This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Drake, W. M. Articles by Shalet, S. M. Search for Related Content PubMed PubMed Citation Articles by Drake, W. M. Articles by Shalet, S. M.

Optimizing GH Therapy in Adults and Children

Departments of Endocrinology, St. Bartholomew’s Hospital, London EC1A 7BE, United Kingdom; and Christie Hospital, Manchester, United Kingdom

Correspondence: Address all correspondence and requests for reprints to: Dr. W. M. Drake, Department of Endocrinology, St. Bartholomew’s Hospital, West Smithfield, London EC1A 7BE, United Kingdom. E-mail: w.m.drake{at}mds.qmw.ac.uk

	Abstract

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
Until the advent of modern neuroradiological imaging techniques in 1989, a diagnosis of GH deficiency in adults carried little significance other than as a marker of hypothalamo-pituitary disease. The relatively recent recognition of a characteristic clinical syndrome associated with failure of spontaneous GH secretion and the potential reversal of many of its features with recombinant human GH has prompted a closer examination of the physiological role of GH after linear growth is complete. The safe clinical practice of GH replacement demands a method of judging overall GH status, but there is no biological marker in adults that is the equivalent of linear growth in a child by which to judge the efficacy of GH replacement. Assessment of optimal GH replacement is made difficult by the apparent diverse actions of GH in health, concern about the avoidance of iatrogenic acromegaly, and the growing realization that an individual’s risk of developing certain cancers may, at least in part, be influenced by cumulative exposure to the chief mediator of GH action, IGF-I. As in all areas of clinical practice, strategies and protocols vary between centers, but most physicians experienced in the management of pituitary disease agree that GH is most appropriately begun at low doses, building up slowly to the final maintenance dose. This, in turn, is best determined by a combination of clinical response and measurement of serum IGF-I, avoiding supraphysiological levels of this GH-dependent peptide. Numerous studies have helped define the optimum management of GH replacement during childhood. The recent requirement to measure and monitor GH status in adult life has called into question the appropriateness of simplistic weight- and surface area-based dosing regimens for the management of GH deficiency in childhood, with reliance on linear growth as the sole marker of GH action. It is clear that the monitoring of parameters other than linear growth to help refine GH therapy should now be incorporated into childhood GH treatment protocols. Further research will be required to define the optimal management of the transition from pediatric to adult GH replacement; this transition will only be possible once the benefits of GH in mature adults are defined and accepted by pediatric and adult endocrinologists alike.

I. Introduction

II. Physiology of GH Secretion and Action

A. Changes in GH and IGF-I levels with age

III. Pharmacokinetics of Administered GH

IV. Review of Pediatric Practice

A. Diagnosis of GHD and patient selection

B. GH treatment: dose and schedule

C. Puberty

D. Side effects of GH therapy

E. Predictors of response to therapy

V. Adult GH Replacement: Historical Perspective

VI. Rationale and Strategies for GH Replacement in Adults

VII. Tolerability

VIII. Clinical Response to GH Replacement

A. Body composition

B. Quality of life and well-being

C. Bone density and bone remodeling

D. Cardiovascular risk factors and cardiac structure and function

E. Conclusions

IX. Is Overtreatment Acceptable in the Asymptomatic Patient?

X. Biochemical Monitoring of GH Replacement

XI. Gender Differences in GH Responsiveness

XII. Adult GHD and Vascular Disease

A. GH replacement therapy and dyslipidemia

XIII. Alternative Mechanisms for Accelerated Vascular Disease

XIV. Transition from Pediatric to Adult Clinic

XV. Effects of Discontinuation of GH Treatment at Final Height

A. Body composition

B. Bone mineral density

XVI. Dosing Strategies for the Adolescent Patient

XVII. Influence of Adult GH Replacement Studies on Pediatric Practice: Reevaluation of Pediatric Practice

	I. Introduction

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
IT IS NOW more than a decade since the publication of the earliest double-blind placebo-controlled studies of the use of recombinant human GH (rhGH) in patients with adult-onset (AO) GH deficiency (GHD) (1, 2, 3, 4, 5). In the intervening years, increasing attention has been devoted to devising treatment protocols that maximize the potential clinical benefit of treatment with GH, while trying to minimize the risks that may result from prolonged excessive GH exposure. In the original trials of GH replacement therapy for adult hypopituitarism, GH dose was calculated on the basis of weight and/or surface area (1, 2, 3, 4, 5). This was essentially done as an extension of pediatric practice, as there was little or no experience of GH therapy outside the pediatric setting at that time. With time and shared clinical experience, it has become apparent that such a dosing strategy was too simplistic and that the dose of GH needs to be individually tailored for each patient (6, 7). In retrospect, this was predictable, given that all other endocrine replacement therapy needs to be carefully adjusted for each patient, reflecting the wide variation in secretion and/or clearance rates, of glucocorticoids, thyroid hormone, and gonadal steroids in healthy subjects. Paradoxically, it is the increasing use of GH for the treatment of AO hypopituitarism that has led to a more thorough examination of the physiological role of GH after the completion of linear growth. Inevitably, the recent experience in adult clinical practice has prompted a critical reevaluation of the strategies used for the management of GHD in childhood. This article will review the current clinical practice of GH therapy for the attainment of final height in children. It will then discuss methods of GH dosing in adults with specific reference to our current state of knowledge about the normal physiology of GH secretion and action after the completion of linear growth. The issues that surround reassessment of GH status in GHD children once final height has been achieved and the decision as to whether to continue with GH therapy in this patient group will also be reviewed. The article will conclude by examining, in the light of recent clinical experience in GH dosing for adults, whether the time has come for a closer examination of GH dosing strategies used in pediatric clinical practice.

	II. Physiology of GH Secretion and Action

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
In any discussion about optimizing GH dosing schedules, it is important to remember that reproduction of normal physiological patterns of GH secretion and action is limited by available modes and routes of administration of exogenous GH. Hence, as with other forms of endocrine replacement therapy (such as glucocorticoid replacement for patients with primary or secondary adrenal failure), the aims of the treating physician are limited to a maximization of clinical benefit while minimizing the risks of excessive exposure.

