Global nutrition transition and the pandemic of obesity in
developing countries
Barry M Popkin, Linda S Adair, and Shu Wen Ng
Decades ago, discussion of an impending global pandemic of obesity was thought
of as heresy. But in the 1970s, diets began to shift towards increased reliance upon
processed foods, increased away-from-home food intake, and increased use of
edible oils and sugar-sweetened beverages. Reductions in physical activity and
increases in sedentary behavior began to be seen as well. The negative effects of
these changes began to be recognized in the early 1990s, primarily in low- and
middle-income populations, but they did not become clearly acknowledged until
diabetes, hypertension, and obesity began to dominate the globe. Now, rapid
increases in the rates of obesity and overweight are widely documented, from urban
and rural areas in the poorest countries of sub-Saharan Africa and South Asia to
populations in countries with higher income levels. Concurrent rapid shifts in diet
and activity are well documented as well. An array of large-scale programmatic and
policy measures are being explored in a few countries; however, few countries are
engaged in serious efforts to prevent the serious dietary challenges being faced.
© 2011 International Life Sciences Institute
INTRODUCTION
Several decades ago, it was heresy to talk about an
impending global pandemic of obesity. However, diets
and activity patterns were clearly changing drastically in
the United States, and by the 1980s it was understood that
dietary quality in the United States was worsening, physi-
cal activity was drastically declining, and obesity was
rising across the United States and Europe. At that point,
only the United States was considered to have an obesity
problem, with more than half of adults in some age-,
gender-, race-, or ethnic-specific subpopulations being
overweight or obese. Home economics was dying as a
taught art in US schools, consumption of processed food
and prepared meals was increasingly common, eating
away from home, particularly in fast food restaurants, was
becoming a major part of many people’s lives, and
concern about a wider array of health conditions related
to obesity was increasing. In reviewing the early publica-
tions on the nutrition transition published in Nutrition
Reviews, several key themes were found
1,2
: urbanization
was a major driving force in global obesity, and over-
weight and obesity were emerging in low- and middle-
income countries. Changes in edible oil production had
created cheap vegetable oils that allowed people in low-
and middle-income countries to increase their energy
consumption at very low levels of income.
1
However, at
that point in history, it was assumed that global hunger
and malnutrition were the dominant concerns in low-
and middle-income countries; it was thus very difficult to
draw attention to the important roles being played by
shifts in diet and physical activity, which were increasing
the threat of obesity in these settings.
Over the past several decades, a dramatic shift has
occurred, in stages, regarding the way the entire global
population eats, drinks, and moves, and these changes
have clashed with human biology to create major changes
in body composition. The primary mismatches between
human biology and modern society that we have identi-
fied in our research are highlighted in Table 1.
3–12
Affiliations: BM Popkin, LS Adair, and SW Ng are with the Department of Nutrition and Carolina Population Center, University of North
Carolina at Chapel Hill, North Carolina, USA.
Correspondence: BM Popkin, Carolina Population Center, University of North Carolina, 123 W. Franklin St., Chapel Hill, NC 27516, USA.
E-mail: popkin@unc.edu, Phone: +1-919-966-1732, Fax: +1-919-966-9159.
Key words: nutrition transition, obesity, obesity-prevention programs, policies
Then and Now
doi:10.1111/j.1753-4887.2011.00456.x
Nutrition Reviews® Vol. 70(1):3–21 3
Downloaded from https://academic.oup.com/nutritionreviews/article-abstract/70/1/3/1829225
by University of North Carolina at Chapel Hill user
on 03 April 2018
In 2008, overweight and obesity were estimated to
afflict nearly 1.5 billion adults worldwide. One estimate,
which a recent analysis of new data shows is an underes-
timate, predicted that in 2030 an estimated 2.16 billion
adults worldwide will be overweight and 1.12 billion will
be obese.
13
Below, trends data are provided from repeated
surveys conducted over the 1990–2010 period in over 40
countries using the same methods; these data suggest that
2 or more billion people are possibly already overweight
or obese today. The implications of these trends for
health, quality of life, productivity, and healthcare costs
are staggering. The burden is greater for much of Asia,
Latin America, the Middle East, and Africa due to differ-
ences in fat patterning and body composition and the
cardiometabolic effects of body mass index (BMI) at
levels far below the standard BMI cutoff of 25 for over-
weight.
