Radiation-induced hypopituitarism
- Department of Endocrinology, Oxford Centre for Diabetes, Endocrinology and Metabolism, Churchill Hospital, Old Road, Headington, Oxford OX3 7LE, UK
1The Institute of Cancer Research and The Royal Marsden National Health Service Foundation Trust, Sutton, Surrey, UK
2Department of Anatomy, Histology and Anthropology, Faculty of Medicine, Vilnius University, Vilnius, Lithuania
- (Correspondence should be addressed to J A H Wass; Email: john.wass{at}noc.anglox.nhs.uk)
Abstract
The hypothalamic–pituitary unit is a particularly radiosensitive region in the central nervous system. As a consequence, hypopituitarism commonly develops after radiation treatments for sellar and parasellar neoplasms, extrasellar brain tumours, head and neck tumours, and following whole body irradiation for systemic malignancies. Increasing tumour-related survival rates provide an expanding population at risk of developing hypopituitarism. In this population, long-term monitoring tailored to the individual risk profile is required to avoid the sequelae of untreated pituitary hormonal deficiencies and resultant decrease in the quality of life. This review analyses the pathogenesis, prevalence and consequences of radiation-induced hypopituitarism (RIH) in diverse subgroups at risk. Also discussed is the impact of modern radiotherapy techniques in the prevalence of RIH, the spectrum of endocrine disorders and radiation-induced brain conditions that also occur in patients with RIH.
Introduction
Radiation therapy (RT) is widely used in the management of intracranial and systemic disorders. While three-dimensional treatment planning and accurate radiation delivery enhance efficacy and safety of radiotherapy, the hypothalamic–pituitary unit often receives significant doses of radiation with the tendency to develop long-term radiation-induced hypopituitarism (RIH) and its sequelae.
Hypopituitarism is a common consequence of radiation treatments and is associated with increased morbidity and mortality (Tomlinson et al. 2001). RIH is a chronic onset disorder that can result in a variable and progressive impairment of pituitary function. The clinical consequences of RIH are present across the lifespan and need adequate lifelong monitoring and optimal replacement to avoid the well-known sequelae of pituitary failure. Furthermore, patients who have received radiotherapy involving the pituitary and hypothalamic areas are at risk of occasionally developing radiation optic neuropathy (Kline et al. 1985), second-brain tumours (Minnitti et al. 2005), vascular problems and other rare but serious radiation-related adverse effects that need specific monitoring. Survivors of non-pituitary neoplasms who develop RIH (mainly patients with extrasellar brain tumours, postnasal space cancer and some systemic malignancies) are also present with direct damage to the pituitary target organs, besides other systemic sequelae of treatment (Cohen 2005). As cancer survival rates steadily increase (Verdecchia et al. 2007), the forthcoming decades are expected to bring an upsurging number of survivors prone to developing endocrine sequelae of antitumoral treatments. These patients do not always undergo routine surveillance for endocrine complications when required. However, adequate management and treatment of RIH represents an important tool for optimising outcomes and improving the quality of life (QoL) in such patients.
This review analyses the pathogenesis, prevalence and clinical consequences of RIH. The impact of modern radiation techniques on its incidence, and strategies for optimising early diagnosis and treatment are also discussed.
Common indications of radiotherapy resulting in RIH
RIH is a common consequence of RT for the conditions shown in Table 1. Tumours developing in the sellar and parasellar regions have a distinct clinical behaviour when compared with other brain tumours; nasopharyngeal carcinomas (NPC) and haematological malignancies are separately described.
Neoplasms associated with treatment-related endocrine sequelae
The most common sellar neoplasms requiring radiation treatment are pituitary adenomas, which include non-functioning tumours, somatotroph, corticotroph adenomas and prolactinomas. Radiotherapy prevents tumour growth either when surgery is contraindicated or in the presence of a significant tumour remnant, usually in the cavernous sinuses. Radiotherapy is also used to treat refractory hormonal hypersecretion when medical and surgical approaches have not been successful (Plowman 2002, Petrovich et al. 2003a,b). The most common radiotherapy regimen for pituitary adenomas is fractionated radiotherapy using a total dose of 45 Gy administered in 25 fractions of 1.8 Gy. Although single-fraction radiotherapy (SFR), also known as radiosurgery, with doses averaging 15–30 Gy has also been used, its value is debated (Brada & Jankowska 2008). The combination of tumour compression of the pituitary parenchyma, surgical damage and the vicinity of the hypothalamic–pituitary unit to the radiation target makes RIH a common event, present in 50–100% of patients. Tables 2 and 3 show the incidence of RIH in diverse series of pituitary tumours depending on the radiation technique and the hypersecretion features.
Incidence of radiation-induced hypopituitarism in pituitary tumours treated with different radiation techniques
Incidence of radiation-induced hypopituitarism in different pituitary tumours classified according to secretory features
RT for other sellar and parasellar tumours also results in RIH. Craniopharyngiomas are tumours with a remarkable incidence of hypopituitarism in different treatment stages (Karavitaki et al. 2005). The tumour invasiveness, the difficulty of total removal, and the contribution of radiotherapy (average total dose 50–60 Gy), make hypopituitarism virtually universal in these patients. RIH also presents the following RT for meningiomas, gliomas, chordomas, pineal tumours and chondrosarcomas lying in parasellar location (Smith et al. 2006).
