2.3.7 Adult growth hormone deficiency

It has been known for many years that growth hormone is essential for normal linear growth, but over the past few years, with the advent of recombinant human growth hormone therapy, the importance of growth hormone during adult life has been described in detail. The growth hormone peptide was first isolated from bovine pituitaries in the 1940s (1), but was found to be species specific and inactive in humans. In 1956, growth hormone was extracted from human cadaveric pituitary tissue (2) and a year later was administered to a 13-year-old boy with hypopituitarism, resulting in an increased growth velocity (3). The first report suggesting growth hormone could have beneficial actions in adulthood was published in 1962 in which a 35-year-old woman with hypopituitarism reported increased vigour, ambition, and wellbeing after 2 months treatment with cadaveric growth hormone (4). However, the limited supply of pituitary-derived growth hormone confined its use to the treatment of children with severe growth failure caused by proven growth hormone deficiency (GHD). In 1985, the association of cadaveric growth hormone treatment with Creutzfeldt–Jakob disease led to its withdrawal from use worldwide (5). Since then, all growth hormone in clinical use has been produced using recombinant DNA technology.

The first placebo-controlled trials of growth hormone replacement therapy in adults with GHD were published in 1989 (6, 7). These and subsequent studies have led to the recognition of adult GHD as a specific clinical syndrome and the impact of GHD and replacement therapy in adults with GHD has been studied in detail.

Growth hormone is a 191 amino acid single chain polypeptide hormone synthesized and secreted by somatotrophs in the anterior pituitary in a pulsatile fashion, with peaks separated by nadirs during which growth hormone levels fall below the sensitivity of routine assays and are only detectable with sensitive chemiluminescent assays (8). Two hypothalamic hormones are the predominant regulators of growth hormone secretion: growth hormone-releasing hormone (GHRH), which stimulates both growth hormone synthesis and growth hormone release, and somatostatin, which inhibits growth hormone release (but not its biosynthesis) (9). A third factor is thought to be ghrelin, the natural ligand of the endogenous growth hormone secretagogue receptor, distinct from that for GHRH (10). The stomach is the principal source of circulating ghrelin, and as well as its growth hormone-releasing role, ghrelin has orexigenic activity among other functions (11). It has been shown that ghrelin stimulates growth hormone secretion synergistically with GHRH, which is required for ghrelin to exert its effect as a growth hormone secretagogue (12, 13), however, the physiological role of endogenous ghrelin in growth hormone regulation remains to be determined. Multiple neurotransmitter pathways as well as a variety of metabolic and hormonal factors are also involved in the regulation of growth hormone secretion, either by acting directly on the somatotrophs and/or modulating GHRH or somatostatin release. It is through these pathways that stress, sleep, exercise, hypoglycaemia, and high levels of circulating amino acids (such as arginine) stimulate growth hormone secretion, while high levels of glucose and free fatty acids inhibit its secretion. Other hormones influence growth hormone secretion: oestradiol increases growth hormone secretion when administered orally and glucocorticosteroids and thyroid hormone excess impair growth hormone release (9). Growth hormone secretion is known to be higher in premenopausal women than in age-matched men (14) and is inversely associated with increasing age and adiposity (15, 16), in particular abdominal obesity (17).

Growth hormone circulates bound to a growth hormone-binding protein, which is homologous to the extracellular domain of the growth hormone receptor (GHR) (18). Growth hormone exerts its effects directly by binding to the extracellular domain of the GHR, at the cell surface which causes dimerization of two GHR molecules and initiates the intracellular signalling pathway which includes Janus kinase and the signal transducers and activators of transcription (STAT) pathway (19). The metabolic effects of growth hormone are mediated through the subsequent production of insulin-like growth factor 1 (IGF-1). The liver is the dominant source of circulating IGF-1 although it can also be generated in many other tissues, where it appears to act in an autocrine/paracrine fashion (20). A small proportion of IGF-1 circulates in a free state, but the majority is associated with a tertiary complex consisting of IGFBP3 and the acid-labile subunit (ALS), significantly prolonging its half-life, maintaining stable circulating levels, and regulating its availability to target tissues (21). These three peptides, IGF-1, IGFBP-3, and ALS, are all growth hormone dependent. Many factors affect both hepatic and local tissue IGF-1 production in response to a given growth hormone stimulus, including sex steroids, thyroxine, glucocorticoids, insulin, and liver disease. Androgens enhance the IGF-1 response to growth hormone, and oestrogens attenuate growth hormone action by reducing IGF-1 in a dose-dependent fashion following oral administration (22).

Somatostatin, GHRH, growth hormone, and IGF-1 are maintained homoeostatically in the hypothalamus, pituitary, and circulation by a complex interplay of feedback signals. Growth hormone inhibits its own secretion indirectly by regulation of GHRH and somatostatin release from the hypothalamus, as well as directly acting on the somatotrophs. IGF-1 inhibits growth hormone secretion through a negative feedback action on the pituitary and less clearly on the hypothalamus (9).

Adults with GHD are primarily divided into those who develop the condition during childhood and those diagnosed during adulthood. Patients who previously had childhood GHD account for 15–27% of the patients with adult GHD (23, 24), idiopathic being the most common cause in this group (25). Isolated, idiopathic GHD of childhood appears to be reversible in a large proportion of cases when retested at the completion of linear growth. Up to 80% of cases demonstrate a normal growth hormone response when retested in young adult life. Individuals with structural disease of the hypothalamic pituitary axis are less likely to recover their growth hormone axis and in many cases the severity of GHD may increase (25). It has been shown that between 40% and 60% of young adults who have completed growth hormone replacement therapy in childhood continue to have a degree of GHD. Therefore, reassessment of growth hormone status once linear growth is completed is mandatory (26).