GH is released from anterior pituitary somatotrope cells in a pulsatile fashion, with surges of GH release punctuating long periods when GH levels in plasma are very low and detectable only by sensitive chemiluminescence assays (8) (Fig. 1B). GH release, in turn, is stimulated by GHRH and inhibited by somatostatin (SST), both of which are produced by the hypothalamus (for review see Ref. 9). A separate receptor exists, the GH secretagogue receptor (10), the ligand for which (Ghrelin) has recently been cloned (11). The details of the neuroendocrine mechanisms by which these various inputs interact to regulate GH release, and in particular the precise role of Ghrelin, are not fully elucidated, but the simplest model postulates that a simultaneous drop in SST tone, together with bursts of GHRH secretion, is responsible for the generation of a GH pulse (12).

View larger version (14K): [in this window] [in a new window]   Figure 1. Plasma time-curve for GH 1.2 IU given by injection to an adult with GHD (a) compared with the physiological pattern of GH secretion in a healthy middle-aged male (b). [Courtesy of Dr. Y. Janssen, personal communication].

The GH secretory pattern, hepatic GH receptors, and circulating GHBP levels are closely interrelated. In the rat, linear growth is most sensitive to pulsatile GH exposure and peak amplitude, whereas GHBP and hepatic GH receptor levels are regulated separately by the level of continuous GH baseline exposure (17). As in rats, baseline GH levels are higher in females than males (18, 19, 20, 21), although the magnitude is less apparent. Short-term comparisons of continuous vs. pulsatile GH treatment in man have so far revealed only minor differences in metabolic parameters (22, 23), but longer treatment in GH-deficient children shows induction of GHBP after continuous, but not pulsatile, GH treatment (24). It has been suggested that the GH pattern is important for growth in man, since increasing the frequency of GH therapy to daily injections improves its growth-promoting effect (25, 26, 27), although in the short term once daily subcutaneous injections stimulated growth equally well as continuous subcutaneous infusion in GH-deficient children (24).

GH has some direct effects on peripheral tissues, most notably epiphyseal chondrocytes (28), but the majority of its actions are mediated through the peptide hormone insulin-like growth factor-I (IGF-I), a member of the insulin-like gene family. Almost all IGF-I in the circulation is bound to one of several IGF binding proteins (IGFBPs), the most abundant of which is IGFBP-3. Together with the acid-labile subunit (ALS), IGF-I and IGFBP-3 (levels of which are all GH dependent) form a ternary complex of 150 kDa. This prolongs the half-life of circulating IGF-I and ensures that levels in a given individual remain stable throughout the day. However, the simplistic use of serum IGF-I measurements as a precise marker of overall GH status is flawed because of the many variables that affect both hepatic and local tissue IGF-I production in response to a given GH stimulus. Most strikingly, GH-mediated IGF-I production varies with gender. Analysis of 24-h GH profiles in normal weight, middle-aged, healthy volunteers shows that to maintain an equivalent serum IGF-I level, mean daily production is approximately 3 times greater in women than in men, due largely to an amplitude-specific divergence in the pulsatile mode of GH secretion (29). Most circulating IGF-I is derived from the liver, but it is also generated in nonhepatic tissues where it appears to function in an autocrine/paracrine fashion (30). IGF-I generation in response to a given GH stimulus may be modulated by local tissue-specific factors, of which gonadal steroids are an important example. Testosterone administration to normal men and those with hypogonadotropic hypogonadism increases serum IGF-I levels, while oral estrogen therapy improves the signs and symptoms of acromegaly (31) and lowers serum IGF-I levels in normal postmenopausal women (32). Furthermore, estrogens have different effects on GH secretion and action depending on the route of administration. Oral ethinyl estradiol attenuates IGF-I production despite a 3-fold increase in mean 24-h GH, whereas transdermal 17ß-estradiol does not alter overall GH secretion but causes a slight increase in circulating IGF-I (33, 34). Such changes are almost certainly physiologically important, on the basis of changes in markers of connective and bone tissue activity that parallel the changes in serum IGF-I levels (35).

As with other hormonal systems, GH, IGF-I, and the hypothalamic peptides SST and GHRH form a complex feedback system at various levels. For example, exogenous GH administration attenuates the size of subsequent GHRH-mediated GH secretion, apparently independently of circulating IGF-I levels (9, 36), and there is evidence that hypothalamic GH receptor expression is suppressed by GH (37). There are also data to suggest feedback regulation of GH secretion by IGF-I (38).

A. Changes in GH and IGF-I levels with age
GH secretion continues throughout life (39, 40) and this, together with the clinical features of adult GH deficiency and the observed favorable effects of GH on hypopituitary patients, forms a persuasive argument that GH may have important physiological functions after the completion of linear growth. Rates of GH secretion increase with the onset of sexual maturation, reaching approximately 2–3 times their prepubertal levels by mid-late puberty (39, 40). Thereafter, GH secretion rates decline by approximately 14%, and the half-life of GH shortens by 6% with each passing decade (39, 40). The reasons for this fall are not entirely clear, although a reduction in responsiveness to injected GHRH and an increase in SST have been documented in vivo in the rat (41, 42). In humans, repetitive administration of GHRH to elderly men partially reverses the age-related decline in responsiveness to this peptide (43), suggesting that failure of hypothalamic GHRH release may, at least in part, underlie the age-related fall in spontaneous GH secretion. Furthermore, concurrent with the suggestion that SST is increased in older individuals, coadministration of arginine enhances the GH-releasing activity of a GH-releasing hexapeptide in healthy elderly, but not young, subjects (44). These age-related changes in GH secretion are largely paralleled by a decline in serum IGF-I levels in both men and women (45). However, the positive relationship between serum IGF-I and spontaneous 24-h GH secretion becomes less striking with age (46). This may, in part, reflect increasing obesity with advancing years. An inverse correlation has been demonstrated between adiposity and plasma IGF-I levels (47, 48), although separation of this from the known effects of obesity on GH secretion (40) is difficult.

