Context One key target of the United Nations Millennium Development goals is
to reduce the prevalence of underweight among children younger than 5 years
by half between 1990 and 2015.
Objective To estimate trends in childhood underweight by geographic regions of
the world.
Design, Setting, and Participants Time series study of prevalence of underweight, defined as weight 2
SDs below the mean weight for age of the National Center for Health Statistics
and World Health Organization (WHO) reference population. National prevalence
rates derived from the WHO Global Database on Child Growth and Malnutrition,
which includes data on approximately 31 million children younger than 5 years
who participated in 419 national nutritional surveys in 139 countries from
1965 through 2002.
Main Outcome Measures Linear mixed-effects modeling was used to estimate prevalence rates
and numbers of underweight children by region in 1990 and 2015 and to calculate
the changes (ie, increase or decrease) to these values between 1990 and 2015.
Results Worldwide, underweight prevalence was projected to decline from 26.5%
in 1990 to 17.6% in 2015, a change of –34% (95% confidence interval
[CI], –43% to –23%). In developed countries, the prevalence was
estimated to decrease from 1.6% to 0.9%, a change of –41% (95% CI, –92%
to 343%). In developing regions, the prevalence was forecasted to decline
from 30.2% to 19.3%, a change of –36% (95% CI, –45% to –26%).
In Africa, the prevalence of underweight was forecasted to increase from 24.0%
to 26.8%, a change of 12% (95% CI, 8%-16%). In Asia, the prevalence was estimated
to decrease from 35.1% to 18.5%, a change of –47% (95% CI, –58%
to –34%). Worldwide, the number of underweight children was projected
to decline from 163.8 million in 1990 to 113.4 million in 2015, a change of
−31% (95% CI, −40% to −20%). Numbers are projected to decrease
in all subregions except the subregions of sub-Saharan, Eastern, Middle, and
Western Africa, which are expected to experience substantial increases in
the number of underweight children.
Conclusions An overall improvement in the global situation is anticipated; however,
neither the world as a whole, nor the developing regions, are expected to
achieve the Millennium Development goals. This is largely due to the deteriorating
situation in Africa where all subregions, except Northern Africa, are expected
to fail to meet the goal.
At the Millennium Summit in 2000, representatives from 189 countries
committed themselves toward a world in which sustaining development and eliminating
poverty would have the highest priority.1
The increased recognition of the relevance of nutrition as a basic pillar
for social and economic development placed childhood undernutrition among
the targets of the first Millennium Development goal to "eradicate extreme
poverty and hunger."2 The specific target goal
is to reduce by 50% the prevalence of being underweight among children younger
than 5 years between 1990 and 2015. Childhood underweight is internationally
recognized as an important public health problem and its devastating effects
on human performance, health, and survival are well established.3-8 A
recent study estimated that about 53% of all deaths in young children are
attributable to underweight, varying from 45% for deaths due to measles to
61% for deaths due to diarrhea.8
Monitoring progress toward the goal requires reliable data sources based
on agreed on international standards and best practices, and standardized
data collection systems that enable comparison over time. The World Health
Organization (WHO) Global Database on Child Growth and Malnutrition was established
in the late 1980s with the objective to collect, standardize, and disseminate
anthropometric data on children using a standard format.9 With
419 national population-based surveys, the database covers more than 90% of
children younger than 5 years worldwide. An earlier analysis of trends from
1980 to 2005 based on 241 national surveys indicated that childhood malnutrition
(as measured by stunting or low height for age) remains a public health problem
worldwide, with stunting rates declining in the majority of countries at about
1% per year or less. Moreover, in some countries, rates of stunting were rising
while in many others they remained high.10 The
aims of the current analysis are to (1) quantify the magnitude of the problem
and describe where malnourished children live and (2) to identify geographical
regions that based on projected estimates are unlikely to achieve the goal
of a 50% decrease in the 1990 prevalence of underweight by 2015.
