OBJECTIVE: Our goal was to describe worldwide growth-faltering patterns by using the new World Health Organization (WHO) standards.
METHODS: We analyzed information available from the WHO Global Database on Child Growth and Malnutrition, comprising data from national anthropometric surveys from 54 countries. Anthropometric data comprise weight-for-age, length/height-for-age, and weight-for-length/height z scores. The WHO regions were used to aggregate countries: Europe and Central Asia; Latin America and the Caribbean; North Africa and Middle East; South Asia; and sub-Saharan Africa.
RESULTS: Sample sizes ranged from 1000 to 47 000 children. Weight for length/height starts slightly above the standard in children aged 1 to 2 months and falters slightly until 9 months of age, picking up after that age and remaining close to the standard thereafter. Weight for age starts close to the standard and falters moderately until reaching approximately −1 z at 24 months and remaining reasonably stable after that. Length/height for age also starts close to the standard and falters dramatically until 24 months, showing noticeable bumps just after 24, 36, and 48 months but otherwise increasing slightly after 24 months.
CONCLUSIONS: Comparison of child growth patterns in 54 countries with WHO standards shows that growth faltering in early childhood is even more pronounced than suggested by previous analyses based on the National Center for Health Statistics reference. These findings confirm the need to scale up interventions during the window of opportunity defined by pregnancy and the first 2 years of life, including prevention of low birth weight and appropriate infant feeding practices.
WHAT'S KNOWN ON THIS SUBJECT:
Children from low- and middle-income countries are born with weights and lengths that are close to those of children in the NCHS growth reference. Growth faltering happens mostly from 3 to 18–24 months of age.
WHAT THIS STUDY ADDS:
Children from low- and middle-income countries are born with weights and lengths below WHO growth standards, and early growth faltering is even faster than currently assumed. The window of opportunity for preventing undernutrition ends at 2 years of age.
Child undernutrition is expressed in different forms, including stunting (low length/height for age), wasting (low weight for length/height), underweight (low weight for age), and deficiencies of essential vitamins and minerals.1 It remains as one of the main public health challenges of the 21st century, particularly in low- and middle-income countries.2,–,6 Recent estimates suggest that stunting, wasting, and intrauterine growth restriction are responsible for 2.2 million deaths and 21% of disability-adjusted life-years lost among children younger than 5 years.3 Undernutrition also has long-term consequences, which include short adult height, reduced intellectual development and economic productivity, and low offspring birth weight.6 In 2001, on the basis of data from 39 nationally representative samples from Africa, Asia, and the Americas, we reported on the worldwide timing of growth faltering.7 For all surveys taken together, weight for age (WAZ) faltered rapidly from 3 to 12 months of age, with a slower decline up to 18 months and followed by some catch-up growth. Weight for length/height (WHZ) declined until 15 months of age, and slowly increased again after that age. Length/height for age (HAZ) was close to the reference value at birth, but declined sharply until the age of 2 years, after which mean values remained stable between 1.5 and 2 SDs below the reference.7 Faltering was most intense in Asian countries, followed by African and Latin American countries. These comparisons were conducted by using the National Center for Health Statistics (NCHS) growth reference8 that was internationally recommended for use at that time. In 2006, on the basis of a multicountry study involving exclusively breastfed children in 6 sites in different regions of the world, new World Health Organization (WHO) Growth Standards9 were introduced. These new standards differed substantially from the NCHS reference.10 The present analyses describe growth-faltering patterns by using the new WHO standards instead of the old NCHS reference. We rely on a larger and more recent set of national surveys than used in the previous analyses. Also, in a previous publication, the authors reported on broad groupings of countries (Asia, Africa, and the Americas), but there is reason to believe that there may be substantial variation within each region11; therefore, more restricted regional classifications were used in the present analyses.