	III. Pharmacokinetics of Administered GH

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
rhGH is almost universally administered subcutaneously (Fig. 1A). Most studies report the time to reach peak level (C_max) is around 4–6 h after an injection, with a length of disappearance of 20–24 h (49, 50); these figures correlate well with the kinetics of insulin absorption in patients with diabetes (51). Studies comparing the bioavailability of GH by continuous infusion and bolus injection show significantly lower GH levels with subcutaneous administration compared with intravenous, suggesting a degree of local subcutaneous degradation (52, 53). The volume of injection may also be important. Injection of 6 IU GH in three separate volumes on separate occasions to normal individuals showed that the use of a larger volume resulted in a higher C_max and a greater area under the curve, implying greater overall bioavailability (54).

It is apparent, therefore, that the GH/IGF-I system, like other endocrine systems, is a dynamic one, the activity of which changes with age, sexual maturation, body composition, and other factors. Clearly, it is not possible to recreate normal physiology with a single subcutaneous injection of GH, so the goal of treatment of GHD is correction of the associated clinical syndrome. In children, failure of linear growth is an almost universal presenting feature, whereas in adults the diagnosis of GHD almost invariably is made in patients with a background of known hypothalamo-pituitary disease.

	IV. Review of Pediatric Practice

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
GH deficiency during childhood is associated with severe growth retardation resulting in marked impairment of adult height. GH therapy for GH-deficient children was first used in the 1950s with hormone extracted from the pituitary glands of cadavers (55). Because of the limited supply, suboptimal doses of pituitary GH were administered initially in a standard fixed dose, independent of size, two to three times weekly, and the response to therapy was modest. Burns et al. (56) reported on the final height of 55 GHD children treated before 1981. GH was administered at a dose of either 10 IU twice weekly or 5 IU three times a week, irrespective of weight or surface area. Treatment was begun between the ages of 9 and 14 yr in most of the children, and therefore a total weekly dose of 15–20 IU represented a suboptimal GH dose by modern standards. The smaller children, however, would have initially received a fairly high dose relative to their size. At completion of growth, average height was more than 2 SD below the population mean, with stature in more than half the children failing to exceed the third centile. The response to therapy did not differ between those treated with two or three injections per week, despite a significant difference in total dose received, suggesting that frequency of injections is important. Other early final height studies also reported a modest response to GH replacement, with adult heights not always significantly greater than those observed in untreated individuals (56, 57, 58, 59). In 1985, the first reports emerged of the link between pituitary-derived GH therapy and Creutzfeldt-Jakob disease (60), and human GH was promptly withdrawn. However, within a year GH produced by recombinant DNA technology became available (61). In addition to negating the risk of Creutzfeldt-Jakob disease, synthetic GH also transformed the practice of GH replacement by providing a potentially limitless supply of therapy. This allowed the treatment of GHD children with higher, more appropriate doses, and also enabled GH therapy to be offered to a much wider range of patients. Subsequent studies have helped define which patients benefit from GH therapy and the optimal dosing schedules.

A. Diagnosis of GHD and patient selection
In pediatric practice, the diagnosis of GH deficiency is usually suspected on the grounds of auxological data, although a minority of children are diagnosed because of known pituitary disease or after radiotherapy involving the hypothalamo-pituitary region. Selection of the patients most likely to derive clinical benefit from GH therapy demands a sensitive and specific test that distinguishes GH-deficient from normal subjects. A number of approaches have been used to confirm GHD in childhood, including GH stimulation tests, 24-h GH profiles, urinary GH, and measurements of IGFs and IGFBPs (62, 63, 64). There is no single test, however, that can invariably distinguish normal children from GH-deficient individuals, particularly those with less severe degrees of GHD.

Provocative tests remain the standard method of confirming a diagnosis of GHD in a child in whom the diagnosis is suspected on the basis of auxological data, or a predisposing factor (64). A number of different agents have been used including insulin, arginine, glucagon, propranolol, clonidine, and L-dopa (63). Physiological stimuli, such as sleep, fasting, and exercise, have also been employed. The definition of what constitutes a normal rise in serum GH concentration after either physiological or pharmacological stimulation is largely arbitrary. In early reports, a stimulated peak GH level of 5 ng/ml or more was considered to be indicative of normal GH reserve. This definition has gradually changed, influenced to some extent by the increased availability of therapy after the advent of rhGH, such that a level of between 7 and 10 ng/ml is now generally accepted as the cutoff. IGF-I and IGFBP-3 levels have also been used as a measure of GH status, and it has been suggested that the use of a combination of tests improves diagnostic accuracy (62). However, there remains no method of reliably separating all GH-deficient children from GH-replete subjects, and this difficulty is reflected in the retest data, which suggest that a substantial proportion of children treated for GHD during childhood have entirely normal GH responses to standard tests in adult life (65). In practice, the diagnosis of GHD is based on careful clinical assessment, augmented by a number of tests, which reflect GH status (63).

B. GH treatment: dose and schedule
There is good evidence that GH therapy should begin as soon as possible to optimize long-term growth (66, 67, 68). Prompt initiation of therapy is particularly important in young children in whom fasting hypoglycemia may complicate GHD (69). While there is broad agreement about the optimal timing of the start of therapy, the selection of the appropriate GH treatment dose is less clear.

There are several methods that may be employed for defining the optimal replacement regimen in GH-deficient patients. Attempts can be made to mimic normal physiology by administering GH at doses that approximate to normal production rates, or by attempting to achieve serum levels of GH or GH-dependent markers that are close to normal levels. Alternatively, treatment may be selected on the basis of the reversal of the biological endpoints of GHD, while minimizing any adverse effects of therapy. In pediatric practice, the selection of optimal GH doses and treatment schedules has rested almost entirely on the response to therapy in terms of linear growth.

