Contents - Previous - Next

Introduction: Causes and mechanisms of linear growth retardation (stunting)


J.C. Waterlow

Department of Public Health and Policy, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK

A workshop held 6 years ago examined the epidemiology and natural history of stunting in Third World children. Although this condition is extremely common and is usually regarded as a manifestation of chronic malnutrition, in fact almost nothing is known about its causes and mechanisms in nutritional, biochemical, or metabolic terms. The objective of the present workshop is to fill these gaps and to identify, if possible, critical mechanisms by which environmental factors might affect linear growth. An example is described of a longitudinal study which attempted to examine, in more detail than has been done before, what is happening to children during the process of becoming stunted.

Correspondence to: J.C. Waterlow, 15 Hillgate Street, London W8 7SP, UK.

In this workshop we have tried to bring together people who are interested in linear growth from many different points of view. The basic problem that we hope will be illuminated by this multi-disciplinary approach is the mechanism of linear growth retardation in Third World children. For those who have not been directly concerned with this problem it may be useful to say a little about the background.

Until about 20 years ago, anthropometric assessment of the nutritional state of children was based on weight-for-age, in relation to a supposedly normal standard or reference. Then Latham and I independently made a separation between low weight-for-height and low height-for-age (Seoane & Latham, 1971). Latham called these acute and chronic malnutrition. I disliked these terms because they imply more than we really know: specifically, is low height-for-age really an indicator of malnutrition? That is what this workshop is about. I prefer to use names that simply describe what we see wasting and stunting (Waterlow, 1973). I proposed these terms instead of just 'thin' and 'short', because wasting and stunting are supposed to imply deviation beyond the range of thinness and shortness that might be regarded as normal.

When this classification began to be used, it soon became apparent that stunting is generally far commoner than wasting, and that in some populations, particularly in Asia, over 50 per cent of children could be classified as stunted (WHO, 1987). Immediately this gave rise to a host of questions: could it really be true that more than half the children in a country such as Nepal, for example, are malnourished? What are the functional implications of this reduced growth? Is it a kind of adaptation? - and so on.

To examine some of these questions, a workshop was organized in Thailand in 1987, sponsored by the Nestlé company, which was focused mainly on the epidemiology and natural history of stunting (Waterlow, 1988a). Two factual points emerged that are relevant for this meeting. The first was that although children generally did not reach the stage of being classified as stunted, according to the conventional definition (- 2 SD below the international reference height-for-age) until the age of 2 or 3 years, the process of slowing of linear growth actually started much earlier, perhaps as early as 2 or 3 months of age (Waterlow, 1988b).

The second point was the remarkable seasonal variation in the linear growth of children in developing countries (Nabarro et al., 1988). This variation, far greater than any observed in affluent populations, must be due to environmental effects; it shows that linear growth is more plastic than many of us had previously supposed, and it implies that, when conditions are favourable, the process of stunting can be halted or reversed.

At the meeting in 1987 some attention was also given to possible causal factors. The conventional definition of stunting, used by WHO (1983), related the observed height of a child to data from an American reference population, collected by the US National Center for Health Statistics (NCHS). Could the definition be inappropriate, and the high prevalence of stunting in Asian children reflect simply a genetic difference in growth potential? Davies (1988), who formerly worked in Hong Kong, proposed that the answer to this question is 'yes', and Ulijaszek will be reexamining this subject. Davies' data, collected from healthy children of well-to-do parents in various Asian countries, showed that mean length-for-age fell steadily from 3 months until, at 2 years, it was 1.5 SD below the NCHS median.

Ought we, then, to have a separate reference for the growth of Asian children? If so, use of the NCHS data as a reference would indeed be inappropriate, leading to exaggerated estimates of the prevalence of stunting. Nevertheless, such a genetic effect, if indeed it exists, is less important than that of the environment. In poor Third World children the mean height-for-age is shifted a good deal more to the left than in the groups described by Davies.

The other major problem which was broached at the Thailand meeting, but without any firm conclusions, may be stated very simply and naively: if stunting is considered to be an indicator of malnutrition, what is the nutritional cause? In Thailand Golden (1988) reviewed earlier studies on the effects of food supplements on height growth. These, together with some epidemiological studies, such as Rutishauser & Whitehead's (1969) comparison of meat-eating and vegetarian tribes in Uganda, point us in the general direction of protein or factors associated with protein in foods.