RIH can also result from radiation used for non-parasellar brain tumours and systemic malignancies. Endocrine sequelae in cancer survivors may be due to the direct effect of the malignancy and radiation treatment on hypothalamic, pituitary, thyroid and gonadal function and due to the effects of chemotherapy (Gleeson & Shalet 2004). Childhood cancer survivors have a decreased QoL, and their 30-years cumulative incidence of chronic health conditions is 73.4% with a 24% incidence of endocrine sequelae (Oeffinger et al. 2006), although other series report incidence figures as high as 41% (Stevens et al. 1998). The severity of RIH is related to the radiation doses given while whole body irradiation regimens to a dose of 18 Gy result merely in isolated GH deficiency (GHD; Clayton & Shalet 1991), higher dose schedules, such as those used for NPC (40–60 Gy falling on the hypothalamic–pituitary unit) result in 62% incidence of pituitary deficiencies, which involve the GH, gonadotroph, corticotroph and thyrotroph axes (Lam et al. 1991).
Pathophysiology of radiation-induced toxicity in the hypothalamic–pituitary unit
Radiation exerts effects on neoplastic and healthy tissues through ionisation of the cell DNA. Damage can occur directly, through the generation of dysfunctional conformational changes in cell DNA, or indirectly through the production of free radicals from water molecules, resulting in damage to the cell DNA (Belka et al. 2001). Radiation damage can induce immediate lethal effects or accumulate sublethal damage that limits the potential for cell replication, thus halting tumour growth (Shrieve 2006). Cell death after radiation, which occurs in the mitotic phase of the cell cycle (late G2-early S), may be the reason for a delayed onset of chronic complications of radiation in slowly replicating tissues, such as those present in the hypothalamic–pituitary unit (Dewey & Bedford 1998, Kunkler 2003).
Ionising radiation also induces degenerative changes in glial cells, leading to a lack of trophic neural support and demyelination, which in turn causes subacute and chronic neural damage in the hypothalamus and other central nervous system (CNS) structures. Radiation also causes vascular derangements (lifting of endothelia, vacuolation of the cytoplasm and nuclear swelling) that ultimately lead to endothelial cell death (Tofilon & Fike 2000, Belka et al. 2001). Vascular changes with alteration in vessel permeability may explain oedema in the acute phase of high-dose radiation treatments (Smith et al. 2006), which is rarely seen following RT for benign sellar and parasellar tumours. In the chronic phase, the vascular lining, exposed to the progressive loss of endothelia, responds by increasing endothelial proliferation, thickening the basement membrane and increasing the collagen synthesis. These changes lead to obliteration of the small vessels with necrosis of irradiated tissues and enhance atherogenesis in major brain arteries. The exposed vascular matrix in great vessels accelerates the formation of platelet aggregates and thrombi. Thrombotic vessel occlusion culminates the derangement in the vascular wall initially induced by radiation atherogenesis. The reported risk of cerebrovascular accidents in irradiated cohorts (Brada et al. 2002) has been attributed by some authors to the increased atherogenicity of the irradiated vascular lining.
Owing to the efficacy of radiotherapy in halting the growth and hormonal hypersecretion of sellar neoplasms, further surgery is rarely required and data on the histological changes occurring in the dysfunctional pituitary gland are scarce. The most commonly reported findings are fibrosis of the pituitary parenchyma, hyperplasia of folliculostelate cells, squamous metaplasia, oncocytic changes and increased cytokeratin immunoreactivity (Nishioka et al. 2002).
In radiation treatments for sellar and parasellar tumours, hypothalamic damage is considered to be present in the early pathophysiology of RIH, while pituitary atrophy has been classified as a late-onset contributing factor (Romer et al. 1991, Darzy et al. 2007). Studies on the effects of external radiotherapy on the lactotroph axis as well as the differential effects of pituitary brachytherapy compared with external radiation have provided evidence in favour of this pathophysiological hypothesis. Lactotroph cells, unlike any other pituitary-cell lines, receive predominantly inhibitory signals from the hypothalamus; this inhibitory control explains the increase in prolactin (PRL) levels that results from either radiation damage to the hypothalamus or hypothalamic–pituitary disconnection. By contrast, hyperprolactinaemia does not occur when direct damage to lactotroph cells occurs. Thus, measuring PRL levels in irradiated cohorts can provide clues as regards the pathophysiology of RIH. In two long-term follow-up studies, 96% of previously normoprolactinaemic patients with acromegaly (Ciccarelli et al. 1989) and up to 50% of patients after receiving external radiotherapy for extrasellar brain tumours developed hyperprolactinaemia (Constine et al. 1993). Radiation-induced hyperprolactinaemia would not be commonly expected if pituitary atrophy was the primary phenomenon in RIH. Furthermore, follow-up studies of patients with pituitary adenomas treated with brachytherapy (instillation of Yttrium-90 in the sella turcica, total dose 50–150 Gy) revealed a reduced risk of hypopituitarism when compared with a total dose of 37.5–45 Gy delivered through external radiation treatments (Sandler et al. 1987, Littley et al. 1989). Meanwhile, observational studies have failed to show an increase in PRL levels after sellar brachytherapy (Clark et al. 1983). The lower incidence of RIH after sellar brachytherapy despite higher total doses when compared with external radiotherapy, together with the preservation of normal immunoactive serum PRL levels in the former group suggests a lower threshold for radiation-induced damage in hypothalamic tissue (preferentially injured by external radiotherapy). Subsequently, an earlier role of hypothalamic secretory derangements in the pathophysiology of RIH has been inferred. Pituitary atrophy, either through direct radiation damage or lack of hypothalamic stimulatory input, would contribute to RIH at later stages.
Controversy surrounds the nature of the hypothalamic derangements leading to pituitary dysfunction. While some authors suggest a reduced release of hypothalamic releasing factors as the primary phenomenon, others argue in favour of an increased basal stimulatory tone from the hypothalamus, which results in reduced hormonal release in the pituitary gland (Darzy et al. 2007). The initial derangements in the pituitary function of irradiated cohorts are mild and subclinical, but can progress over time and become clinically apparent, requiring early recognition and therapy by regular yearly assessment.