Some attempts have been made to define physiological GH requirements by examining GH production in normal subjects, or by measuring biochemical markers of GH status in treated patients. Studies in healthy children have estimated daily endogenous GH production to be approximately 20 µg/kg/d (equivalent to 0.14 mg/kg/week) (70, 71), rising to 35 µg/kg/d in late puberty (71), although considerable interindividual variability exists (72, 73, 74, 75). Extrapolating this to replacement doses is hampered by the different pharmacodynamics of exogenous GH administered subcutaneously, but it does provide an estimate of GH requirements. Few data exist regarding the measurement of GH markers during replacement therapy in children. Hibi et al. (76) measured plasma IGF-I, urinary IGF-I, and urinary GH in a cohort of GH-deficient children and suggested that a GH dose of 0.16 mg/kg/wk was close to a physiological replacement dose. A more recent study has documented serum IGF-I and IGFBP-3 levels within the normal range in a group of GHD children receiving a mean GH dose of 0.17 mg/kg/wk. Furthermore, in a large multicenter randomized US trial (n = 139), girls with Turner syndrome received either 0.27 mg/kg/wk or 0.36 mg/kg/wk of GH in combination with either low-dose estrogen or oral placebo, and only 1.4% had serum IGF-I levels greater than 2 SD above the mean for age (77). In contrast, in a smaller study of 31 patients, Tillman et al. (78) reported two of 20 GHD and three of seven children with Turner syndrome showed supraphysiological IGF-I levels during the first year of GH treatment. As the measurement of IGF-I during childhood GH therapy becomes commonplace, more information will become available regarding IGF-I levels achieved with different GH doses. It is clearly important that normative data should be accumulated across the childhood age range.

The vast majority of clinicians have used the growth response to GH therapy to define the optimal replacement dose. The endpoints used to define the benefits of different doses have been both short-term growth and growth velocity and, more importantly, final height. It was recognized before the advent of rhGH that a dose-response relationship exists between GH dose and growth rate (79, 80), and more recent data have confirmed that GH dose influences the short-term growth response to GH replacement (81, 82). In addition, dose frequency has been shown to be an important factor in determining the response to therapy. Changing from three times a week to a daily subcutaneous injection results in an increased growth rate for a given total GH dose (83, 84), although no further growth advantage has been demonstrated with more frequent injections (84, 85, 86). With the greater availability of rhGH, higher GH doses have generally been used, and this has resulted in more favorable final height data. Despite reasonable improvements in height SD scores (the number of SD scores by which an individual’s height differs from the mean for his/her age and sex) during treatment with pituitary-derived GH, average final height in these children was only -2.3 SD scores (87). The vast majority of patients treated with rhGH, however, achieve a final height within the normal range, with an average final height of -1.4 SD score (87). In the last 5 yr a number of groups have published data from cohorts of children treated with rhGH to final height (66, 68, 87, 88, 89, 90, 91, 92). Height gain in these studies, as assessed by the difference between final height and predicted adult height or initial height SD score, ranged between 0.8 SD and 2.4 SD, with average final height ranging between -2.1 SD and -0.7 SD. The criteria used to diagnose GHD and therapeutic regimens employed varied between the studies and probably account for much of the variation in the results.

These data confirm the benefits of treatment with a weekly dose of at least 0.15 mg/kg but controversy remains concerning the additional benefits of higher doses. Analysis of final height data suggests that a dose between 0.17 and 0.3 mg/kg/wk is a reasonable replacement dose (70). The most reliable data are taken from large multicenter studies such as the Kabi International Growth Study (KIGS), the National Co-operative Growth Study (NCGS), and the Genentech Growth Study group (GGSG). Interpretation of data from these large cohorts of patients is complicated by the fact that they were collected over a number of years from many different centers, and there is therefore a degree of variability in the treatment protocols employed. Thus, a proportion of patients have received treatment that is now considered suboptimal, and the responses to therapy need to be assessed with this in mind. In addition, some of the differences observed in the response to therapy may be due to differences in the study populations. Data from the GGSG (66) suggest that treatment with 0.3 mg/kg/wk is associated with significantly greater improvements in final height than those observed in patients enrolled into KIGS, who received a lower average dose of 0.16 mg/kg/wk (88). The GGSG cohort achieved a final height of -0.7 SD compared with -1.5 in the KIGS group. However, the midparental height of the GGSG cohort was significantly greater than that in the KIGS group, and, after correcting for this, there was no difference in final height achieved (88). In addition, data from the NCGS in a larger cohort of GH-deficient children treated with 0.3 mg/kg/wk demonstrated a more modest response (68) similar to that seen in the KIGS patients (final height -1.4 SD). However, analysis of a separate cohort of Swedish patients within KIGS treated with an intermediate dose of 0.22 mg/kg/wk demonstrated complete normalization of final height, indicating that higher doses may result in a better response to therapy (88). Furthermore, a recent report suggested a significant short-term growth advantage from a GH dose of 0.35 mg/kg/wk over that observed with 0.17 mg/kg/wk (93). A further increase in the dose to 0.7 mg/kg/wk conferred no additional benefit, and this concurs with other studies that have examined the use of higher GH doses (70, 94). In accordance with all these data, recent internationally agreed guidelines for the treatment of GH deficiency in childhood suggest a dose of 0.17–0.35 mg/kg/wk (95).

Ultimately, the selection of the replacement dose is based on interpretation of the available data, local availability of GH, and also on financial grounds. Given the variability in GH production in normal individuals, it is likely that GH requirements will vary from patient to patient, and a more individual approach may eventually be required. This would necessitate an accurate method for assessing and monitoring the appropriateness of a given GH dose in each patient.

C. Puberty
GH production in normal individuals rises during puberty (71). In addition, a positive correlation has been found between total pubertal height gain and mean GH dose during puberty (96). It has therefore been suggested that the dose of GH replacement should be increased at the onset of puberty to mimic normal physiology. Stanhope et al. (97), however, demonstrated no increase in growth rate on increasing GH dose from 15 IU/m²/wk to 30 IU/m²/wk during puberty in a small cohort of GH-deficient children, compared with a control group who continued on 15 IU/m²/wk. Indeed, their data suggested that a high GH dose accelerated the progression through puberty and may therefore be detrimental to final height outcome. It should be noted, however, that their conclusions were based on the short-term growth response, and these children were not followed to final height. More recently, Albertsson-Wikland et al. (98) demonstrated no increase in total pubertal height gain in boys treated with 0.42 mg/kg/wk compared with boys treated with 0.21 mg/kg/wk. In addition, MacGillivray et al. (70) compared data between several large studies of GH replacement employing differing doses of GH. Pubertal height gain did not differ significantly between the cohorts, suggesting no additional benefit from a higher replacement dose during puberty (70). Thus, while some centers still advocate an increase in GH dose at puberty, many clinicians continue treatment at a similar dose (calculated per kg or m²) throughout childhood.