I want to end by mentioning a small study with which I was concerned, that was stimulated by the Thailand workshop' even though the results were on the whole negative. This research was carried out by a group of young paediatricians and biochemists in the medical school of Khon Kaen, in North-East Thailand (Chusilp et al., 1992). Most of the previous work with supplementary feeding had been done on older children, even school-children. We wanted to try to get a better idea of what is happening during the process of becoming stunted.

Eighty infants were selected at birth, all of whom had either one or two sibs, mostly two. These index babies were divided into 3 groups, according to whether the 2 sibs were born within normal limits (above -2 SD) of height-for-age both stunted or one stunted and one normal. The referees to this paper were very critical of the fact that the selection was not random. I was able to persuade them that we do not have to be slaves to the epidemiological approach. For a physiologist trying to understand a process, it is perfectly legitimate to manipulate the selection criteria. The objects of the manipulation were: first, to ensure as far as possible that in this relatively small cohort, which was going to be studied intensively, a reasonable number of children - i.e. those with stunted sibs - would become stunted; secondly, the infants with normal sibs would serve as a comparison group - from an environment that, as far as we could see, was identical.

The children were examined monthly for the first year and 3-monthly for the second year, after which the study ended. At 2 years they were divided into 3 tertiles of height-short, medium and tall. Table 1 shows that there was a significant correspondence between the grouping at selection and the grouping at 2 years, so that one could to a certain extent predict the development of stunting. Table 2 shows the results when the children were classified in tertiles of length at 2 years. Even the most stunted group had on average a satisfactory birth weight, so that it seems doubtful that one could regard the postnatal stunting as a result of prenatal growth retardation. Moreover, the mothers of the short children were significantly shorter than those of the taller children. Thus we might say that we were dealing with stunted families. I will return to this point later.

Table 1. Fate of children from stunted families

Initial classification


At two years: Proportion (%)stunted

I Sibs N. N


5/30 (17)

II Sibs N. S or S. N


11/29 (38)

III Sibs S. S


11/21 (52)

Table 2. Results of classification by tertiles of length at 2 years

Maternal height (cm)

Birth weight (kg)

Birth length (cm)

Z score for length at 2 years
















At each visit there were 24-h recalls of dietary intake, inaccurate though we know this method to be, and spot urine samples were taken. At 1 and 2 years 24-h intakes were weighed and blood samples were taken-to do that more frequently was not thought to be ethical.

The only thing that came out of the dietary data was that calcium intakes and Ca/P ratios were very low by accepted standards (Table 3), and at 2 years the Ca intakes were significantly lower in the shorter children. In the urine there was a significant, though not very strong negative correlation between both Ca and P related to creatinine, and attained height-forage. These negative correlations might suggest that the taller children were retaining dietary Ca and P more efficiently. In blood we measured calcium and inorganic phosphate, alkaline phosphatase, albumin, ferritin, T4, somatomedin C and vitamin D, the last by the courtesy of Dr. Angela Fearney of St. Mary's Hospital and Dr. Nigel Belton of Edinburgh. There were virtually no differences, either between the tall and short children, or between the mean values and accepted normal concentrations, with one exception: the mean serum P was 4.5 mg/dl, compared with the supposedly normal concentration at this age of 6 mg/dl (Specker et al., 1986). I think that the possibility of P deficiency has to be taken seriously, because in this region of Thailand bladder stones used to be very common and Thai and American workers have provided a good deal of evidence that this may be related to a low intake of P (Valyasevi & Dhanamitta, 1967). Although the supposed P intakes of our children were not too bad compared with estimated requirements, absorption may have been less than that from Western diets. Thus the very extensive biochemical work and the analyses of it that were done by the Khon Kaen group sadly led to no very clear-cut results. However, there are possibly some hints or clues that might point us in the direction of calcium or phosphate.

Table 3. Khon Kaen study. Estimates of calcium intake and Ca/P ratio by tertiles of length at 2 years




Estimated Requirement*

At 1 year

Ca, mg/d










At 2 years

Ca, mg/d










* From Department of Health, 1991.

Table 4. Length (cm) of male Thai children compared with the NCHS reference



Age, months











B. Bangkok middle class






C. Bangkok slum






D. Chiang Mai






E. Khon Kaen


















Reproduced by permission from the European Journal of Clinical Nutrition. For sources of data see Chusilp et al. (1992).

I interpret these findings as follows: first, the cohort is a homogeneous population, in which are represented the genetic or familial differences in height that are found in any group of children. It is noteworthy that the dispersion of heights in our cohort is no greater than that of the NCHS data.

