References and standards for infant and child growth

Sir,

Geoffrey CannonReference Cannon¹ says kind things about my contributions in this field, but on one point he goes far astray. He writes ‘the idea that reference values are not normative is an obvious contradiction in terms’. Not so. The original paperReference Waterlow, Buzina, Keller, Lane, Nichaman and Tanner² recommending the NCHS growth charts as an international reference said very clearly: ‘A reference is a device for grouping and analyzing data and for enabling comparisons between different populations. It implies nothing about values or targets… A standard embodies the concept of a norm or target – that is, a value judgement’. Inevitably the two concepts have been confused in practice and the reference used as a norm.

In 1976 there was an urgent need for a means of assessing and comparing different groups of children. The NHCS was chosen as a reference, in spite of its well-known disadvantages, because it included measurements of height and length, and was well worked out statistically. There followed an enormous amount of work and discussion about whether it was realistic to use it as a normative standard, particularly for height, for different populations. Now, 30 years later, the NCHS has been superseded by a new internationally based reference which can reasonably be used as a standard or norm as well as a referenceReference de Onis, Garza, Victora, Bhan and Norum³.

From my point of view, the most important contribution of the NCHS reference was that it enabled the traditional index of weight-for-age to be separated into two biologically different components: weight-for-height and height-for-weight. I proposed the terms ‘wasting’ and ‘stunting’ for extremes of deficits in these two components, because they describe what one actually sees, in a more graphic way than more speculative names such as ‘acute’ and ‘chronic’ malnutrition. Certainly these two names imply a value judgement or norm, since they are defined as deviations of more than 2SDs below the reference mean.

Nevertheless, in spite of uncertainty about the validity of the reference, I believe that a high prevalence of stunted children in a population is an indicator of a disadvantaged environment, though precisely what the disadvantage is, whether nutritional, repeated infection or whatever, we do not know. An economist has described stunting as a beneficent adaptation, because a stunted child needs less food. That may be so, but the ‘adaptation’ comes at a huge cost. The stunted child is impaired in mental as well as in physical development, as shown by the studies of Grantham-McGregor et al. In a recent series of papers in the Lancet Reference Grantham-McGregor, Cheung, Cueto, Glewwe, Richter and Strupp⁴ some workers have found that stunting is reversible when the child is transferred to a better environment, others not. A fascinating paradox is described by Satyanarayana et al. In IndiaReference Satyanaranya, Prasanna Krishna and Narasiga Rao⁵ poor children at 5 years of age had a very large height deficit compared with their well-to-do peers; between 5 and 18 they grew as much in stature as children in California, but they never made up the deficit with which they started.

Thanks to the NCHS we know a good deal about the natural history of stunting. I am not well up on the literature; I know of little work on the biochemical or metabolic defect that is holding back growth. Perhaps there may be a hint in the finding of Millward's group that rats on a low-protein diet had decreased synthesis of the proteoglycans of cartilageReference Yahya, Tirapequi, Bates and Millward⁶, but that is only a beginning.

Why do I go on about this? I ask myself does the ‘new nutrition science’ provide any stimulus to tackle the old but very important problem of stunting – a problem that involves nutritionists at all levels: the biochemist, the epidemiologist, the administrator? I can't see that it does.