The most likely explanation for the lack of a significant growth advantage with an increased GH dose during puberty is that this is associated with an advance in bone maturation. This will lead to earlier fusion of the epiphyses and therefore shorten pubertal growth, and it has thus been suggested that pharmacological delay of pubertal development may improve the overall growth response to GH replacement. Delaying the progression through puberty by the administration of GnRH analogs has been standard practice for the treatment of precocious puberty for a number of years. The main long-term goal of this therapy is to prevent reductions in adult height, which will occur if puberty is allowed to progress at an early age, because of the reduced time available for linear growth. Improvements in final height have been achieved with GnRH analog therapy in precocious puberty (99, 100, 101), although some authors have suggested that the addition of GH therapy may further improve growth. A few studies have examined the effects of combined treatment with GH and GnRH analogs in children with precocious puberty and normal GH levels (102, 103). In addition, there are a number of reports of combined treatment in children with both GHD and early puberty (104, 105). The use of combined therapy has also been investigated in short normal children (103, 106, 107, 108). The results of these studies have been variable, with many showing little improvement in growth velocity or final height. Nonetheless, it has been postulated that GnRH analog therapy may augment the growth response to GH therapy in GHD. A few studies have suggested improvements in final height prognosis with a combination of GH and GnRH therapy (109, 110, 111, 112). Some of these reports have been limited by the use of final height predictions based on the short-term response to therapy. The recent report from Mericq et al., however, followed 21 GH-deficient subjects to near-final height defined as a bone age of 14 yr in girls and 16 yr in boys (111). Patients were randomly assigned to GH therapy plus LHRH analog of GH alone. A significant gain in near-final height was demonstrated for those receiving combination treatment compared with those treated with GH alone (mean height SD score -1.3 vs. -2.7), with no alteration in body proportions. This was achieved at the expense of significantly delaying puberty, with the mean age at menarche in the girls treated with LHRH being 18.2 yr compared with 15.9 yr in the GH-alone group. Thus, while these data are promising in terms of potential height gain, the psychosocial implications of pubertal delay need to be balanced against the growth advantage that is potentially conferred by the addition of GnRH analogs to GH therapy. In addition, the number of patients who have been followed to final height remains small, and further data are required before the addition of GnRH analog therapy can be routinely recommended for use in GH-deficient children in the absence of coexisting precocious puberty.

D. Side effects of GH therapy
There are a number of adverse effects that have been attributed to GH replacement during childhood. The most comprehensive data are available from large international surveillance studies that have been specifically designed to monitor safety of treatment. Idiopathic (benign) intracranial hypertension was first reported in 1992 (113), and a number of subsequent reports have confirmed the relationship with GH therapy (114, 115, 116, 117, 118). Data from the NCGS and KIGS databases have revealed 35 cases of idiopathic intracranial hypertension from a total of more than 40,000 patients receiving more than 109,000 yr of GH therapy (114, 117). The condition improves after withdrawal of therapy, and GH can often be restarted without a recurrence of the problem.

The question of the impact of GH therapy on tumor growth has often been raised, particularly with reference to populations of children previously treated for childhood malignancies. Because interpretation of tumor recurrence data is complicated by biases introduced by the selection of children for GH therapy, careful matching of control populations is necessary. The available data do not suggest any increase in the risk of tumor recurrence in the children after GH treatment; both single-center studies (119) and large-multicenter surveillance studies (114, 120) have failed to show any increase in the incidence of de novo malignancies during GH replacement. Fasting glucose levels have been shown to rise after the commencement of GH therapy in children (121), and there have been reports of the development of diabetes mellitus during treatment (122). Data from the NCGS (117) and KIGS have not suggested an increase in the incidence of type 1 diabetes, but the KIGS database did demonstrate a higher than expected incidence of type 2 diabetes in a heterogeneous population including children treated with GH for short stature not due to GHD (123). The incidence was, however, very small (34 cases per 100,000 yr of GH treatment), and it was postulated that the higher rates may indicate an acceleration of the disorder in predisposed individuals.

A number of other adverse events occur more commonly during GH therapy in children. Slipped capital femoral epiphysis (117, 124), gynecomastia (116), and juvenile osteochondritis (114) have all been reported during treatment, although a direct causal relationship with GH has not been established. Interestingly, edema and carpal tunnel syndrome only occur only very rarely in pediatric practice (117), although these are commonly reported in adult-onset GH-deficient patients receiving GH replacement (125). The reason for this discordance is not clear.

E. Predictors of response to therapy
Several reports from large cohorts of GH-deficient children have provided some information on factors that influence the growth response to GH replacement (51, 66, 72, 73, 76, 82, 88, 89, 92, 108, 126). Analysis of the KIGS database suggested that first year height velocity was negatively correlated with age and height SD score and positively correlated with birth weight, weight at beginning of therapy, GH dose, frequency of injection, target height SD score, and degree of GHD, as judged by the peak GH response to a stimulation test (81, 127). Analysis of final height data demonstrated no effect of GH dose on adult height, although the duration of GH therapy was a significant factor (88). This underscores the need to begin GH therapy as early as possible to attain the maximum final height, and also suggests that, within the dose range used (10th–90th centiles; 0.11–0.24 mg/kg/wk), variations in weekly GH dose has little effect on final height. Data from the NCGS are consistent with these findings, suggesting that the initial response to GH therapy may be predicted by age, degree of GHD, weight adjusted for height, GH dose, injection frequency, and midparental height (82). Final height was dependent on pretreatment height and age, duration of treatment, sex, and first year growth rate (66). Thus, knowledge of a number of baseline parameters will help predict the response to therapy. From these data, models have been developed that allow reasonably accurate prediction of the first year growth velocity after GH therapy. However, although a greater initial response to treatment will be psychologically important to the patient and is likely also to improve compliance, the final height achieved is generally considered to be the most important goal of therapy. The model developed from the KIGS database (127) has been extended to examine second, third, and fourth year growth response and has demonstrated that first year height velocity is the most important predictor of subsequent growth. Extrapolation of these results would suggest that first year height velocity is likely to be an important determinant of final height; however, this has yet to be established, and at present these models can predict only the initial growth after the institution of GH replacement and not the overall response to therapy.