To estimate trends in childhood underweight, national prevalence of
low weight for age, which was defined as weight 2 SDs below the mean weight
for age of the National Center for Health Statistics and WHO reference population,
were derived from the WHO Global Database on Child Growth and Malnutrition.9 A total of 419 nationally representative surveys,
which included about 31 million children, were available from 139 countries.
Of the 419 surveys, 398 were conducted in developing countries and 21 in developed
countries. For 42 countries, national survey data were available from only
1 survey, 36 countries had 2 surveys, and 61 countries had 3 or more surveys.
More than half of the surveys (227) were conducted between 1991 and 1999,
33% (137) dated back to 1990 and earlier, and 13% (55) were performed in 2000
or later (Table 1). The earliest
survey dates back to 1965 (from Colombia), while the most recent surveys were
conducted in 2002 (Eritrea, Jordan, and Romania). All surveys included boys
and girls, and the age groups ranged from birth to 5 years. The complete set
of surveys included in the analysis is available from the authors on request.
A data file was constructed and consisted of region, subregion, country,
survey year, sample size, prevalence of underweight, and population of children
younger than 5 years during the survey year. To obtain comparable prevalences
of underweight children across countries, surveys were analyzed following
a standard format using the National Center for Health Statistics and WHO
international reference population, the same cut-off point (ie, 2 SDs below
the mean weight for age), and the same reporting system (ie, z score). The steps followed to analyze the surveys in a standard way
have been described elsewhere.9
For this analysis, all available childhood underweight prevalence estimates
from national surveys were used. Countries providing data were regarded as
a representative sample of all countries within their subregions (ie, covering
at least 80% of the population younger than 5 years), which were nested within
regions. For example, the region of Asia comprised Eastern, South Central,
Southeastern, and Western Asia. Trends were also derived for larger units.
The group of developing regions included Africa, Asia, Latin America (including
the Caribbean), and Oceania. The developed countries were grouped in 1 unit.
We obtained global estimates by combining the developing regions and developed
countries. Country groupings followed the United Nations (UN) country classification.11
The method of linear mixed-effect models, as described by Laird and
Ware,12 was applied to model the data set at
subregional levels with the country's effect being defined as random. By using
this model, the fact that not all countries had available underweight prevalence
data was incorporated assuming countries without data were missing at random.
This also allowed all data points for each country to be included in the estimates
for subregions.
Using the surveys' underweight prevalence estimates, a linear mixed-effect
model was considered for each group of subregions belonging to the same region.
The dependent variable was the logit of the prevalence (ln [prevalence/{1−prevalence}]).
The basic model contained subregion, year of survey, and
the interaction between year and the subregion as fixed effects and country
as a random effect. This model is part of a more general class of models—the
multilevel models. In multilevel modelling literature, notably in Goldstein,13 the same model is called a 2-level model, counting
the levels of variation (ie, level 1 being the survey and level 2 being the
country). Consequently, we obtained from each model an estimate of the change
in prevalence for every subregion between the years 1990 and 2015. Subregion
prevalence estimates and their respective confidence intervals (CIs) were
derived by back-transformation.
To account for the population differences among countries and to ensure
that the influence on the regional trend analysis of a country's survey estimate
was proportional to the country's population, we performed weighted analysis.
The population weights were derived from the UN Population Prospects.11 For each data point, we obtained the respective population
estimate of children younger than 5 years for the specific survey year. If
a survey was performed over an extended period, for example 1995-1997, the
mean year (ie, 1996) was used as the year from which to choose the respective
population estimate. For countries with multiple data points, the weights
were calculated by dividing the mean of the country's population of children
younger than 5 years (over the observed years) by the sum of the population
means for countries in the entire region. Weights of countries with single
data points were derived by dividing the population of children younger than
5 years at the time of the survey by the sum of the countries' population
means in the whole region.
We considered 3 different structures for modeling the covariance: compound
symmetric, unstructured, and autoregressive.14(pp231-273) The compound symmetric model for a region allowed the country
to have its own intercept (influencing prevalence estimate) and forced all
countries to have a common slope in prevalence over time. The unstructured
model for a region allowed each country to have its own intercept and slope.