The causes of GHD that occur in adulthood are summarized in Table 2.3.7.1, using data derived from 1034 patients enrolled in the KIMS database, a multinational, pharmacoepidemiological survey of patients receiving growth hormone replacement (23). Although GHD may occur in isolation, it is often observed in the context of multiple pituitary hormone deficiencies, where growth hormone is typically the first hormone to become clearly deficient. In this way, in patients with organic hypothalamic-pituitary disease, the likelihood of GHD increases with the increasing number of pituitary hormone deficits from approximately 45% if no other deficits are present to 95% if three or four pituitary hormone deficiencies are present (27, 28) (Fig. 2.3.7.1). Adult-onset GHD usually results from damage to the pituitary gland or the hypothalamus, the most common cause being a pituitary adenoma or its treatment with surgery or radiotherapy (25). Pituitary microadenomas are rarely associated with hypopituitarism, and patients with this condition may not need assessment of GHD unless other pituitary hormone deficits are present or unless there is a strong clinical suspicion of GHD (27). Radiotherapy induces damage to the neuroendocrine axes and growth hormone axis is the most susceptible, being the earliest and most frequent pituitary hormone affected following radiotherapy. The severity and speed of onset of radiation-induced GHD is dose and time (elapsed postirradiation) dependent (29). Children appear to be more radiosensitive than adults and GHD is found less frequently following irradiation in adulthood (30).

Traumatic brain injury (TBI) and subarachnoid haemorrhage (SAH) have been reported to produce some degree of hypopituitarism which can be permanent in 27.5 and 47% of patients respectively (31). In several of these studies, GHD has been identified as the most frequent pituitary deficiency, with GHD reported in 12.4% and 25.4% of the patients after TBI and SAH, respectively (31). Severe GHD has also been reported in 8% to 21% of TBI patients (32). Isolated GHD is rarer in adults than in children, accounting for approximately 10% of cases of GHD in adulthood (23, 33). In this group of patients, nonfunctioning and secreting pituitary adenomas are the most common primary diagnosis (33). Idiopathic isolated GHD occurring de novo in adults is not a recognized entity (34).

Adult GHD is recognized as a clinically relevant syndrome associated with a variety of symptoms and signs, which are summarized in Box 2.3.7.1. It causes abnormalities that affect multiple systems, but three areas are most relevant clinically: quality of life (QoL), cardiovascular risk (including disturbance of body composition), and skeletal health.

Adults with GHD report reduced psychological wellbeing and quality of life compared with matched healthy controls, particularly low energy levels, social isolation, greater emotional liability, impaired socioeconomic performance, and greater difficulties forming relationships (35). The aspect of quality of life most frequently affected by GHD is energy and vitality (25). Initial studies evaluating quality of life in GHD used a variety of generic questionnaires designed for subjects with long-term illness, but condition-generated questionnaires focusing on quality of life issues that are relevant to patients with GHD have been developed (35). One of these questionnaires which is now widely used for the baseline and follow-up of patients, is the Quality of Life Assessment in Growth Hormone Deficient Adults (QoL-AGHDA), a disease-generated questionnaire developed from interviews with patients with GHD (36). It consists of 25 questions, with a yes/no response, the score being determined by the number of positive answers. A score of 25/25 indicates poor quality of life, scores of 4/25 or less have been reported in a normal control population (37). However, quality of life evaluations of GHD have shown a high degree of variability, and while some GHD patients report severe impairment of quality of life, others demonstrate a normal quality of life (25). This disparity may be because of many other possible influences on the quality of life, such as the age of onset of GHD. Impairment of quality of life is less frequently found in patients with childhood-onset GHD (38) and in older patients (39).

Growth hormone has important effects on lipid, protein, and carbohydrate metabolism (40, 41). Consequently, GHD is associated with substantial changes in body composition; body fat mass is increased (by approximately 7%) (42), with a propensity towards visceral fat deposition (43). This central obesity is known to increase cardiovascular risk via several mechanisms including atherothrombotic and proinflammatory abnormalities, insulin resistance, hypertension, and dyslipidaemia (43). The cause of central obesity in adult GHD is not clear, although it has been suggested that increased local tissue exposure to cortisol could play a role. IGF-1 inhibits the enzyme 11B-hydroxysteroid dehydrogenase type 1 (11β-HSD1), responsible for conversion of inactive cortisone into active cortisol in liver and adipose tissue (44). Patients with GHD in the context of hypopituitarism have an increased cortisol/cortisone metabolite ratio and reduction in circulating cortisol concentrations in patients receiving hydrocortisone replacement (44). In GHD, central obesity (as well as insulin resistance and other cardiovascular abnormalities) could be a consequence of exposure to raised cortisol levels at key target tissues, and therefore the reported benefits of growth hormone replacement on cardiovascular risk factors may be an indirect effect of alterations in cortisol metabolism (44).

Lean body mass has been shown to be reduced by 7% in GHD (42), which may explain the reductions in muscle strength and exercise capacity observed in these patients. Some of the decrease in lean mass could be due to the reduction in total body water seen in adults with GHD, as growth hormone has antinatriuretic properties (45). Growth hormone also affects lipoprotein metabolism, and adult GHD is associated with an atherogenic lipid profile. Increased levels of total and low-density lipoprotein (LDL) cholesterol, triglycerides, and apolipoprotein B-100, with normal or decreased high-density lipoprotein (HDL) cholesterol have been reported in GHD adults (25, 46).