14,15
For example, these results are seen in India in
relation to the prevalence of diabetes and impaired
fasting glucose and in China in the prevalence of hyper-
tension and diabetes.
This review documents the changes that have been
and are occurring in global obesity and describes what
some countries are doing in response; a focus is placed on
the potential options being considered by low- and
middle-income countries. This review also provides a
comprehensive examination of the current state of the
science on the diet-related changes occurring in the low-
and middle-income countries of the world.
OBESITY IN LOW- AND MIDDLE-INCOME COUNTRIES
Recent studies have used data from a large number of
countries to estimate current prevalence rates and to
project increases in obesity in all regions of the world.
13,16
However, little detailed information exists on longitudi-
nal trends for low- and middle-income countries, aside
from Brazil, China, India, and Mexico.
17,18
In addition,
none of the recent studies have focused on within-
country trends related to urban-rural or income/wealth
differences. The general impression has been that in
higher-income countries greater obesity rates are often
found in rural areas and among the poor, which is the
reverse of what is seen in lower-income countries.
However, new evidence suggests that these patterns are
changing, and the increasing rate of obesity among the
poor has important implications for the distribution of
health inequalities.
19
In the past three decades, the age-
standardized mean BMI, the most widely used metric for
defining overweight and obesity, has increased by 0.4–0.5
kilograms/meter
2
/year.
13
The major gaps in this literature relate to lack of data
and superficial examinations of patterns and trends
without sufficient attention to the extant literature and the
dynamics of change,rather than simplistic cross-sectional
perspectives. For example,in recent papers, Subramanium
et al.,
16
using just one wave of data and ignoring dynamics,
point out that the rich are far more likely than the poor to
be obese. This is a very different conclusion from that
reached by Jones-Smith et al.,
20,21
who use similar data but
longitudinal analysis. Jones-Smith et al. studied repeated
cross-sectional data from women between the ages of 18
and 49 years in 37 developing countries to assess within-
country trends in overweight/obesity inequalities by
socioeconomic status (SES) between 1989 and 2007
(n = 405,550). Meta-regression was used to examine the
associations between gross domestic product (GDP) and
disproportionate increases in overweight prevalence by
SES with additional testing for modification by country-
level income inequality. In 27 of 37 countries, higher SES
(versus lower) was associated with higher gains in over-
weight prevalence; in the remaining 10 countries, lower
SES (versus higher) was associated with higher gains in
overweight prevalence. GDP was positively related to a
faster increase in overweight prevalence among the lower
wealth groups.Among countries with a higher GDP,lower
income inequality was associated with faster growth in
overweight prevalence among the poor.
Another limitation of current research is the focus on
women of childbearing age and preschoolers. This reflects
the availability of data from multiple countries that have
relied on demographic and health surveys, which focus on
women of childbearing age and their children. A few
studies, particularly some national surveys from Mexico
and Brazil and a few large-scale longitudinal studies
including the China Health and Nutrition Survey, the
Indonesia Family Life Survey, and the Mexico Family Life
Survey, cover all age and gender groups.
19
Using inclusive
data, one sees quite different gender-specific patterns of
change and differentials by SES. According to the limited
research and data available, men with higher SES have
higher rates of overweight and obesity than men with
lower SES.
22
These figures for men contrast with the
trends found among women as highlighted by the two
Jones-Smith et al. papers described above.
20,21
Table 1 Technological clashes with human biology.
Biology Technology
Sweet preferences Cheap caloric sweeteners,
food processing benefits
Thirst and hunger/satiety
mechanisms not linked
Caloric beverage revolution
Fatty food preference Edible oil revolution;
high-yield oilseeds;
cheap removal of oils
Desire to eliminate exertion Technology in all phases of
movement/exertion
Data from references 3–12.
Nutrition Reviews® Vol. 70(1):3–214
Downloaded from https://academic.oup.com/nutritionreviews/article-abstract/70/1/3/1829225
by University of North Carolina at Chapel Hill user
on 03 April 2018
While we know obesity prevalence appears to be
rising across all low- and middle-income countries, it is
not clear what urban-rural difference may exist. This
review brings together some comprehensive data that
were recently published elsewhere.