We analyzed information available from the WHO Global Database on Child Growth and Malnutrition, comprising data from cross-sectional anthropometric surveys worldwide.12 We included national data from 52 countries where demographic and health surveys (DHS) were used13 and from 2 countries where multiple-indicator cluster surveys (MICS) were used.14 We included the most recent data set available for each country. Table 1 shows the survey year and type, the age range, and the sample size. The oldest data set comes from the Central African Republic (1994), whereas the most recent ones are from Jordan and Liberia (2007). All but 7 surveys were conducted since 2000. Data are available from birth to 59 months of age in 51 countries. Exceptions were Kyrgyzstan, Central African Republic, and Comoros, for which the upper age limit was 35 months. Of the 54 countries, 32 are classified as low income, 17 as lower middle income, and 5 as upper middle income.15
Children younger than 24 months of age were measured in supine position (recumbent length), whereas older children were measured standing up (height). Each child in the database had the 3 separate z scores calculated by using the 2006 WHO child growth standards9: WAZ, HAZ, and WHZ. For every monthly age group in each national sample, the mean WAZ, HAZ, and WHZ were calculated. Individual z scores based on WHO standards are calculated by using age in days, and the results are grouped by using completed months. The WHO regions were initially used to aggregate countries: Europe and Central Asia (EURO); Latin America and the Caribbean (PAHO/AMRO); North Africa and the Middle East (EMRO); South Asia (SEARO); and sub-Saharan Africa (AFRO). Because only 2 countries from the Western Pacific region (WPRO) were part of the data set, we decided not to analyze this as a separate region, but instead to group Mongolia with the EURO region (which already included the Central Asian republics) and Cambodia with SEARO region, which includes South Asian countries. The EURO samples do not include any of the high-income nations in Western Europe because there are no DHS surveys in such countries; our data set includes Turkey plus a number of ex-socialist republics, all in the middle-income group. The global average of all studies was calculated, as well as national averages at birth, 1, 2, 3, 4, and 5 years of age. All analyses are unweighted.
Forty-seven national databases included between 1000 and 10 000 children, 6 from 10 000 to 15 000, and the Indian data set included 47 000 children. The average number of children per survey in each month of age grouping across all surveys was 103 (92 if the large Indian survey is excluded). The distribution was very homogeneous for all ages except for children aged <1 month, for whom the average number per survey was 47 (43 if India is left out). The average numbers of 1-month-olds were in the expected range (99 for all surveys and 90 without India). In light of the undersampling of children aged <1 month, these were excluded from the analyses.
Figure 1 shows mean WAZ, WHZ, and HAZ for all countries combined. WHZ is slightly above the standard in children aged 1 to 2 months, then falters slightly until 9 months of age, and picks up after that, reaching the standard mean at ∼24 months and remaining above the standard mean until ∼48 months, when it starts to decline slightly until the end of the age range of the studies. WAZ starts close to the standard and falters moderately until reaching approximately −1 z at 24 months, remaining reasonably stable after that. HAZ already starts below the standard and falters dramatically until 24 months, showing noticeable bumps just after 24, 36, and 48 months, but otherwise increasing slightly after 24 months.
Figure 2 presents the mean WAZ by WHO region. In SEARO, the average z score starts at approximately −1.25 and declines steadily to −1.75 until 18 months of age, after which it continues to decline slowly (at rates between −0.01 and −0.02 z score per month) until 5 years of age when the average is close to −2 z scores. In AFRO, WAZ starts close to the standard but marked faltering takes place in the first year of life when the speed of decline is −0.09 z score per month. After that age, only slight faltering is observed (between −0.01 and −0.02 z score per month). Children in PAHO/AMRO and EMRO are born close to the standard and show slight faltering throughout the first 5 years, with rates of decline not exceeding −0.02 z score per month in PAHO/AMRO (in the first and second year) and −0.04 z in EMRO (second year of life). At 5 years of age, the mean WAZ in PAHO/AMRO is −0.6 and in EMRO −0.9. In EURO, the line shows marked fluctuations, reflecting the greater variability among countries. Mean WAZ starts slightly above the standard and shows very discreet faltering until 5 years, at rates of decline at approximately −0.02 z score per month.
Data on WHZ by region are shown in Fig 3. Unlike WAZ or HAZ, faltering patterns are not consistent. In SEARO, mean WHZ starts at −0.75, falls to below −1 by 12 months, and remains at that level for the next 4 years. In AFRO, infants start above the standard and WHZ declines steadily until 12 months of age. After that, mean values increase until ∼30 months and decline very slightly after that age. In PAHO/AMRO, mean values were between 0.25 and 0.5 z at virtually all ages without any clear increase or decline. Infants from EMRO and EURO start with similar mean WHZ, but whereas in EURO WHZ increases steadily until ∼30 months of age and declines slightly thereafter, in EMRO WHZ was very close to the standard (remaining between −0.25 and 0.25 z) at all ages.