Although low, normal and high basal levels of insulin have been found in GHD adults, probably reflecting different degrees of obesity, it has also been demonstrated using the hyperinsulinaemic euglycaemic clamp method that there is a twofold to threefold reduction in insulin sensitivity in GHD patients compared with controls, despite normal fasting glucose and insulin levels (43). Blood pressure has not consistently been shown to be increased in patients with GHD, and the possible mechanism of hypertension is not clear (35).

GHD is associated with decreased fibrinolysis, with augmented levels of fibrinogen, plasminogen activator inhibitor 1 (PAI-1), and tissue plasminogen activator antigen (tPA) (46). These changes in thrombogenic proteins may contribute to the development of atherosclerosis and cardiovascular events. Patients with GHD have endothelial dysfunction, which is an early step in the atherogenic process and is associated with decreased nitric oxide production (47). Indeed, IGF-1 has a direct stimulatory effect on nitric oxide synthesis, and nitric oxide formation is decreased in GHD patients (46, 47). Furthermore, GHD is associated with elevated levels of inflammatory markers also associated with cardiovascular risk, such as C-reactive protein (CRP) and interleukin 6 (IL-6) (35). GHD patients also have abnormal parameters of vascular integrity, such as increased arterial intima-medial thickness (independent risk marker for myocardial infarction and cerebrovascular accident (48)), reduced endothelium-derived flow-mediated dilation, reduced muscle blood flow and increased vascular resistance associated with increased sympathetic nerve activity (49). Adults with GHD in addition to the above effects on large- and medium-sized vessels show reduced capillary density, microvascular perfusion, and capillary leakage (49).

Finally, GHD is associated with changes in cardiac size and function, and although there are conflicting results in patients with adulthood-onset GHD, reductions in left ventricular mass (50, 51) and left ventricular systolic dysfunction (51, 52) have been described in some studies. GHD exerts adverse effects upon several cardiovascular risk factors, and it has been shown that both cerebrovascular and cardiovascular morbidity are increased in GHD patients (53). Several studies have reported a twofold increase in the overall mortality rate (mainly due to increased cardiovascular mortality) in hypopituitary adults without growth hormone replacement (53–56). Although GHD may contribute to the increase in morbidity and mortality observed in these patients, multiple factors are likely to be involved, such as the underlying disease, untreated hypogonadism, excessive glucocorticoid or thyroxine replacement, and previous treatment with surgery or radiotherapy.

Growth hormone has an important effect on bone metabolism. Initially stimulating bone resorption followed by bone deposition the overall effect of growth hormone-dependent bone remodelling is anabolic. Growth hormone is involved in the promotion of linear growth in childhood, achievement of peak bone mass after cessation of linear growth, and maintenance of bone density through life (57). Adult patients with GHD have a marked reduction in bone turnover, demonstrated by bone biopsies which show reduced osteoid and mineralization surfaces and decreased bone formation rate and by the finding of diminished levels of both markers of osteoclastic and osteoblastic activity (57).

Bone mineral density (BMD) in adults with GHD is approximately 1 SD score below those of age- and sex-matched controls, even when the possible effect of hypogonadism or glucocorticoid over-replacement are corrected for (25). The severity of the observed bone loss is related to the age of the patient the severity of GHD and the age of onset of GHD (25, 57). Patients with childhood-onset GHD have a more severe reduction in bone mass than those developing GHD in adulthood. Achievement of optimal peak bone mass requires growth hormone. In young adults, GHD prevents the acquisition of optimal bone mass resulting in a relatively severe osteopenia (57). As age increases the severity of osteopenia associated with GHD declines until it is no longer apparent after the age of 55 or 60 years (58, 59).

There have been reports of an increased risk of nonvertebral fractures (frequently localized in the radius) as well as radiological vertebral deformities (suggesting vertebral fractures) in GHD patients (57). The prevalence of bone fractures is related to the degree of GHD and seems not to be affected by concomitant hypopituitarism or by replacement of other pituitary hormones (57).

There are conflicting data suggesting resting energy expenditure (REE) is affected by GHD, as decreased REE and REE expressed in terms of lean body mass have been described in some, but not all studies (60). In contrast, GHD patients show reduced exercise capacity, evidenced by cycle ergometry, with a 20–30% reduction in maximum oxygen uptake compared to those predicted for age, gender, and height. This could be attributed to the reduction in lean body mass that leads to decreased muscle strength, the diminished cardiac capacity, the decreased oxygen transport capacity (as IGF-1 stimulates erythropoiesis), as well as an impaired ability to dissipate heat by sweating after heat or exercise (60). Adults with GHD have reduced isometric muscle strength (static contraction of a muscle without any visible movement in the angle of the joint) and reduced or low normal isokinetic muscle strength (muscle contracts and shortens at constant speed) when compared with normal controls (35). Moreover, local muscle endurance has also been found to be either reduced or in the lower normal range in adults with GHD (35).

In addition to the effects of GHD described above, GHD patients have been reported to have a dry, thin, and cool skin, probably related to the loss of direct anabolic actions of growth hormone on skin, decreased cardiac output, and decreased sweating (61). Finally decreased glomerular filtration rate has also been described in GHD patients (62).