Understanding the pathophysiology of RIH is important for the choice of diagnostic strategies used to assess the pituitary function. This requires the use of appropriate stimulation tests acting at the hypothalamic level. The GH-releasing hormone (GHRH) plus Arginine stimulation test has proven to be significantly less sensitive, as compared with the insulin tolerance test (ITT) in patients with radiation-induced GHD (Darzy et al. 2003). Futhermore, a flat TSH response to thyrotroph releasing hormone (TRH) stimulation has been traditionally postulated as a sensitive marker of hidden impairment of TRH secretory reserve (Trejbal et al. 1994, Rose et al. 1999b), although the value of these tests has recently been questioned (Darzy & Shalet 2005a,b). The absence of pituitary atrophy in the initial stages of RIH and derangements in the secretory rate of hypothalamic releasing factors could explain lower sensitivities of functional tests directly dependent on the integrity of the pituitary tissue. Table 4 displays the most commonly used pituitary function tests in the assessment of irradiated cohorts.
Diagnosis of hypopituitarism
The mechanisms underlying the characteristically evolving pattern of RIH, in which GHD and gonadotrophin deficiency appear more commonly and usually earlier than corticotroph and thyrotroph insufficiency (Littley et al. 1989) is not known. The thyrotroph and corticotroph axes are known to be less sensitive to radiation damage and develop later; diabetes insipidus does not occur as a consequence of radiation treatments. Although 61% of patients present the usual temporal sequence of pituitary failure (GHD, gonadotrophins, ACTH and, last, TSH deficiency), RIH can present with any temporal sequence of hormonal deficiencies (Littley et al. 1989). Therefore, routine testing for all the pituitary axes is warranted. RIH, as other effects of radiation, is a concept of probability, and this probability increases with time. Although RIH tends to be present in the first 5 years after exposure, new deficiencies can appear even 20 years later, necessitating lifelong surveillance (Shalet et al. 1976, Littley et al. 1989).
RIH is more common in patients with previous insults to the pituitary gland (Duffner et al. 1985). Compressive damage by an adjoining tumour causes ischaemic derangements in the normal pituitary parenchyma and a degree of pituitary dysfunction, which is dependent on tumour size and location, duration of pituitary compression and tumoral invasiveness. Tumours located in the hypothalamic region tend to produce a higher degree of hypopituitarism, presumably by depriving the pituitary gland of trophic factors and blood supply from the hypophyseal portal system. Surgical insults to the pituitary gland are also risk factors for the ultimate development of RIH, with 7.2–20.6% of patients presenting new pituitary deficiencies after a transsphenoidal approach in experienced centres (Ciric et al. 1997, Ahmed et al. 1999). The excision of pituitary parenchyma and the compromise in blood supply following surgery in the sellar and parasellar regions could account for surgically-induced hypopituitarism. Remarkably, this phenomenon, as well as the surgery success rates, is highly dependent on surgical expertise (Bates et al. 2008). Postsurgical pituitary dysfunction occurs more commonly in patients undergoing craniotomy, as compared with transsphenoidal approach (Nomikos et al. 2004) and this is related to size and invasiveness of the tumour, which may also influence the development of postsurgical hypopituitarism.
RIH can also ensue after RT for non-pituitary brain tumours, such as gliomas, meningiomas, pinealomas and medulloblastomas. A cross-sectional pituitary function evaluation of adult brain tumour survivors, performed at an average age of 6 years after radiation exposure, found that 41% of the patients had evidence of some degree of hypopituitarism. This included 16% with a single hormonal deficiency, 25% with multiple deficiencies and 7% with panhypopituitarism. The hormonal deficiencies presented in the same frequency order as previously shown with RIH from other causes (32% had GHD, 27% gonadotrophin failure, 21% ACTH deficiency and 9% central hypothyroidism), but were less common than following RT for parasellar tumours (Agha et al. 2005). The hormonal deficiencies correlated with the length of follow-up after radiation exposure and with the biologically effective dose (BED). The high-incidence of pituitary dysfunction in adult brain tumour survivors treated with radiotherapy is not dissimilar to RIH found in childhood brain tumour survivors. These findings further highlight the need for regular surveillance of pituitary function in patients with non-pituitary brain tumours treated with radiation fields including the hypothalamic–pituitary unit.
The hypothalamic–pituitary unit commonly falls within radiation fields in the treatment of NPC, due to the vicinity of the postnasal space to the sellar region. Radiotherapy for NPC brought about a 62% incidence of RIH after 5 years of follow-up following mean doses of 40 Gy to the hypothalamus and 62 Gy to the pituitary gland (Lam et al. 1991) with 63.5% GH deficiency, 30.7% gonadotroph deficiency, 26.7% ACTH and 14.9% TSH deficiency. The shielding of the pituitary gland may reduce the risk of developing hypopituitarism apparently without decreasing antitumoural efficacy (Sham et al. 1994), although further studies are needed to confirm this.
Radiotherapy techniques
The purpose of radiation treatments to the sellar region is to prevent tumour growth and normalise hormonal hypersecretion while sparing the surrounding normal structures. The last decades have seen advances in the techniques of irradiation. These include improved imaging and three-dimensional visualisation of tumour and normal tissue, better immobilisation and more localised means of radiation delivery. The term ‘stereotactic’, borrowed from neurosurgery, indicates the use of more precise patient immobilisation and the use of a coordinate system to better define the three-dimensional spatial orientation of radiation beams. We discuss the principal technical developments and their clinical impact on the management of pituitary adenoma.
Radiotherapy equipment
External beam radiotherapy, which is the principal means of delivering radiation to targets in the brain, is given through multiple beams from a high-energy radiation source focused on the tumour. Local delivery of radiation with brachytherapy (interstitial radiotherapy) is rare, if ever employed in this location.