19,20
Thesedataare
from repeated nationally representative cross-sectional
surveys of 441,916 rural and 364,267 urban (806,183
total) adult women (18–49 years old) from 42 countries
in Asia, the Middle East, Africa (East, West, central, and
southern), and Latin America. The absolute and relative
changes in prevalence of overweight and obesity for
women in these countries and regions are shown in the
supporting information for this article available online as
Table S1. The combined prevalence of overweight and
obesity (overweight = BMI ⱖ 25, obesity = BMI ⱖ 30;
called overweight/obesity hereafter), grew for all 42 coun-
tries at about 0.7 percentage points per year on average.
Using population weights, it is estimated that 19% of rural
women and 37.2% of urban women are overweight or
obese.
Urban versus rural differences and the shifting burden
of obesity toward the poor
Figure 1 and Figure S1 in the supporting information
online summarize the weighted absolute annual change
and relative annual change in the prevalence of
overweight/obesity and obesity only, respectively, among
rural versus urban women by region. On average, urban
women have higher baseline prevalence and larger
increases in prevalence of overweight/obesity compared
to rural women in the 42 countries surveyed (0.8 versus
0.5 percentage points for overweight; 0.4 versus 0.2 per-
centage points for obesity). However, there are regional
differences, with rural women in Latin America, the
Middle East, and North Africa having much higher
increases in prevalence compared to their urban counter-
parts. The relative annual change in weighted prevalence
is higher for rural (3.9%) than urban (2.5%) women. In
other words, women in rural areas are quickly catching
up to their urban counterparts. Figure S1 (available in the
supporting information online) shows the statistics for
obesity prevalence, and the results are consistent. The
higher relative annual rates of change for obesity com-
pared to overweight suggest that obesity prevalence, in
particular, is changing very quickly.
We also looked at the data for each of the 42 coun-
tries in our study ranked by GDP per capita (see Figure S2
in the supporting information online). There appears to
be little association between residence type and preva-
lence of overweight or obesity in countries with higher
GDP. Among lower-GDP countries, urban women are
Figure 1 Absolute and relative annual percentage-point changes in weighted prevalence of overweight and obesity
(BMI ⱖ 25) among women in rural and urban areas of 42 countries by region (N = 42).
Nutrition Reviews® Vol. 70(1):3–21 5
Downloaded from https://academic.oup.com/nutritionreviews/article-abstract/70/1/3/1829225
by University of North Carolina at Chapel Hill user
on 03 April 2018
more likely to be overweight or obese, and countries
around the middle and the bottom of the GDP distribu-
tion range have a higher proportion of urban women who
are overweight or obese compared to rural women. Sta-
tistical analyses show that increased per-capita GDP is
associated with increases in the absolute annual change in
prevalence of overweight and obesity only in rural areas.
KEY FACTORS EXPLAINING INCREASED OBESITY:
POTENTIAL ROLE IN FUTURE PROGRAMMATIC AND
POLICY MEASURES
Ultimately, obesity reflects energy imbalance, so the
major areas for intervention relate to dietary intake and
energy expenditure, for which the main modifiable com-
ponent is physical activity. It is clear that large shifts in
access to technology have reduced work-related energy
expenditure in the more labor-intensive occupations,
such as farming and mining, as well as in the less energy-
intensive sectors such as service and manufacturing.
23
Changes in transportation,
24
leisure, and home produc-
tion (cooking, cleaning, child care, etc.)
25
have also led to
reduced physical activity. In addition, the complex inter-
play between biological factors operating during fetal and
infant development and the related energy imbalances
exacerbate many health problems.
26
Such changes have
been well documented for China and are also found in
varying manifestations in many countries.