The manifestations of GHD in adults, in contrast to the reduced growth velocity observed in children, are subtle and none is pathognomonic. In the absence of a good clinical indicator that will reliably discriminate a GHD patient from the normal population the diagnosis of adult GHD is based on the results of biochemical testing in an appropriate clinical context (27). Current consensus guidelines identify two groups of patients who should be tested for GHD in adult life; those at risk of hypothalamic-pituitary dysfunction, in whom there is an intention to treat with growth hormone replacement, and those with childhood-onset of GHD who have reached final height (34) (Fig. 2.3.7.2). The first group includes patients with evidence of hypothalamic-pituitary disease (endocrine, structural, and/or genetic causes), those who have received cranial radiotherapy, which impacts on the hypothalamic-pituitary axis, and those with TBI or subarachnoid haemorrhage. For those patients with childhood-onset GHD, the need for continuation of growth hormone replacement should be assessed once final height has been achieved, although repeat growth hormone testing is not required for those with a transcription factor mutation, those with more than three pituitary hormone deficits, and those with isolated GHD associated with an identified mutation (34). All other patients should undergo growth hormone testing after at least 1 month off growth hormone therapy (34). Idiopathic isolated GHD in adulthood is not a recognized clinical entity, so adults who do not fulfil the criteria outlined above should not undergo assessment of their growth hormone status.

Growth hormone is secreted in a pulsatile fashion and has a short half-life, which renders assessment of random serum growth hormone concentrations worthless for the diagnosis of GHD, although a single growth hormone measurement taken fortuitously at the time of a secretory peak may exclude GHD (63). Because of this pulsatile secretion, multiple sampling of growth hormone levels would be ideal (24-h profile with a minimum of 20-min sampling), but is impractical in routine clinical practice (25), so endocrinologists rely on dynamic function tests to determine growth hormone status.

A number of agents have been used to stimulate the release of a growth hormone pulse. The insulin tolerance test (ITT) is regarded as the ‘gold standard’, as it distinguishes GHD from the reduced growth hormone secretion that accompanies normal ageing and obesity, provided adequate hypoglycaemia is achieved (27). The ITT should be carried out in dedicated units, under supervision by experienced staff (27). The ITT is contraindicated in patients with electrocardiographic evidence or a clinical history of ischaemic heart disease or seizures and in elderly people (34). The ITT should be performed with caution in brain-injured patients (64). There are other accepted and validated alternative tests when the ITT is contraindicated, including the glucagon stimulation test, GHRH with arginine, and GHRH with growth hormone-releasing peptide 6 (GHRP-6). The glucagon stimulation test is as reliable as the ITT, and has the same validated cut-off for GHD in adults: a peak growth hormone response of <3 μg/l (34, 65) (Table 2.3.7.2). The other tests consist of the combined administration of GHRH with either arginine (that seems to reduce hypothalamic somatostatin secretion) or a synthetic growth hormone secretagogue such as GHRP-6 (34, 65), although the latter is not available for use in normal clinical practice. Both tests provide a potent stimulus to growth hormone secretion and constitute appropriate alternatives to the ITT, with a very good safety profile and relatively fewer contraindications (65). It must be noted though, that these combined tests may miss GHD due to hypothalamic disease (e.g. those having received irradiation of the hypothalamic-pituitary region), as they stimulate both the hypothalamus and the pituitary. Therefore, they are not recommended for the diagnosis of severe GHD in patients with cranial irradiation, the ITT being the preferred test to be used in this situation, as this test shows the greatest sensitivity and specificity within the first 5 years of irradiation (25, 34, 65) following which all tests seem to be reliable and generally concordant (65).

As growth hormone secretion is affected by age, gender, and body mass index (BMI), the majority of tests are limited by the lack of validated normative data based on these parameters. Obesity is particularly important, as both spontaneous and stimulated growth hormone secretion is negatively associated with BMI, and sometimes the growth hormone response to provocative stimuli is as impaired to the same degree as in hypopituitary patients with severe GHD. This constitutes a clinical problem in the interpretation of the growth hormone response to provocative tests in obese patients in whom GHD is suspected, particularly as GHD is often associated with obesity. It is recommended that diagnostic thresholds appropriate to lean, overweight, and obese subjects are used, in order to achieve an appropriate diagnosis in obese adults and lean GHD adults (65). For the ITT, the diagnostic threshold is <3 μg/l and can be applied to all as it has been shown to distinguish normal subjects (including those who are obese) from patients with severe GHD (65). Cut-off levels of growth hormone response validated by BMI are not available for the glucagon test, but are clearly defined in the GHRH + arginine test (34, 65) (see Table 2.3.7.2).

Serum levels of the growth hormone-responsive molecules IGF-1 and IGFBP3 are stable through the day, show minimal diurnal variation and need not be drawn fasting (61). IGF-1 and IGFBP3 effectively provide integrated markers of growth hormone secretion that can be utilized as indicators of growth hormone status. Of the two, IGF-1 is the most sensitive marker of growth hormone action, but its diagnostic value in adults is limited by the significant overlap in IGF-1 values between GHD and normal controls; this overlap is more prevalent in GHD patients with adult-onset than those with childhood-onset GHD (65, 66). Therefore, a normal age-adjusted IGF-1 value does not exclude the diagnosis of GHD in adults (25, 27, 65) (Fig. 2.3.7.3). Nonetheless, in the absence of conditions that may decrease IGF-1 generation, including malnutrition, hepatic disease, hypothyroidism, or poorly controlled diabetes mellitus, an IGF-1 below the reference range in the presence of multiple pituitary hormone deficits confirms the diagnosis of GHD (27). Thus, although it is widely accepted that the diagnosis of adult GHD is established by provocative testing of growth hormone secretion, it is accepted that, in patients at risk of GHD (childhood-onset GHD, severe GHD, or multiple hormone deficits acquired in adulthood), a serum IGF-1 below the age-specific normal range is diagnostic (65) (see Fig. 2.3.7.2).