Radiation treatments are most commonly delivered through photons (high energy X-rays), generated by a linear accelerator (LINAC). Cobalt 60 as the source of high-energy gamma radiation has largely been replaced with the exception of a multiheaded Cobalt Unit (gamma knife). Charged particle beams in the form of protons, and more recently Helium and Carbon ions have been tested as therapeutic radiation sources. The principal theoretical advantage of protons is marginally more localised radiation delivery. Helium and carbon ions also have higher linear energy transfer properties (Bolsi et al. 2003; Ronson et al. 2006) and theoretically produce a more potent biological effect.
In modern pituitary RT, the tumour is localised on unenhanced magnetic resonance imaging coregistered with planning computed tomography (CT). The treatment delivery is planned in three-dimensions (Littmann et al. 2006) with improved visualisation of dose distribution within the target and the organs at risk of radiation toxicity. Localised irradiation is achieved by giving treatment in 3–4 beams each shaped to conform to the shape of the tumour using a multileaf collimator (MLC). This is described as conformal radiotherapy and is part of the standard practice in departments specialised in the treatment of benign brain tumours.
MLC leaves, apart from altering the shape of the radiation beam, can also be used to modulate the intensity of radiation and this is described as intensity modulated radiotherapy (IMRT). Studies of dose distribution in tumour and normal tissue show no clear benefit for IMRT compared with fixed-field stereotactic conformal radiotherapy (SCRT) in the treatment of pituitary adenomas.
Stereotactic techniques
Stereotactic techniques are a refinement of conformal radiotherapy with further improvement in immobilisation using relocatable or fixed frames, improved imaging and more precise treatment delivery. Stereotactic irradiation can be given as SFR using either multiheaded cobalt unit (gamma knife) or a LINAC or as SCRT delivered as fractionated treatment using a LINAC. The theoretical benefit of stereotactic compared with conventional RT is further reduced in the amount of normal tissue treated to high doses with the hope of reducing long-term treatment-related morbidity.
Neither conventional nor stereotactic RT's are currently able to avoid treating the hypothalamus and the pituitary itself receives equally high radiation doses. Based on spatial dose distribution alone, it is unlikely that modern RT will be associated with a significantly lower risk of RIH unless techniques are developed for selective hypothalamic avoidance.
Dose fractionation
Conventional external beam radiotherapy is given in small daily fractions (fractionated RT) over an extended period of time. Although fractionated RT developed largely empirically the rationale is based on differential recovery of radiation-induced damage due to altered repair mechanism in neoplastic as compared with healthy tissues that selectively favour recovery of normal tissue over tumour.
In most sellar neoplasms (Bauman et al. 2006), replication is a relatively slow process, which enables fractionation strategies to provide antitumour and antisecretory efficacy with preferential recovery of normal tissue without decreasing tumour control. Radiotherapy for pituitary adenomas is generally delivered in 25–30 fractions over five to six weeks, with a total dose of 45–50 Gy. Individual fraction doses above 1.8 Gy and total doses above 45 Gy are independent risk factors for RIH (Littley et al. 1989).
SFR can be delivered through either a LINAC or a 60-Cobalt unit (gamma knife). SFR, despite smaller total doses than fractionated RT, is associated with a higher risk of radiation damage to normal CNS such as the optic apparatus (Pouratian et al. 2006, Jagannathan et al. 2007) and can only be employed for lesions away from the optic chiasm and nerves. It is less effective in achieving tumour control than fractionated treatment with no faster decline in hormone levels (Brada et al. 2004, Jagannathan et al. 2007, Brada & Jankowska 2008). Although some studies report the reduced prevalence of RIH (Vladyka et al. 2003) there is currently no evidence of reduced risk of RIH following SFR compared with conventional or stereotactic fractionated RT (Pollock & Young 2002, Pollock et al. 2002, Degerblad et al. 2003, Petrovich et al. 2003a,b).
RIH as a dose-related complication
Hypopituitarism, similar to other radiation-related complications, is a dose-dependent event. RIH occurs more commonly following total doses >45 Gy or fraction doses >1.8 Gy (Littley et al. 1989). From these thresholds, the risk of pituitary deficiencies is proportional to the increase in the total and fraction dose. RIH is virtually universal in the high-end of the dose spectrum, with total doses as high as 60 Gy occasionally used for craniopharyngiomas and other aggressive parasellar neoplasms. Early SFR studies suggest a dose-threshold for GH and gonadotroph deficiency of 15 and 18 Gy for ACTH and TSH deficiencies (Vladyka et al. 2003).
For a given total dose, the effects of radiation on tissues can vary markedly depending on the fractionation schedule and tissue radiosensitivity. These variables are included in the concept of BED, which predicts antitumoural efficacy and side effects of radiation treatments more accurately than total doses. The BED has been shown to be a close predictor of GHD (Schmiegelow et al. 2000), but data on other pituitary axes are lacking.
Clinical consequences of RIH
In the presence of hypopituitarism, mortality is increased by almost twofold (Tomlinson et al. 2001). The clinical consequences of a given radiation schedule vary according to patient's age and sex. In childhood, RIH can impair somatic, sexual and psychological development by undermining GH and thyrotroph secretory reserve and disrupting pubertal timing processes. Furthermore, somatotrophs are reported to be more radiosensitive in children as compared with adults: 59% of childhood cancer survivors receiving 18 Gy of total body irradiation for haematological malignancies show a blunted response on GH stimulatory tests (Costin 1988) which is more than twice the frequency of adult GHD after the same dose (Brennan et al. 1998). This discrepancy has been explained by the need of greater GH secretory reserve in children (Shalet et al. 1976), rather than through variations in radiosensitivity of somatotroph cells through the lifespan (Clayton & Shalet 1991). A direct effect of age on the incidence of other pituitary hormonal deficits has not been demonstrated. Age-related changes in radiosensitivity have nevertheless been reported, when the gonads and the thyroid gland fall within the radiation fields (Sklar et al. 2000, Wallace et al. 2005).
