also at age10; E: Wage earners with short vocational education, living in towns; and F: Wage earners with
long vocational or academic education, living in towns.
A Hunter/fishermen
B Other villagers
D Other blue collar
E Intermediate
F Professionals
p values for group
F relative to group A*
Test for trend**
p values****
Daily PA
n ¼ 670
Median hours
(IQR)
Work domain***
0.01
0.075
Total daily
PA n ¼ 1,171
Median hours
(IQR)
3.1 (1.5–4.3)
3.3 (1.5–5.1)
3.3 (1.7–5.6)
3.0 (1.3–5.3)
3.0 (0.9–5.3)
1.0 (0.3–3.0)
0.001
Daily PA
n ¼ 541
Median hours
(IQR)
5.0 (2.6–6.7)
4.2 (2.5–6.7)
4.7 (2.1–6.9)
4.2 (2.3–6.6)
4.1 (1.9–6.3)
2.6 (1.2–4.3)
<0.001
Total daily
PA n ¼ 875
Median hours
(IQR)
Leisuretime domain
[6.9.2011–1:26pm]
(SJP)
Transition
groups
(women)
A Hunter/fishermen
B Other villagers
D Other blue collar
E Intermediate
F Professionals
p values for group
F relative to group A*
Test for trend**
p values****
Transition
groups
(men)
Work domain***
Table II. Daily hours spent on physical activity (PA) in total and separately in the four domains of daily life and proportion of participants reporting no physical activity stratified by groups of
social transition. Results from the PA analysis were only presented for those participants reporting PA within each domain. Greenland, 2005–2009.
XML Template (2011)
[1–9]
[PREPRINTER stage]
5
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
6
[1–9]
[PREPRINTER stage]
Hours of moderate and vigorous intensity activity
8
6
4
*
*
**
2
**
0
Hunters/fishermen
Professionals
Hunters/fishermen
Men
Professionals
Women
Moderate intensity activity
Vigorous intensity activity
Figure 1. Time presented in median hours/day with interquartile range and 2.5% and 97.5% percentiles. Only data from participants
reporting physical activity are presented.
*p < 0.001 for group F relative to group A derived from linear regression on transformed data and adjusted for age; **p ¼ 0.02 for men
and women for group F relative to group A derived from linear regression on transformed data and adjusted for age. Likelihood- ratio
test for trend adjusted for age only significant for vigorous activity for women p ¼ 0.02. Moderate activity, men n ¼ 777 and women
n ¼ 1,087, vigorous activity, men n ¼ 496 and women n ¼ 393.
men and women in the most modern group were
found to spend more than five hours a day on
sedentary activity compared with those in the most
traditional group spending three hours a day.
Discussion
By stratifying the population into transition groups
we were to some extent able to evaluate how physical
activity patterns change when society is changing.
The stratification of the population provided us with
the possibility not only to study the differences
between town and village (urban vs rural) but the
differences in educational level and occupation
status, which are known to have implications for
physical activity [16].
The stratification into six population groups is only
a proxy for longitudinal information and hence there
remains the question to what extent it represents true
temporal changes in physical activity patterns. In any
case its use in other populations and with other
cardiovascular risk factors must be done with caution. A further study of its plausibility and usefulness
in the study of behavioural risk factors for cardiovascular disease among the Inuit is underway.
The decrease in total physical activity by social
change was mostly explained by less time used on
physical activity in the occupational domain for men
and in the domestic domain for women. Moreover a
greater proportion of men in the latest stage of social
change were reporting no physical activity at work
compared with men in the early stage of social
change. This is most likely a result of differences in
type of work, from the traditional physically demanding work, such as hunting and fishing, to servicebased and sedentary occupations. Also the increase
of labour-saving household goods and improved
sanitation facilities could have contributed to a
lower amount of physical activity in the domestic
domain for women in the latest state of social change.
Increasing time spent on sedentary activities, such
as watching television, is seen in many Western
countries. But some research teams in Western
countries have identified an increase in leisuretime
physical activity [17] and an upward trend in sports
participation over time [18]. However studies of
temporal trends in physical activity are sparse. In our
study we found more time used on sedentary activity
among individuals in the late stages of social change,
but no indication of increased physical activity during
leisure time to compensate for less physical activity at
work and for domestic chores and more sedentary
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
[1–9]
[PREPRINTER stage]
7
Daily hours of sedentary activity
15
*
10
5
0
A.Hunters/
fishermen
B. Other
villagers
C. Blue collar
migrants
D. Other blue
collar
E Intermediate F. Professionals
Figure 2. Sedentary time presented in median hours with interquartile range and 2.5% and 97.5% percentile and stratified by groups
of social transition (men and women). Only data from participants reporting sedentary activity are presented. Greenland 2005–2009.
p value group F relative to group A: p < 0.001. Test for trend p ¼ 0.01. n ¼ 1,996. A: Hunters and fishermen in villages; B: Inhabitants of
villages who were not hunters or fishermen; C: Wage earners with no vocational education, currently living in towns but having lived in
villages at age 10; D: Wage earners with no vocational education, currently living in towns and having lived in towns also at age 10;
E: Wage earners with short vocational education, living in towns and F: Wage earners with long vocational or academic education, living
in towns.
behaviour. However, the proportion of individuals
reporting no physical activity in their leisure time was
smaller among individuals in the most Westernized
group compared with individuals in the most traditional group. The tradition for promoting leisuretime
physical activity might not yet be that well-established compared with Western countries, although
activities for promoting physical activity in the larger
towns of Greenland have increased. Most large towns
have a number of sports facilities and activities and an
enormous potential for being outside and other
outdoor activities. However, it is likely that time is
spent on watching television or DVDs and on a
computer rather than using these facilities. A study of
trends in physical activity of Greenlandic schoolchildren showed that about one-quarter of the children were heavy television watchers, spending four
hours a day or more [19].
In Greenland there are no roads connecting the
towns and villages, but nevertheless there has been an
increase in the number of cars, buses, taxis and snow
mobiles in the larger towns over the last 40 years,
which may explain our finding of a less physically
active way of transportation among the most
Westernized individuals. Access to motorized transportation has been found to be associated with low
levels of physical activity [20].