Finally, it is important to note that the actual value reported for the growth hormone concentration in a specific’s serum sample is determined to a great extent by the assay method used, limiting the applicability of international consensus guidelines to local clinical practice. Some of the reasons for the difficulty in growth hormone assays standardization are the heterogeneity of the analyte itself, the availability of different reference preparations for calibration and the interference from matrix components such as GHBP (67). In a recent study more than twofold variation in growth hormone and IGF-1 values measured in different laboratories was found, probably due to variability in assay performance, coupled with use of inappropriate conversion factors and reference ranges (68). The Growth Hormone Research Society (GRS) recommends the adoption of universal growth hormone and IGF-1 calibrators (recombinant 22 kDa growth hormone calibrator, International Reference Preparation 98/574 and a recombinant human IGF-1 of the highest purity) by all growth hormone and IGF-1 assays manufacturers, in order to reduce this substantial heterogeneity among existing assays (34).

The rationale for treating adults with GHD is primarily based on the improvement of the features discussed above. To date, there is no evidence that treatment of adult GHD results in normalization of the increased mortality observed in this patient population (37, 63). In contrast to other endocrine replacement used in hypopituitarism, for which there is general agreement regarding its use and efficacy, there remains considerable variation in the use of growth hormone between centres and between countries. Some endocrinologists advocate the blanket approach, adopted by the GRS consensus guidelines, which suggest all patients with documented severe GHD are eligible for growth hormone treatment and should be treated (27, 34). The approach adopted by the UK and other European countries is to select patients for growth hormone replacement based on impaired quality of life or to optimize bone mass. (37) (Box 2.3.7.2). In the UK, current guidance for growth hormone replacement in adults was issued by the National Institute of Clinical Excellence (NICE) in 2003 (Box 2.3.7.3), (69). These guidelines restrict access to growth hormone replacement therapy to patients with an impaired quality of life, defined as a baseline QoL-AGHDA score of 11 or more.

The benefits of growth hormone replacement therapy on psychological wellbeing and perceived health have been described in some double-blind, placebo-controlled trials, and in a meta-analysis of open studies of growth hormone replacement (35). The findings of double-blind, placebo-controlled studies are not consistent; some have shown a definite benefit in quality of life, others have reported a more limited improvement or no change in quality of life after growth hormone therapy (25). On the other hand, some long-term postmarketing surveillance studies that have compared the quality of life in GHD adults treated with long-term growth hormone replacement with general population data from several European countries have shown a sustained improvement in quality of life towards the normative country-specific values (35). Finally, a sustained improvement in overall psychological wellbeing in GHD patients after 10 years of growth hormone therapy has also been demonstrated (70)

In general, the degree of improvement in quality of life is proportional to the degree of impairment at the outset, but shows no correlation with the degree of improvement in IGF-1 levels; this means that if the quality of life at baseline is normal, no improvement will be seen with growth hormone replacement (25). It is worth noting, that concomitant obstructive sleep apnoea syndrome (which is a common finding in hypopituitary adults) may confound quality of life evaluation; therefore in severe GHD patients not responding adequately to growth hormone treatment this syndrome should be ruled out (35).

In addition to the beneficial effects upon quality of life, improvements on cognitive performance in GHD patients during growth hormone replacement have also been described, particularly in attention and memory (35). In practice, there are patients who perceive a benefit from growth hormone replacement therapy and report an improvement in their quality of life which can be measured, while others perceive no benefit and will chose to discontinue therapy.

The beneficial effects of growth hormone replacement on body composition have been consistently observed. Total body fat (especially visceral fat) is reduced, and lean mass is increased (Table 2.3.7.3) (35, 71). Some of the increase of lean body mass is accounted for by growth hormone-mediated fluid retention (61); however, there is also a genuine increase in skeletal muscle mass (61). Long-term studies of growth hormone replacement (over 5–10 years) have confirmed a sustained improvement in lean body mass, although in some of these studies the increase in total body nitrogen, which reflects total body protein, seemed to be transient. After 10 years of growth hormone therapy, body weight has been found to increase, and total body fat returns towards baseline values. These changes may reflect the changes in body composition associated with normal ageing rather than a waning of the effects of growth hormone replacement (35).

Short-term studies of growth hormone replacement therapy in GHD adults have shown a reduction or no change in serum total cholesterol concentrations, a reduction of serum LDL cholesterol and apoprotein B concentrations (35). Serum HDL cholesterol concentrations are generally unchanged although some have demonstrated an increase. Serum triglyceride levels have mostly been unchanged (35) (Table 2.3.7.3) (71). The magnitude of the reduction of both total and LDL cholesterol is greater in those patients with higher baseline serum cholesterol levels, and it occurs even in those patients receiving concurrent lipid-lowering agents (HMG CoA-reductase inhibitors) (72). Long-term growth hormone replacement studies also show that the beneficial changes in serum lipoprotein profile are maintained following 10 years of growth hormone therapy, and that the improvement of serum lipid levels may even be progressive over that period (35). Serum lipoprotein(a), a proposed independent risk factor for cardiovascular disease, is increased after initiation of growth hormone replacement, although this observation has not been uniform (73). Although this contradictory data could be due to lipoprotein (a) assay differences (73), its overall significance regarding cardiovascular risk remains to be determined.

During short-term (<6 months) growth hormone replacement therapy, there is an initial period during which insulin resistance is adversely affected, due to increased lipolysis with elevated circulating free fatty acid concentrations (35, 47). Insulin sensitivity returns towards baseline values after 3 months of growth hormone treatment as the beneficial effect upon body composition become apparent (35). The long-term (≥1 year) effect of growth hormone replacement therapy on insulin sensitivity is not clear. Some studies have demonstrated unchanged insulin sensitivity compared to baseline and others reported that insulin sensitivity is still compared to baseline assessments (35). As individual patients have differential sensitivity in these parameters, it is not surprising that with growth hormone administration, some show a worsening of insulin sensitivity, while others show little change (25). A meta-analysis of blinded, randomized, placebo-controlled trials of growth hormone treatment in adults with GHD showed that growth hormone therapy significantly increases both fasting insulin levels and plasma glucose, although mean glucose levels remained in the normal range (71) (Table 2.3.7.3). However, this meta-analysis included very few studies with a prolonged follow-up (≥12 months), and insulin sensitivity was assessed in only one long-term trial (74) showing that the initial increase of both insulin levels and insulin-to-glucose ratios reported after growth hormone therapy, was not maintained at 18 months, and the HbA_1c did not change significantly.

