Debate: Idiopathic short stature should be treated with growth hormone


  • Geoffrey R Ambler,

    Corresponding author
    1. Institute of Endocrinology and Diabetes, Sydney Children's Hospitals Network (Westmead Campus)
    2. Discipline of Paediatrics and Child Health, The University of Sydney, Sydney, New South Wales
    Search for more papers by this author
  • Jan Fairchild,

    1. Department of Paediatric Endocrinology and Diabetes, Women's and Children's Hospital
    Search for more papers by this author
  • Dominic JC Wilkinson

    1. The Robinson Institute, Discipline of Obstetrics and Gynecology, University of Adelaide, Adelaide, South Australia, Australia
    2. Program on Ethics and the New Biosciences, Oxford Uehiro Centre for Practical Ethics, The University of Oxford, Oxford, United Kingdom
    Search for more papers by this author

Professor Geoffrey Ambler, Institute of Endocrinology and Diabetes, The Children's Hospital at Westmead, Locked Bag 4001, Westmead, Sydney, NSW 2145, Australia. Fax: +61 29845 3170; email:


In this paper we outline the case for and against the treatment of idiopathic short stature with growth hormone. Drs Ambler and Fairchild argue that many of those with ‘idiopathic’ short stature are not ‘short, normal children’ and will ultimately receive molecular diagnoses. They also argue that there is a subset of children who suffer negative psychosocial consequences of their stature for whom growth hormone therapy is effective. Growth hormone has a very good safety record and is likely to be as cost-effective in idiopathic short-stature as in some other conditions that are currently funded. Dr Wilkinson counters that short stature is not associated with physical or psychological illness, and that there is no evidence that growth hormone improves psychological or physical wellbeing. Moreover, growth hormone for idiopathic short stature represents a form of enhancement rather than treatment, and is not a fair use of resources. Socially mediated disadvantage should be treated by attention to prejudice and not by hormone treatment.

The case for considering growth hormone therapy in children with ‘idiopathic’ short stature

Geoffrey Ambler and Jan Fairchild

Short stature is a common reason for referral to paediatric endocrinologists and paediatricians. In those children with severe short stature, a variety of medical conditions need to be excluded by careful history, examination and investigation. However, in a significant number, depending on definitions and the rigour and availability of modern molecular diagnostics, no cause is identified and these children are labelled as having idiopathic short stature (ISS).1

ISS can be defined as a condition in which the height of an individual is more than 2 standard deviation scores below the mean height for age and sex in the population, without evidence of systemic, endocrine, nutritional or chromosomal abnormalities.2 Specifically, children with ISS have normal birthweight and are GH sufficient. This statistical definition includes children with familial short stature and constitutional delay of growth and puberty (CDGP); however, when these normal variants are excluded, we are left with a heterogeneous group of children with less severe defects within the growth hormone–insulin-like growth factor 1 (GH-IGF1) axis and other as yet unidentified causes of short stature.

Determining growth hormone deficiency (GHD) based on pharmacological stimulation tests has considerable difficulties and limitations. There is no agreed standard that adequately represents the full spectrum of the disorder, and in practice, the definition of GHD is not based on a gold standard but on the combination of short stature, slow growth and suboptimal levels of GH during two stimulation tests. Assay variability, intra-individual variability, the arbitrary definition of normal, as well as the influence of age, sex, pubertal staging, nutritional status and BMI, all complicate assessment.3 With integrated GH secretion being a continuous variable, it is impossible to define a precise cut-off point that separates normal from abnormal, and only severe GH deficiency can be diagnosed with any confidence. In addition, accurate diagnosis is further complicated by our limited understanding of physiological and pathological variations in GH responsiveness at the growth plate and other tissues.

Acknowledging these difficulties, the Australian GH programme introduced both biochemical and auxological eligibility criteria for GH treatment in September 1988. Those children with severe GHD, defined as a peak GH < 10 mIU/L on two stimulation tests, are eligible under biochemical GHD criteria. Those children with a peak GH > 10 mIU/L but short stature and slow growth (height < 1st centile and growth velocity for bone age < 25th percentile) are eligible under auxological criteria, provided other causes of short stature have been excluded and they do not have CDGP. Those qualifying under auxological criteria in Australia include many who would be considered to have GHD in other countries, where GHD is defined as peak GH < 20–30 mIU/L. The remainder include children with short stature due to being born small for gestational age (SGA), various syndromes, chronic disease and ISS.