In our study we used a questionnaire measuring
physical activity in four domains of daily life. This
differs from the majority of other questionnaires that
only evaluate leisuretime physical activity. We found
that for all groups relatively less time was spent on
physical activity in leisure time compared with time
spent on physical activity in the occupational and
domestic domains. That is consistent with studies of
physical activity in developing countries, which have
shown that leisure time in general is the domain with
the lowest activity level because people’s lifestyles do
not give them the time and the physical capacity for
leisuretime activities.
The international physical activity questionnaire
used in this study was developed to be used in
different cultural settings [8] and has been widely
used and tested for reliability and validity in several
countries [21–23]; However, IPAQ and self-reported
methods in general also have some limitations.
Studies have found both over- and underestimation
of physical activity using the long form of IPAQ
compared with other measurement methods
[22,24,25]. A study among Inuit in Nunavut found
very high levels of physical activity using the short
form of IPAQ and suggested an additional need for
validation of the questionnaire for the Inuit population [26]. Therefore we chose not to evaluate whether
the population met the national recommendations
for physical activity or to compare the activity level
with other countries. Even though IPAQ is a questionnaire designed to measure physical activity in
different cultural settings, some degree of misclassification of activity is likely. In particular traditional
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
8
[1–9]
[PREPRINTER stage]
activities such as hunting and fishing could potentially be difficult to classify according to domains and
intensity. However, we tried to minimize this misclassification by using trained Greenlandic interviewers instead of using a self-administered
questionnaire and by adapting the activity examples
to Greenlandic living conditions. A Swedish study
used a qualitative approach in order to examine the
concepts of physical activity and exercise and different intensity modifiers as used in IPAQ and found
that the interpretation of physical activity was dependent on differences in people’s experiences [27].
Such an approach could contribute to knowledge of
the interpretation of physical activity in a cultural
setting such as Greenland and also contribute to the
development of appropriate activity examples in
questionnaires measuring activity as IPAQ to
reduce potential misclassification of activity.
In order to gain as much information on physical
activity as possible and to prevent overlooking important differences in physical activity patterns in relation to social transition among indigenous
populations it is recommended that all four domains
of physical activity should be examined. In addition a
number of studies have indicated domain-specific
health effects worth further examination in populations undergoing rapid transition [28]. Our results
suggest that Westernization of a country means more
time spent on sedentary activity. Sedentary activity
and watching television proved to be independent
risk factors for metabolic health [29,30] and we
suggest that not only physical activity transition is
studied but also changes in sedentary activity in
countries in the circumpolar north.
Conclusion
In this study we found differences in the level of
occupational and domestic physical activity, physical
activity during transportation and time spent on
sedentary activity in relation to social transition
among Inuit in Greenland. Greenland is an example
of an indigenous population in the circumpolar north
undergoing rapid transition. Differences in physical
activity patterns across the social transition groups
may be interpreted as reflecting temporal changes in
physical activity. As a part of the prevention of
chronic lifestyle diseases it is suggested that the
promotion of physical activity during leisure time
and transportation should be intensified and that
sedentary behaviour should be reduced.
Knowledge of changes in physical activity patterns
along with social transition can be useful in other
indigenous populations in the circumpolar north
going though rapid social transition.
Funding
The work was supported by Karen Elise Jensen’s
Foundation.
References
[1] Bjerregaard P, Young TK, Dewailly E, Ebbesson SO.
Indigenous health in the Arctic: an overview of the circumpolar Inuit population. Scand J Public Health
2004;32(5):390–5.
[2] Jørgensen ME, Bjerregaard P, Borch-Johnsen K. Diabetes
and impaired glucose tolerance among the inuit population
of Greenland. Diabetes Care 2002;25(10):1766–71.
[3] Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of
cardiovascular diseases: part I: general considerations, the
epidemiologic transition, risk factors, and impact of urbanization. Circulation 2001;104(22):2746–53.
[4] Katzmarzyk PT, Mason C. The physical activity transition.
J Phys Act Health 2009;6(3):269–80.
[5] Snodgrass JJ, Leonard WR, Tarskaia LA, Schoeller DA.
Total energy expenditure in the Yakut (Sakha) of Siberia as
measured by the doubly labeled water method. Am J Clin
Nutr 2006;84(4):798–806.
[6] Rode A, Shephard RJ. Physiological consequences of acculturation: a 20–year study of fitness in an Inuit community.
Eur J Appl Physiol Occup Physiol 1994;69(6):516–24.
[7] Young TK, Bjerregaard P. Health transitions in Arctic
populations. Toronto: University of Toronto Press; 2008.
[8] Guthold R, Ono T, Strong KL, Chatterji S, Morabia A.
Worldwide variability in physical inactivity a 51-country
survey. Am J Prev Med 2008;34(6):486–94.
[9] Sobngwi E, Mbanya JC, Unwin NC, Kengne AP, Fezeu L,
Minkoulou EM, et al. Physical activity and its relationship
with obesity, hypertension and diabetes in urban and rural
Cameroon.
Int
J
Obes
Relat
Metab
Disord
2002;26(7):1009–16.
[10] Monda KL, Gordon-Larsen P, Stevens J, Popkin BM.
China’s transition: the effect of rapid urbanization on adult
occupational
physical
activity.
Soc
Sci
Med
2007;64(4):858–70.
[11] McDade TW, Adair LS. Defining the ‘‘urban’’ in urbanization and health: a factor analysis approach. Soc Sci Med
2001;53(1):55–70.
[12] Dahly DL, Adair LS. Quantifying the urban environment:
a scale measure of urbanicity outperforms the urban-rural
dichotomy. Soc Sci Med 2007;64(7):1407–19.
[13] Bjerregaard P. Inuit health in transition. Greenland survey
2005–2008. Population sample and survey methods. 1–13.
National Institute of Public Health; 2009.http://www.sifolkesundhed.dk/upload/inuit_health_in_transition_greenland_methods_5_002.pdf (accessed 20. January 2011).
[14] IPAQ-group. http://www.ipaq.ki.se/downloads.htm (accessed
20 January 2011).
[15] IPAQ-group. IPAQ analyse guidelines. http://www.ipaq.ki.se/
scoring.pdf (accessed 20 January 2011).
[16] Ford ES, Merritt RK, Heath GW, Powell KE, Washburn RA,
Kriska A, et al. Physical activity behaviors in lower and higher
XML Template (2011)
[6.9.2011–1:26pm]
(SJP)
[1–9]
[PREPRINTER stage]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
socioeconomic status populations. Am J Epidemiol
1991;133(12):1246–56.