Finally, there are contradictory results of glucose metabolism in several long-term trials of growth hormone replacement; while two of them reported no changes in glucose homoeostasis (assessed by glucose tolerance and insulin concentrations on the first one (75), and by fasting glucose, insulin and C peptide on the second (70)) after 7 and 10 years, respectively, of growth hormone therapy, others showed an increase (76) and reduction (77) of HbA_1c, respectively, after 10 years’ of growth hormone replacement.

There are discordant data about the medium and long-term effects of growth hormone replacement on cardiovascular abnormalities (78). However, in a meta-analysis reviewing 16 trials (nine blinded, placebo-controlled trials and seven open studies) of growth hormone treatment in GHD adults, growth hormone treatment was found to have a positive effect on many cardiac parameters assessed by echocardiography, such as a significant increase in left ventricular mass, interventricular septum thickness, left ventricular posterior wall, left ventricular end-diastolic diameter, and stroke volume (79).

GHD is associated with endothelial dysfunction with impairment of nitric oxide production. Although growth hormone therapy is not associated with a very impressive effect on circulating markers of endothelial dysfunction (35), it does normalize urinary nitrate excretion (47), endothelium-derived flow mediated dilation, and brachial artery blood flow (49). There are conflicting results of growth hormone replacement effect on blood pressure, as some studies have reported reduction of blood pressure, while others have shown no change or even an increase of blood pressure after growth hormone therapy (35) (Table 2.3.7.3) (71). On the other hand, administration of growth hormone can, at least partly, reverse the pathological fibrinolysis and restore the augmented sympathetic nerve activity found in untreated GHD adults (35). In addition, growth hormone replacement therapy has also been shown to reduce circulating CRP and IL-6 levels (35), as well as normalizing the microvascular abnormalities described in patients with GHD (49). Growth hormone replacement has been shown to reverse early atherosclerotic changes in the carotid arteries in GHD adults such as the increase of carotid intima media thickness(35). Moreover, it has been reported that this beneficial effect is sustained even after 10 years of treatment (70) (Fig. 2.3.7.4).

Although growth hormone replacement therapy improves most of the adversely affected cardiovascular risk factors observed in GHD patients, there are still limited data regarding the effect of growth hormone therapy on cardiovascular morbidity and mortality. In fact, Svensson et al. have reported an increased rate of cerebrovascular events in GHD patients receiving growth hormone therapy (53), although the rate of myocardial infarctions as well as the overall mortality were lower and similar respectively to those of the normal background population. This increased risk ratio for cerebrovascular events could be related to radiation angiopathy, which, again, demonstrates that factors other than growth hormone may have an important effect on outcome in patients with hypopituitarism (53).

In conclusion, current data suggest that the standardized mortality rate is not increased in adults receiving growth hormone replacement, although the duration of therapy is relatively short. However it remains unclear whether growth hormone replacement will normalize the mortality rate observed in hypopituitary patients over the longer term.

Growth hormone replacement in GHD patients produces a biphasic increase in bone turnover, causing a maximal effect on bone resorption after 3 months, and on bone formation after 6 months, which leads to a net gain of bone mass after 6–12 months in children, and after 18–24 months in adults (as the initial bone loss must first be replaced) (35, 57). Furthermore, studies of up to 10 years of growth hormone replacement show a sustained increase of BMD. It has been demonstrated that BMD continues to increase 18 months after discontinuation of growth hormone therapy (57).

Patients with lower BMD prior to growth hormone therapy respond with a higher increase in BMD, and the increase in BMD is slower in women than in men (35). The impact of growth hormone replacement on bone is greater on cortical than on trabecular bone (35). Unfortunately, measurement of BMD in GHD may not be a reliable predictor of fracture risk, and there is a lack of prospective studies documenting a reduction in fracture rates (57). A few cross-sectional studies have shown that growth hormone therapy reduces the risk of morphometric vertebral and nonvertebral fractures in GHD patients, even in those with untreated hypogonadism (57). This beneficial effect of bone fracture reduction seems to only occur in patients receiving growth hormone treatment shortly after being diagnosed with GHD (57). Finally, the addition of conventional osteoporosis therapy to GHD patients with confirmed osteoporosis, who already receive growth hormone replacement therapy, is also beneficial (35).

The GRS recommends assessment of BMD before initiating growth hormone therapy, and subsequently every 2 years (34) although this may be excessive. In order to monitor the skeletal response to growth hormone therapy, some authors also recommend an independent baseline assessment of fractures, as BMD may not be a good predictor of fractures in GHD (57). Additionally, the measurement of serum calcium, phosphate, alkaline phosphatase activity, and osteocalcin levels after the initiation of growth hormone treatment could be useful to evaluate the achievement of a therapeutic response (57).

