Growth Hormone Outgrows Growth

[~MaryO~]

From Clinical Endocrinology

Growth Hormone Outgrows Growth

Posted 07/26/2004

S. M. Shalet

Abstract and Introduction

Abstract

Growth hormone (GH) replacement has been offered to GH-deficient (GHD) children for approximately 40 years whereas it has only been a licensed indication for the treatment of GHD adults since 1996. Nonetheless, the advent of GH replacement for adult GHD patients (Jorgensen et al., 1989; Salomon et al., 1989) has proved informative about the overall management of the GHD child and teenager; equally, the longstanding experience of the paediatric endocrinologist with GH replacement has provided some guidance about potential pitfalls in the diagnosis and management of the adult GHD patient. It is the knowledge exchange at this professional interface that forms the focus of this article.

Introduction

The type of underlying pathophysiology differs in childhood-onset (CO) compared with adult-onset (AO) GHD. In childhood the commonest aetiology is isolated idiopathic GHD (August et al., 1990; Vanderschueren-Lodeweyckx, 1994), a blanket term that includes some children with distinctive pathophysiology that may be demonstrated radiologically, and others in whom the pathological insult is unknown and the explanation for the GHD ill-understood. By contrast, AOGHD is most frequently due to a pituitary adenoma and/or treatment with surgery or radiotherapy (Rosen & Bengtsson, 1990; Toogood et al., 1994). In GHD children poor growth dominates the clinical imperative to offer GH replacement whereas in adult life there is no single symptom or sign that is pathognomonic of GHD; nonetheless, GHD in adult life is associated with increased fat mass, particularly distributed in the truncal region (Salomon et al., 1989; DeBoer et al., 1992), reduced lean mass (Salomon et al., 1989; DeBoer et al., 1992), osteopenia (Johansson et al., 1992; Kaufman et al., 1992; Holmes et al., 1994), an adverse cardiovascular risk factor profile (Amato et al., 1993; Beshyah et al., 1994; Johansson et al., 1994, 1995), reduced exercise capacity (Cuneo et al., 1991; Nass et al., 1995) and reduced quality of life (Rosen et al., 1994). In addition, the standardized mortality rate is increased in hypopituitary adult patients on full replacement therapy with glucocorticoids, thyroxine and sex steroids but in whom GHD remains untreated; thus the increased mortality has been attributed to the GHD to a varying degree dependent on the interpretation of the literature by different endocrinologists (Murray and Shalet, 2000).

Dose requirements and the biochemical severity of the hormone deficiency also vary between GHD children and adults; children with all grades of GHD from mild to severe, but only adults with severe GHD, are considered for GH replacement and the replacement dose of GH is significantly greater in children than in adults, the variation in dose requirement reflecting the evolution of change in endogenous GH secretion with age in normal individuals.

Diagnosis

GH secretion is a continuum between normality and abnormality; therefore, with rare exception the diagnosis of GHD must be made on arbitrary grounds. The more severe the GHD, the less arbitrary the diagnosis, whereas the lesser degrees of GHD (GH insufficiency) merge into normality. In recent years more normative data about the GH response to the entire range of provocative stimuli in childhood have become available (Zadik et al., 1990; Marin et al., 1994; Ghigo et al., 1996); however, over the years the threshold level used to characterize a normal GH response has been defined arbitrarily. Initially it was set at 5 ng/ml (considered equivalent to 10 mU/l at the time) based on the results of studies such as those performed by Kaplan et al. (1965, 1968), in which 80 of 91 (88%) children who were not GHD showed a peak GH response to an insulin tolerance test (ITT) of > 5 ng/ml whereas all 53 GHD children had a peak GH response below 5 ng/ml to the same stimulus. The GH response setpoint was gradually moved to 7 ng/ml (Frasier, 1974) as GH testing became more common, and finally reached 10 ng/ml (20 mU/l) with the increased availability of biosynthetic GH. Not only is normality defined arbitrarily, but the same threshold GH level is used inappropriately to define normality irrespective of the pharmacological stimulus applied.

The gradual 'slippage' in the definition of GHD in childhood from 5 ng/ml to 10 ng/ml brought with it certain consequences; given the fact that GH secretion is a continuum between normality and abnormality, the higher the threshold for diagnosing GHD is raised, the more likely there will be overlap between normality and abnormality. Thus the relaxing of the biochemical definition of childhood GHD causes diagnostic problems and, at the same time, introduces therapeutic dilemmas because of the increasing mildness of the degree of GHD of some of the children now being considered for GH replacement; the milder the deficiency the less the expected impact of GH replacement on biological endpoints such as growth. In other words, how much difference does it make to offer standard GH replacement to a child with a peak GH response of 9 ng/ml rather than another child with a peak GH response to provocation of 11 ng/ml?