Petersen CB, Thygesen LC, Helge JW, Gronbaek M,
Tolstrup JS. Time trends in physical activity in leisure time
in the Danish population from 1987 to 2005. Scand J Public
Health 2010;38(2):121–8.
Stamatakis E, Ekelund U, Wareham NJ. Temporal trends in
physical activity in England: the Health Survey for England
1991 to 2004. Prev Med 2007;45(6):416–23.
Vindfeld S, Schnohr C, Niclasen B. Trends in physical
activity in Greenlandic schoolchildren, 1994–2006. Int J
Circumpolar Health 2009;68(1):42–52.
Dombois OT, Braun-Fahrlander C, Martin-Diener E.
Comparison of adult physical activity levels in three Swiss
alpine communities with varying access to motorized transportation. Health Place 2007;13(3):757–66.
Craig CL, Marshall AL, Sjostrom M, Bauman AE,
Booth ML, Ainsworth BE, et al. International physical
activity questionnaire: 12-country reliability and validity.
Med Sci Sports Exerc 2003;35(8):1381–95.
Hagstromer M, Oja P, Sjostrom M. The International
Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr
2006;9(6):755–62.
Hallal PC, Victora CG. Reliability and validity of the
International Physical Activity Questionnaire (IPAQ). Med
Sci Sports Exerc 2004;36(3):556.
Boon RM, Hamlin MJ, Steel GD, Ross JJ. Validation of the
New Zealand Physical Activity Questionnaire (NZPAQ-LF)
[25]
[26]
[27]
[28]
[29]
[30]
9
and the International Physical Activity Questionnaire
(IPAQ-LF) with accelerometry. Br J Sports Med 2008
[online] 2010;44:741–6.
Maddison R, Ni MC, Jiang Y, Vander HS, Rodgers A,
Lawes CM, et al. International Physical Activity
Questionnaire (IPAQ) and New Zealand Physical Activity
Questionnaire (NZPAQ): a doubly labelled water validation.
Int J Behav Nutr Phys Act 2007;4:62.
Hopping BN, Erber E, Mead E, Roache C, Sharma S. High
levels of physical activity and obesity co-exist amongst Inuit
adults in Arctic Canada. J Hum Nutr Diet 2010;23(suppl.
1):110–14.
Johnson I, Tillgren P, Hagstromer M. Understanding and
interpreting the concept of physical activity – a focus group
study among Swedish women. Scand J Public Health
2009;37(1):20–7.
Abu-Omar K, Rütten A. Relation of leisure time, occupational, domestic, and commuting physical activity to
health indicators in Europe. Preventive Medicine
2008;47(3):319–23.
Dunstan DW, Salmon J, Owen N, Armstrong T, Zimmet PZ,
Welborn TA, et al. Associations of TV viewing and physical
activity with the metabolic syndrome in Australian adults.
Diabetologia 2005;48(11):2254–61.
Healy GN, Wijndaele K, Dunstan DW, Shaw JE, Salmon J,
Zimmet PZ, et al. Objectively measured sedentary time,
physical activity, and metabolic risk: the Australian Diabetes,
Obesity and Lifestyle Study (AusDiab). Diabetes Care
2008;31(2):369–71.
Paper III
PAPER III
Physical activity energy expenditure is associated with 2-h insulin
independently of obesity among Inuit in Greenland
Dahl-Petersen IK
Bjerregaard P
Brage S
Jørgensen ME
Diabetes Research and Clinical Practice, 2013; article in press
Included in this thesis in submitted form
Title: Physical Activity Energy Expenditure is associated with 2-hour Insulin independently of
obesity among Inuit in Greenland
Running title: Physical Activity and metabolic health in the Arctic
Authors: MSPH IK Dahl-Petersen1, Prof P Bjerregaard 1, PhD S Brage 2, PhD ME Jørgensen 3
1
National Institute of Public Health, University of Southern Denmark
2
MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge, United Kingdom
3
Steno Diabetes Center, Gentofte, Denmark
Corresponding author:
Inger Katrine Dahl-Petersen, National Institute of Public Health, University of Southern Denmark
Address: University of Southern Denmark
Øster Farimagsgade 5A,2
DK 1353 Copenhagen K, Denmark
tlf:+45 65507777, fax:
E-mail: [email protected]
This study was funded by Karen Elise Jensen´s foundation, NunaFonden, Danish Medical Research
Council, and Greenland Medical Research Council. The authors declare that there are no conflicts
of interest.
Word count abstract: 249, Word count main text: 2981, Tables:4, Figures:1 references:30
1
Aims: Indigenous populations throughout the Arctic are experiencing a rapid increase in the
prevalence of obesity and type 2 diabetes. The role of physical activity in relation to glucose
metabolism in Arctic populations is not well studied. We examined the association between
objectively measured physical activity energy expenditure (PAEE) and glucose metabolism in a
population-based study of adult Inuit in Greenland.
Methods: Cross-sectional data were collected by combined accelerometry and heart rate monitoring
(ACC+HR) among Inuit (18+years) in Greenland during 2005–2010 (n=1545). PAEE was
calculated and the associations with fasting glucose, 2-h glucose, fasting insulin, 2-h insulin
concentrations and body composition were analysed by linear regression.
Results: An inverse association between PAEE and fasting insulin, 2-h insulin, 2-h glucose, fat
percentage, BMI and Waist Circumference (WC) was found after adjustments by age and sex. Only
the association between PAEE and 2-h insulin remained significant after adjustment by WC
(p=0.01), most pronounced at low levels of PAEE indicating a threshold around 35-40 kJ/kg/day.
No overall linear trend was found for fasting glucose and 2-h glucose.
Conclusions: This population-based study found that PAEE was associated with 2-h insulin
independently of weight in an inverse dose response relation. Insufficient physical activity may
contribute to impaired glucose tolerance through a pathway including alterations in obesity and fat
distribution. Both obesity and low levels of PAEE may be important contributing risk factors for the
increasing prevalence of type 2 diabetes among Inuit in Greenland, but additional risk factors
should be examined in this indigenous population.