Finding ways to increase physical activity across all
age groups is important for public health, but options for
increasing energy expenditure through physical activity
may be limited in low- and middle-income countries. For
instance, to offset any increase of about 110 kcal of food
or beverage in average daily energy intake, a woman
weighing 54 kg must walk moderately fast for 30 min and
a man weighing 82 kg for about 25 min. Such levels of
physical activity may be too much to expect, so dietary
modification is a key approach to lowering obesity preva-
lence, particularly with the ongoing decline in physical
activity and increase in sedentary time (unpublished
data). The dietary dynamics represent a major set of
complex issues. On the global level, new access to tech-
nologies (e.g., cheap edible oils, foods with excessive
“empty calories,” modern supermarkets, and food distri-
bution and marketing) and regulatory environments
(e.g., the World Trade Organization and freer flow of
goods, services, and technologies) are changing diets in
low- and middle-income countries. Accompanying this
are all the critical issues of food security and global access
to food sufficient to meet adequate levels of nutrient
intake. Many populations focus on basic grain and
legume food supplies, while the overall transition has
shifted the structure of prices and food availability and
created a nutrition transition linked with obesity as well
as hunger. We have used detailed data on time use along
with data on energy expenditures and other factors to
examine past patterns and trends in order to predict, until
2020 and 2030, patterns of physical activity and sedentary
behavior in the United States,the United Kingdom, Brazil,
China, and India.
27
Prior to exploring the dietary dimension, we con-
sider an important biological factor affecting obesity and
chronic diseases in rapidly developing countries in Asia
and Africa. This factor is the biological insults suffered
during fetal and infant development that may influence
susceptibility to the changes described above, thus influ-
encing trends in the development and severity of chronic
disease in these countries.
Developmental origins of health and disease: Special
concerns for low- and middle-income countries
The patterns of change in dietary intake and energy
expenditure related to the global nutrition transition are
particularly important in the context of current theories
of the developmental origins of adult disease. Based on
three decades of research, it is now recognized that sus-
ceptibility to obesity and chronic diseases is influenced by
environmental exposures from the time of conception to
adulthood. An extensive body of literature demonstrates
that fetal nutritional insufficiency triggers a set of ana-
tomical, hormonal, and physiological changes that
enhance survival in a “resource-poor” environment.
28
However, in a postnatal environment with plentiful
resources, these developmental adaptations may contrib-
ute to the development of disease. Some of the strongest
evidence on the long-term effects of moderate to severe
nutrition restriction during pregnancy comes from
follow-up of infants born after maternal exposure to
famine conditions, such as those experienced in parts of
Europe during World War II. For example, AC Ravelli
et al.
29
found higher rates of obesity in 50-year-old men
and women whose mothers were exposed to the Dutch
famine in the first half of their pregnancies, and GP
Ravelli et al.
30
found obesity among 19-year-old men
whose mothers experienced famine during their preg-
nancies. Similarly, a follow-up of Hmong refugee immi-
grants revealed higher rates of central obesity among
those raised in a war zone, with the effects amplified
in those who migrated to the United States compared to
those living in a traditional rural setting.
31
The developmental origins theory of mismatch fits
closely with the broader issues of mismatch discussed
below, which emerged in our early research
1,2
and later
work.
32–34
The mismatch theory of early nutritional
deficits followed by later excesses
35
may be particularly
important in low- and middle-income countries under-
going rapid social and economic changes, because
Nutrition Reviews® Vol. 70(1):3–216
Downloaded from https://academic.oup.com/nutritionreviews/article-abstract/70/1/3/1829225
by University of North Carolina at Chapel Hill user
on 03 April 2018
economic progress amplifies mismatch. Much of the lit-
erature on the developmental origins of health and
disease focuses on chronic diseases. However, given the
strong association between chronic diseases and obesity,
particularly central obesity, this evidence is highly rel-
evant and provides a strong rationale for obesity preven-
tion efforts in populations that have experienced
dramatic changes in the nutritional environment as a
consequence of nutrition transition.
Mechanisms may include effects on the number of
nephrons in the kidney,
36
glucocorticoid exposure subse-
quent to maternal stress, or poor nutritional status poten-
tially programming the insulin and hypothalamic-
pituitary axes for high levels of metabolic efficiency
28
and
epigenetic changes. Maternal stress and specific aspects of
diet (for example, intake of folate and other methyl
donors) can affect DNA methylation and gene
expression.
37–39
Ongoing studies in places such as India
are examining the role of maternal micronutrient intake
on epigenetic changes that affect child adiposity.