The autoregressive model allowed for correlations between observations within
the same country to weaken as the time between them increased. Year was centered
at 1995, around which there was a high concentration of available survey data
points. Models were fitted using restricted maximum likelihood.14(pp43-47) We used the sandwich variance estimator14(pp61-62) to obtain SEs of estimates. The decision on how to choose
the best model among different covariance structures was based on the Akaike
information criterion,14(pp74-76) which
penalizes the log likelihood values for the number of parameters in the model.
In parallel, we examined the graphed display of the fitted trend line against
the survey data points and discarded models that did not present a reasonable
fit with respect to the empirical data.
The prevalence estimates for 1990 and 2015 for each subregion were derived
using the final regression model chosen according to the criteria described
above. For the subregions of Africa and Latin America, the unstructured covariance
model with random intercept and slope at country level was used. The compound
symmetric covariance structure was chosen to model the subregions of Asia,
imposing a common slope for all countries in that region. The estimates on
the logit scale were back-transformed to provide the prevalence estimates.
Using the resulting prevalence estimates, the total number of affected
children was calculated by multiplying the prevalence estimates and respective
CIs by the subregional population of children younger than 5 years.11 The UN population estimates and projections are derived
incorporating all new and relevant information regarding the past demographic
dynamics of the population of each country or area of the world; and formulating
detailed assumptions about the future paths of fertility, mortality, and migration.11
For the regional level, the prevalence was derived by using the sum
of the numbers of affected estimates in the subregions divided by the total
population of children younger than 5 years in that region. To construct a
CI for the overall regional prevalence estimate, we obtained approximate SEs
associated with the subregions' prevalence estimates using the delta method15(pp56-58) and summarized them to compute
SEs corresponding to the overall weighted prevalence estimates.
For the developed countries, 21 observations were available. Considering
the relative homogeneity of the group, we fitted a linear mixed-effects model
with year being a fixed effect and country being a random effect.
The relative change in prevalence was calculated as the 2015 prevalence
minus the 1990 prevalence divided by the 1990 prevalence. The CIs for the
change in prevalence were constructed using the ratio between the 2015 and
the 1990 prevalences on the log scale.
Trends and Percentage Change in Prevalence of Underweight Children,
1990-2015
Table 1 presents the surveys
included in the analysis and Table 2 presents
estimates of underweight children for 1990 and projections for 2015 with 95%
CIs of the prevalence and the relative percentage change. Worldwide, prevalence
of childhood underweight was projected to decline from 26.5% in 1990 to 17.6%
in 2015, a change of –34% (95% CI, –43% to –23%). In developed
countries, the prevalence was estimated to decrease from 1.6% to 0.9%, a change
of –41% (95% CI, –92% to 343%). In developing regions, the prevalence
was forecasted to decline from 30.2% to 19.3%, a change of –36% (95%
CI, –45% to –26%).
In Africa, the prevalence of underweight was forecasted to increase
from 24.0% in 1990 to 26.8% in 2015, a change of 12% (95% CI, 8%-16%). The
prevalence of childhood underweight was estimated to increase in sub-Saharan
Africa by 9% (from 26.8% to 29.2%) and in Eastern Africa by 25% (from 26.7%
to 33.3%). The prevalence of childhood underweight was projected to be reduced
by 15% for Middle Africa; 5%, Southern Africa; and 6%, Western Africa. Only
Northern Africa, with a forecasted reduction in the prevalence of childhood
underweight from 9.5% to 4.2%, was estimated to reach the Millennium Development
goal. Figure 1 presents the 2015
projections of the prevalence of underweight children for the African subregions
compared with the Millennium Development goal for those subregions.
In Asia, between 1990 and 2015 the prevalence was estimated to decrease
from 35.1% to 18.5%, a change of –47% (95% CI, –58% to –34%).
The largest decline was estimated in Eastern Asia, where the prevalence of
underweight children was forecasted to decrease by 84% in the same period.