2
Introduction
Indigenous populations throughout the Arctic have experienced rapid social, economic and cultural
changes over the past decades (1). Parallel to this, a substantial rise in prevalence of obesity and
type 2 DM similar to the characteristics of the epidemiologic transition in developing countries has
been seen (2-4). Research in developing countries and countries undergoing rapid transition has
shown that the risk of impaired glucose metabolism increases with urbanization (5;6). Physical
inactivity has been found to contribute to this risk (7;8). Research among Arctic populations has
shown that physical activity is higher among individuals with a traditional lifestyle compared with
individuals with modern lifestyles (9;10) and a study in an isolated subarctic Native Canadian
populations showed that both physical activity and fitness were independently associated with
fasting insulin concentrations (11);. In Greenland, a higher prevalence of type 2 DM and glucose
intolerance was found in rural areas compared with towns and among people with a traditional
dietary pattern, despite a higher physical activity level in these groups (12), Jeppesen et al, personal
communication). Relative beta cell dysfunction rather than insulin resistance has been shown to be
associated with type 2 DM in this population. However, knowledge about PA and glucose
metabolism among indigenous populations in the Arctic is limited. Moreover, only few studies
(7;13-15), all conducted in non-Arctic populations, have reported on objectively measured freeliving PA and its association with glucose metabolism.
In order to design appropriate prevention strategies in the Arctic it is important to obtain more
knowledge about the population-specific association between PA and various glucose metabolism
outcomes.
The aim of this study was to examine the association between objectively measured physical
activity energy expenditure (PAEE) and glucose metabolism outcomes in a population-based study
of adult Inuit in Greenland.
3
Methods
Study population
Data are from the Inuit Health in Transition study (16). In brief, data was collected as part of a
countrywide cross-sectional health survey in Greenland. The total population of Greenland is
57,000 of whom 90% are ethnic Greenlanders (Inuit). Genetically, Greenlanders are Inuit with a
mixture of European, mainly Scandinavian genes. They are genetically and culturally related to the
Inuit in Canada and Alaska and the Yupiit of Alaska and Siberia. Greenland’s 80 communities are
all located on the coast and are divided into towns ( population between 469 and 15,469) and
villages ( population from less than 10 to about 550). Participants were selected as a stratified
random sample. Greenland was divided into strata based on geography (South-West coast; CentralWest coast; North-West coast; East Greenland; North Greenland) and community size (towns with
≥ 2000 inhabitants; towns with < 2000 inhabitants; and villages). From each of these strata one or
more towns and at least two villages were selected for the study as representative of the stratum
with regard to living conditions. A random sample was drawn from the population register to obtain
around 300 participants from each town. Villages were chosen at random in the strata and all adults
in selected villages were invited to participate. Information on adults aged 18 years and older, born
in Greenland or Denmark, was collected during 2005-2010 in 9 towns and 13 villages. Ethnicity as
Greenlander or Dane was determined at enrolment based on the primary language of the participant
and self-identification. The current study focuses on Greenlanders only. Data was collected using
clinical procedures, sampling of biological media and questionnaires. Questionnaires were
developed in Danish, translated into Greenlandic, back-translated and revised where necessary.
Interview and self-administered questionnaires gave information about socio-demographic factors,
self-reported health, and lifestyles including diet, physical activity, smoking and alcohol use. The
4
study was approved by the ethical review committee for Greenland. Written informed consent was
obtained from all participants.
Anthropometric measurements
Height (nearest 0.1cm) and weight (nearest 0.1 kg) were measured with the participants wearing
light clothing and no shoes. BMI was calculated as weight/height2 (kg/m2). Waist circumference
was measured midway between the rib cage and the iliac crest, hip circumference at its maximum
on the standing participant. Weight was measured on a standard electronic clinical scale.
Bioimpedance and calculation of fat percentage was performed on a leg-to-leg Tanita TBF-300MA.
Based on a single reading, fat percent was calculated by the internal algoritm of the device, which is
based on height, weight, sex, impedance and age; body type was set as standard.
Metabolic measures
After a minimum of 8 hours of fasting, participants underwent a standardized 2-hour oral glucose
tolerance test (75 g) except for those with known diabetes at the time of health examination. Plasma
glucose was measured fasting, spun at 20 °C, 3000 rpm for 10 minutes. Plasma was separated,
frozen at –20°C and transported to one laboratory for measurement of plasma glucose. Fasting and
2-hours insulin were allowed to stand for >30 minutes and <1.5 hour before centrifugation at 20 °C,
3000 rpm for 10 minutes and stored frozen at -20ºC. Non-fasting participants or participants with
known diabetes were not included in the further analysis involving glucose and insulin parameters.
Impaired fasting glucose (IFG), impaired glucose tolerance (IGT) and type 2 DM were classified
according to WHO criteria (WHO 1999).
5
Combined accelerometry and heart rate monitoring
A combined accelerometer and heart rate monitor (ACC+HR) (Actiheart®, CamNtech Ltd,
Papworth, UK), described in technical detail elsewhere (17), was provided to a subgroup of the
participants (n=2055). The monitor was set up to measure acceleration and heart rate in 30-seconds
intervals and attached to the participant´s chest by two standard ECG electrodes (MXC55
MediMax, UK). The participants wore the monitor 24 hours a day up to 5 days from the day of the
health examination. Due to study logistics, only a limited time was available at each study location,
especially for data collection in villages. Together with a finite stock of monitors, this explains why
not all participants were given a monitor and why the length of recordings from some participants
was shorter. A subgroup of participants (n=135) conducted an individual calibration test (8-minute
step test) as described previously (18). The step tests were used to define a population-specific
calibration equation of the heart rate – activity energy expenditure relationship. Caloric intensity of
PA was estimated by combining the acceleration-based estimate of intensity (18) with the heart
rate-based estimate from the population-specific equation in a branched equation modelling
framework (19). Briefly, this method predominantly uses the accelerometer estimate during low
levels of heart rate and movement and the heart rate estimate when both heart rate and acceleration
levels are high, with equal weighting for other conditions. Resulting time-series of activity intensity
(in J/min/kg) were summarised into total physical activity energy expenditure (PAEE) in kJ/kg/day,
whilst minimising diurnal bias from potentially unbalanced data accumulated over hours of the day.
Intensity categories were defined using multiples of 1 MET (Metabolic Equivalent Task) as derived
using the Oxford equations for resting metabolic rate (20).