In addition to the effects of growth hormone replacement discussed above, growth hormone has effects on many other areas of the body. Skin thickness and sweat secretion increase (60). Growth hormone affects deiodinase activity, increasing circulating triiodothyronine (T₃) levels (60) (35). This together with increase protein synthesis and fat oxidation (60) may explain the rapid 12–18% increase in REE which cannot be accounted for by increased lean body mass alone (61). Some, but not all studies, have shown increases in isometric and isokinetic strength (25); these changes become apparent after approximately 1 year and persist after 5 years of growth hormone therapy (35). This increase in muscle strength seems to be caused by an increment in muscle volume, and not by changes in muscle morphology or metabolism (35). Moreover, the degree of augmentation of muscle strength seems to be greater in childhood-onset GHD patients (35). In addition, local muscle endurance returns to baseline values after long-term growth hormone replacement (over 5 years), despite an initial decrease observed during the first 2 years of growth hormone treatment (35).

Finally, some but not all short- and long-term growth hormone replacement studies have shown an improvement in exercise capacity and physical performance, with marked increments in maximum oxygen uptake as well as maximum work capacity (25). Growth hormone therapy seems to increase exercise capacity through several mechanisms, including an increment in muscle mass, increased cardiac capacity, decreased fat mass, augmented red cell volume by IGF-1 stimulated erythropoiesis, and possibly by improved sweating (61).

The goal for growth hormone replacement in adults is to correct the abnormalities associated with adult GHD, maximizing benefits and minimizing side effects (34). As growth hormone secretion is greater in younger individuals than older ones, and in women than men, it is recommended that the starting dose of growth hormone in young men and women be 0.2 and 0.3 mg/day, respectively, and in older patients 0.1 mg/day (34) (see Box 2.3.7.2). Dose determination based on body weight is not recommended due to large interindividual variation in absorption and sensitivity to growth hormone, as well as the lack of evidence that a larger replacement dose is required for heavier individuals in adulthood (34). Dose escalation should be gradual, individualized, and guided by clinical and biochemical response (34). Measurement of serum IGF-1 provides the most useful marker for growth hormone dose titration in adults and it should be measured at least yearly. During the initial stages of dose titration, frequent measures are required; following a change in dose, IGF-1 assessment should be undertaken after 6 weeks (27, 34). The aim of dose titration is to achieve a serum IGF-1 level within the upper half of the age related normal range (34).

Despite IGF-1 being considered the most sensitive serum marker of growth hormone action, it may not reflect appropriately the growth hormone status of the patient, because as already described, normal serum IGF-1 concentrations are found in a considerable proportion of severe GHD patients. Moreover, the relationship between serum IGF-1 response during growth hormone therapy and other treatment effects such as metabolic endpoints and body composition is poor (35). Thus, it can be observed that the same dose of growth hormone can be suboptimal in one patient but cause side effects of over-dosage in another, and that normalization of serum IGF-1 can induce side effects attributable to growth hormone excess in some individuals (35). Hence, it is recommended to monitor other aspects of growth hormone therapy to assess both the efficacy and side effects of growth hormone treatment, and therefore to perform an individual dose titration (growth hormone dose titration against both clinical features of GHD and evidence of over-treatment determined by serum IGF-1 and the appearance of side effects) (35).

Since changes in body composition have been consistently found in growth hormone replacement trials in adults with GHD, the assessment of body composition, in particular extracellular water, could be used to monitor growth hormone replacement (35, 80). The GRS recommends performing a careful clinical examination and recording of anthropometric measures (weight, height, and BMI) before the start of the growth hormone replacement therapy, and the assessment of body composition to monitor growth hormone treatment response (34). The GRS considers that quantification of body composition changes (including bone mineral density) should preferably be made by dual X-ray absorptiometry (DEXA) where available, at baseline, and every 2 years after (34). However, body composition can also be assessed using anthropometric measures (including measurement of waist and hip circumferences), as well as with bioelectrical impedance evaluation (35). In addition, the GRS considers that cardiovascular risk markers should be measured yearly (34). Thus, some authors recommend measuring plasma lipids before growth hormone therapy starts, and subsequently on a regular basis, particularly in patients with baseline abnormalities, or those with other cardiovascular risk factors (35). Insulin sensitivity monitoring should be undertaken, which can be done by calculation of the homoeostasis model assessment (HOMA), by measurement of fasting levels of glucose, insulin, and HbA_1c (35). In contrast to the NICE guidelines, GRS usually reserves the use of disease-specific quality of life questionnaires for research purposes, although it recommends undertaking a detailed history with attention to quality of life parameters to monitor the efficacy of growth hormone replacement therapy (34). However, other authors recommend the use of a specific questionnaire before starting growth hormone treatment, which could be repeated every year in the follow up of GHD patients to evaluate the sustained response of quality of life to growth hormone therapy (63). In reality, the monitoring of patients receiving growth hormone replacement undertaken in clinical practice is determined by the rationale for treatment and safety.

Both GRS and NICE guidelines agree that adult patients receiving growth hormone replacement should be followed by an endocrinologist with special experience in pituitary disease/special interest in the management of growth hormone disorders, although it can be managed in partnership or ‘shared-care’ agreement with an internist or general practitioner. Growth hormone replacement is considered to most likely be for life; GRS recommends that a trial of withdrawal should be considered if any patient perceives no benefit (34), while NICE advises that growth hormone treatment should be discontinued in those patients who demonstrate inadequate improvement in quality of life score (<7 points on the QoL-AGHDA scale) after the first 9 months of therapy (69). There are concerns over the sole use of quality of life as a determinant of who receives growth hormone therapy and for the evaluation of its efficacy in GHD patients, as this practice fails to consider the other benefits of growth hormone replacement such as the improvement in markers of cardiovascular risk or bone health. Strategies for growth hormone replacement based solely on quality of life may deny patients these benefits which may have a consequence for their long term health.