Southeastern and South Central Asia were also forecasted to experience substantial
improvements, with reductions in the prevalence of underweight of 49% and
42%, respectively. However, both subregions are projected to still have high
levels of childhood underweight in 2015. Western Asia was estimated to be
the Asian subregion with the lowest reduction in the prevalence of childhood
underweight (29%).
In Latin America, the prevalence of underweight children was forecasted
to decrease from 8.7% in 1990 to 3.4% in 2015, a change of –61% (95%
CI, –77% to –35%). All subregions in Latin America were estimated
to experience decreasing trends with changes of −72% for the Caribbean,
−54% for Central America, and −65% for South America. There were
no sufficient data to derive estimates for the region of Oceania.
Trends and Percentage Change in Numbers of Underweight Children, 1990-2015
Table 3 presents estimates
for 1990 and projections for 2015 of the number of underweight children, and
the relative percentage change with corresponding 95% CIs by geographical
region. Worldwide, the number of underweight children was projected to decline
from 163.8 million in 1990 to 113.4 million in 2015, a change of –31%
(95% CI, –40% to –20%). In developed countries, the number was
estimated to decline from 1.2 to 0.6 million, a change of –54% (95%
CI, –94% to 244). In developing regions, the number was forecasted to
decline from 162.6 to 112.8 million, a change of –31% (95% CI, –40%
to –20%).
In Africa, the number of underweight children was forecasted to increase
from 25.8 million in 1990 to 43.3 million in 2015, a change of 68% (95% CI,
63%-74%). The subregions of sub-Saharan, Eastern, Middle, and Western Africa
were all forecasted to experience substantial increases in the number of underweight
children (77%, 102%, 72%, and 54%, respectively). Only Southern and Northern
Africa were estimated to reduce the number of underweight children by 14%
and 59%, respectively.
In Asia, where the largest number of underweight children live, the
number was estimated to decrease from 131.9 to 67.6 million between 1990 and
2015, a change of –49% (95% CI, –59% to –36%). The largest
decline was projected for Eastern Asia (−87%), followed by Southeastern
Asia (−52%) and South Central Asia (−39%). Western Asia was estimated
to remain stagnant.
In Latin America, the number of underweight children was forecasted
to decline from 4.8 million in 1990 to 1.9 million in 2015, a change of –60%
(95% CI, –76% to –34%). All subregions in Latin America were forecasted
to experience decreasing trends in childhood underweight (Caribbean, −74%;
Central America, −52%; and South America, −64%).
Geographical Patterns of Underweight Children
Figure 2 shows the geographical
distribution of the prevalence of underweight children based on the latest
survey data. Prevalences were categorized as less than 10%, 10% to 19%, 20%
to 29%, and 30% or more.16 Very high levels
of childhood underweight were found in 12 African countries (Angola, Burkina
Faso, Burundi, Democratic Republic of the Congo, Eritrea, Ethiopia, Madagascar,
Mali, Mauritania, Niger, Nigeria, and Sudan) and 13 Asian countries (Afghanistan,
Bangladesh, Cambodia, India, Lao People's Democratic Republic, Maldives, Myanmar,
Nepal, Pakistan, Philippines, Sri Lanka, Vietnam, and Yemen). Prevalences
of childhood underweight for each country were based on available information
from the WHO Global Database on Child Growth and Malnutrition.17 Prevalence
data by urban and rural residence, age groups, sex, and subnational level
administrative regions can be found at the database's Web site (http://www.who.int/nutgrowthdb).
Based on national surveys of past prevalences of underweight children,
we have estimated prevalence rates and numbers of underweight children by
geographic region for the years 1990 and 2015 to assess the likelihood of
meeting the Millennium Development Goal of reducing 1990-level prevalence
by 50% in the year 2015. According to our analysis, despite an overall improvement
on the global situation, neither the world as a whole nor the developing regions
are expected to achieve the goal. This is largely due to the deteriorating
situation in Africa where all subregions, with the exception of Northern Africa,
are expected to fail to meet the goal. Moreover, sub-Saharan and Eastern Africa
are forecasted to experience an increase in the prevalence of underweight
children during the 25-year period. In Asia, Eastern Asia (mainly driven by
China), and Southeastern Asia are forecasted to reach the goal, while South
Central and Western Asia are not. Moreover, our estimates project that in
2015, most subregions in Africa and South Central Asia will continue to have
very high prevalences of underweight children. According to our projections,
all subregions in Latin America will achieve the Millennium goal.
