Cover Image
close this book4th Report on the World Nutrition Situation - Nutrition throughout the Life Cycle (ACC/SCN, 2000, 138 p.)
close this folderCHAPTER 1: NUTRITION THROUGHOUT THE LIFE CYCLE
View the document(introduction...)
View the document1.1 Intrauterine Growth Retardation (IUGR)
View the document1.2 Undernutrition in Preschool Children
View the document1.3 The Growth of School-Age Children
View the document1.4 Adolescent Nutrition
View the document1.5 Adult Malnutrition
View the document1.6 Nutrition of Older People in Developing Countries
View the documentSummary

1.2 Undernutrition in Preschool Children

This section describes the estimated prevalence and number of preschool (under five years old) children suffering from stunting, underweight, and wasting at global, regional, and sub-regional levels.

Defining Indices and Indicators

· Stunting. The anthropometric index height-for-age reflects linear growth achieved pre - and postnatally with its deficits indicating long-term, cumulative effects of inadequate nutrition and/or health. Shortness in height refers to low height-for-age that may reflect either normal variation in growth or a deficit in growth. Stunting refers to shortness that is a deficit or linear growth that has failed to reach generic potential as a result of poor diet and disease. Stunting is defined as low height-for-age at < -2 standard deviations (SD) of the median value of the National Center for Health Statistics/World Health Organization (NCHS/WHO) international growth reference.2 Severe stunting is defined as < -3 SD.

· Underweight. The anthropometric index weight-for-age represents body mass relative to age. Weight-for-age is influenced by the height and weight of a child and is thus a composite of stunting and wasting, making interpretation of this indicator difficult. In the absence of wasting, both weight-for-age and height-for-age reflect the long-term nutrition and health experience of the individual or population. Underweight refers to a deficit and is defined as low weight-for-age at < -2 SD of the median value of the NCHS/WHO international reference.2

· Wasting describes a recent and severe process that has produced a substantial weight loss, usually as a consequence of acute shortage of food and/or severe disease. Chronic dietary deficit or disease can also lead to wasting. The anthropometric index weight-for-height reflects body weight relative to height. Wasting refers to low weight-for-height at < -2 SD of the median value of the NCHS/WHO international weight-for-height reference. Severe wasting is defined as < -3 SD. The statistically expected prevalence of wasting (as with underweight and stunting) is between 2 and 3%, given the normal distribution of wasting rates.2 This indicator is used extensively in emergency settings.

BOX 1.1

The Foetal Origins of Disease

Genes provide a general recipe for making a human being, but the human being is determined by the ingredients provided by the mother.

- David Barker

The “foetal origins of adult disease” hypothesis originated in the 1980s when Professor David Barker of the University of Southampton noted a link between low birth-weight and the incidence of cardiovascular disease among middle-aged men and women born in the United Kingdom. More than 30 studies around the world have indicated that low-birthweight term infants have a higher incidence of hypertension later in life than those with a normal birthweight, independent of adult social class and other adult risk factors as smoking, drinking, and overeating. Low birthweight, as well as thin-ness at birth, has also been correlated with glucose intolerance in childhood and noninsulin-dependent diabetes in later life.

In one stark example, semi-starved Dutch women in the closing stages of World War II gave birth to children who as adults were especially vulnerable to diabetes, high blood pressure, and coronary heart disease. This relationship was found to be particularly strongly associated with pregnancies that were subjected to food shortages in the third trimester of pregnancy.

The “Barker hypothesis” posits that maternal dietary imbalances at critical periods of development in utero can trigger an adaptive redistribution of foetal resources, including growth retardation. Such adaptations affect foetal structure and metabolism in ways that predispose the individual to later cardiovascular and endocrine diseases. The correlation between low birth-weight and later cardiovascular disease and diabetes may arise from the fact that nutritional deprivation in utero programmes a newborn for a life of scarcity. Problems arise when the child’s system is later confronted by a world of plenty.

Recent studies have shown a link with immune system development and subsequent risk of infection-related mortality in adulthood. A 1997 analysis of over a thousand deaths in one Gambian community has shown that infants born in the wet season were ten times more likely than infants born in the dry season to die prematurely in adulthood, mainly from infections. The difference was manifested only after adolescence. This phenomenon may be due to abnormal growth of the thymus gland (immune cell producer) or the lymph system (immune cell transporter) during pregnancy.31

With regard to future research, there is a need to progress beyond epidemiological associations to greater understanding of the cellular and molecular processes that underlie them. We need to know what factors limit the delivery of nutrients and oxygen to the human foetus, how the foetus adapts to a limited supply, how these adaptations programme the structure and physiology of the body, and by what molecular mechanisms nutrients and hormones alter gene expression. Further research requires a strategy of interdependent clinical, animal, and epidemiological studies.