Figure 4 presents data on HAZ. Massive faltering is observed in the 5 regions, although its magnitude differs by region. In EURO, mean HAZ values are close to the standard until 12 months, drop by −0.05 z score per month in the second year, and remain nearly stable thereafter. EMRO and PAHO/AMRO show very similar faltering patterns with average declines of −0.04 and −0.06 z score per month, respectively in the first year of life, and of −0.05 and −0.07 in the second year, children in both regions remaining practically stable throughout ages 2 to 4 years. Similar, but more accentuated, patterns were observed in AFRO and SEARO. Although both regions start at different levels, their faltering patterns are very similar from 3 to 24 months (monthly rates of decline −0.10 z score in AFRO and −0.08 in SEARO). After 24 months, there is evidence of some slight catch-up in AFRO (0.005 z score per month increase) but not in SEARO (0.000 z score per month).
Tables 2 to 4 present mean anthropometric scores for children at 1, 12, 24, 36, 48, and 59 months of age for the countries included in the analyses. At 24 months, Bangladesh presented the lowest mean WAZ of −2.02 z, whereas Montenegro had the highest mean of 1.34 z. WHZ at 24 months birth was the highest in Guatemala (2.19 z) and the lowest in Comoros (−1.39 z). HAZ at 24 months ranged from −2.59 z in Niger to 0.70 z in Montenegro.
Our main objective was to describe patterns of growth faltering according to the WHO child growth standards, and discuss the policy implications of the analyses based on it in comparison with a previous study based on the NCHS curves. The earlier analyses7 were based on 39 nationally representative surveys conducted between 1987 and 1997, whereas we now report on 54 surveys dating from 1994–2007.
HAZ curves were quite similar by using the NCHS reference and the WHO standards. Rapid growth faltering was observed in children until 24 months of age, regardless of the curve used. The main difference is that length z scores for children aged <3 months were lower when using the WHO standards than those obtained in the previous analyses in which NCHS standards were used. The length curves in the first 3 months are notably different, particularly for the −2 and −3 z-score lines, with the WHO curve being above the NCHS reference. This is largely because NCHS only measured length at birth and at 3 months, and curve-fitting was used for the intermediate ages. The WHO standards, on the other hand, had empirical data at birth, weeks 2, 4, and 6, and monthly thereafter in the first year.10 As a consequence, the results presented here on early growth faltering are more precise than with the previous reference. The WHO curves, therefore, provide a better description of physiologic growth, showing that intrauterine retardation in length is a greater problem than previously believed, and revealing the need for greater efforts to improve nutritional status of pregnant women and women in childbearing age.3
WAZ patterns, on the other hand, differed markedly from the earlier analyses. With NCHS, rapid decline was observed in children from 3 to 12 months of age when the mean z score was approximately −1.25, with stability or slight catch-up afterward. In the current analyses, WAZ did not show such a rapid decline in the first year (mean z score at 12 months: −0.75) but there was progressive and slow faltering throughout the first 5 years.
WHZ patterns were similar in both analyses, but faltering lasted until 18 months of age with NCHS and 12 months with the WHO curves. With the latter, the mean WHZ was close to 0 from 24 months of age onward.
Differences across regions were marked. Although the shapes of the WAZ curves were quite similar across regions, absolute values were consistently lowest in SEARO and highest in the EURO region. Likewise, HAZ patterns show faltering in all regions in children up to ∼24 months and stability thereafter, with almost parallel curves for all regions. Again, children from SEARO were the shortest, closely followed by those from AFRO. The tallest children were those from EURO, who were ∼1.5 z scores taller than SEARO and AFRO children from 24 to 59 months of age. The pattern of WHZ above the median in EURO and PAHO/AMRO reveals the growing problem of childhood overweight in these regions, and potentially also in the EMRO region.
Some limitations of our analyses should be considered. First, some degree of measurement error is expected in the different national surveys, particularly in terms of measuring young infants. Children aged <1 month were excluded from the analyses because of such measurement issues and also because they are underrepresented in the samples; undercounting of newborns in population-based surveys is a common finding. Age rounding and heaping (ie, tending to report age in full years rather than in exact months) may also explain the bumps observed in the HAZ and, to a lesser extent, in the WAZ curves at approximately 24, 36, and 48 months, but this does not affect our overall conclusions.