The number of short children labelled as having ‘idiopathic’ short stature will continue to fall as our understanding of growth disorders increases.4 It is inevitable that specific molecular diagnoses will be increasingly identified, which are likely to include a range of defects in the GH-IGF-1 axis and response pathways. Short stature homeobox-containing gene (SHOX) disorders are an example of a recently identified subgroup of ISS in which GH secretion is normal but GH responsiveness is impaired.5 There is also increasing information suggesting that epigenetic changes have an important impact on growth;6 however, the mechanisms remain unclear.

The function of the GH-IGF1 axis may be disturbed by defects at many points along this axis, with the range of defects representing a continuum from GHD to GH resistance.7 Understanding more about these defects will help guide appropriate investigation and management in the future. It is important to note that children qualifying under auxological criteria are as short or shorter than others treated within the GH programme for GHD, Turner syndrome and chronic renal insufficiency.8 It would seem short-sighted to disqualify this group from consideration of therapy based on the inability of medical science to define their pathology.

Unsurprisingly, the response of children with ISS to GH therapy is highly variable reflecting the heterogeneous nature of this group and is dose dependent. The mean increase in adult height attributable to GH therapy (average duration 4–7 years) in children with ISS is 3.5–7.5 cm compared with historical controls, pretreatment height prediction and non-treatment or placebo control groups.8–13 It is likely that adult height outcome would be further improved by earlier treatment and higher GH doses, as has been shown in other conditions such as Turner syndrome. Multiple factors affect growth response; however, children who are younger, heavier, short relative to their family or on higher GH doses respond best. Australian data published under the OZGROW programme (National Growth Hormone database of the Australasian Paediatric Endocrine Group) in 1996 and 20118,14 supports the need for higher than physiological replacement doses in this group. A better growth response was demonstrated in those with clear GHD than in those qualifying on auxological grounds when treated with physiological replacement doses.8 Pharmacogenomic approaches to GH therapy are also being actively investigated and are likely to influence patient selection and dosing in the future.15

It is important to consider whether short stature in itself is associated with negative psychosocial or quality of life consequences and whether there is improvement with GH therapy. If not, why are significant numbers of otherwise well children with short stature referred to growth physicians by well-intentioned general practitioners and paediatricians at the request of the children and families themselves? These questions are methodologically difficult to study and have not yet been conclusively answered.16–18 Some studies have reported chronic psychosocial stress in children with short stature although without clear evidence of clinical psychological disorder. Reports on low self-esteem and emotional problems associated with teasing or bullying have been variable. A clearer conclusion from several studies is that parents report that short children have lower social competence and show more social problems than normal statured children;18 whether or not this reflects broader societal bias or insightful individual concern is unclear. It is also likely that children with short stature and negative psychosocial consequences are a subset, because there is reasonable evidence that such problems are absent in short children who are not medically referred.

Data are also conflicting on the consequences of short adult stature and confounded by comorbidities. Some data, but not all, suggest lower rates of marriage and employment and lower income associated with shorter stature. A strong association between short stature and increased rates of suicide has been reported in a large cohort of men based on Swedish national registry data; short men had twice the risk of suicide than tall men and each 5 cm greater height was associated with a 9% lower risk of suicide.19

Two studies of interest were published in 2011. Stephen et al.20 reported comprehensive health-related quality of life (HRQOL) data on short stature and GH treatment in a clinic-based population. Short children reported lower emotional and social functioning than healthy peers, with greater deficits in the short children not treated with GH. Parent reports were in agreement. After 1 year of treatment, GH-treated subjects also reported an improvement in physical functioning. These authors have recommended application of certain HRQOL scales to proactively identify short children at risk of negative consequences. Also, in a referred clinic population in Sweden, Chaplin et al.21 reported adverse behavioural, mood and self-esteem findings in short children that improved after 3 months of GH treatment and were maintained over 24 months of treatment. At baseline, 14% of short children had a depression rating scale within the recommended zone for clinical intervention, falling to 5% after 2 years of GH treatment and compared with 2% in the reference population.21 Improvements were better for those with ISS than GHD, suggesting other organic confounders in GHD. The rapid improvements seen in this study suggest that acknowledgement and action to address the child's concern of short stature are an important factor, even before this has translated into height gain.

Based on the available evidence, albeit incomplete, we argue that there is a subset of children with ISS who suffer significant negative psychosocial and quality of life consequences as a result of their stature and that these can be improved by GH therapy. This subset appears to be self-identifying through seeking clinical referral and intervention. While alternative approaches, such as cognitive behavioural therapy and assertiveness training, have been proposed,22 we are not aware of data on their efficacy in this situation.