6
Statistical methods
Descriptive characteristics of the study sample are presented as means and SD or proportions,
stratified by sex and demographic and clinical variables by quartiles of PAEE (age and sex-adjusted
means (SE)). Linear trends across PAEE quartiles were assessed by linear regression adjusted by
age and sex. Association between deciles of PAEE and health outcomes are presented graphically
(age- and sex adjusted means). Associations between PAEE and health outcomes were examined by
multiple linear regression models (A–C). Interaction was tested. A test for non-linearity was applied
by adding PAEE2 to the regression model. Outcome variables with a skewed distribution were logtransformed for the analysis and presented as percent increase or decrease in outcome. Logistic
regression analyses were applied to examine the association between PAEE and impaired fasting
glucose (IFG), (IFG vs. normal glucose tolerance (NGT)); impaired glucose tolerance (IGT), (IGT
vs. IFG and NGT), and Diabetes (Diabetes vs. IGT, IFG and NGT). Gaussian Process regression of
heart rate was performed in JAVA using a MySQL database, and all other analyses were carried out
in STATA version 12.
Results
Data on PA were obtained from 2055 Inuit participants. After excluding recordings with insufficient
valid PA data (<48 hours), 1545 participants with complete data from ACC+HR monitoring were
available for analysis. The characteristics of the study population are presented in table 1. Mean age
was 43.8 years and 22.5% of the participants were classified as obese (BMI>30) and 7.5% with type
2 DM. A fifth of the participants lived in the capital, Nuuk, and more than half in other towns.
Median PAEE (IQR) was higher among men (56.4, IQR:40.5;75.3) compared with women (45.7,
IQR:34.2;60.7).
7
Table 2 presents socio-demographic and clinical characteristics of the study population stratified by
quartiles of PAEE. Almost twice as many men were highly physically active compared with women
(Q4: 34.2% and 18.0% for men and women) and a larger proportion of women were in the lowest
quartile of PAEE compared with men (Q1: 29.4% and 19.2% for men and women). A significant
negative linear trend across quartiles of PAEE was found for BMI, waist circumference and fat
percentage. A larger proportion of men in the lowest quartile of PAEE (Q1) was classified with
IGT compared with the most physically active (Q4) (8.7% vs. 1.5%, p=0.04). For women, this
pattern was found for IFG with 18.6% in the lowest quartile (Q1) compared with 6.9% in quartile 4
(p=0.001). The largest proportion of women and men with type 2 DM were found in the group with
the lowest level of PAEE (Q1).
Fasting and 2-h concentrations of insulin and glucose across deciles of PAEE are displayed in
Figure 1. Inverse associations were found for fasting insulin and 2-h insulin concentrations, the
latter being the strongest relationship (p=0.01), most pronounced at low levels of PAEE. The results
indicated a threshold around 35-40 kJ/kg/day. No overall linear trend was found for fasting glucose
and 2-h glucose; however a step-wise decrease was seen for 2-h glucose at the lowest level of
PAEE indicating a threshold effect around 30-35 kJ/kg/day. Inverse associations were also found
for BMI and waist circumference as shown in Figure 1 (p<0.01), and similar trend was found for fat
percentage (data not shown).
In Table 3, unadjusted (Model A) and age- and sex adjusted (Model B) linear regression analysis
showed that PAEE was strongly and significantly inversely associated with both BMI, waist
circumference and fat percentage. For every 10kJ/kg/day increase in PAEE, WC and fat percentage
decreased with 1cm and 0.6% respectively and BMI with 0.3. A significant inverse association
between PAEE and fasting glucose, 2-h glucose, fasting insulin and 2-h insulin was found in the
crude model (Model A). After adjustments for age and sex (model B) the association remained
8
unchanged for insulin measures and 2-h glucose but became insignificant for fasting glucose.
Further adjustments for waist circumference showed that only 2-h insulin remained significantly
negatively associated with PAEE (p=0.01) (Model C). Waist circumference was strongly associated
with all glucose measures in the regression analysis (data not shown). Further adjustments for
smoking and family history of type 2 DM did not materially change these associations. No
interactions between sex, WC and PAEE were found. Table 4 presents associations between PAEE
and Diabetes, IGT and IFG. Inverse associations between PAEE and non-normal glucose tolerance
states were found; however when analyses were adjusted by age, sex and WC, the associations were
no longer significant and a positive association between PAEE and IFG was found.
Lastly, we conducted sensitivity analyses to ascertain the impact of restricting our analyses to
participants providing at least 72 hrs of valid PA data. Except for the association between PAEE
and IFG which remained insignificant after adjustments for WC the results were very similar and
therefore not presented.
Discussion
Studies in populations undergoing rapid social transition have demonstrated an inverse association
between physical activity and insulin resistance independently of weight (7;11;21). A study found
that PA was negatively associated with insulin concentrations both among Pima Indians who tended
to be overweight and in Mauritians who were leaner and suggested a beneficial role of PA not
influenced by body composition (21). This finding is in accordance with our study. Since studies
have shown that both fasting and post-load insulin concentrations are correlated with measures of
insulin resistance from the insulin clamp our findings suggest an inverse association between PAEE
and insulin resistance in our population (22;23). Contrary to most previous research we were not
able to identify an association between PAEE and 2-h plasma glucose and IGT when measures of
9
body composition were included in the analysis (11;13;24;25). However the evidence from the
literature is not clear and not all studies have included body composition measures as potential
confounding or mediating factors. Research examining the patho-physiology and aetiology of
impaired fasting glycaemia (IFG) and impaired glucose tolerance (IGT) showed that Impaired
Glucose Tolerance (IGT) was predominantly related to physical inactivity, unhealthy diet and short
stature (26). Although, we adjusted our analysis for factors known to be related to both PA and
glucose metabolism outcomes, such as smoking and family history of type 2 DM, residual
confounding might be present. For example, diet, early life factors or genetic disposition not
captured by family history, which we were not able to adjust for, could play a significant role in the
development of type 2 DM in our population. Our regression analysis showed that waist
circumference was associated with glucose and insulin and that PAEE was inversely associated
with all measures of overweight and obesity. It is suggested that overweight or obesity have a
significant role in explaining differences in 2-h insulin and fasting insulin in our study population.