Perceived wisdom, at the time adult GH replacement was introduced, suggested that establishing a diagnosis of GHD in the adult might provide even greater diagnostic dilemmas than in the child, given that auxology was of no use in the adult. In practice this has not proved to be the case for the majority of adults considered for GH replacement. This is for two reasons; first, only adults with biochemically severe GHD are even considered for GH replacement; and second, there is a natural history of evolution of hypopituitarism in patients with a mass lesion of the hypothalamic-pituitary region or in those who have undergone surgery and/or radiotherapy to this region. GH is usually the first of the anterior pituitary hormones to be affected by these various pathological insults, which means that in a patient with multiple pituitary hormone deficits the probability of severe GHD being present is extremely high (Klibanski, 1987; Littley et al., 1989). Thus Toogood et al. (1994) showed that the median peak GH response to an ITT in four age-matched groups of adult patients with isolated GHD and GHD plus one, two and three additional pituitary hormone deficits was 3·8, 1·5, 0·8 and 0·7 ng/ml (1 ng/ml = 2·6 mU/l), respectively (Fig. 1). Of patients with two or three additional pituitary hormone deficits, 91% had a peak GH response less than 1·9 ng/ml. Similar conclusions regarding the relationships between GHD and additional pituitary hormone deficits have been drawn irrespective of whether GH status has been assessed by ITT (Weissberger et al., 1994), urinary GH (Bates et al., 1995) or 24-h spontaneous profile (Toogood et al., 1997b).

Figure 1. (click image to zoom) The distribution of the peak serum GH levels in response to an ITT in 190 patients divided into groups according to the degree of hypopituitarism present, that is the number of anterior pituitary hormone deficiencies, in each patient. Horizontal bars represent medians; 1 ng/ml = 2·6 mU/l. (a) GHD0; ( GHD1; © GHD2; (d) GHD3. Reproduced from Toogood et al. (1994).

The pattern of evolution of hypopituitarism in children with hypothalamic-pituitary disease is similar to that seen in adults; the observation is, however, less useful in the child in whom the commonest cause of GHD is 'isolated idiopathic'. Thus in the investigation of GH status in a child, a strategy consisting of two tests of GH status evolved; the rationale being based on the observation that any normal child might 'fail' any single GH provocative test. Additionally, in recent years with the advent of IGF-I and IGFBP-3 databases derived from normal prepubertal and pubertal children, the use of IGF-I/IGFBP-3 estimations has been increasingly incorporated into the investigation of the short child, either to replace one of the two GH provocative tests or to be performed in addition to the GH provocative tests (Juul & Skakkebaek, 1997). In the meantime, however, the 'two GH test' approach has never been formerly validated. This left some uncertainty as to how best to investigate GH status in an adult with partial hypopituitarism, a problem subsequently studied by Lissett et al. (1999); the latter group reviewed the results of GH provocative tests in 103 adult patients with documented or potential hypothalamic-pituitary disease and 35 adult normal volunteers. All patients and normal volunteers underwent an ITT and an arginine stimulation test (AST). Severe GHD was defined in a stimulus-dependent manner. Patients were categorized into groups according to the number of anterior pituitary hormone deficits present other than GHD. The concordance between the AST and the ITT (percentage of patients in whom both tests confirmed or refuted the biochemical diagnosis of severe GHD) was 100%, 76·8%, 66·6%, 83·3% and 92·3% in the controls and in those GHD patients with 0, 1, 2 and 3 additional deficits, respectively. Thus 16/69 GHD0, 5/15 GHD1, 1/6 GHD2 and 1/13 GHD3 patients were misclassified by one or other test.

In addition, the magnitude of the difference between the GH responses to the ITT and AST increased with the underlying mean GH value (mean of the peak responses to the two tests) (Fig. 2). Thus it is a constant ratio that links the GH response to an ITT and AST in an individual rather than a constant difference and the difference between the GH responses to two provocative stimuli is greater in those patients with milder degrees of GHD. There are implications from this study for investigation of GH status in both children and adults. In children it lends support to the idea that the two-test approach is increasingly worthwhile the higher the threshold definition of GHD is set, and in adults, while a single GH provocative test can be used with confidence in patients with two or three additional pituitary hormone deficits, in patients with suspected isolated GHD or with only one additional pituitary hormone deficit, two tests of GH status are required.