The growth hormone dose may need to be reduced during long-term treatment, mimicking the decline in growth hormone secretion associated with ageing, reflected by the fall in the upper limit of the age specific normal range for IGF-1 (35). Growth hormone requirements may also change because of initiation or discontinuation of oral oestrogen (35). Other endocrine replacement may need to be adjusted. Initiation of growth hormone replacement may modify the dose of thyroxine or unmask the presence of central hypothyroidism (by increasing conversion of T₄ to T₃). More importantly the action of growth hormone and IGF-1 on 11-βHSD type 1 can lead to cortisol deficiency, particularly in patients on fixed-dose glucocorticoid replacement, which should be increased if indicated clinically (35).

Growth hormone replacement therapy in adults appears to be safe, when standards of care are followed (34). Recombinant human growth hormone is identical to the endogenous hormone and therefore it does not produce hypersensitivity reactions (61) although some patients may be sensitive to components of the diluent used. Absolute contraindications for growth hormone treatment include active malignancy, benign intracranial hypertension, and proliferative or preproliferative diabetic retinopathy (27). Although early pregnancy is not a contraindication, growth hormone should be discontinued in the second trimester as growth hormone is produced by the placenta (27).

Most adverse effects are dose related and are rarely seen in clinical practice if the dose of growth hormone is titrated carefully (25). Fluid retention, caused by the antinatriuretic effect of growth hormone, is the most frequent side effect, occurring in 5–18% of patients and includes paraesthesias, joint stiffness, peripheral oedema, arthralgia, and myalgia (25). In addition, 2% of treated GHD adults develop carpal tunnel syndrome (25). These fluid retention complications are more frequently seen in older, heavier, and female GHD adult patients, and most of them will resolve with dose reduction (25).

Growth hormone replacement therapy is not associated with an increased incidence of either type 1 or type 2 diabetes mellitus in adults (34), although data suggests it may be in children (35). However, growth hormone increases insulin resistance, so GHD patients with high risk of developing type 2 diabetes (positive family history, obese or older) require careful monitoring (34). These patients should be given a very low dose of growth hormone at initiation of therapy, which should be followed by a gradual increase in dose based on the clinical response (35). If type 2 diabetes is diagnosed, it should be managed similarly to any other patient with this disease, and growth hormone replacement therapy can be continued (34). Patients who have pre-existing diabetes mellitus require careful monitoring as their requirements for hypoglycaemic agents may increase during initiation of growth hormone therapy (25).

Other reported side effects of growth hormone replacement include atrial fibrillation, gynaecomastia, congestive heart failure and benign intracranial hypertension, all of which are more likely to occur in the elderly (61) (with the exception of intracranial hypertension, which occurs primarily in children and adolescents) (25). Again, all these mentioned side effects are dose related (61). Retinopathy is an extremely unusual complication, but can also improve after growth hormone therapy withdrawal (25).

There is no evidence that hypothalamic or pituitary tumour recurrence is influenced by growth hormone replacement therapy (34). Thus, although data of tumour recurrence and regrowth during growth hormone replacement is still limited (small studies and limited follow-up periods), the findings are reassuring (35). Consensus guidelines for the diagnosis and treatment of adults with GHD (34) recommend undertaking pituitary imaging before starting growth hormone therapy with appropriate follow-up imaging determined by the nature of the underlying condition. The use of growth hormone replacement does not require additional monitoring of residual disease (34).

There are conflicting results regarding the incidence of malignancies in hypopituitary patients not receiving growth hormone, as decreased and increased rates of malignancies have been reported in these patients (53). Some second tumours, such as meningioma, may be attributable to treatment with radiotherapy rather than hypopituitarism per se (53).

There is also an important concern regarding the possibility of increased risk of de novo cancer with growth hormone treatment, due to the mitogenic and growth-promoting actions of growth hormone and IGF-1. However, to date there is no evidence that growth hormone replacement in adults increases the risk of de novo malignancy, although growth hormone treatment during childhood slightly increases the relative risk of secondary neoplasia among cancer survivors (34). However, as reported by authors of the study that suggested an increased incidence of second neoplasms in survivors of acute leukaemia (81), the data need to be interpreted with caution given the small number of events (3 osteogenic sarcomas in 122 leukemia/lymphoma survivors treated with growth hormone vs 2 cases in 4545 leukemia/lymphoma survivors not treated with growth hormone). In the same study, growth hormone replacement did not appear to increase the risk of disease recurrence or death in survivors of childhood cancer. In a long-term follow-up study of 1848 patients treated in childhood and early adulthood with growth hormone (82), two patients died from colorectal cancer and two from Hodgkin’s disease.

Extensive long-term, postmarketing surveillance of thousands of children and adults treated with growth hormone has not shown any increase in cancer rates (83). If growth hormone replacement treatment does result in a small increase in cancer risk compared with untreated patients with GHD, it is unlikely that, with careful dosing and monitoring, it will exceed that observed in the general population (83). Patients treated with growth hormone do not require screening for malignant disease beyond that recommended for the normal population.

Over the past 20 years understanding of the impact of GHD in adults has increased and it is now a recognised clinical entity that affects a wide range of pathophysiological parameters. Although growth hormone replacement has become routine in modern endocrine practice there are still many questions that need to be answered. Long-term observational studies will determine whether the observed increased mortality in hypopituitarism will be reduced to levels seen in the normal population. Other work is required to determine whether the ‘treatment for all’ approach should be adopted universally or whether a more focused, symptom related approach is more appropriate. As new causes of pituitary dysfunction are identified, e.g. TBI, the role of growth hormone in rehabilitation needs to be explored. Finally, the features of GHD and the response to treatment described above are derived from patients with severe GHD. Future studies are required to determine whether patients with partial GHD would benefit from growth hormone replacement.