The foetal origins of disease provide even greater justification for prioritizing nutrition of girls and women, for avoiding in utero and post utero nutritional imbalances, and for smoothing nutrition transitions. Essentially it calls for a long-term life cycle approach to nutrition improvement.

BOX 1.2

A Growth Curve for the 21st Century

Infants fed according to WHO recommendations and living in conditions that favour the achievement of genetic growth potential grow less rapidly than the NCHS/WHO International Growth Reference,2 particularly after 4-6 months. A significant discrepancy of approximately half a standard deviation in estimated height status arises immediately before and after 24 months of age. The distributions of weight-for-age and weight-for height are skewed, reflecting a substantial level of childhood obesity. These drawbacks led a WHO Expert Committee on Physical Status in 1995 to recommend the development of a new growth reference.

A multicountry growth reference study specifically designed for this purpose was launched in 1997. Depending on the availability of funds, data collection is expected to be complete in 2003. The study is being undertaken in seven countries in diverse geographical areas, using samples of infants and children whose caregivers follow recognized health recommendations.32 The research design combines a longitudinal study from birth to 24 months of age of 300 newborns per country, with a cross-sectional study of 1,400 children aged 18-71 months per site. More than 13,000 healthy infants and children will be involved in the study.

The new international growth reference will achieve several aims. Most important, it will provide a scientifically reliable descriptor of physiologic growth and a powerful tool for child health advocates. This will be achieved by applying the highest scientific rigor in a complex field-based project. Another objective will be to stress that human growth during the first five years of life is very similar across groups of children of different ethnic backgrounds. Equally important, the new growth reference will set growth of the breastfed infant as the standard to match. Future prevalence estimates based on the new reference will clearly be affected to the extent that the new reference differs from the current one.

Comparing Prevalences and Numbers

Table 1.2 and Table 1.3 show the estimated prevalences and numbers of stunted and underweight children, respectively, for the UN regions and sub-regions.c Multilevel modeling - the same statistical method described in the Third Report - was used for this report to develop the trend projections for both stunting and underweight. These tables were prepared for the ACC/SCN by the Department of Nutrition for Health and Development of WHO. All survey data were taken from the WHO Global Database on Child Growth and Malnutrition.33 The latest available national prevalence rates appear in Appendix 5. The national survey data are of variable quality, and some surveys are more than ten years old.

c The majority of the estimated prevalences cannot be compared with those published in the Third Report because of the different composition of regions. However, the estimated prevalence for all developing countries and for South America can be compared with the Third Report. In both cases, the confidence intervals of the estimates do include the midpoint estimates; thus there is consistency between the two reports.

Data were available from at least one survey for 107 countries for the estimation of stunting prevalences and from 108 countries for the estimation of underweight prevalences. Data were available from at least two surveys for 65 countries for estimation of trends in, stunting. For trends in underweight, data were available from two surveys for 68 countries. There are no differences in prevalence rates for boys and girls for stunting, underweight, or wasting, so results are not disaggregated by gender.

Estimates of the numbers of undernourished preschool children were derived by applying the estimated prevalences to the estimated total preschool population for each region and sub-region taken from UN population projections.9 Thus, estimates of numbers cover all countries within regions, including those that did not have a survey to contribute to the prevalence estimate.

STUNTING

In 2000, it is estimated that 32.5% of children under five in developing countries are stunted. There has been a steady improvement since 1980 when the estimated global prevalence was nearly one half (47.1%). By 2005, the estimated global prevalence will be further reduced to about 29.0%. Still, the numbers are extremely high. Some 182 million preschool children will be stunted in 2000, decreasing to about 165 million in 2005. More than two-thirds (70%) of these children live in Asia (of which 61% are in South Central Asia), while some 24% live in Sub-Saharan Africa.

