The 2006 WHO child growth standards

In April 2006, the World Health Organization released its new WHO child growth standards,1 16 years after a WHO working group on infant growth recommended that these standards should describe how children should grow rather than how they actually grow.2 The basis for the new growth standards was six population based studies of infants and children from Ghana, India, Norway, Brazil, Oman, and North America, undertaken between 1997 and 2003. Participants were fed according to accepted international nutritional standards (including breast feeding), and their mothers were adequately nourished and avoided known adverse factors such as tobacco exposure.

The new growth standards show that children born in different regions of the world can and should grow equally well, and they also show that sex and ethnic origin are minor determinants of growth compared with adequate nutrition, environment, and health.2 However, as expected, important differences in the diagnosis of malnutrition emerge when the standard cut-offs are applied using either the National Center for Health Statistics (NCHS)-WHO reference or the WHO 2006 growth standards.

In this week's BMJ, Seal and Kerac report the implications of adopting the new WHO child growth standards in emergency and non-emergency child feeding programmes using secondary data analysis from three nutritional surveys in emergency settings.3 Nutritional status can be expressed using either z scores or percentage of the median values. WHO recommends that weight for height should be expressed as a z score. However, while not recommended by WHO, many agencies working in emergency settings use weight for height expressed as a percentage of the median as the criterion for admission to feeding programmes.

Seal and Kerac tabulated and compared the weight for height z score and percentage of the median cut-offs for moderate and severe acute malnutrition from both the NCHS-WHO growth reference and the new WHO standards. With the new WHO standards, a marked increase (0.5-2.7) in the prevalence of severe wasting was seen if weight for height was expressed as a z score (weight for height ≤3 z scores), confirming a previous report that showed an increase of 1.5-2.5.4 Paradoxically, however, the new WHO standards showed a significantly lower prevalence of severe wasting when expressed as a percentage of the median (weight for height <70% of the median). These findings have serious programmatic and resource implications.

In emergencies, growth standards are used for various reasons, including interpreting the results of nutritional surveys, estimating potential beneficiary figures, and calculating entry and exit criteria for feeding programmes. Increased prevalence of severe wasting varies in different settings when using either the new WHO standards or the NCHS-WHO references. This fact makes the interpretation of nutritional surveys difficult and means that results cannot be converted from one to the other by using a simple algorithm. During the transition period, therefore, both the WHO standards and the NCHS-WHO references should be used until the implications of these differences are better understood.

The choice of using the percentage of the median or the z score for admission of children to feeding centres is more complicated and should be related to functional outcomes. Weight for height expressed as a percentage of the median is used partly because it is a better predictor of mortality than weight for height expressed as a z score when using the NCHS reference.5 We urgently need to study the functional significance of the weight for height indicators and investigate the suitable z score cut-offs for therapeutic and supplementary feeding in relation to mortality and other functional outcomes.

WHO, the United Nations World Food Programme, and Unicef are in the process of producing a joint statement on community based management of severe acute malnutrition. New evidence using ready to use foods suggests that large numbers of children with severe acute malnutrition can be treated in their communities without being admitted to therapeutic feeding centres or health facilities.6 7

It is difficult and costly for community health workers or volunteers to measure weight for height, and the mid-upper arm circumference is a good predictor of mortality. A mid-upper arm circumference less than 110 mm has therefore been recommended as an indicator of severe acute malnutrition in community based management of malnutrition. It would be useful to investigate whether this measurement can also be used as an indicator for admission to feeding programmes.

The international nutrition community working in emergencies has welcomed the introduction of the new WHO growth standard. Many will agree with Seal and Kerac, however, that a full assessment of the appropriate use of the new WHO standards in the diagnosis of acute malnutrition is urgently needed, and that this should be completed before they are adopted by agencies engaged in running nutritional programmes in emergencies.