The vast majority of underweight children live in developing regions,
mainly in Asia and Africa. The projected trends in the prevalence of underweight
children combined with the different population growth these regions are experiencing
(increasing in Africa, decreasing in Asia)11 will
narrow the gap between their respective contributions to the total number
of underweight children. While in 1990, of 100 underweight children, 80 were
estimated to live in Asia and 16 in Africa; in 2015, these numbers are expected
to change to 60 and 38, respectively, if recent trends continue.
Our study has a number of limitations. First, the availability of trend
data is limited for a number of countries and some have not yet conducted
national surveys. Second, surveys were not done randomly. Depending on where
and when surveys were conducted, this may have biased our estimates of past
and future prevalences. Third, although the surveys included in the WHO database
undergo data quality control that results in the exclusion of surveys with
obvious flawed data,9 there are variations
in data quality between the different surveys included in the analysis. Fourth,
when estimating prevalence trends we did not account for uncertainty in each
survey's prevalence estimate, that is, the estimate of the variance of each
prevalence was not included in the regression analysis. As a result, our CIs
are likely to be too narrow. Similarly, for constructing the CIs for the number
of underweight children, the uncertainty around the population estimates was
not considered. Lastly, the precision of the estimates of the prevalence and
numbers of underweight children, as expressed by the 95% CIs, varies depending
on the availability of data for each region. The developed countries and the
subregion of Western Asia present large CIs for the 2015 projections, which
result in wide CIs for the relative percentage change between 1990 and 2015.
Despite these limitations and the inherent speculative nature of extrapolations
to 2015, the present estimates provide a useful base for monitoring progress
toward the achievement of the goal.
The deteriorating situation in Africa is likely to be partly due to
the effect of the human immunodeficiency virus and AIDS epidemic, together
with the political and social instability experienced in many African countries.
The disease has both a direct and indirect effect: infected children are more
likely to be underweight, but also AIDS orphans or children of parents affected
by AIDS are at increased risk of becoming malnourished. In sub-Saharan Africa,
an estimated 333 000 children younger than 5 years died in 1999 with
human immunodeficiency virus infection18 and
11 million are estimated to be orphaned because of AIDS.19 The
predictions of childhood underweight made for 2015 might be underestimates
if the human immunodeficiency virus and AIDS epidemic worsens in Africa.
Economic progress is a major determinant for improving childhood nutrition.
Higher purchasing power enables improved dietary intake in terms of both quality
and quantity. Countries in Asia, such as China and Vietnam, have experienced
rapidly growing economies and changes in lifestyles.20,21 Consequently,
childhood nutritional status has been improving in these countries. The improvement
in some areas has even moved beyond what is desirable and has lead to an increase
in childhood overweight.20,22 An
analysis of childhood overweight in developing countries reported that 16
of the 38 countries with more than 1 national survey showed a rising trend
in overweight.23
The Millennium Development goals are intended to focus attention on
the most critical problems and to maintain that focus by monitoring progress
toward the achievement of specific goals. For childhood underweight, there
are many countries for which national data are still not available. For this
group, surveys would need to be conducted to have a baseline against which
to assess progress. Moreover, because subnational differences in the prevalence
of underweight children can be substantial, it is recommended to map underweight
hot spots for better targeting of interventions aimed at preventing and treating
childhood undernutrition. These interventions are particularly important during
the period from birth to age 3 years—the critical time in which growth
failure and malnutrition occur.24
1. UN General Assembly. UN Res A/55/2 (2000).
3.Chang SM, Walker SP, Grantham-McGregor S, Powell CA. Early childhood stunting and later behaviour and school achievement.