TABLE 1.2: Estimated prevalence and number of stunted children, 1980-2005

UN regions and sub-regions

Prevalence of stunting (%)

Number stunted (million)


1980

1985

1990

1995

2000

2005

1980

1985

1990

1995

2000

2005

Africa

40.5

39.2

37.8

36.5

35.2

33.8

34.78

38.51

41.68

44.51

47.30

49.40


Eastern

46.5

46.9

47.3

47.7

48.1

48.5

12.88

14.83

17.13

19.28

22.03

24.41


Northern

32.7

29.6

26.5

23.3

20.2

17.0

6.01

6.01

5.55

4.90

4.44

3.86


Western

36.2

35.8

35.5

35.2

34.9

34.6

9.04

10.51

11.99

13.47

14.74

16.03

Asia

52.2

47.7

43.3

38.8

34.4

29.9

173.37

169.72

167.66

143.49

127.80

110.19


South Central

60.8

56.5

52.2

48.0

43.7

39.4

89.36

93.45

93.36

83.62

78.53

72.28


South-East

52.4

47.5

42.6

37.7

32.8

27.9

27.71

26.47

24.24

21.51

18.94

15.78

Latin America and the Caribbean

25.6

223

19.1

15.8

12.6

9.3

13.19

11.87

10.38

8.59

6.82

5.11


Caribbean

27.1

24.4

21.7

19.0

16.3

13.7

0.92

0.86

0.81

0.71

0.61

0.51


Central America

26.1

25.6

25.0

24.5

24.0

23.5

3.87

3.81

3.87

3.94

3.92

3.82


South America

25.1

21.1

17.2

13.2

9.3

5.3

8.38

7.35

6.05

4.55

3.16

1.84

Oceania

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

All developing countries

47.1

43.4

39.8

36.0

32.5

29.0

221.35

220.10

219.73

196.59

181.92

164.70

Source: 34.

Notes: Stunting is defined as low height-for-age at < -2 standard deviations of the median value of the NCHS/WHO international growth reference, n/a = not available.

As previously reported,35 levels of stunting vary across regions. The highest levels of stunting are estimated for Eastern Africa, where on average 48.1% of preschool children are affected in 2000. This region includes Ethiopia, where a national survey in 1992 found that 64.2% were stunted. National surveys in Malawi (1995) and Zambia (1996/97) found stunting prevalences of 48.3% and 42.4%, respectively. In Eastern Africa, stunting is increasing at 0.08 percentage points per year. This trend, together with high population growth rates, translates into larger numbers of East African children stunted each year. Over the period 2000 to 2005, numbers are expected to increase from about 22 to 24.4 million preschool children stunted.

Although stunting is widespread in South Central Asia, the trend in this region is towards improvement. South Central Asia includes Afghanistan, Bangladesh, India, and Pakistan, which all have high levels of child undernutrition. Central Asian countries, formerly part of the Soviet Union, are also included in this region. These countries report surprisingly high prevalences of stunting: Kyrgyzstan, 24.8% in 1997, and Uzbekistan, 31.3% in 1996. The estimated prevalence for South Central Asia as a whole in 2000 is 43.7%. This rate is decreasing by 0.85 percentage points per year. The number of stunted children declined over the 1990s. If this trend continues, about 6.25 million fewer children will be stunted in South Central Asia by 2005 compared with 2000.

The Western African sub-region has a much lower estimated prevalence than either Eastern Africa or South Central Asia: 34.9% in 2000. However, the prevalence has stagnated over recent years, and population numbers are increasing. The number of stunted children in this region therefore continues to rise. Between 2000 and 2005, the number of stunted children will increase by about one and a half million. To a large extent the trend in this sub-region will be driven by Nigeria, which has by far the largest child population in this group of countries. For Southern Africa, the sub-regional prevalence of stunting is 23.7%, according to surveys carried out in Lesotho, Namibia, and South Africa.d It was not possible to estimate a trend for this sub-region, owing to lack of repeated surveys. National prevalences documented in the most recent surveys available for four of the five countries in this sub-region range from 22.5% in South Africa (1994 - 95) to 44% in Lesotho (1996).

d This sub-regional prevalence is based on survey data, not on the model employed for all other estimates, because there is insufficient recent survey coverage in this sub-region.

About one-third (32.8%) of South-East Asian preschool children are stunted in 2000. This region has been experiencing the highest rate of improvement, at 0.98 percentage points per year, or a 10-percentage-point reduction between 1990 and 2000. This means that the number of children stunted is falling steadily and will continue to do so; it is projected to drop by more than 3 million between 2000 and 2005. Still, some 19 million children in this region are stunted in 2000. The effects’ of the financial crisis in this region are discussed in section 4.2.