This conclusion is in accordance with a study of Rana et al. demonstrating that obesity and physical
inactivity independently contributed to the development of type 2 DM; however, the magnitude of
risk contributed by obesity was much greater than the lack of PA (27), although relative importance
of exposures measured with different degree of precision must be interpreted with great caution
(28). In our study adjustments by age and sex eliminated the inverse association between PAEE and
fasting glucose, a result supported by several previous studies (13;24). In a study examining
different pathophysiologic mechanisms of impaired fasting glucose and impaired postprandial
glucose tolerance it was concluded, that fasting glucose was a marker of beta cell dysfunction and
hepatic glucose production rather than peripheral insulin resistance, and predominantly related to
genetic factors, smoking and male sex (26), which could be a plausible explanation for our findings
as well.
10
Our results showed that an increase in PAEE, in particular for those participants with the lowest
level of PAEE were associated with a lower 2-h insulin concentration, which indicates a doseresponse relation of the volume of PAEE; on average fasting and 2-h insulin levels were 3% and
9% lower for every 10kJ/kg/day difference in PAEE. This difference could be achieved with an
extra hour of gentle walking each day. Moreover, a potential threshold for PAEE was found around
35 kJ/kg/day; however, further examination of a threshold for the association between PAEE and
glucose metabolism is warranted since the cross-sectional observational design of our study does
not allow firm conclusions on causality. Let alone the significance of this threshold seemingly
dividing this sample into insulin sensitive and insulin resistant participants.
This is the first study to report on associations between objectively measured PAEE and glucose
metabolism among Inuit. It has been shown to be particularly important to monitor both HR and
movement in the estimation of PAEE in rural populations due to a higher number of activities that
cannot be fully measured by motion sensing alone. Moreover the use of monitor-based measures of
PA avoids issues of recall bias and is therefore considered a strength in this study (17;29). For
logistic reasons, not all participants were objectively monitored for 48 hours or more but this
subsample still represents the largest study conducted to date in this population. Moreover,
participants included in the subsample differed only very little in age, sex and residence compared
to the entire study population, which implies that the results of this present analysis applies to the
population of Greenland. Measures of glucose tolerance and insulin concentrations in our
population-based sample were based on blood samples, which is a strength compared with selfreported measures of type 2 DM.
The study also has limitations. The cross-sectional design did not allow us to make conclusions
about the direction of associations or any strong inference on causality; however, several
intervention studies have shown the importance of physical activity in the prevention of type 2 DM
11
and obesity. The PAEE estimation is based on individual recordings from 48 hours to 5 full days.
Rennie et al. estimated that 3 days of recording yielded a validity coefficient at 0.85 for the
assessment of energy expenditure in a European sample (30). In our study, only 858 of the
participants had more than 3 days of wear data, but our sensitivity analyses showed similar results
when applying this stricter inclusion criteria. Ideally, more days of objective recording would have
been preferable to capture variations in PA during the week but logistics made this unfeasible.
We found that PAEE was associated with 2-h insulin independently of weight and an inverse dose
response relation. No such association was found for fasting glucose, 2-h glucose or fasting insulin.
Insufficient physical activity may contribute to impaired glucose tolerance through a pathway
including alterations in obesity and fat distribution. Our results suggest that both obesity and low
levels of PAEE may be important contributing risk factors for the increasing prevalence of type 2
DM among Inuit in Greenland. Nevertheless the study also points out the importance of examining
other factors than just those related to current lifestyle, such as genetic or early life factors, which
could play a role in the development of impaired glucose metabolism in this indigenous population.
The results of this study can be incorporated in strategies in order to prevent the rising prevalence of
type 2 DM and obesity among indigenous populations in the Arctic.
Author Contributions: IDP wrote the manuscript; IDP,PB, SB and MEJ designed the study; and
IDP, SB and MEJ were responsible for data management and analysis, IDP, PB, SB and MEJ
helped to draft the manuscript. All authors interpreted the data and contributed to the final
manuscript. IDP is responsible for the contents of the article.
12
Acknowledgement
This study was funded by Karen Elise Jensen´s foundation, Denmark, NunaFonden, Danish
Medical Research Council, and Greenland Medical Research Council. The authors are grateful to
the participants and the participating communities. The authors would also like to thank Kate
Westgate and Stefanie Mayle at the MRC Epidemiology Unit, Cambridge, United Kingdom for
assistance in data processing. The authors declare that there are no conflicts of interest.
13
Reference List
(1) Bjerregaard P, Young TK, Dewailly E, Ebbesson SO. Indigenous health in the Arctic: an overview of
the circumpolar Inuit population. Scand J Public Health 2004;32(5):390-5.
(2) Galloway T, Blackett H, Chatwood S, Jeppesen C, Kandola K, Linton J, et al. Obesity studies in the
circumpolar Inuit: a scoping review. Int J Circumpolar Health 2012;71:18698.
(3) Jørgensen ME, Bjerregaard P, Borch-Johnsen K. Diabetes and impaired glucose tolerance among the
inuit population of greenland. Diabetes Care 2002 Oct;25(10):1766-71.
(4) Omran AR. The epidemiologic transition: a theory of the epidemiology of population change. 1971.
Milbank Q 2005;83(4):731-57.
(5) Ebrahim S, Kinra S, Bowen L, Andersen E, Ben-Shlomo Y, Lyngdoh T, et al. The effect of rural-tourban migration on obesity and diabetes in India: a cross-sectional study. PLoS Med 2010
Apr;7(4):e1000268.
(6) Mohan V, Deepa M, Anjana RM, Lanthorn H, Deepa R. Incidence of diabetes and pre-diabetes in a
selected urban south Indian population (CUPS-19). J Assoc Physicians India 2008 Mar;56:152-7.
(7) Assah FK, Ekelund U, Brage S, Mbanya JC, Wareham NJ. Urbanization, physical activity, and
metabolic health in sub-Saharan Africa. Diabetes Care 2011 Feb;34(2):491-6.
(8) Sobngwi E, Mbanya JC, Unwin NC, Kengne AP, Fezeu L, Minkoulou EM, et al. Physical activity and its
relationship with obesity, hypertension and diabetes in urban and rural Cameroon. Int J Obes Relat
Metab Disord 2002 Jul;26(7):1009-16.
(9) Snodgrass JJ, Leonard WR, Tarskaia LA, Schoeller DA. Total energy expenditure in the Yakut (Sakha)
of Siberia as measured by the doubly labeled water method. Am J Clin Nutr 2006 Oct;84(4):798806.