Figure 2. (click image to zoom) Magnitude of the difference between each individual's GH response to the ITT and AST is plotted against mean GH value: black square, normal individuals; black circle, patients. Spearman's rank correlation, r = 0·88; P

Thus far, the discussion about the number of tests that are required to establish GH status assumes that the information gained from each of the tests is the same and independent of the nature of the pathophysiological changes affecting the hypothalamic-pituitary region. The latter belief has been challenged by the data arising from recent studies; Darzy et al. (2003) investigated GH status in 49 adult patients, all of whom had received cranial irradiation for nonpituitary brain tumours or leukaemia, and 33 sex- and age-matched controls. A combined GH releasing hormone (GHRH) plus AST and an ITT were performed in all patients and controls on separate occasions. GH responses to either test were significantly lower in the patients than in the controls; in patients and controls the median peak GH response to the GHRH + AST was significantly greater than the response to the ITT. However, the ratio of the peak GH response to the GHRH + AST over that achieved with the ITT (discordancy ratio) was significantly higher in the patients compared with normals, consistent with dominant hypothalamic damage and relatively preserved somatotroph responsiveness.

The peak GH response to the ITT fell significantly within 5 years of irradiation with little further change over the subsequent 10 years. By contrast, the peak GH response to the GHRH + AST barely changed within 5 years of irradiation but subsequently declined significantly over the next 10 years (Fig. 3). Thus the evolution of change in GH responsiveness to the two different stimuli over time was markedly different, resulting in a significantly raised discordancy ratio of 6 within the first 5 years postradiotherapy, which then normalized (3-4) over the next 10 years (Darzy et al., 2003).

Figure 3. (click image to zoom) Box and whisker plots representing the peak GH responses to the ITT (a), the GHRH + AST ( and the discordancy ratio © in normals and patients according to the time interval since irradiation. The lower boundary of the box indicates the 25th percentile, a line within the box marks the median, and the upper boundary of the box indicates the 75th centile. Error bars above and below the box indicate the 90th and 10th percentiles, respectively. P-values are derived from comparison of each patient group with the normal control group. Note the BED (biological effective dose of irradiation reaching the hypothalamic-pituitary region) was not different among the groups. Reproduced from Darzy et al. (2003).

At a practical clinical level the discordancy between the GH test results is important; 50% of patients classified as severely GHD by the ITT were judged normal or only GH insufficient by the GHRH + AST. The bigger question arising from the observation that for the first 5 years after cranial irradiation the definition of GH status utilizing pharmacological tests is stimulus dependent must centre around the broader application of this principle; in other words are there other hypothalamic pathologies, such as infiltrative disorders, in which discordant GH responses may be seen to different pharmacological stimuli?Therapy The biological endpoint that has dominated GH replacement for GHD children is auxology. Growth is easily quantified in terms of current height, and height velocity over 12 months; it is an endpoint that is defined in the context of the normal population as well as the target height for the individual child, providing the heights of both parents are known. The rationale underpinning GH treatment for short children, however, has been less well substantiated -'whilst individual short children may show psychological stress, as groups (statistically) they do not appear to have clinically significant behavioural or emotional problems and it needs to be established whether being made taller produces measurable benefit in terms of academic or material success or psychological contentment. Currently there is no strong evidence that GH therapy improves psychological adaptation in short stature children' (Kelnar, 2003). Exactly the same comments might have been written about the GH-insufficient child treated with GH replacement.

A major question that arises from our inability to substantiate the rationale for childhood GH treatment is do we need to optimize final height in GHD children receiving GH replacement? In previous times, prior to the introduction of recombinant GH preparations, the supply of pituitary-derived GH was limited; therefore it was often considered adequate for the GH-replaced child to attain a final height within the normal population (3rd to 97th centile) range but not necessarily within the target height range. With the unlimited availability of recombinant GH preparations a much greater possibility exists for the child to achieve his or her target height range, but how much does such an achievement matter?