There are also a number of other markers that may help predict the initial growth response. Markers of bone turnover are significantly reduced in GH-deficient children and increase after GH replacement (128). The increase in serum bone alkaline phosphatase levels (a marker of bone formation) after 3 months of GH replacement has been shown to correlate with improvements in the height SD score in the first year of therapy. Serum leptin levels also alter with GH status, predominately as a result of changes in fat mass and distribution. Leptin levels reduce during GH replacement (129), and changes in leptin concentration 1 and 3 months after the beginning of GH therapy have been shown to correlate with growth in the first year of treatment (130). These observations of changes in bone alkaline phosphatase and serum leptin indicate that metabolic markers are potential predictors of the short-term growth response to GH therapy.

Standard GH replacement therapy in GH-deficient children thus consists of daily injections of rhGH, usually administered at a weight-based dose of between 0.17 and 0.3 mg/kg/wk. Treatment is initiated as soon as possible once the diagnosis has been made and is continued until the attainment of final height. This is usually defined as either a slowing of growth to an annualized height velocity of less than 1 cm/yr or the demonstration of fusion of the long bone epiphyses (131). Improvements in linear growth have been almost the sole indication for the use of GH in pediatric practice. There are some data, however, concerning the use of GH in normally growing GH-deficient patients after surgery for craniopharyngioma (132), which demonstrated beneficial metabolic effects of treatment resulting in advantageous changes in body composition and suggested that GH replacement is indicated in these children despite their normal growth. In addition, data from the use of GH therapy in children with Prader-Willi syndrome have demonstrated beneficial effects of treatment on body composition, muscle strength, and respiratory function (133, 134, 135). These unusual situations highlight the potential benefits of GH replacement in childhood other than linear growth.

GH replacement has evolved since the pioneering work of the 1950s using cadaveric pituitary-derived GH. Numerous studies have helped define the optimal management of GH replacement during childhood. The recognition of the importance of GHD during adult life has necessitated a more detailed study of GHD and the impact of treatment, which has resulted in a reevaluation of pediatric practice. The monitoring of parameters other than linear growth to help refine GH therapy should now be incorporated into childhood GH replacement. Further research will be required to define the optimal management of the transition from pediatric to adult GH replacement, and a smooth changeover will only be accomplished once the benefits of GH after the completion of growth are accepted by pediatric and adult endocrinologists alike.

	V. Adult GH Replacement: Historical Perspective

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
The earliest report of the beneficial effects of GH in the treatment of adult hypopituitarism was in 1962, when increased vigor and well being were reported by a 35-yr-old hypopituitary patient treated with cadaveric GH (136). This was followed, approximately 30 yr later, by a series of randomized, placebo-controlled trials, in which it was convincingly demonstrated that treatment of GHD adults with rhGH led to significant improvements in body composition, well being, and serum lipoprotein levels (1, 2, 3, 4, 5, 6). Almost coincident with these studies came the first (137) of several reports (138, 139, 140) to suggest that hypopituitarism is associated with decreased life expectancy compared with age-matched healthy controls, despite adequate replacement with glucocorticoids, thyroid hormone, and sex steroids. Although it remains an intriguing possibility that the increased visceral adiposity and abnormal cardiovascular risk profile associated with adult GHD contribute to the demonstrable decreased life expectancy, there are at present no solid data to support this hypothesis or that treatment with rhGH improves mortality outcome in adult hypopituitarism.

	VI. Rationale and Strategies for GH Replacement in Adults

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
In the absence of evidence that GH replacement therapy for adult GHD is associated with an improvement in mortality outcome, it is the development of symptoms of a characteristic clinical syndrome, discussed in detail below, that is the most frequent trigger for consideration of treatment of hypopituitary adults with rhGH. Unlike other hormones used for the treatment of hypopituitarism, for which the benefits of replacement therapy are universally accepted, there is considerable national variation in the clinical indications for the prescription of GH. This variation relates in part to financial constraints, but it is also indicative of an area of clinical practice in which the sudden availability of an unlimited supply of drugs has necessitated a rapid evaluation of the indications for its use and, hence, a strategy for its monitoring. Essentially three main approaches to the practice of GH replacement have been used. One is that because of its cost and the lack of evidence of its long-term efficacy in improving cardiovascular risk and reducing mortality, GH replacement should not be offered to hypopituitary patients A second approach, adopted by some countries, is that all hypopituitary patients with GHD require GH replacement, simply on the principle of complete hormone replacement therapy. In most countries, however, a third approach is adopted involving selection of patients for GH replacement. In the United Kingdom and many other European countries, selection is made largely on the basis of quality of life and/or bone mineral density considerations, but alternative strategies could include selection of patients with a particularly high cardiovascular risk profile.

The potential improvement of several of the adverse features of the adult GHD syndrome with rhGH therapy means, in turn, that changes in these parameters are used as clinical indicators of GH efficacy during replacement therapy. During the double-blind placebo-controlled trials of GH replacement, and in the immediate period after its license for use in adults, it was thought, by analogy with pediatric practice, that clinical monitoring would be sufficient and that markers such as body composition would simply substitute for linear growth. However, with time and shared clinical experience, it has become apparent that individual responsiveness to GH is highly variable and that the dose should be adjusted to suit each individual patient. This, in turn, is accomplished using a combination of tolerability (i.e., the occurrence of side effects), clinical response, and measurement of biochemical indices of GH action. In the debate surrounding the optimization of GH dosing schedules, supportive evidence comes from a combination of placebo-controlled studies (both single- and multicenter) and information collected through international outcomes-based multicenter research databases, in which data are recorded during longitudinal follow-up in a conventional clinical setting. Although patient numbers may be limited in a single center, it is often the case that those patients have been treated by a single physician or group of physicians in an identical manner, such that it becomes possible to draw conclusions about specific treatment protocols. In contrast, dosing strategies vary between centers, but large databases permit the identification of subtle trends regarding individual susceptibility, together with early detection of important safety issues that may not be possible from a single center or even a single country.

