Previous - Next
1. Introduction
2. Study design and methods
3. Factors determining early and late faltering
4. Conclusions
References
Discussion
Reference
D. Skuse, S. Reilly and D. Wolke
Behavioural Sciences Unit, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK
Correspondence to: D. Skuse.
For many years it has been widely accepted that an adverse family and social environment can retard physical growth (Department of Health and Social Security, 1976), but there has been relatively little discussion of the mechanisms by which this happens. What are the processes by which social disadvantage results in stunted children in relatively prosperous societies such as our own? Such knowledge is essential if assistance is to be provided for those in the community whose children are at risk. What forms of intervention are likely to be effective in ameliorating the consequences of such social deprivation (Skuse, 1989)? During the 1940s there were reports that emotional deprivation, as such, could influence growth, linked to claims that deprivation during the preschool years usually led to irreversible effects upon psychosocial functioning in later childhood and even adult life. Spitz (1945) described the growth of infants in a foundling home where they had received 'excellent food' yet little individual attention outside feeding times. Growth in both weight and length were severely retarded and Spitz reports that "The physical picture of these children (when 2-4 years of age) impresses the casual observer as that of children half their age". Talbot et al. (1947) described a series of 51 children (aged 21/2-15 years) referred to Massachusetts General Hospital who had stunting "with a height age no more than 80% of the actual age". A subsample of 29 children were studied in more detail, from the point of view of their dietary intakes. The nutritional histories of the majority indicated that they had major feeding difficulties which persisted for most of their lives, in many cases since infancy. These were associated, in 60% of the children, with some degree of socioemotional adversity within their families. Attempts to improve the diets of the 66% of the subjects who were underweight for their height met with little success because the children could not be induced to become interested in the idea of eating more food; however, the investigators entertained the hypothesis that nevertheless a number of them had been chronically undernourished. Following 'psychiatric therapy', three subjects showed an improved appetite and, following that change in appetite, demonstrated catch-up growth. An example is given of such catch-up growth commencing between 7 and 8 years of age. Although it was not possible at the time to test the hypothesis, the investigators proposed that the ultimate cause of the children's abnormally slow growth had been a deficiency of growth hormone and that they had tended to adapt themselves to nutritional privation in early life by becoming 'relatively hypopituitary', but the hypopituitarism did not usually reverse when energy intakes were improved. It is of interest to note that 24% of this subsample were considered to have been emotionally rejected by their mothers.
A third account, from the same period, on the growth of children in conditions of psychosocial deprivation or adversity was that of Widdowson (1951). The story of her serendipitous observations is well known. In 1948 she was stationed with an army medical unit in a town in the British zone of occupation in Germany where two small municipal orphanages were located. Each housed around 50 boys and girls between 4-14 years. The children had nothing except rations to eat and they were well below normal in both height and weight. The medical unit instituted a programme of physical examinations of the orphans every two weeks and continued these observations for 12 months. During the first six months the orphans received only the official rations. During the last six months the children in orphanage A received in addition unlimited amounts of bread, an extra ration of jam and a supply of concentrated orange juice. The matron of orphanage A at the start of the study was a cheerful young woman who was fond of the children in her care. In contrast, the matron in charge of orphanage B was older, stern and a strict disciplinarian towards all the children in her care except for a small group of favourites. It so happened that, at the end of the first six months, the cheerful matron left orphanage A for other employment, and the disciplinarian was transferred from orphanage B to orphanage A, bringing her eight favourites with her.
The physical examinations of the children's growth revealed that, during the first six months, the weight gain by the children in orphanage A was substantially more than the weight gain by the children in orphanage B, although the strict matron's favourites did much better than the rest of the children. The shift in matrons then occurred, and this coincided with the provision of extra food for orphanage A. During the next six months the children of orphanage B, whose food supply had not increased but who no longer had the strict disciplinarian in charge of them, showed a rapid rise in weight. Yet, in spite of their improved nutrition, those children who were now subject to the regime of the disciplinarian matron in orphanage A continued to gain weight at about the same rate as before. Trends in weight and height were very similar; the matron's favourites exceeded all others in terms of their rate of growth.