A recent study by Attanasio et al. (2002) has shed light on this question by determining body composition parameters in 92 COGHD patients, mean age 20·9 years, who had been treated to final height with GH for just under 9 years and had stopped treatment a mean of 1·57 years previously, but who remained severely GHD with an IGF-I level below the first centile of the age-matched normal range; these were compared with 35 age-matched GH-naïve hypopituitary patients with AOGHD. Lean body mass (LBM), fat mass (FM) and total bone mineral content (BMC) were assessed by dual-energy X-ray absorptiometry (DEXA) and corrected for actual height. Within genders, COGHD patients had about 20% less total body mass, LBM, FM and BMC than AOGHD patients. In addition, statistically significant correlations were present between final height SD score minus target height SD score and LBM and BMC in both genders (Attanasio et al., 2002). The implications from these observations are important; they indicate that the achievement of target height in the GH-replaced child does matter and that failure to reach target height has detrimental consequences in terms of skeletal health and body composition. Height is not necessarily meaningful per se, but as a surrogate for normalization of body composition it is crucial!

Influence of Timing of Onset of GHD in Adult Life

While the clinical and biochemical picture of the GHD adult is essentially the same in broad terms irrespective of whether the onset of GHD occurred in childhood or adult life, there are important differences that are of practical as well as biological interest.

Dyslipidaemia (Murray et al., 2002) and significant impairment of quality of life (QOL) (Murray et al., 1999) are less frequently seen in adults with COGHD compared with AOGHD patients; nonetheless the QOL response to GH replacement is not significantly influenced by the timing of onset of GHD (Murray et al., 1999); in other words for those with severe impairment of QOL, the response to GH replacement is similar in both groups of patients. Interestingly, however, in the early days of adult GH replacement when supraphysiological GH dosage dominated the clinical studies, the side-effects related to fluid retention occurred predominantly in AOGHD patients and rarely in COGHD patients (Holmes & Shalet, 1995); the latter observation sits easily with the rarity of such complications during childhood GH replacement.

The IGF-I status of the GHD adult is significantly affected by the timing of onset of GHD (Lissett et al., 2003); thus if all other relevant variables are controlled then the IGF-I SD score is 1·4 SDS lower in age-matched patients with COGHD compared with those with AOGHD (Fig. 4). There are practical implications from this observation in that the dominant variable influencing the adult GH replacement dosage required to normalize the IGF-I level in GHD adults is the pretreatment IGF-I level (Murray et al., 2000); therefore, the replacement dose utilized in adults with COGHD is significantly greater than in those with AOGHD. The scientific explanation for the lower IGF-I status of COGHD adults is also of interest; does it reflect inappropriate programming of IGF-I production in childhood? If this is the case does it simply imply suboptimal GH dosage during childhood or could it be influenced by the nonphysiological nature of the once-daily GH subcutaneous injection?

Figure 4. (click image to zoom) A box and whisker plot of IGF-I SDS, subdivided by age at onset of pituitary disease and timing of onset of pituitary disease. Subjects with onset of pituitary disease at age 16-20 years occurred in both groups but had significantly different IGF-I SDS, P

GH is intricately involved in bone growth and turnover; this is supported by the finding of reduced serum and urinary markers of bone turnover in GHD adults (Toogood et al., 1997a; Colao et al., 1999a), and the increase in these markers following GH replacement therapy (Binnerts et al., 1992; Bengtsson et al., 1993; Vandeweghe et al., 1993; Johannsson et al., 1996). Bone turnover is a coupled process that occurs continuously throughout life, bone resorption being followed in time by bone formation. With ageing it has been proposed that at the level of the remodelling unit, this process becomes increasingly inefficient. Thus in the elderly at the end of each remodelling cycle, small deficits in bone mass are accrued (Marcus, 1997, 1998), which lead to the observed age-related loss of bone mass. Predictably, therefore, the effect of GHD on the skeleton is heavily influenced by the patient's age; GHD during adolescence and young adult life before attainment of peak bone mass, when bone mass is being accrued, slows this acquisition and results in osteopenia. However, in the elderly in whom bone turnover is inefficient, a reduction in bone turnover, as a consequence of GHD, may reduce the rate of bone mineral loss, resulting in a normal bone mineral density (BMD).

Observational studies (Murray et al., 2004b) support the latter conclusions; osteopenia is seen in young adult GHD patients, with approximately 20-30% of those under 30 years of age having BMD z-scores below – 2 at the lumbar spine, hip and radius. By contrast, those adult GHD patients over the age of 60 years are no more frequently osteopenic than the normal age-matched population.

Thus the skeletal indication for GH replacement is primarily a consideration for young adult patients who acquired GHD either during childhood or early after completion of linear growth. The clinical occasion when this becomes an extremely important concern is during the transfer of management of the severely GHD teenager from paediatric to adult endocrine care.