About one in five preschool children in Northern Africa is stunted, translating into some 4.4 million children. The steady decline in both prevalences and numbers is forecast to continue, resulting in a 3-percentage-point decline between 2000 and 2005. This region comprises seven North African countries, as well as the Sudan. (There are about 4 million internally displaced persons in the Sudan, families normally not reached during national nutrition surveys.) Egypt, with the largest child population in this sub-region, may drive the overall pattern of improvement in the coming years for this group of countries. The most recent survey in Egypt (1997-98) shows that about 24.9% of preschool children are stunted.

The estimated prevalence for Latin America and the Caribbean as a whole (12.6%) continues to decline, by an average of 0.79 percentage points per year in South America. By 2005 only 5.3% of South American preschool children will be stunted. As discussed later in this report, undernutrition is being replaced by overweight in some South American countries. Central America, however, has an estimated prevalence of 24.0% and no significant improvement forecast over the next five years. Indeed the numbers of stunted children in Central America have remained about constant from 1980 to 2000.

UNDERWEIGHT

Underweight, due to chronic undernutrition or to wasting or to both, affects fewer children globally than stunting. However, underweight is still widespread among developing-country children. In 2000 it is estimated that 26.7% of preschool children in developing countries are underweight. Underweight has declined steadily since 1980, when 37.4% of the world’s pre-school children were underweight. The global prevalence will reach 24.3% by 2005. Global numbers will decrease from 150 million to 138 million between 2000 and 2005. The majority of underweight children (52%) live in South Central Asia.

TABLE 1.3: Estimated prevalence and number of underweight preschool children, 1980-2005

UN regions and sub-regions

Prevalence of underweight (%)

Number underweight (million)


1980

1985

1990

1995

2000

2005

1980

1985

1990

1995

2000

2005

Africa

26.2

26.7

27.3

27.9

28.5

29.1

22.47

26.30

30.11

34.03

38.32

42.45


Eastern

24.9

27.7

30.4

33.2

35.9

38.7

6.92

8.76

11.03

13.42

16.47

19.48


Northern

17.5

16.4

15.6

14.8

14.0

13.2

3.22

3.32

3.27

3.11

3.08

2.99


Western

30.1

31.7

33.3

34.9

36.5

38.1

7.51

9.29

11.23

13.34

15.41

17.66

Asia

43.9

40.2

36.5

32.8

29.0

25.3

145.95

142.95

141.31

121.03

107.91

93.16


South Central

58.1

54.5

50.9

47.3

43.6

40.0

85.35

90.06

90.90

82.40

78.49

73.48


South-East

43.5

39.9

36.2

32.6

28.9

25.3

23.00

22.21

20.60

18.56

16.68

14.27

Latin America and the Caribbean

14.2

12.2

10.2

8.3

6.3

4.3

7.32

6.50

5.57

4.48

3.40

2.35


Caribbean

22.9

20.1

17.2.

14.4

11.5

8.7

0.78

0.71

0.65

0.54

0.43

0.32


Central America

15.1

15.2

15.2

15.3

15.4

15.4

2.24

2.26

2.36

2.46

2.52

2.51


South America

13.2

10.7

8.2

5.7

3.2

2.3

4.40

3.71

2.88

1.96

1.08

0.80

Oceania

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

All developing countries

37.4

54.7

32.1

29.2

26.7

24.3

175.74

175.75

176.99

159.55

149.63

137.95

Source: 34.

Notes: Underweight is defined as low weight-for-age at < -2 standard deviations of the median value of the NCHS/WHO international growth reference, n/a = not available.

South Central Asia is the worst affected sub-region, with some 43.6% of children underweight. This translates into almost 79 million children underweight. However, prevalence and numbers continue to decline. By 2005 the estimated prevalence will drop a further 3.6% to 40.0%. The number of underweight children has been decreasing since about 1990 and is expected to continue. By 2005 some 5 million fewer children will be underweight.

Both Western and Eastern Africa have lower prevalences (36.5% and 35.9% respectively) than South Central Asia, but the situation is deteriorating. Countries of Eastern Africa are experiencing a rise in underweight of 0.56 percentage points per year, or a full 5-percentage-point increase between 1995 and 2005. During this period numbers have increased by about 6 million. The trend in Eastern Africa is very worrying. Western Africa has seen an increase of 0.32 percentage points per year in recent years. The increase in underweight among Western African children is explained in part by the high rates of wasting in this region, discussed below.