(10) Dahl-Petersen IK, Jorgensen ME, Bjerregaard P. Physical activity patterns in Greenland: A country in
transition. Scandinavian Journal of Public Health 2011 Nov;39(7):678-86.
(11) Kriska AM, Hanley AJ, Harris SB, Zinman B. Physical activity, physical fitness, and insulin and glucose
concentrations in an isolated Native Canadian population experiencing rapid lifestyle change.
Diabetes Care 2001 Oct;24(10):1787-92.
(12) Jorgensen ME, Borch-Johnsen K, Witte DR, Bjerregaard P. Diabetes in Greenland and its
relationship with urbanization. Diabet Med 2012 Jun;29(6):755-60.
(13) Assah FK, Ekelund U, Brage S, Mbanya JC, Wareham NJ. Free-Living Physical Activity Energy
Expenditure Is Strongly Related to Glucose Intolerance in Cameroonian Adults Independently of
Obesity. Diabetes Care 2009 Feb;32(2):367-9.
(14) Ekelund U, Franks PW, Sharp S, Brage S, Wareham NJ. Increase in physical activity energy
expenditure is associated with reduced metabolic risk independent of change in fatness and fitness.
Diabetes Care 2007 Aug;30(8):2101-6.
14
(15) Healy GN, Wijndaele K, Dunstan DW, Shaw JE, Salmon J, Zimmet PZ, et al. Objectively measured
sedentary time, physical activity, and metabolic risk: the Australian Diabetes, Obesity and Lifestyle
Study (AusDiab). Diabetes Care 2008 Feb;31(2):369-71.
(16) Bjerregaard P. Inuit Health in Transition. Greenland survey 2005-2008. Population sample and
survey methods. 1-13. 2009. National Institute of Public Health (article online)2009. Available from
http://www.si-folkesundhed.dk/upload/metoderapport_endelig.pdf.
(17) Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart
rate and movement sensor Actiheart. Eur J Clin Nutr 2005 Apr;59(4):561-70.
(18) Brage S, Ekelund U, Brage N, Hennings MA, Froberg K, Franks PW, et al. Hierarchy of individual
calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol 2007
Aug;103(2):682-92.
(19) Brage S, Brage N, Franks PW, Ekelund U, Wong MY, Andersen LB, et al. Branched equation
modeling of simultaneous accelerometry and heart rate monitoring improves estimate of directly
measured physical activity energy expenditure. J Appl Physiol 2004 Jan 1;96(1):343-51.
(20) Henry CJK. Basal metabolic rate studies in humans: measurement and development of new
equations. Public Health Nutrition 2005 Oct;8(7A):1133-52.
(21) Kriska AM, Pereira MA, Hanson RL, de Court, Zimmet PZ, Alberti KG, et al. Association of physical
activity and serum insulin concentrations in two populations at high risk for type 2 diabetes but
differing by BMI. Diabetes Care 2001 Jul;24(7):1175-80.
(22) Hanson RL, Pratley RE, Bogardus C, Narayan KM, Roumain JM, Imperatore G, et al. Evaluation of
simple indices of insulin sensitivity and insulin secretion for use in epidemiologic studies. Am J
Epidemiol 2000 Jan 15;151(2):190-8.
(23) Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol 1993 May
1;137(9):959-65.
(24) Kriska AM, LaPorte RE, Pettitt DJ, Charles MA, Nelson RG, Kuller LH, et al. The association of
physical activity with obesity, fat distribution and glucose intolerance in Pima Indians. Diabetologia
1993 Sep;36(9):863-9.
(25) Pereira MA, Kriska AM, Joswiak ML, Dowse GK, Collins VR, Zimmet PZ, et al. Physical inactivity and
glucose intolerance in the multiethnic island of Mauritius. Med Sci Sports Exerc 1995
Dec;27(12):1626-34.
(26) Faerch K, Borch-Johnsen K, Holst JJ, Vaag A. Pathophysiology and aetiology of impaired fasting
glycaemia and impaired glucose tolerance: does it matter for prevention and treatment of type 2
diabetes? Diabetologia 2009 Sep;52(9):1714-23.
(27) Rana JS, Li TY, Manson JE, Hu FB. Adiposity compared with physical inactivity and risk of type 2
diabetes in women. Diabetes Care 2007 Jan;30(1):53-8.
(28) Wareham NJ, Wong MY, Day NE. Glucose intolerance and physical inactivity: the relative
importance of low habitual energy expenditure and cardiorespiratory fitness. Am J Epidemiol 2000
Jul 15;152(2):132-9.
15
(29) Crouter SE, Churilla JR, Bassett DRJ. Accuracy of the Actiheart for the assessment of energy
expenditure in adults. Eur J Clin Nutr 2007 Apr 18.
(30) Rennie KL, Wareham NJ. The validation of physical activity instruments for measuring energy
expenditure: problems and pitfalls. Public Health Nutr 1998 Dec;1(4):265-71.
16
Tables
Table 1 Characteristics of the study population, Inuit in Greenland.
Men
Women
(n=672)
(n=873)
44.9 ±14.1
43.0 ±13.8
Age (years)
Anthropometric measures
BMI (kg/m2)
25.8 ±4.6
26.5 ±5.5
Waist circumference (cm)
92.9±12.9
90.8±13.6
Fat percentage (%)
22.0±7.9
33.0±9.1
Metabolic parameters
Fasting glucose (mmol/l) *
5.8±0.8
5.6±0.7
2-h glucose (mmol/l)*
5.5±2.4
6.1±2.2
Fasting insulin (pmol/l)*
33.0 (22.0;50.0)
41.0 (29.0;59.0)
2-h insulin (pmol/l)*
66.5 (30.0;140.0) 148.0 (79.0;251.0)
IGT n(%)
22 (3.5)
64 (7.8)
IFG n(%)
99 (15.9)
82 (10.0)
Diabetes n(%)
55 (8.8)
53 (6.4)
10 (1.6)
15 (1.8)
Known diabetes n (%)
59 (8.8)
94 (11.0)
Family history with diabetes n (%)
442 (65.9)
602 (69.1)
Smoking n (%)
75.1± 17.7
75.0±18.0
Wear time (min)
59.9±27.6
48.6±21.2
Physical Energy Expenditure (kJ/kg/day)
Values are mean ±SD or median (IQR) unless otherwise noted. * Based on fasting participants and participants not
diagnosed with diabetes.