The justification of GH therapy also needs to include consideration of safety and cost. Since the availability of recombinant human GH, there has been detailed safety surveillance, particularly through post-marketing surveillance registries (summarized by Miller23). Particular interest for all GH indications has focused on risks of intracranial hypertension, scoliosis, diabetes, primary or secondary malignancy and mortality. Of these, only intracranial hypertension as a short-term and reversible complication is proven. Long-term surveillance studies continue such as the Safety and Appropriateness of Growth Hormone Treatments in Europe study. Preliminary data from the French subset of this study have suggested a possible increase in all-cause mortality in young adults with ISS or isolated GHD previously treated with higher GH doses (higher than ever used in Australia);24 however, the Swedish, Belgian and Dutch dataset reports no such increase.25 These data remain to be further assessed and validated, and regulatory authorities have not altered their safety recommendations.26 Overall, we contend that given its widespread use in excess of 25 years in a wide variety of indications, GH has a very good safety record and that carefully collected longer term safety data is likely to bear this out.

GH is currently an expensive medication that has prompted debate about its justification and cost-effectiveness for various indications. This has been recently considered in the UK by The National Institute for Health and Clinical Excellence27 whose committee concluded that somatropin represents a cost-effective treatment for children with growth failure associated with GHD, Turner syndrome, Prader Willi syndrome, chronic renal insufficiency, small for gestational age with growth failure at >4 years age and SHOX deficiency. The committee also noted that cost-effectiveness could improve considerably if the price of GH falls. Specifically, ISS was not included in this analysis, but we argue that the cost-effectiveness should be similar to any other condition in which height is the primary measured outcome (all those above, except Prader Willi syndrome). As we argue above, it is erroneous to exclude those currently classified as ISS from such considerations because of the major diagnostic overlaps, likelihood of hitherto unrecognized pathologies and lack of any differentiation in the severity of short stature or response to GH treatment.

In summary, we argue that GH therapy is justifiable in selected individuals with ‘idiopathic’ short stature based on:

  • 1the knowledge that children with short stature labelled ‘ISS’ are as short as many groups treated for known pathology and that treatment with GH has similar efficacy compared with other pathologies (except for severe GHD);
  • 2the high likelihood that ‘idiopathic’ short stature will be proven not to be idiopathic as further knowledge is gained and many or most of these children will be found to have defects in the GH-IGF1 axis;
  • 3there are sufficient data to be concerned that short stature has adverse psychosocial and quality of life consequences in some individuals and these are generally the group that present clinically;
  • 4GH has an enviable safety record in standard recommended doses and good mechanisms are in place for longer term surveillance; and
  • 5cost-effectiveness will increase if the cost of GH falls as economies and new developments occur in manufacturing. New pharmacogenomic approaches may also increase cost-effectiveness.

We conclude that GH programmes should continue to allow consideration of GH therapy for such abnormally short children provided that they have had appropriate evaluation by growth specialists, including full discussion about the current state of knowledge on risks and benefits.

Better not bigger: ISS should not be treated with GH

Dominic Wilkinson

Why should not we treat children with ISS? Drs Ambler and Fairchild have summarised the potential effects of GH on adult height and allude to the widely held view that short stature is associated with social and psychological disadvantage. However, there are both medical and ethical arguments against the use of GH in children with ISS. The medical arguments are reasonably straightforward.

1. Short stature is not associated with psychological or physical illness

The belief that short stature leads to problems with psychosocial well-being, self-esteem and achievement is widespread. In a recent national survey of US paediatric endocrinologists, more than more than 50% believed that reduced height sometimes impaired emotional well-being, while 25% believed that it often or always did so.28 Yet longitudinal community-based studies such as the Wessex Growth Study provide little or no support for this view.29 Children less than the 3rd centile at school entry were followed-up through childhood and compared with controls of average height. There were no significant differences in self-esteem or self-perception during adolescence in short children and no differences in psychological functioning in adulthood.30,31 The most likely reason for the pervasive impression that short stature causes problems is the effect of referral bias.29,32 Psychological problems or poor educational attainment may be blamed on short stature, when in fact the cause lies elsewhere.

2. GH does not improve psychological or physical well-being

Recombinant GH was approved for children with ISS by the Food and Drug Administration in the USA in 2001 and is widely prescribed for this indication in the USA. In other parts of the world, however, uptake has been slower. The European Medicines Evaluation Agency has still not approved GH for ISS.33 In Australia, some 154 children with ISS were treated between 1994 and 2007.14

However, randomised controlled trial evidence of substantial benefit is lacking. Systematic reviews found an average increase in adult height of 4 cm in GH-treated children compared with controls.9,34 The most recent review found only three randomised controlled trials of low or moderate quality.34 This effect is less than that seen in Turner syndrome or children born small for gestational age (6–7 cm).34 More importantly, there is no evidence that GH treatment improves psychological functioning or educational achievement.9,35

What about the ethical arguments?