GH deficiency
GH has consistently shown to be the most radiosensitive pituitary axis, with series reporting a prevalence of GHD between 50 and 100% after radiotherapy for sellar masses. Radiation-induced GHD is also progressive with time, developing more frequently in the first 10 years after radiation delivery (Shalet et al. 1976, Toogood et al. 1995). GHD severity and speed of onset closely correlate with lower final height, decreased lean body mass (LBM) and higher fat mass in children with RIH. GHD in the adult conveys an increased risk of fracture, a deranged cardiovascular risk profile and, mainly, a decreased QoL. These abnormalities reverse partially or completely with adequate replacement (Molitch et al. 2006), and early diagnosis is therefore very important for optimising outcomes.
The diagnosis of GHD in irradiated patients is based on a low insulin-like growth factor-I (IGF-I) level in patients with multiple pituitary hormonal deficiencies or structural abnormalities in the pituitary region, which has 96% accuracy in predicting GHD (Hartman et al. 2002). However, 31–60% of patients with GHD have normal IGF-I values, thus requiring a stimulation test for the diagnosis of GHD (Biller et al. 2002, Hartman et al. 2002). The ITT is considered the gold standard for the diagnosis of GHD in irradiated cohorts (Molitch et al. 2006). Tests using glucagon or GHRH–arginine are less sensitive for the detection of hypothalamic damage, but can be used when epilepsy or cardiac disease contraindicates the ITT (Aimaretti et al. 1998, Gomez et al. 2002).
Notwithstanding the test of choice, obesity is increasingly being considered an important confounding factor in the assessment of GHD. Both patients with systemic malignancies and brain tumours are at increased risk of excess body weight through various mechanisms (Reilly et al. 2000, Harz et al. 2003, Oeffinger et al. 2006, Brouwer et al. 2007, Garmey et al. 2008). Radiation treatments, supraphysiological glucocorticoid therapy (Reilly et al. 2001), young age at exposure, decreased sleep duration, diminished physical activity and the effects of pituitary hormonal deficiencies represent some of the risk factors for obesity in this population. Particularly in brain tumour survivors, hypothalamic involvement (either induced by the tumour itself or surgical or radiation treatments) conveys a significantly increased risk of excess body weight (Müller et al. 2001, Lustig et al. 2003). Notwithstanding the underlying cause of increased adiposity, it is well known that GH secretion is attenuated by obesity (Williams et al. 1984, Iranmanesh et al. 1991). Recently, studies by Corneli & Makimura have established GH cut-offs in the GHRH–arginine stimulatory test adjusted according to body mass index (Corneli et al. 2005) and waist circumference (Makimura et al. 2008). By contrast, currently available evidence suggests that IGF-I reference values are not affected by increased adiposity (Nam et al. 1997, Maccario et al. 1999a,b). As regards the diagnostic value of ITT in obese patients, to our knowledge only a comparative study between 8 hypopituitary patients, 10 obese subjects and 10 normal individuals has been performed (Cordido et al. 2003); this survey found no overlap in somatotroph responsiveness between normal obese and hypopituitary individuals. In previous studies on the diagnostic value of ITT, GH dynamics were analyzed comparing subjects with a putative insult to the pituitary gland with healthy controls that were adjusted for body mass index; these studies defined stringent cut-offs for the diagnosis of GHD in order to minimise false positive results (Hoffman et al. 1994, Biller et al. 2002). As a result, although further research aimed to define if adiposity-related cut-offs are needed for the ITT as a diagnostic test for GHD, the stringency of current GH thresholds should minimise any possible diagnostic interference.
GHD in childhood
Children who receive doses >18 Gy for the treatment of haematological malignancies, sellar neoplasms and other brain tumours have an increased risk of developing the classical child sequelae of GHD (Shalet et al. 1976). Patients with untreated childhood-onset GHD (CO-GHD) attain average final heights 1 s.d. below ethnically, age and sex matched controls (Adan et al. 1996, Melin et al. 1998). In GH-deficient children, height outcomes are directly proportional to the severity of GHD, age of onset and whether adequately-dosed and timely GH replacements were undertaken (Darzy & Shalet 2006). In the Childhood Cancer Survivor Study (Brownstein et al. 2004), which followed 183 survivors of childhood tumours treated with cranial or craniospinal irradiation, the s.d. score of children who received GH therapy improved from −2.08±0.08 to −1.48±0.10 (P<0.0001). Children with lower final height received spinal radiotherapy and lower doses of GH replacement. GH treatment is generally well tolerated and the use of human recombinant GH is not considered to be associated with significant adverse effects. GH replacement does not increase the risk of tumour recurrence in pituitary adenomas, craniopharyngiomas (Karavitaki et al. 2006) and childhood cancer survivors (Sklar et al. 2002). These data altogether underline the need for timely and careful consideration of GHD assessment in cancer survivors in order to optimise height outcomes and health-related QoL.
Substantial research has also emphasised the role of factors not directly related to the somatotroph axis in determining height outcomes; among these, direct irradiation of the spine (through arrest of vertebral growth), nutritional imbalances, deficiencies in other pituitary hormones, early puberty, chemotherapy and the transient effect of the tumour itself should be considered (Cohen 2005). Although GH therapy partially prevents height loss in irradiated children, both timely and weight-adjusted replacements and a comprehensive assessment of additional factors involved play a relevant role in decreasing disparities between height potential and outcome in cancer survivors.