If I am correct that it is an essentially homogeneous group, one would not expect to find much in the way of biochemical or endocrine differences between the taller and shorter children. Secondly, in the group as a whole, height is shifted to the left by nearly 2 SD compared with the NCHS reference. I suggest that this is an environmental and not a genetic or ethnic effect, because, contrary to the findings of Davies (1988), the growth of middle class Thai children in Bangkok comes close to that of the reference (Table 4). Of course there might be true ethnic differences between the different groups in the Table. However, a recent study suggests that small differences between the children of two distinct Thai tribal groups could be attributed to cultural rather than ethnic factors (Omori & Greksa, 1993). As far as one can tell from that paper, there is virtually no difference between the heights of the Karen hill-tribe children and those of Chiang Mai (lowlands of North Thailand) or of Khon Kaen. I have thought it worthwhile to mention this study, because it relates to later papers in this workshop, and I will end by summarizing what I think are the workshop's main aims: first, to look at new data that have accumulated in the last 5 years, to see whether they give us any stronger clues to the nutritional components of stunting; secondly, to look at the physiology and biochemistry of linear bone growth, to identify, if possible, points of entry at which environmental effects might be most likely to distort the process.


Chusilp K, Somnasang P, Kirdpon W, Wongkham S, Sribonlue P, Mahaverawat U, Yongvanit P, Sawakontha S & Waterlow J (1992): Observations on the development of stunting in children of the Khon Kaen region of Thailand. Eur. J. Clin. Nutr. 46, 475-488.

Davies DP (1988): The importance of genetic influences on growth in early childhood with particular reference to children of Asiatic origin. In Linear growth retardation in less developed countries, ed. JC Waterlow, pp. 75-86. Vevey: Nestlé Nutrition/New York: Raven Press.

Department of Health (1991): Dietary reference values for food energy and nutrients for the United Kingdom. Report on Health and Social Subjects No. 41. London: HMSO.

Golden MHN (1988): The role of individual nutrient deficiencies in growth retardation of children, as exemplified by zinc and protein. In Linear growth retardation in less developed countries, ed. JC Waterlow, pp. 143-164. Vevey: Nestlé Nutrition/New York: Raven Press.

Nabarro D, Howard P, Cassels C, Pant M, Wijga A & Padfield N (1988): The importance of infections and environmental factors as possible determinants of growth retardation in children. In Linear growth retardation in less developed countries, ed. JC Waterlow, pp. 165-184. Vevey: Nestlé Nutrition/New York: Raven Press.

Omori K & Greksa LP (1993): Growth and health status of pre-school Karen highlanders. Ann. Hum. Biol. 20, 21-30.

Rutishauser IHE & Whitehead RG (1969): Nutritional status assessed by biological tests. Br. J. Nutr. 23, 1-13.

Seoane N & Latham MC (1971): Nutritional anthropometry in the identification of malnutrition in childhood. J. Trop. Pediatr. Environ. Child Health 17, 98-104.

Specker BL, Lichtenstein P, Mimouni F, Gormley C & Tsang, RC (1986): Calcium-regulating hormones and minerals from birth to 18 months of age: a cross-sectional study. II: Effects of sex, race, age, season and diet on serum minerals, parathyroid hormone and calcitonin. Pediatrics 77, 891-896.

Valyasevi A & Dhanamitta S (1967): Studies of bladder-stone disease in Thailand. VII. Urinary studies in newborn and infants of hypo- and hyperendemic areas. Am. J. Clin. Nutr. 20, 1369-1377.

Waterlow JC (1973): Note on the assessment and classification of protein-energy malnutrition in children. Lancet ii, 87-89.

Waterlow JC (ed.) (1988a): Linear growth retardation in less developed countries. Nestle Nutrition Workshop Series no. 14. Vevey: Nestle Nutrition/New York: Raven Press.

Waterlow JC (1988b): Observations on the natural history of stunting. In Linear growth retardation in less developed countries, ed. JC Waterlow, pp. 1-16. Vevey: Nestle Nutrition/New York: Raven Press.

World Health Organization (1983): Measuring change in nutritional status. Annex 3. Reference data for the weight and height of children. Geneva: WHO.

World Health Organization (1987): Global surveillance through anthropometric measurements. WHO Weekly Epidemiological Record, March 1987, 1-20.

Contents - Previous - Next