39,40
Research in India has provided other important insights.
For example, Indian infants with poorly nourished
mothers are born with weight deficits, but in relative
terms the deficits in lean mass are greater than those in
adiposity. In later life, when modern high-energy and
high-fat diets are consumed, the previously “thin-fat”
babies also have greater central adiposity.
41,42
It is apparent from studies of the developmental
origins of disease that there is a strong intergenerational
component to health. While much of the literature on the
early origins of obesity and associated risk has focused on
undernutrition, there is also substantial evidence that
maternal overweight and obesity in pregnancy influence
disease risk among offspring. For example, gestational
diabetes is related to offspring body composition and
increased risk of insulin resistance and diabetes in off-
spring.
43,44
Thus, there is concern about an intergenera-
tional amplification of diabetes risk. Women who were
malnourished as children are at increased risk of being
centrally obese and having impaired glucose tolerance as
adults.If these conditions affect a woman’s pregnancy, her
offspring will be at increased risk of early development of
obesity and diabetes. As obesity develops at younger and
younger ages, the likelihood that adolescents and young
women who become pregnant will experience complica-
tions associated with gestational diabetes and hyperten-
sion will increase dramatically. There is growing evidence
that maternal obesity,even without gestational diabetes, is
a risk factor for child obesity through a pathway related to
fetal overnutrition (see the review by Fall
45
).
On the other end of the nutrition spectrum, short
maternal stature acts as a physical constraint on
fetal growth,
46,47
and stunting in offspring may, in turn,
relate to increased obesity risk.
48
Beyond the fetal period,nutrition and other contribu-
tors to health in infancy, childhood, and adolescence are
important determinants of adult body composition and
obesity risk. In light of the large increases in overweight
and obesity in children as well as adults, attempts have
been made to determine the ages at which faster weight
gain relates to later obesity. A large body of literature
relates “rapid growth” during infancy to risk of obesity in
later childhood and into adulthood.
49
In addition, rapid
weight gain, particularly from mid-childhood onwards, is
related to increased risk of elevated blood pressure or
impaired fasting glucose in young adulthood in low- and
middle-income countries.
50
Concerns have been raised
about the promotion of rapid weight gain in children who
are malnourished. In low-income countries catch-up or
compensatory growth following a period of faltering
growth is desirable, because it is associated with reduced
morbidity and improved survival
51–53
as well as better cog-
nitive development.
54
A key concern, however, is whether
the benefits of faster growth in these settings outweigh the
possible long-term risks.Based on the COHORTS analysis
of children from five low- and middle-income countries,
faster weight gain in the first 2 years of life has a number of
benefits.It is associated with the development of lean body
mass but not with increased risk of impaired fasting
glucose or diabetes in young adulthood (data submitted
for publication: e.g., Kuzawa et al.
55
). In light of observa-
tions that patterns of child growth have important conse-
quences for the development of obesity and chronic
diseases, another line of research focuses on factors that
contribute to or protect against early development of adi-
posity. In this regard, the potential programming roles of
early diet have been explored,including the roles of breast-
feeding and high intakes of dietary protein, fat, and
sodium.These topics are important in light of the dramatic
changes in diet composition that characterize many popu-
lations in the developing world.
Of course, early feeding issues are important. Some
studies show a protective effect of breastfeeding on later
development of obesity and chronic diseases,
56,57
while
other studies show no effects.
58
Similarly, consistently
high protein intake during complementary feeding in the
first 2 years of life has been associated with a higher mean
BMI and body fat percentage at 7 years of age in cohort
studies of German children,
59
and other researchers have
suggested a strong link between high protein intake and
obesity.
60
Dietary fat (both the amount and the composition)
may also play a role in the development of non-
communicable diseases. The STRIP study in Finland
demonstrated that lower total and saturated dietary fat
intake in infancy results in lower serum cholesterol,
LDL cholesterol, and triglyeride levels (as well as
lower blood pressure) in children up to 14 years of age,
Nutrition Reviews® Vol. 70(1):3–21 7
Downloaded from https://academic.oup.com/nutritionreviews/article-abstract/70/1/3/1829225
by University of North Carolina at Chapel Hill user
on 03 April 2018