Transitional Care of the GHD Teenager

It is the GHD teenager that has done more to bring the adult and paediatric endocrine communities together than any other category of patients, with several practical questions facing these communities. Which diagnostic criteria should be used to confirm severe GHD at completion of linear growth? In all those confirmed to be severely GHD, should GH replacement be continued throughout teenage years and the rest of adult life or would there be disadvantages if GH replacement was stopped for a few years and only reintroduced later on an individual basis? Furthermore, if GH replacement is continued seamlessly, what is the optimal dose – that used in adult or paediatric practice?

Currently the criteria used for the diagnosis of severe GHD in a teenager are identical to those adopted for severe GHD in adult life. GH secretion, however, is far greater in the normal teenager than in the middle-aged healthy adult; therefore, the threshold adopted for the diagnosis of severe GHD in the teenager may be inappropriately set too low. This raises further important questions about how to manage and follow those teenagers who retest GH insufficient using such criteria. Relatively little information is available to answer these questions, and in a practical world in which a health-economic battle to provide GH replacement for all adults with severe GHD prevails, these are questions for the future.

Recent discontinuation studies and placebo-controlled randomized therapeutic trials, however, are beginning to provide guidance about the need for continuation of GH replacement for GHD throughout the transition phase (Johannsson et al., 1999; Norrelund et al., 2000a,, and the choice of GH dose. More recently, we have completed a large multinational controlled 2-year study in patients who had terminated paediatric GH at final height (Shalet et al., 2003). Patients were randomized to GH at 25 µg/kg/day (paediatric dose, n = 58) or 12·5 µg/kg/day (adult dose, n = 59) or no GH treatment (control, n = 32); all patients had severe GHD with an IGF-I value less than the first centile of age-related normative values. Bone mineral content (BMC) and BMD were measured by DEXA and evaluated centrally. After 2 years significant increases were seen with both GH treatments compared with control in bone-specific alkaline phosphatase (BAP) and type 1 collagen C-terminal telopeptide : creatinine (ICTP/creatinine) ratio, but there were no significant dose effects.

Total BMC increased by 9·5 ± 8·4% in the adult group, 8·1 ± 7·6% in the paediatric dose group and 5·6 ± 8·4% in controls (anova, P = 0·008), with no significant GH dose effect. BMC increased predominantly at the lumbar spine rather than at the femoral neck or hip. There were no gender-related differences in BMC changes with either dose whereas the IGF-I increase was significantly higher with the paediatric than with the adult dose in females but not males.

Highly significant negative correlations were found between the time since the last GH injection of paediatric treatment and the baseline values for BAP (r = 0·31) and for the ICTP/creatinine ratio (r = 0·31) (Fig. 5). Even more interesting, the pattern of response in total BMC to GH treatment was also related to the time since withdrawal of paediatric GH therapy (Shalet et al., 2003). In the control group there was a significant negative correlation (r = –0·44) between the 2-year change in BMC and time since last GH injection, and for both treatment groups similar correlations were also found with BMC changes in the first year of GH treatment (r = –0·35).

Figure 5. (click image to zoom) Correlation of serum BAP and urinary ICTP/creatinine ratio with time since stopping paediatric GH treatment in patients with childhood GHD. Reproduced from Shalet et al. (2003).

The results of this study confirm that in patients with severe COGHD accrual of bone mass continues after GH cessation, and the increase in BMC observed in the control group during the 2-year study period is consistent with the 4·5% increase described by Fors et al. (2001) in a smaller group of COGHD patients after discontinuation of GH treatment. These findings support the concept that the GH effect on bone persists for 1-2 years after cessation of GH treatment, but eventually disappears.

Without GH treatment the net gain in total BMC was about 5% and with GH treatment about 10%, confirming the hypothesis that continued GH treatment after attainment of final height induces significant additional bone maturation in patients with severe GHD (Shalet et al., 2003). The study period was 2 years, but, given the prolonged effect of GH on bone, it is likely that further progression to peak bone mass would have been observed with a longer follow-up period.

The observation that the overall treatment effect was substantially the same with both GH dosages was unexpected. One possible explanation of this finding is that different dosages may have had different effects on bone turnover. In fact, although differences were statistically nonsignificant, the increase in ICTP/creatinine was greater with the higher dose and the opposite was the case for BAP. It has been shown that high GH doses in young GHD adults cause desynchronization of bone turnover with predominance of bone resorption over bone formation (Balducci et al., 1995); the trends seen in the present study suggest a similar mechanism and indicate that the paediatric dose was inappropriately high for adequate bone mass accumulation.