There is much less underweight among North African children (14.0%). This region is more similar to the Caribbean (11.5%) and Central America (15.4%) than to either Africa or Asia. Underweight in South America will have been eliminated by 2005, when the regional prevalence estimate reaches 2.3%. Similar progress is not being achieved in Central America, where there has been no improvement over the past 20 years in either prevalence or numbers. In South-East Asia the estimated prevalence, which has been falling steadily since 1980, is forecast to decrease further by 2005 to 25.3%. Still, some 14 million children in South-East Asia will be underweight.

The World Summit for Children set a global goal of halving severe and moderate malnutrition among children under five between 1990 and 2000. Our analysis indicates that only South America will have achieved this goal. In this region the overall rate has decreased from 8.2% in 1990 to 3.2% in 2000. Progress has been steady and significant in South Central Asia (from 50.9% to 43.6%), but the rate of progress is all too slow. Northern Africa, with higher mean household incomes, has seen very slow progress, from 15.6% to 14.0%. In other parts of Africa, 8.4 million more children are underweight now than in 1990.

WASTING

An estimated 50 million preschool children were wasted in 1995 (Table 1.4). Wasting is not as common as stunting or underweight in any region; the global prevalence is about 9.4%. Wasting rates can change rapidly, however, especially in situations of emergency food shortage and population displacement. This is discussed further in section 5.1.

TABLE 1.4: Prevalence and number of wasted preschool children, 1995

UN regions and sub-regions

Survey countries/total countriesa

Population covered by surveys (%)

Prevalence of wasting (%)

Number wasted (million)

Africa

43/53

94.5

9.6

11.06


Eastern

16/17

95.8

7.0

2.74


Middle

5/9

84.6

8.6

1.36


Northern

6/6

99.8

7.2

1.46


Southern

4/5

95.9

2.9

0.17


Western

12/16

94.2

15.6

5.33

Asia

31/46

93.7

10.4

37.87


Eastern

2/4

94.4

3.4

3.73


South Central

12/14

99.2

15.4

27.27


South-East

5/10

84.0

10.4

5.75


Western

12/18

70.7

5.1

1.12

Latin America and the Caribbean

21/31

97.2

2.9

1.59


Caribbean

4/13

64.8

n/a

n/a


Central America

7/8

99.8

4.9

0.79


South America

10/12

99.6

1.8

0.64

Oceania

n/a

n/a

n/a

n/a

Developing countries

99/147

94.1

9.4

50.59

Source: 34.

Notes: Wasting is defined as low weight-for-height at < -2 standard deviations of the median value of the NCHS/WHO international growth reference, n/a = not available because of insufficient data.

a Number of countries that have national surveys out of the total number of countries for each sub-region.

These updated results are similar to those published by WHO in 1997,e but there has been a substantial increase in wasting among Western African children. This increase helps explain the high rates of underweight in these countries. Western Africa and South Central Asia have the highest prevalences of wasting (both about 15.5%) followed by South-East Asia (10.4%). Central America presents low levels, and South America and Southern Africa are close to the statistically expected prevalence, implying that wasting is not common in these children.

e In 1997, WHO reported sub-regional, regional, and global estimates for the prevalence and number of wasted children under five years old for 1995.33 These estimates were derived from nationally representative data using a weighted prevalence approach for sub-regions where the proportion of children covered by national surveys was at least 70%. For this Fourth Report, WHO has updated these 1995 estimates on the basis of recent data (as of June 1999), although the median survey year for the countries that have nationally representative data does not permit an estimate beyond 1995.

Exploring the Changes in Preschool Nutrition in Sub-Saharan Africa

The region that has seen the least overall improvement in the nutritional status of its children is Sub-Saharan Africa, where one-half of the 25 countries with more than one national survey indicated a rise in stunting rates.35

What explains the differential progress across countries in reducing child malnutrition and the slow progress of this region as a whole? The causes of child malnutrition are complex, multidimensional, and interrelated, ranging from factors as fundamental as political instability and slow economic growth to those as specific in their manifestation as respiratory infection and diarrhoeal disease. This is well illustrated by the framework in Appendix 1. Determinants also differ considerably across geographical areas. Nonetheless, using cross-country regression analysis it is possible to gain a general sense of the relative importance and contribution of some broad causal factors for the developing world as a whole. A recent study undertook such an analysis using national underweight prevalence rates collected from 1970 to 1995 for children under five years of age.36 This IFPRI study, which is described in Appendix 6, found that women’s educational and social status, national per capita food availability, and access to safe water were important underlying determinants of child nutritional status at a global level.

In a further analysis for Sub-Saharan Africa, countries for which data over rime were available for some portion of the 1970 to 1995 period were divided into two groups: (1) those with periods of decreased prevalence (“improving”) and (2) those with periods of increased underweight prevalence (“deteriorating”). Appendix 7 lists the countries and time periods.