The prevailing view then, during the 1950s and 1960s, was that socioemotional deprivation could indeed be the cause of some cases of abnormally short stature, and that the most likely aetiology was 'deprivation' (or the lack of mothering in a broader sense) (Elmer, 1960; Patton & Gardner, 1962; Coleman & Provence, 1957) and that the failure to thrive was mediated either through diminished intestinal absorption, inefficient utilisation of energy, or possibly some abnormality of endocrine function (Coleman & Provence, 1957; Blodgett, 1963; Leonard, Rhymes & Solnit, 1966).
Then, in 1969, Whitten and colleagues (1969) set out to test the hypothesis that, rather than some subtle influence of deprivation or neglect upon metabolic functioning, the ultimate aetiology of children's growth failure in conditions of socioemotional adversity was simply an inadequate energy intake. Whitten observed that former researchers had all assumed, but not determined, the adequacy of the children's energy intakes during the period when "unfavourable emotional forces were operative". The study was of a series of 16 infants aged between 3-23 months, who were seriously failing to thrive. Their mean length at the time of being taken into the study was -3.06 Z (±1.25 SD) and their mean weight was -3.26 Z (±0.9 SD). By means of an ingenious design they showed, reasonably convincingly, that despite the protestations of the mothers that they had been feeding the children sufficiently, if they were indeed fed 'adequate energy', either during a hospital admission or at home (before or after the hospital admission) they did gain weight, whether or not they received additional stimulation or maternal attention. The authors conclude that "maternally deprived infants are underweight because of undereating, which is secondary to not being offered adequate food or to not accepting it, and not because of some psychologically induced defect in absorption or metabolism". However, they concede that "our evidence that underfeeding is the aetiologic factor in the growth failure of maternally deprived infants is based entirely upon weight gain". In fact the children's average weight for length at the time of admission to the study was -1.8 Z (±0.92 SD); they were clearly not simply stunted and the great majority (88%) were seriously underweight for their length.
The findings of Whitten et al.'s study had the effect of changing the emphasis of research on the association between growth impairment and emotional deprivation and abuse, to focus upon nutritional influences. Now it was reckoned that mothers' accounts of what they fed children who were suspected cases of deprivation could not be true, and that such children were probably being starved. Reviewing the current state of knowledge on the effects of emotional disturbance and deprivation on somatic growth MacCarthy (1974) concludes "hormonal dysfunction appears to be minimal... if there is an explanation, it is probably more in terms of nutrition than of endocrine disturbance".
Krieger (1974) also wrote that maternal rejection could cause a 'psychosocial deprivation syndrome' in which linear growth failure and retarded bone age were associated with persistent restriction of food by mothers who abused their children physically and who had characteristic personality traits. MacCarthy (1981) stated that 'deprivation dwarfism' caused by malnutrition in early childhood, particularly in the first two years of life, was likely to be due to undernutrition", adding "there are strong indicators, short of absolute proof, that this undereating is due to 'under-giving' by the mother and the additional element is the child, who being underfed, becomes undemanding". A reasonable consensus, then, existed in the 1970s that failure to thrive in the early preschool years was due to chronic undernutrition, although there was some perplexity about the child's role. As MacCarthy (1981) put it "why did not these babies cry in hunger, or if they did why were their cries ignored?".