Another limitation is the inference of growth patterns from cross-sectional data in which each child was measured only once. Survival bias could affect our interpretation; however, undernourished children are more likely to die than those who are well nourished,3 and survival bias would tend to result in fewer, not more, undernourished children as ages advance.
As in the 2001 article,7 we opted not to weight the data for analyses, because data are not available for all countries in a given region and similar curves were obtained from the different surveys. The small number of samples from SEARO and EMRO is another limitation. It is important to note although the WHO regional classification has been applied to group countries, none of the regions can be considered representative. We repeated the analyses dividing countries according to the World Bank classification of low, lower-middle, and upper-middle income countries. However, because most countries with available data were in the low-income category, and the regional classification showed important variability even among low-income countries, we opted for presenting only the latter.
In terms of policy implications, the new analyses confirm the importance of the first 2 years of life as a window of opportunity for growth promotion. Also, monitoring of length/height in addition to weight throughout infancy and childhood, as suggested previously,6,7 seems essential because (1) faltering patterns are clearly different for HAZ and WAZ, and (2) short stature is associated with deleterious long-term outcomes.6 Another important feature of the WHO standards is that they reveal a much greater problem of undernutrition during the first 6 months of life than previously believed, bringing coherence between the rates of undernutrition observed in young infants and the prevalence of low birth weight and early abandonment of exclusive breastfeeding. The NCHS reference, because of its technical and biological deficiencies, did not pick up this early faltering and incorrectly made it seem that growth was protected during the first months of life. These findings highlight the need for prenatal and early-life interventions to prevent the growth failure that mainly happens during the first 2 years of life, including the promotion of appropriate infant feeding practices.2
We thank Dr Allen Shoemaker for the reanalysis of the national data sets by using the WHO child growth standards.
- Accepted September 2, 2009.
- Address correspondence to Cesar Gomes Victora, MD, PhD, Universidade Federal de Pelotas, Postgraduate Program in Epidemiology, Rua Marechal Deodoro 1160, 96020-220, Pelotas, RS, Brazil. E-mail:
The views in this article are those of the authors and do not necessarily represent the views of the World Health Organization.
The authors have indicated they have no financial relationships relevant to this article to disclose.
- WAZ =
- weight for age •
- WHZ =
- weight for length/height •
- HAZ =
- length/height for age •
- NCHS =
- National Center for Health Statistics •
- WHO =
- World Health Organization •
- DHS =
- demographic and health surveys •
- MICS =
- multiple-indicator cluster surveys •
- EURO =
- Europe and Central Asia •
- PAHO/AMRO =
- Latin America and the Caribbean •
- EMRO =
- North Africa and the Middle East •
- SEARO =
- South Asia •
- AFRO =
- sub-Saharan Africa •
- WPRO =
- Western Pacific region
- Shrimpton R,
- Victora CG,
- de Onis M,
- Lima RC,
- Blössner M,
- Clugston G
- 8.↵National Center for Health Statistics. NCHS Growth Curves for Children, Birth to 18 Years. Washington, DC: National Center for Health Statistics, Department of Health, Education, and Welfare; 1977. PHS publication No. 78–1650
- 9.↵WHO child growth standards based on length/height, weight and age. Acta PaediatrSuppl.2006;450:76–85
- de Onis M,
- Onyango AW,
- Borghi E,
- Garza C,
- Yang H
- Frongillo EA Jr.,
- de Onis M,
- Hanson KM
- de Onis M,
- Blossner M
- 13.↵Demographic and Health Surveys. Available at: www.measuredhs.com/aboutsurveys/dhs/start.cfm. Accessed January 12, 2009
- 14.↵United Nations Children's Fund. Multiple indicator cluster surveys. Available at: www.unicef.org/statistics/index_24302.html. Accessed January 12, 2009
- 15.↵World Bank. World Bank groupings. Available at: http://info.worldbank.org/etools/wti2008/docs/Countrylist.htm#Income. Accessed January 15, 2009
- Copyright © 2010 by the American Academy of Pediatrics