J Child Psychol Psychiatry.2002;43:775-783.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12236612&dopt=Abstract
Google Scholar 4.Martorell R, Rivera J, Kaplowitz H, Pollitt E. Long-term consequences of growth retardation during early childhood. In: Hernandez M, Argente J, eds. Human Growth:
Basic and Clinical Aspects. Amsterdam, the Netherlands: Elsevier; 1992:143-149.
5.Walker SP, Grantham-McGregor SM, Powell CA, Chang SM. Effects of growth restriction in early childhood on growth, IQ, and
cognition at age 11 to 12 years and the benefits of nutritional supplementation
and psychosocial stimulation.
J Pediatr.2000;137:36-41.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10891819&dopt=Abstract
Google Scholar 6.Pelletier DL, Frongillo EA. Changes in child survival are strongly associated with changes in malnutrition
in developing countries.
J Nutr.2003;133:107-119.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12514277&dopt=Abstract
Google Scholar 7.Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S, Murray CJL.and the Comparative Risk Assessment Collaborating Group. Selected major risk factors and global and regional burden of disease.
Lancet.2002;360:1347-1360.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12423980&dopt=Abstract
Google Scholar 8.Caulfield LE, de Onis M, Blössner M, Black RE. Undernutrition as an underlying cause of child deaths associated with
diarrhea, pneumonia, malaria and measles.
Am J Clin Nutr.In press.Google Scholar 9.de Onis M, Blössner M. The World Health Organization Global Database on Child Growth and Malnutrition:
methodology and applications.
Int J Epidemiol.2003;32:518-526.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12913022&dopt=Abstract
Google Scholar 10.de Onis M, Frongillo EA, Blössner M. Is malnutrition declining? an analysis of changes in levels of child
malnutrition since 1980.
Bull World Health Organ.2000;78:1222-1233.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11100617&dopt=Abstract
Google Scholar 11.Department of Economic and Social Affairs, UN Population Division. World Population Prospects: the Sex and Age Distribution
of the World Population, the 2000 Revision. New York, NY: United Nations; 2001.
12.Laird NM, Ware JH. Random-effects models for longitudinal data.
Biometrics.1982;38:963-974.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7168798&dopt=Abstract
Google Scholar 13.Goldstein H. Multilevel Statistical Models. 2nd ed. London, England: Arnold; 1995.
14.Verbeke G, Molenberghs G. Linear Mixed Models for Longitudinal Data. New York, NY: Springer-Verlag; 2000.
15.Agresti A. Categorical Data Analysis. New York, NY: John Wiley and Sons; 1990.
16.WHO Expert Committee Report. Physical Status: the Use and Interpretation of Anthropometry. Geneva, Switzerland: World Health Organization; 1995. Technical report
series 854.
18.Walker N, Schwartländer B, Bryce J. Meeting international goals in child survival and HIV/AIDS.
Lancet.2002;360:284-288.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12147371&dopt=Abstract
Google Scholar 19.UNICEF. The State of the World's Children 2004. New York, NY: UNICEF; 2004.
20.Luo J, Hu FB. Time trends of obesity in pre-school children in China from 1989 to
1997.
Int J Obes Relat Metab Disord.2002;26:553-558.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12075583&dopt=Abstract
Google Scholar 21.Hop le T. Programs to improve production and consumption of animal source foods
and malnutrition in Vietnam.
J Nutr.2003;133(11 suppl 2):4006S-4009S.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=14672303&dopt=Abstract
Google Scholar 22.Sakamoto N, Wansorn S, Tontisirin K, Marui E. A social epidemiologic study of obesity among preschool children in
Thailand.
Int J Obes Relat Metab Disord.2001;25:389-394.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11319637&dopt=Abstract
Google Scholar 23.de Onis M, Blössner M. Prevalence and trends of overweight among preschool children in developing
countries.
Am J Clin Nutr.2000;72:1032-1039.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11010948&dopt=Abstract
Google Scholar 24.Shrimpton R, Victora CG, de Onis M, Lima RC, Blössner M, Clugston G. Worldwide timing of growth faltering: implications for nutritional
interventions.
Pediatrics.2001;107:E75.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11331725&dopt=Abstract
Google Scholar