GHD in the transition period
In subjects with childhood-onset RIH, GH replacement has been customarily used for the attainment of height potential, and subsequently stopped. By contrast, substantial research performed over the last decade has focused its attention on the additional benefits of continuous GH therapy after cessation of growth. In the transition age, defined as the period of time between final height acquisition and full adult maturation (usually 6–7 years after final height has been achieved), GH promotes the attainment of peak bone mass and stimulates LBM gain and fat mass reduction (Bonjour et al. 1991, Theintz et al. 1992, Matkovic et al. 1994, Mukherjee et al. 2003, Clayton et al. 2005, Veldhuis et al. 2005, Giustina et al. 2008). Conversely, GHD in the transition period results in derangements of these processes (Mukherjee et al. 2003, Baroncelli et al. 2004, Radovick & DiVall 2007). Therefore, reassessment of GH requirement is required at final height completion (Clayton et al. 2005).
The choice of biochemical reassessment method for GHD in the transition period is highly dependent on the pretest likelihood of persistently deranged secretory reserve. Patients with multiple pituitary hormonal deficiencies, with structural disease or radiotherapy-induced GHD are very likely to need GH continuation, and thus an IGF-I measurement after GH withdrawal suffices to confirm the need for continued GH replacement (Clayton et al. 2005).
The changes in body composition during the transition period are an important hallmark of subsequent cardiovascular risk. Patients with CO-GHD have 80% of the LBM and increased fat mass content, as assessed by height-normalised dual energy X-ray absorptiometry (DEXA), when compared with patients with adult-onset GHD (AO-GHD; Attanasio et al. 2002). Randomised trials confirmed that GH replacement in the transition period restores body composition to normal: LBM and LBM/height ratio increased and body fat mass decreased, and these changes were dependent on GH dose in three out of four clinical trials analyzed. LBM response was more marked in males, probably reflecting the recovery of sexual dimorphism in LBM gain on GH replacement (Vahl et al. 2000, Underwood et al. 2003, Attanasio et al. 2004, Mauras et al. 2005).
Adults with GHD are ∼1 s.d. below the age and sex matched reference ranges for bone mineral content, as assessed by DEXA absorptiometry (Holmes et al. 1994). However, when interpreting bone density tests in untreated GHD patients, the effects of age of GHD onset (Kaufman et al. 1992), other pituitary hormonal imbalances (Mancini et al. 2004, Vieira da Costa et al. 2004) and their treatment strategies (Prince et al. 1991, Kayath et al. 1993, Behre et al. 1997, Colao et al. 2000, Rossouw et al. 2002, Jódar et al. 2003, Naliato et al. 2005) should be considered. GHD predominantly affects cortical sites, such as the radius, as compared with the spine. Bone size is also significantly smaller, but trabecular density is relatively preserved (Murray et al. 2004, 2006). Discontinuation of GH replacement in transitional patients with severe GHD limits peak bone mass accrual and places transitional patients at long-term risk for fractures (Drake et al. 2003). A GH replacement trial in the transition period showed an increase in bone mineral density when GH replacement was continued after final height completion. This effect was dose-dependent, and particularly significant in severely GHD patients (Shalet et al. 2003).
Other more controversial benefits of GH replacement in this period involve an improvement in lipid profile and QoL. Severely GHD patients have increased total cholesterol, low-density lipoprotein (LDL) cholesterol and decreased high-density lipoprotein (HDL). The abnormalities in lipid status classically described for adult GHD apply for the transition period. One clinical trial showed a significant decrease in LDL and LDL:HDL ratio following GH replacement (Underwood et al. 2003), but these findings have not been confirmed by other authors (Attanasio et al. 2004). The multifactorial nature of cardiovascular derangements in the irradiated cohorts, including increased obesity, sedentarism and possible changes in diet, should be taken into account when analyzing changes in lipid markers in this population. The effect of GH on carbohydrate metabolism is dual, improving insulin sensitivity by inducing favourable changes in body composition, but bearing a direct antagonistic effect on insulin action. The outcome is dependent on baseline insulin resistance parameters (influenced by puberty) and GH dose.
The effects of GHD on QoL in the transition period are not as clear as in adults. Significant impairments of QoL in patients with CO-GHD in the transitional period have not been confirmed by all authors, and the effect of GH therapy on this variable remains controversial in transitional patients (Vahl et al. 2000, Underwood et al. 2003, Attanasio et al. 2004, 2005, Mauras et al. 2005).
GHD in the adult
Adults who receive cranial or craniospinal radiotherapy have a high prevalence of GHD, present in 50–100% adult patients irradiated for sellar neoplasms and up to 32% of patients with extrasellar brain tumours receiving radiotherapy. In AO-GHD, decreased QoL, adverse changes in body composition, increased risk of fracture and worsening cardiovascular risk profile are the main sequelae (Molitch et al. 2006).
GHD patients have increased fat mass that preferentially accumulates in the visceral compartment (Hoffman et al. 1995, Snel et al. 1995), and this pattern of fat deposition has adverse consequences on long-term cardiovascular risk. GHD is also associated with decreased LBM (Binnerts et al. 1992, Hoffman et al. 1995) that is associated with impaired exercise capacity. These parameters are at least partially reversible on GH therapy (Jorgensen et al. 1989).
Severe GHD has been related to impaired vasodilatation responses to stress and exercise (Boger et al. 1996), which could explain why patients with severe GHD tend to be more hypertensive. Increased inflammatory markers (Sesmilo et al. 2001) have been found to correlate inversely with IGF-I levels in GHD patients, and it is estimated that 26–45% of adults with GHD harbour abnormalities in lipoprotein metabolism: increased total and LDL cholesterol levels and decreased HDL are the main reported abnormalities (Bengtsson et al. 1999). The effect of GHD on insulin resistance is, as in the transition period, dose-dependent (Al-Shoumer et al. 1998, Yuen et al. 2002). High doses increase insulin resistance parameters, while doses close to physiological ranges could even prevent age-related increases in insulin resistance (Svensson et al. 2002). Overall, routine monitoring of glucose abnormalities in replaced GHD patients remains a cautionary measure.