3. GH for ISS is enhancement not treatment

Children with ISS do not have a measurable or detectable abnormality of physiological or organ functioning. Their problem, if it is a problem, is that they are at the left end of the bell curve of normal growth. The use of GH in this group of children is not treatment for an illness. Instead it is what is sometimes referred to as an ‘enhancement’– an intervention to improve a ‘normal’ function or physical feature.36

The conventional view is that the proper role of medicine is to treat illness or disease but not to enhance or improve normal function. In fact, there is considerable debate within the bioethics literature about whether there is or should be a clear distinction between treatments and enhancements.36 Labelling something a disease or disorder involves a value judgement about that condition. There has been considerable attention in the last decade to the way in which pharmaceutical industries have been involved in the generation of new disorders. Medical diagnoses are created at least partly in order to generate a market for new pharmaceutical agents.37,38 We could create a definition for the shortest individuals in our population and call it a disorder, and arguably the medical definition of ‘idiopathic short stature’ does just that.38,39

Why should doctors provide GH as a treatment but not as an enhancement? One reason is provided in the previous section – that GH may be less effective in patients without a defined illness. However, even if it were effective, there are two reasons why it should not be used for this indication.

4. Socially mediated disadvantage should not be treated with hormones

There is some doubt that ISS leads to any significant reduction in psychosocial function overall. It is possible, though, that some children are disadvantaged by their stature. Short boys in the Wessex Growth study were twice as likely as controls to report being bullied, and significantly more likely to report being upset by bullying.40 Children with short stature may also be treated as if they are younger than their age. One survey in a clinic population suggested that ‘juvenilization’ of children with short stature was common.41

However, if some children do fare worse because they are smaller than their peers, it is far from clear that the appropriate response to this is to attempt to make them taller. The disability rights movement has pointed to the ways in which society's attitudes and behaviour creates disadvantage for those with physical impairment.42 This provides a strong argument for addressing those attitudes and that behaviour rather than attempting to change the physical features of those who are different from the norm. Short children and adults may be disadvantaged by prejudice by what we could call ‘heightism’.36 An analogous practice (that is hard to imagine today): in the 1950s and 1960s, tall girls were given oestrogens in an attempt to reduce their final height to less than 5′10″ to prevent psychosomatic illness and to increase their chances of finding a husband.39 Those girls were given medical treatment (subsequently demonstrated to impose significant risk) to enforce social values that have changed within a generation and to mitigate a perceived disadvantage that was wholly mediated by prevailing discriminatory attitudes.

We should work towards removing height prejudice in our society rather than attempting to reduce the numbers of those who are potential targets for heightism.

5. Enhancing stature for children with ISS is not a fair use of resources

The cost of recombinant GH has been recently estimated at $US27 000 (£17 000) per year of treatment, or $36 000 per centimetre of adult height gained.34 In 2006, a cost-effectiveness study of GH found an average per-patient cost of $100 000 and noted that the total potential cost of treating all eligible children in the USA with GH was $40 billion.43

The fact that a treatment is expensive does not mean that it is not worthwhile. However, the important question is whether, in a situation of limited healthcare resources, the public funding of treatment for short stature can be justified. By way of comparison, in Australia, the threshold for funding new pharmaceutical agents is a cost per quality adjusted life year of ≤$42 000. Given the lack of evidence of any effect of GH on patients' quality of life, GH for ISS cannot be cost-effective.

None of the above arguments mean that parents who wish to pay for GH for their children with ISS should be prevented from doing so, particularly where a small child has a strong desire to be taller. However, this should be seen as a cosmetic enhancement and not a medical treatment. GH appears to be safe,44 although the burden on children of hundreds or thousands of injections over a period of years is not insignificant, and small or delayed risks of treatment may take a long time to emerge.45

One interesting point to note is that while the above arguments provide a strong case against the public funding of GH for children with ISS, they also (with the exception of argument 3), provide a case against the funding of GH for some other ‘pathological’ causes of short stature, for example children born small for gestational age, or with genetic syndromes that predispose to short stature. If there is a psychosocial disadvantage for some children from being short, it does not seem likely that this will be any greater for a child with a defined cause of their short stature.

GH is not an effective therapy for ISS. There is no clear evidence that it improves the well-being of children who are treated. It achieves only modest gains in adult height and may confer a risk to the child. Parents should be allowed to pay for GH for their child if they choose, but disadvantage arising from ISS, where it exists, is socially mediated and should not be treated pharmacologically. Finally, even if GH were to be shown to produce measurable psychological benefits, it should not be provided as a publicly funded treatment. There are other far more pressing healthcare needs that are currently unfunded or underfunded. The first job of a paediatrician is to make children better, not bigger.