However, another condition was also causing perplexity, for some older children who had excessively short stature and came from homes where it was thought that they were being neglected or abused were reported by their caretakers to have huge appetites, eating far more than their siblings, yet they failed to grow (Powell, Hopwood & Barratt, 1973). This condition has been termed 'psychosocial dwarfism' (Money, 1992), but it is not found as a diagnostic category within ICD-10 (World Health Organization, 1992) or DSM IV (American Psychiatric Association, 1993). The diagnostic criteria for that condition have been summarised by Green (Green, Campbell & David, 1984; Green, Deutsch & Campbell, 1987). The diagnosis has rarely been made before the age of 2 years. In virtually all cases there is said to be a severely disturbed relationship between the primary caretaker and the child, which is the main cause. The most typical behavioural symptoms are bizarre patterns of eating and drinking, with gorging and vomiting, polyphagia and polydipsia. Growth failure usually begins in infancy, but in some cases growth retardation is said to have commenced as late as 6 or 7 years (Ferholt et al., 1985). The children are stunted with a near-normal weight for height. Body proportions are allegedly immature (MacCarthy, 1981) and often there is microcephaly (Marks et al., 1978). Symptoms can be considered under four broad headings: disorders of biological rhythms, of self-regulation, of mood, and of social relationships (Skuse, 1993a). Relevant biological rhythms include sleep, appetite and growth hormone release. Sleep is disrupted, with frequent wakings and night wanderings, often in search of food. Active REM sleep may be increased and stage 3 and 4 NREM sleep decreased (Taylor & Brook, 1986). Appetite is disturbed, normal hunger and satiety rhythms are lost, and there is an apparent inability to achieve satiation. Consequently, these children habitually take food which they have been forbidden, steal food from their peers and-given the opportunity gorge themselves until they vomit. Growth hormone release shows a diminished pulse amplitude (Stanhope et al., 1988), meaning that the cumulated 24 hour circulating levels of the hormone are severely curtailed. Disordered self-regulation means deviant patterns of defecation, urination and attention. It is not uncommon to encounter a history which goes far beyond simple encopresis and enuresis to encompass features in which those bodily functions have acquired the quality of aggressive acts. Deliberate urination over others' belongings, and the concealment of faeces or soiled clothes in public places are typical. Attention span is almost invariably brief. Receptive and expressive language skills are impaired; the children also have poor non-verbal skills and poor practical reasoning abilities (Skuse, 1993b). At school they underfunction, and may require special education for moderate learning difficulties. Mood is almost invariably characterised by depression and poor self-esteem. Social relationships are always severely impaired; these children are disliked by virtually everybody with whom they come into contact including siblings, peers, and schoolteachers as well as their parents. External validation for the condition comes from a fairly well recognised aetiology, primarily emotional abuse (which may be a necessary criterion) but the condition is often associated with physical abuse and sexual abuse too. The natural history is less well documented. A cardinal feature is the potential for reversibility of all symptoms when the child is removed from the abusing environment (Skuse, 1993a; Money, Annecillo & Kelley, 1983; Skuse, 1992; Albanese et al., 1993). The rate of change in mental growth seems to run parallel to that of physical development, and is reflected by a progressive increase in intelligence. The bizarre behavioural features are usually lost within a few weeks of the child being taken into a caring and nurturing family, but if they persist for months or years the possibility that the abuse is also persisting should be considered. The increase in height velocity is often dramatic, and pathognomonic. Children's stature 'catches up' to a trajectory far closer to their genetic potential.
Accordingly, over the past few decades two main subgroups of children with short stature associated with socioemotional deprivation or adversity have been identified. It has come to be appreciated that there could be several quite different ways, involving contrasting mechanisms, in which early experiences can have enduring effects upon linear growth. At one extreme there is growth failure of very early onset, certainly within the first two years of life, in which the aetiology is thought to be undernutrition. The infants with this condition tend not to complain that they are being undernourished, but it has been suggested that the reason they do not complain is that their malnourishment has rendered them apathetic (e.g. Graves, 1976). Such children are said to show reduced activity levels which could be construed as a reductive adaptation to their nutritionally depriving environment (Waterlow, 1984). Once stunting is established within the first three years, so it is hypothesised, the children are unlikely to increase their intake even if offered more food at a later age (Martorell, 1985).
On the other hand, a further subgroup of stunted children from home environments which lack emotional nurture display a very different range of symptoms (Green, Campbell & David, 1984; Green, Deutsch & Campbell, 1987). The children show voracious and bizarre appetites, polyphagia, polydipsia and a variety of other odd behavioural features. Whilst it used to be thought that food restriction was the cause of their obsession with eating (Krieger, 1974), in recent years the view has been expressed that the condition has little to do with malnutrition; although the two disorders may co-exist in the same child, the primary aetiology is a disturbance in growth hormone dynamics (Albanese et al., 1993; Skuse et al., 1993).