GH therapy has shown to decrease fat mass content (Bengtsson et al. 1993), increase flow-mediated vasodilatation responses to stress and/or exercise (Boger et al. 1996), decrease the values of inflammatory markers (Sesmilo et al. 2000), and improve the lipid profile derangements (Boot et al. 1997, Kearney et al. 2003). However, even adequate replacement has failed to demonstrate an effect on mortality (Molitch et al. 2006).
GHD is also associated with decreased left ventricular mass and impaired ejection fraction at rest and after physical exercise testing, as assessed by echocardiography (Colao et al. 2001). These abnormalities are particularly important in young GHD patients, in whom the benefit of GH therapy on cardiac parameters is of greater significance (Colao et al. 2004).
Fracture rates are increased two- to fivefold when compared with non-GHD control populations (Wuster et al. 2001, Vestergaard et al. 2002). These findings have been related to decreased bone mineral density values, ∼1 s.d. below the mean (Rosen et al. 1993, Holmes et al. 1994). The presence of osteoporosis in GHD cohorts varies from 20% in AO-GHD to 35% in CO-GHD (Kaufman et al. 1992). This effect is more pre-eminent in cortical bone sites. GH replacement enhances peak bone mass accrual in young adults and decreases bone resorption in middle-aged adults. However, the biphasic nature of GH bone action, in which bone resorption is preferentially increased during the first year and bone formation is enhanced at later stages, makes necessary a delayed assessment of bone mineral density after starting GH replacement (Hansen et al. 1996). Currently randomised trials that directly assess the effect of GH on fracture risk are lacking.
Deranged QoL remains a milestone of GHD syndrome. Through disease-specific QoL questionnaires (Burman et al. 1995, Holmes et al. 1995), energy and vitality have consistently shown to be decreased in GHD patients. Adults with GHD experience functional limitations in the social and laboural spheres, increased fatigue and deranged vitality (Woodhouse et al. 2006). Decrease in LBM, particularly of muscle mass, diminished erythropoiesis (Christ et al. 1997) and worsened cardiac performance have been related to functional limitations in GHD populations (Colao et al. 2001). Notwithstanding the mechanisms of GH effect on physiological variables, adequate replacement restores diverse aspects of QoL derangements in GHD patients, and the degree of improvement is directly proportional to the level of pre-treatment impairment (McGauley 1989, Drake et al. 1998, Murray et al. 1999, Blum et al. 2003).
GH replacement has associated side effects that need monitoring. The most common are fluid retention and oedema that are present more frequently in older populations (Holmes & Shalet 1995) and reinforce the need for age-related therapy initiation regimens. Possible effects of GH on carbohydrate metabolism should be routinely assessed and GH therapy should only be introduced in patients with clinically and radiologically inactive neoplasms.
Anomalies in gonadotroph axis: early and delayed puberty, adult gonadotrophin deficiency
Gonadotroph deficiency is the second most common clinical consequence of radiotherapy on the hypothalamic–pituitary region. Pituitary-driven secretion of gonadal steroids (testosterone and oestradiol) is essential in the development of secondary sexual characteristics during puberty and the initiation and maintenance of sexual activity and fertility during adult life. Furthermore, gonadal steroids play a role in the achievement of an optimal peak bone mass. Other actions involve the regulation of body composition (muscle-fat proportion), lipid, carbohydrate and mineral metabolism, regulation of hypothalamic thermoneutral zone and dilatory responses of vascular smooth muscle. Gonadal steroids play unique roles in the different stages of life, and their effects clearly require a degree of sexual dimorphism. As a consequence, the clinical expression of radiation-induced gonadotrophin dysfunction is time and sex dependent.
Disorders of pubertal timing
The correct timing in the attainment of secondary sexual characteristics and somatic maturation is essential for a normal physical and psychological progress into adult life. Cohorts receiving cranial radiation treatments have an increased incidence of pubertal timing disorders that require specific surveillance and treatment (Brauner et al. 1984, Quigley et al. 1989, Brauner & Rappaport 1990, Ogilvy-Stuart et al. 1994, Rose 2008). A list of the risk factors for pubertal timing disruption in cancer survivors is provided in Table 5.
Risk factors for cancer-related disruption of pubertal timing
Radiation-induced disorders in pubertal timing can result in:
-
Precocious puberty: initiation of thelarche before the age of 8 or increase in testicle size before the age of 9. Pathophysiologically, it is hypothesised that precocious puberty is caused by late-onset disruption of cortical inhibitory influences in the initiation of puberty (Roth et al. 2001). Central precocious puberty in cancer survivors usually occurs as a result of the use of low-dose cranial irradiation (total dose averaging 18–24 Gy) as a prophylactic treatment for childhood-onset haematological neoplasms (Leiper et al. 1987) and brain tumours distant from the hypothalamic–pituitary area (Lannering et al. 1997). Female sex and young age at treatment onset are considered as risk factors for radiation-induced precocious puberty (Ogilvy-Stuart et al. 1994).
-
Rapid puberty: puberty starts within 2.5 s.d.s of the reference population range, but the transition from thelarche to menarche takes <18 months (Bath et al. 2002) or the increase by >10 ml in testicle size occurs in <12 months.
-
Inadequate pubertal development in the presence of slow linear growth (Toogood 2004, Darzy & Shalet 2005a): in the presence of any factor interfering in growth processes, even a normally-timed puberty can contribute to a suboptimal height outcome. In this context, pharmacological arrest of puberty might be indicated until the factors interfering in somatic growth are identified and addressed.