In summary, this study demonstrated that withdrawal of GH replacement at final height may limit progression to peak bone mass in patients with severe COGHD and that adequate GH replacement is required to continue this process. The effect on bone is obtained with a dose regimen that is in the high adult replacement range and is of clinical relevance for subsequent bone health in adult life. The data also indicate that for optimal progress to peak bone mass, GH treatment after attainment of final height should not be discontinued.

Diagnostic and Therapeutic Studies in Adults Inform About Paediatric GH Practice

The significant differences in body composition and BMC between young adults treated in childhood for COGHD and young adults with untreated AOGHD imply that paediatric GH replacement has often been suboptimal. The observation that there is a significant relationship between the statural amount by which a GH-replaced teenager fails to achieve target height and the reduction in total BMC and LBM implies that auxology deserves its role as the dominant clinical endpoint followed in paediatric GH practice; not because achieving target height is of itself crucial, but as a surrogate for the acquisition of normal body composition, optimization of growth is a very worthy goal. Information obtained from the use of two GH diagnostic tests in adult patients with varying degrees of hypopituitarism has provided some validity for the strategy of testing utilized in the investigation of the child with the putative diagnosis of isolated idiopathic GHD. The numerous studies of body composition changes in GHD teenagers confirm that growth and development do not end at attainment of final height, and show that GH makes a crucial contribution to the end result; thus GH replacement for developmental purposes should continue until adult acquisition of LBM and PBM and not stop once final height is achieved.

Partial GHD

There is no doubt that partial GHD exists; this is hardly surprising given that partial deficiencies of gonadotrophin, ACTH, TSH and vasopressin exist. Up until now, however, there have been relatively few studies of the impact of partial GHD on biological endpoints in adults. Furthermore, the paediatric GH experience has emphasized that it is the continuum in GH secretion between normal children and children with varying degrees of GHD that underlies the major difficulty in diagnosing GHD in childhood.

Initial studies by Colao et al. (1999a,, using the arginine plus GHRH test to categorize GH status, showed dyslipidaemia but normal BMD in a cohort of adults with partial GHD. More recently, increased FM (Murray et al., 2004a), reduced LBM (Murray et al., 2004a) and abnormal insulin secretion (Murray & Shalet, 2001) have been observed in adults with partial GHD, the degree of abnormality tending to lie between that seen in normals and those with severe GHD. This is not very surprising, particularly as the biochemical diagnosis of severe GHD is arbitrarily based.

These observational studies raise the possibility of potentially treating adults with partial GHD with GH; to do so, however, will first require the

endocrine community to diagnose partial GHD in adults.

The hypothetical patient will have a putative lesion of the hypothalamic-pituitary axis and, given the early timing of GHD in the evolution of hypopituitarism, usually have no other pituitary hormone deficits. Truncal obesity will be present, and it is now established in clinically nonobese healthy adults that relative adiposity, in the abdominal region in particular, is a major negative determinant of stimulated GH secretion (Vahl et al., 1996); therefore GH status will be defined as partial GHD. In practice, an IGF-I estimation is unlikely to be helpful. A pathologically low IGF-I level would point towards a diagnosis of GHD but obesity is associated with a normal IGF-I level and the majority of patients with partial GHD are also likely to exhibit a normal IGF-I level. The fundamental dilemma requiring resolution is the following; is this a fat patient in whom visceral obesity is responsible for the biochemical findings of partial GHD? Or is this a patient afflicted by partial GHD, which is itself responsible for the truncal obesity?

Therefore, if we do move in this therapeutic direction the current problems that torment the paediatric endocrinologist, regarding diagnostic criteria for GHD and selection of patients likely to benefit from GH therapy, will move on to plague the adult endocrine community.

CLICK HERE for subscription information about this journal.

Acknowledgements

This review reflects the contents of my Clinical Endocrinology Trust Lecture (2003); its substance owes much to the efforts of numerous research fellows over the past 25 years. I am also grateful to Colin Beardwell for letting me be myself and Andrea Attanasio for many stimulating discussions about physical development during the transition years.

Reprint Address

Correspondence: S. M. Shalet, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK. Fax: + 44 0161 446 3772; E-mail: stephen.m.shalet@man.ac.uk