	VII. Tolerability

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
rhGH has an identical amino acid sequence to endogenous GH (61), such that side effects of GH replacement therapy are almost exclusively due to excess dosing. In the early trials of GH replacement, symptoms related to the antinatriuretic actions of GH, such as edema, arthralgia, and myalgia, were common and necessitated dose reductions in up to 40% of patients (2, 3). Side effects were more common in elderly and obese patients (141), both of whom would have received disproportionately larger doses, in the early studies using weight-based dosing regimens, than would be predicted from the physiological principles of GH secretion outlined above. Side effects were significantly reduced by beginning GH at smaller doses, building up to the (then) weight-based target dose. Conversely, trials involving mostly lean and young adults have been associated with fewer side effects (142). However, even when GH doses are lowered because of adverse symptoms, biochemical markers of GH action, such as serum IGF-I, remain elevated in up to one-third of patients (143), suggesting that the absence of classical symptoms of GH overtreatment is a relatively crude method of judging excess GH exposure.

	VIII. Clinical Response to GH Replacement

Top Abstract I. Introduction II. Physiology of GH... III. Pharmacokinetics of... IV. Review of Pediatric... V. Adult GH Replacement:... VI. Rationale and Strategies... VII. Tolerability VIII. Clinical Response to... IX. Is Overtreatment Acceptable... X. Biochemical Monitoring of... XI. Gender Differences in... XII. Adult GHD and... XIII. Alternative Mechanisms for... XIV. Transition from Pediatric... XV. Effects of Discontinuation... XVI. Dosing Strategies for... XVII. Influence of Adult... References

 
In the absence of conclusive evidence that GH replacement reduces cardiovascular mortality, it is the potential reversal of many of the features of GHD, most notably abnormal body composition, impaired quality of life, osteopenia with increased fracture risk, and cardiac dysfunction, that leads to the initiation of treatment. It might therefore be argued that each of these could be used as a marker of efficacy during GH replacement. In discussing the merits of different approaches to adult GH replacement therapy, it is important to note that supportive evidence comes from both placebo-controlled and open-label studies, each of which provide separate, but complementary, information.

A. Body composition
GHD adults have increased android (abdominal and visceral) fat, decreased lean body mass, and decreased total body water compared with age-matched healthy controls (1, 2, 144). Several double-blind, placebo-controlled studies, all slightly different in design, have shown consistent, beneficial effects on all these parameters with GH replacement therapy (1, 2, 3, 4, 5), attributable to its known lipolytic (145), protein synthetic (146), and antinatriuretic actions. A variety of techniques are available for the assessment of body composition in clinical trials of GH therapy in adult hypopituitarism, including bioimpedance analysis (147), isotope dilution estimation of total body water (TBW) (148), total body potassium (TBK) estimation using a ⁴⁰K counter (149), dual energy x-ray absorptiometry (150), anthropometry (151), and CT scanning (152). Although some groups have used a four-compartment model of body composition (144), most studies have monitored changes in body composition on the basis of a two-compartment model [fat mass (FM) and lean body mass (LBM)], each of which has distinct physico-chemical properties. For example, estimates of LBM from measurements of TBK rely on an assumption of 60 mmol potassium per kg LBM (153). FM is then calculated by subtracting the derived LBM from the total body weight. Isotopic dilution measurements of TBW may be used to calculate LBM on the basis that water constitutes 73% of LBM (hence LBM = TBW/0.73). It is important to note that such calculations are based on models of body composition in healthy, GH-replete individuals and that extrapolation to GHD adults may not be strictly valid. Further, although most techniques measure body fat with accuracy and precision, some techniques, notably bioimpedance analysis and dual energy x-ray absorptiometry, may overestimate LBM changes (154). However, as can be seen from Table 1, the qualitative effects of GH replacement on body composition in adult GHD (both AO and CO) have been strikingly similar in the trials listed, all of which were randomized, double-blind, and placebo controlled. Some of the quantitative differences between studies may, at least in part, be attributed to different GH dosing regimens used and the discrepancies known to exist between the various techniques available for the measurement of body composition (154). Although many of the above techniques are not routinely available outside supervised clinical trials, it should be noted that the simple measurement of waist-hip ratio correlates well with the reduction in visceral FM that occurs with GH replacement and provides a sensitive and reproducible method of monitoring certain aspects of altered body composition during GH therapy (2).

In a separate study (159), assessments of body composition and measurements of biochemical markers of GH action (IGF-I, IGFBP-1, and BP-3 and ALS) were monitored during 12 months of treatment with a weight-based GH dosing regimen. Significant improvements in body composition were observed. Although no individualized dosing was made on the basis of the above measurements, dose reductions were necessary in 7 of 20 patients because of side effects of fluid retention. Even with these dose reductions, serum IGF-I remained elevated in seven patients (35%), while ALS and IGFBP-3 were above the age-related reference range in five (25%) and three (15%) patients, respectively (159).

De Boer et al. (143) conducted a 12-month placebo-controlled trial of GH replacement, randomizing 46 GHD male patients to one of four treatment protocols: placebo for 6 months followed by GH 2 IU/m²/d; or one of three doses of GH for 12 months (1, 2, and 3 IU/m²/d). Some reductions in dose were necessary due to unacceptable side effects of GH excess, but the absence of such symptoms was a poor guide to overtreatment, judged by serum IGF-I levels (Fig. 2). Most of the patients treated with the highest dose of GH had serum IGF-I levels outside the age-related reference range. It should be noted that, in this study, the doses of GH necessary to normalize serum IGF-I were also associated with restoration of normal tissue hydration, emphasizing the potential use of clinical markers of GH efficacy during GH dose titration.

View larger version (18K): [in this window] [in a new window]   Figure 2. Measurement of IGF-I response in 46 men with CO-GHD randomized to receive one of three starting doses of GH with subsequent adjustment according to clinical characteristics. •, Symptoms of GH excess. , No symptoms of GH excess. [Reproduced with permission from H. de Boer et al.: J Clin Endocrinol Metab 81:1371–1377, 1996 (143 ). © The Endocrine Society.]