Table 1.5 reports the differences across the two groups in underweight rates and in the factors identified as important determinants of underweight in the preceding global analysis. The average decrease in underweight prevalence for the “improving” group is 5.5 percentage points, while the average increase for the “deteriorating” group is 6.0 percentage points. When the intergroup differences among the potential determinants were examined, two were found to be statistically significant: differences in women’s relative status and in per capita dietary energy supply

The proxy variable for women’s relative status - the ratio of female life expectancy to male life expectancy - actually declined for both groups, but the decline for the “deteriorating” group was five times larger than that for the “improving” group. One would expect such declines in women’s relative status to have had a negative impact on child nutrition for both groups, wiping out some of the gains made through improvements in the other factors considered. Countries that are better able to protect women’s status relative to men’s will be more likely to experience improvements in child nutritional status.

The female secondary enrollment rate improved for both sets of countries. The increase in the enrollment rate for the “improving” group is more than double that of the “deteriorating” group, although this difference is not statistically significant.

Differences in food availability were very important. “Improving” countries had an average increase in per capita dietary energy supply of 82 kilocalories, compared with an average decrease of 92 kilocalories in the other group. In the case of national income, “improving” countries had an average increase in per capita gross domestic product (GDP) of US$175, compared with an average decrease of US$82 in the other group, although the difference was not statistically significant. National income is a determinant of investment in health environments, education, improvements in women’s status, and food supplies. Slow progress in both food availability and national income in this region is a result not only of rapid population growth, but also of conflict, the debt burden, and the HIV/AIDS epidemic (not measured in this analysis).

TABLE 1.5: Comparison of Sub-Saharan African countries with periods of increased versus decreased underweight rates over 1970-95

Group means

P-value for significance of difference in group means

Countries with periods of decreased underweight prevalence (n = 18)

Countries with periods of increased underweight prevalence (n = 18)


Change in underweight rate (percentage points)

-5.5

6.0

.000*

Change in population with access to safe water (%)

7.4

8.4

.819

Change in female secondary school enrollment rate (%)

5.4

2.0

.122

Change in ratio of female life expectancy to male life expectancy (proxy for women’s relative status)

-.0025

-0.119

.020*

Change in per capita dietary energy supply (kilocalories)

82

-92

.008*

Change in per capita gross domestic product (US$ purchasing price parity)

175

-82

.141

Change in democracy (index from 1 to 7, 1 = least democratic)

0.5

0.8

.488

Source: Based on 36.
* Statistically significant at least at 5% significance level.

BOX 1.3

Overweight in Children

In industrialized countries several studies report increasing prevalence of obesity in children. Some 23.7% of U.S. preschool children are overweight, and 7.4% are obese.37 In developing countries, such studies are scarce. Research in Latin America has concluded that the levels of overweight and obesity in children under five in the region are lower than those in the United States, although prevalences in some countries are higher than expected statistically.38 A clear pattern of change over time in overweight and obesity in Latin American children is not yet discernible.

Sub-regional, regional, and global estimates for the prevalence and number of overweight children under five (> +2 SD of the NCHS/WHO reference median value) have recently been published by WHO.39 These estimates incorporated new data from 160 nationally representative surveys. A weighted prevalence approach was used for sub-regions where the proportion of children covered by national surveys was greater than 70%. The median survey year for countries that have nationally representative data does not allow for estimates beyond 1995. Summary results of this work are presented in Appendix 8.

An estimated 17.6 million children were overweight in the developing world in 1995. Northern and Southern Africa, Eastern Asia, Central America, and South America had prevalences higher than expected, while Western Africa and South Central and South-East Asia did not. A lack of data prevented estimation of prevalences for other sub-regions. Overall the prevalence of overweight in preschool children in developing countries is low (3.3%). Two regions have both overweight and stunting among their children. Northern Africa has a prevalence of overweight of 8.1%, while 20.2% of children are stunted. Central America has a prevalence of overweight of 3.5%, while stunting affects 24.0%. Countries in these regions are undergoing a rapid nutrition transition, including adoption of “western” diets that are high in saturated fats, sugar, and refined foods.30

In sum, this study reveals that the reasons why child undernutrition in many Sub-Saharan African countries has increased over the last 25 years - and why the region as a whole has progressed very little - are associated with declines in women’s relative status, slow-progress in improving women’s educational attainment, and low per capita food availability and income.