The research to be presented
comprises part of a series of epidemiological studies on the subject of individual and
environmental variables associated with failure to thrive. We are not concerned here with
'psychosocial dwarfism', a far rarer condition, the clinical features of which become
manifest in later childhood and, as discussed above, are quite unique. Previous work has
demonstrated that stunting in the preschool period usually begins as growth faltering in
infancy; persistent growth failure from infancy to 4 years of age is associated with a
very poor outcome in terms of cognitive and behavioural development (Dowdney et al., 1987;
Skuse, Wolke & Reilly, 1992). Data will be presented to show that infants whose growth
falters early on in postnatal life, who live within a socioeconomically disadvantaged area
and who are at relatively high risk of inadequate parenting, have contrasting outcomes
depending upon whether that growth failure begins immediately after birth or later. We
will present data to show that each broad trajectory of growth in the first year, measured
in terms of weight gain, is associated with contrasting family circumstances and degree of
psychosocial disadvantage. Those with the later onset of failure to thrive, after six
months or so of age, had mothers who were subject to greater social and psychological
adversity, and who, for a variety of reasons, might have been expected to provide a poorer
quality of parenting experience. Yet developmentally these very same children had a better
outcome in the second postnatal year than the 'early' failure to thrive group.
The investigation was a prospective longitudinal survey of virtually all infants born during one calendar year within an inner city health district in London, England. The district has an ethnically diverse population (140,000), which in socioeconomic terms is relatively homogeneous and quite severely disadvantaged (South East Thames Regional Health Authority, 1984).
The sampling frame comprised all 2510 births that were registered with participating child health clinics or family doctor practices in 1986. The planning and execution of the survey was facilitated by the good relationship built up between the research team and the local community paediatric services in the course of previous research there (e.g. Dowdney et al., 1987).
The selection of subjects who had growth faltering was made from all children who continued to live in the district until 12 months of age, and who were weighed on at least one occasion. During this period 14.4% of the population moved out of the area. There were also missing data on a small minority of subjects (1.2%) who were known to health visitors but were seen only at home. Other families were untraceable (3.3%). A small proportion of infants (0.8%) are known to have died.
Growth trajectories were computed from weight data recorded at clinic visits and expressed as Z scores (Hamill et al., 1979; Jordan, 1986). Scores have been interpolated to target ages 4 weeks, 6 weeks, 3 months, 6 months, 9 months, 12 months and 15 months. In virtually all cases the interpolations were made where at least one of the data points was within 1 month of the target age. Quadratic (three point) interpolation was undertaken by means of the NEWLONG program developed by Carter (1985).
2.1. Definition of growth faltering
Potential cases were limited to singleton deliveries at term, i.e. between 38 and 41 completed weeks gestation. Preterm infants (gestation 37 weeks 6 days or less) were excluded because of the known association with below average growth in the early postnatal period (Brandt, 1986; Ounsted, Moar & Scott, 1982). Also excluded were infants with severe intra-uterine growth retardation (i.e. birthweights at or below the 3rd centile, on charts standardised for gestation, sex, ordinal position, maternal height and mid-pregnancy weight; Tanner & Thomson, 1970). Confirmed cases of growth faltering had to have a weight for age Z score (WAZ) below -1.88 (corresponding approximately to the 3rd centile). This trajectory relative to population norms had to be attained by 12 months of age, and sustained for three months or more. All children with a suspicious weight gain trajectory, on the basis of clinic data, were traced and visited at home for confirmatory anthropometry.
There were 1554 potential subjects remaining after exclusion criteria had been implemented. Of those, 52 (3.3%) were confirmed cases of growth faltering at 12 months of age. Three of these were excluded because of an overt organic disorder that was considered to be making a major contribution to the infant's poor rate of growth. Further information about the study design and outcome can be found in Skuse, Wolke & Reilly (1992).
2.2. Pattern of growth in first year
Preliminary exploration of the data led to the conclusion that the children could be approximately divided into two subgroups, those in whom the onset of growth failure was before 6 months of age, and those in whom it was later.
This was done according to whether the difference in weight between birth and 6 months was greater than or less than half the difference between birth and one year. The former suggests that the greater part of the faltering had occurred in the first 6 months and the child was classified as early faltering (early failure to thrive; FTT) (Fig. 1); the latter implied that it had occurred in the second 6 months, and the child was classified as late faltering (late failure to thrive; FTT) (Fig. 2). When the mean values for those trajectories were plotted (see Fig. 3) significant differences in standardised weight were found at 3, 6 and 9 months. The early FTT infants actually had higher birthweights than the others.