17
Table 2 Clinical and metabolic characteristics by quartiles of Physical Activity Energy Expenditure (PAEE) among
adult Inuit in Greenland.
Quartiles of PAEE
Q1
Q2
Q3
Q4
P value
28.3 ±0.6 43.7 ±0.5 56.9 ±0.5 85.0 ±0.6
<.001
PAEE (kj/kg/day)
Sex
129(19.2) 127(18.9) 186(27.7) 230(34.2)
<.001
Men n(%)
257(29.4) 259(29.7) 200(22.9) 157(18.0)
Women n(%)
54.5±0.6 45.2 ±0.6 40.7 ±0.6 34.9 ±0.6
<.001
Age (years)
27.5 ±0.3 26.0 ±0.3 26.1 ±0.3 25.0 ± 0.3
<.001
BMI(kg/m2)
95.5 ±0.7 91.6 ±0.7 91.7 ±0.7 88.2 ± 0.7
<.001
WC (cm)
<.001
Fat percentage (%) 30.4 ±0.5 28.1 ±0.4 28.3 ±0.4 26.1 ±0.5
IGT*
8(8.5)
5(4.7)
6(4.0)
3(1.5)
0.04
Men n(%)
23(11.5)
21(9.4)
10(5.8)
10(7.1)
0.2
Women n(%)
IFG*
21(24.4)
16(15.7)
29(20.0)
33(17.0)
0.4
Men n(%)
33(18.6)
30(14.8)
10(6.1)
9(6.9)
0.001
Women n(%)
Diabetes*
16(13.5)
9(7.6)
11(6.7)
9(4.3)
0.02
Men n(%)
21(8.7)
8(3.3)
7(3.8)
2(1.4)
0.004
Women n(%)
Data are age and sex-adjusted means (SE) or n(%) not adjusted. P values for linear trend across quartiles of PAEE
(linear regression) or for differences between groups (CHI2 test). * Analysis based on fasting participants and
participants not diagnosed with diabetes.
18
Table 3. Linear associations of objectively measured PAEE (per 10kJ/kg/day) with BMI, Waist circumference (WC),
fat percentage, fasting glucose (F-glucose), 2-h glucose, fasting insulin (F-insulin) and 2-h insulin adjusted by potential
confounders (Model A–C). Inuit in Greenland.
Model A
Model B
Model C
PAEE (10kJ/kg/day)
BMI (kg/m2)
WC (cm)
Fat percentage
F-glucose (mmol/l)
β
95% CI
P
β
95% CI
P
-0.4
-1.1
-1.4
-0.1
-0.5 to -0.3
-1.4 to -0.9
-1.6 to -1.2
-0.2 to 0.09
-0.7 to -0.4
-3.0 to -0.1
<.001
<.001
<.001
<.001
-0.3
-1.0
-0.6
-0.03
-0.4 to -0.2
-1.3 to -0.7
-0.8 to -0.4
-0.08 to 0.02
<.001
<.001
<.001
0.18
β
95% CI
P
-0.06 to 0.69
0.01
0.04
-0.5
-0.2 -0.3 to -0.02
0.2
2-h glucose mmol/l)
<.001
0.03 -0.1 -0.3 to 0.06
-1.0*
-7.0 to 0.6
0.1
F-insulin (pmol/l)*
0.02 -7.0* -10.0 to -2.0 0.003
3.0*
-24.4 to - <.001
- -19.0 to -5.0 <.001
-16.0 to - 0.01
2-h insulin (pmol/l)*
18.0*
8.0
12.0*
9.0*
2.0
Model A: No adjustments; Model B: adjusted by age and sex. Model C: adjusted by age, sex and Waist Circumference.
PAEE2 included. *Percentage decrease in insulin concentrations derived from back transformation. No interaction was
found for the variables sex and WC. Analysis based on fasting participants and participants not diagnosed with diabetes.
19
Table 4. Logistic regression associations of objectively measured Physical Activity Energy Expenditure (PAEE) (per
10kJ/kg/day) with Diabetes, Impaired Glucose Tolerance (IGT) and Impaired Fasting Glucose (IFG) adjusted by
potential confounders (Model A-C). Inuit in Greenland.
Model A
Model B
Model C
PAEE
(10kJ/kg/day)
OR
95% CI
P
0.826
0.742 to 0.919
<.001
OR
95% CI
P
OR
95% CI
P
0.875 to 0.835 0.998
0.884 to 0.979
1.114
1.128
0.915 0.853 to 0.981 0.013 1.078
0.992 to 0.076 1.098
1.009 to 0.030
IFG
1.171
1.195
0.787 0.703 to 0.881 <.001 0.975
0.859 to 0.701 1.014
0.892 to 0.829
Diabetes
1.107
1.153
Model A: No adjustments; Model B: adjusted by age and sex. Model C: adjusted by age, sex and Waist Circumference.
Analysis based on fasting participants and participants not diagnosed with diabetes.
IGT
Fasting Insulin
160
140
120
100
80
60
40
20
0
2-h insulin
Pmol/l
6,4
mmol/l
Fasting glucose
2-h glucose
6,2
6
5,8
5,6
5,4
5,2
kJ/kg/day
28
0.987
BMI
kg/m2
27
26
25
24
23
kJ/kg/day
kJ/kg/day
98
96
94
92
90
88
86
84
82
CM
WC
kJ/kg/day
Figure legend:
Figure 1.Age- and sex adjusted means of insulin and glucose (fasting glucose, 2-h glucose, fasting
insulin and 2-h insulin) and waist circumference (WC) and BMI across deciles of PAEE. Inuit in
Greenland.
20
ISSN: 1601-7765
ISBN: 978-87-7899-256-7
ISBN: 978-87-7899-257-4

Physical activity pattern and its relation to glucose metabolism in

Transcription

Similar documents

Lecture 1: slides - University of Washington, School of Oceanography

VIEW PDF - American Commercial Real Estate

Lecture 12

cognition insert

The Story of the Norse Vikings in Greenland and Why their

Project progress - Northern Periphery Programme

bloodless glucose monitoring

Greenland Garden Centre

Introducing Modz…

A GLIMPSE of Greenland`s future