B. Quality of life and well-being
Reduced quality of life and sense of vitality are well recognized features of the adult GHD syndrome (161, 162). Although this may in part relate to abnormal body composition and impaired muscle strength, there is widespread agreement that the low energy levels, social isolation, increased emotional stress, impaired socio-economic performance, and greater difficulties forming relationships evident in many hypopituitary patients are directly due to GHD. In many countries, availability of GH is limited to patients with severe GHD associated with one or a combination of these symptoms, and their improvement is therefore an important clinical parameter by which to judge the efficacy of GH replacement. The early trials of GH replacement used a variety of generic methods to measure and monitor well-being such as the Nottingham Health Profile (NHP) (163) and the Psychological General Well-Being Schedule (PGWBS) (164). The NHP questionnaire consists of a number of specific questions, with yes/no responses, about energy levels, sleep, relationships, emotional responses, physical mobility, and pain. The Psychological General Well-Being Schedule (PGWBS) involves the use of a rating (from 0, worst, to 5, best) for a series of questions about affective categories such as anxiety, depression, positive well-being, general health, and vitality.

A number of placebo-controlled studies have documented statistically significant improvements in well-being, assessed by various methods, after the initiation of GH replacement (3, 162, 165), although such results have not been universally reported (5, 166). The reasons for such discrepancies are not entirely clear, although it is interesting to note that many patients in the study of Baum et al (166) did not achieve significant increments in serum IGF-I levels, suggesting that compliance with GH therapy may have been suboptimal. It is also important to note that patients who participated in the early trials of GH therapy were frequently the most severely disadvantaged in terms of psychological distress (167) and therefore more likely to wish to continue with GH replacement after a therapeutic trial (168). Hence, because of these findings, caution should be exercised in the interpretation of psychological well being data.

Since these placebo-controlled studies, a number of open-label studies have examined the clinical utility of using well-being scores as a marker of efficacy during GH replacement. In most cases, these studies have used doses of GH that would now be considered inappropriately high, and the proportion of patients with a serum IGF-I above the age-related reference range was, in some cases, as high as 56% (169). In one study (170), the effect of two different GH dosing regimens (0.012 and 0.024 mg/kg/d) on well-being, judged by NHP and PGWB scores, was investigated. Identical improvements in well-being were observed in both groups, yet 45% of the patients in the higher dose group had an elevated serum IGF-I compared with 24% receiving the lower dose. In other words, the extra GH administered resulted in no greater clinical benefit in terms of well-being, but was associated with biochemical overtreatment in nearly twice as many patients. In the same report, of those patients that chose to continue GH therapy after the conclusion of the trial, 33% had a supranormal serum IGF-I compared with 30% of those who elected to discontinue due to a lack of improvement in well-being. If the dose of GH in the "non-improvers" had been increased because of a poor clinical response, it may have pushed serum IGF-I further into the acromegalic range.

More recently, attempts have been made to use scoring systems for psychosocial morbidity that are more specific to GHD. The adult GHD assessment (AGHDA) score (171, 172, 173) provides a sensitive and highly reproducible method of monitoring improvements in the psychosocial consequences of GHD that may accompany GH replacement therapy. The AGHDA questionnaire consists of 25 questions derived from the symptoms most frequently reported by patients with adult-onset GHD. A score of 25/25 represents the worst possible well-being score, while scores of 4/25 or less have been recorded in a normal control population (174). In those patients whose well-being improves with GH, improvements in AGHDA scores occur within 3 months of GH replacement therapy in the majority of patients and are maintained at 6 and 12 months (175). Interestingly, improvements in AGHDA scores may be seen in some patients treated with GH whose dose is insufficient to have caused a significant increment in serum IGF-I, suggesting that improvements in the psychological aspects of GHD may, at least in part, be mediated directly by GH rather than via generation of IGF-I (175). It is not known whether patients exposed to excess GH (either in the context of acromegaly or by overtreatment with GH in hypopituitarism) have AGHDA scores that are different from control populations. However, an interesting comparison can be made between hypopituitary patients treated initially on weight-based dosing schedules, with subsequent dose adjustment during clinical follow-up and patients initially started on low doses of GH with subsequent careful dose titration on the basis of levels of serum IGF-I (175). Maintenance doses of GH and serum IGF-I levels were significantly higher in the patients initially treated with weight-based dosing schedules, yet well-being, as judged by AGHDA score, was no different.

C. Bone density and bone remodeling
It is thought that, in childhood, GH promotes longitudinal bone growth by a combination of a direct effect on epiphyseal chondrocytes (176) and by paracrine generation of IGF-I (177). However, it is now widely accepted that GH also has an important role to play in the achievement of peak bone mass after the completion of linear growth and also in the maintenance of bone mass through adult life. AO hypopituitary patients receiving conventional endocrine replacement therapy are osteopenic compared with age-matched healthy controls (178, 179), an observation that is almost certainly clinically relevant given the increased fracture rate evident in this patient group (180). Furthermore, there are data to suggest that the severity of bone loss is proportional to the biochemical severity of GHD (181). The mechanisms for this disadvantage are not fully understood but are likely to relate, at least in part, to reduced bone remodeling activity. Activity of the bone remodeling unit (i.e., the rate of bone turnover) may be assessed by measuring markers of the activity of the two limbs of the bone remodeling unit. Osteoclasts mediate bone resorption, and indices of their activity include pyridinoline, deoxypyridinoline, and serum type I carboxy-terminal cross-linked telopeptide. Markers of osteoblastic activity (bone formation) include the bone-specific isoenzyme of alkaline phosphatase (BSAP), osteocalcin, and carboxy-terminal propeptides of type I collagen. GH stimulates proliferation and differentiation of osteoblasts in vitro in humans (28) and in mice (182) and further, surrogate, evidence for an important effect of GHD in vivo is supported by the observation of subnormal levels of osteocalcin and BSAP in adults with GHD (142, 183).

Although osteopenia is an important factor in considering a trial of GH replacement, few clinicians would regard it as the sole reason to begin treatment. However, changes in BMD represent an important marker of efficacy of GH therapy, and a review of the data in this regard is appropriate. A number of placebo-controlled trials have examined the effect of GH replacement on bone metabolism and BMD. From these studies it is apparent that GH replacement is frequently associated with a reduction in bone density in the short term (184, 185, 186), probably as a result of an expansion of the bone remodeling space (186). However, with more prolonged treatment increases of bone density of 4–10% above baseline, measurements have been recorded (185, 187). The study of Baum et al. (187) is particularly noteworthy as the increments in BMD were achieved with a dosing regimen of GH that specifically aimed to avoid overtreatment by maintaining serum IGF-I levels within the age-adjusted normal range. The timescale over which the