-
Delayed puberty: if low doses of radiation disinhibit the initiation of puberty, doses above 24–30 Gy irreversibly damage the central input that drives the initiation of puberty (Stubberfield et al. 1995, Müller 2002). Pubertal delay can actually characterise the initial stages of permanent gonadotroph deficiency in children irradiated for the treatment of brain neoplasms (Rappaport et al. 1982). The consequences of delayed or absent puberty involve a decrease in peak bone mass accrual (Finkelstein et al. 1987, 1992, Moreira-Andres et al. 1998, Hawker et al. 2002, Ho & Kung 2005) and the psychological sequelae resulting from delayed pubertal milestones (Graber et al. 1997). Thus, in irradiated children, the absence of breast development by 12 years of age in girls or the failure to increase the size of the testes by 14 years of age should prompt evaluation by a paediatric endocrinologist. If cancer therapy is being administered during the pubertal years, ‘catch-up growth’ and resumption of pubertal progress should resume in <1 year after completion of cancer therapy. At this stage, or previously if there are treatment or tumour-related risk factors for hormonal deficiencies, full endocrinological evaluation should be performed.
-
Arrested puberty: radiation-related damage to GnRH producing cells is slowly progressive. Thus, early, rapid or normally-timed puberty can be followed by the cessation of pubertal progression before treatment is inititiated. Unless nutritional imbalances or psychosocial factors interfere, radiation-related arrested puberty is frequently followed by permanent gonadotroph dysfunction. Some patients might initially experience precocious puberty (induced by radiotherapy or by direct tumoral compression) after which arrested pubertal development with permanent gonadotroph deficiency, usually caused by radiation treatments, might ensue; the fact that radiotherapy can provoke opposite effects in gonadotroph physiology at different stages in the same patient may be explained by the slow onset of radiation-induced damage, with more cumulative damage required for gonadotroph deficiency as compared with precocious puberty (Burstein 1994, Yeung et al. 1998). Time-dependency of RIH is another factor that warrants careful monitoring of childhood cancer survivors by an experienced multidisciplinary team.
The treatment of pubertal timing disorders in patients with childhood cancer is esentially the same as in the general population (Gleeson & Shalet 2004). Early, rapid or inappropriate puberty for growth benefit from GnRH analogues, while delayed puberty is best treated by progressive doses of gonadal steroid replacement.
Radiation-induced hypogonadism in the adult
Primary and secondary hypogonadism are among the most common endocrine sequelae in cancer survivors (Skinner et al. 2006). Hypogonadism might result from damage at the hypothalamic, pituitary or gonadal level, which is readily discernible through the medical history and basic biochemical evaluation. Cranial radiotherapy, tumoral compression and surgical damage are the most common aetiologies of gonadotroph deficiency, while surgical gonadectomy, chemotherapy and radiotherapy (in pelvic, abdominal and total body irradiation) are the most common determinants of primary gonadal failure. Aetiological investigations are necessary for a tailored management of fertility problems, which can be treated with gonadotroph stimulation in secondary hypogonadism but remain a difficult challenge in primary gonadal failure.
In adult women, hypogonadism results in the inability to ovulate, infertility and decreased oestradiol levels. The initial stages of oestradiol deficiency present with menstrual irregularity, hot flushes, vaginal dryness and atrophy, dyspareunia, loss of libido and fatigue. Chronic oestradiol deficiency is reflected in breast tissue atrophy, weaning of secondary sexual characteristics and osteoporosis (Baird 2002, Conway 2002, Verga 2002). A woman receiving cranial irradiation might develop primary or secondary amenorrhoea, depending on the age of exposure. In patients receiving radiotherapy for tumours of the sellar region, the cause of amenorrhoea is often multifactorial including damage to GnRH releasing hypothalamic cells, ablation of pituitary parenchyma by the tumour or previous surgical procedures and hyperprolactinaemia from stalk disconnection, radiation-related damage to hypothalamic dopamine-secreting cells. The approach to a cancer survivor with amenorrhoea is even more complex. Amenorrhoea may be due to RIH, gonadal damage from radiation or chemotherapy, and primary or secondary hypothyroidism. Malnutrition after cancer therapy can also generate a functional hypothalamic amenorrhoea. All require evaluating in a cancer survivor presenting with amenorrhoea.
In men, radiation-induced secondary hypogonadism results in decreased sperm counts (with secondary infertility) and low testosterone levels. Testosterone deficiency causes reduced libido, poor erection, decreased energy, hot flushes and sweats. Decreased energy, motivation, depressed mood and poor concentration and memory are other symptoms. At later stages, male hypogonadism is accompanied by decreased muscle mass and strength, osteoporosis, excess body fat mass and increased cardiovascular risk. Gynaecomastia can occur as a result of an increased oestradiol/testosterone ratio in hypogonadal patients, although this is far more common in primary hypogonadal patients than in RIH, as FSH and LH stimulate aromatase activity (Bhasin et al. 2006). Hypogonadal patients tend to have lower haematocrit values which add to the effects of hypogonadism on functional performance (Ferrucci et al. 2006).
ACTH and TSH deficiency
ACTH and TSH deficiencies are the least common among RIH. They are more frequently described in patients with previous tumour or surgical insults to the pituitary gland, and following radiation doses above 50 Gy (Littley et al. 1989).
ACTH deficiency in irradiated patients can be mild and subclinical, but without timely recognition it can progress and lead to catastrophic consequences. The most common symptoms are anorexia, weight loss, fatigue, tiredness, weakness, dizziness and postural hypotension, gastrointestinal symptoms, arthralgia and myalgia, intolerance to stress and infection, decreased axillary and pubic hair and reduced libido in women. During stressful situations, cortisol deficiency manifests with hypotension and tachycardia due to vascular collapse and reduced response to inotropic factors. In routine biochemistry, hypoglycaemia, eosinophilia and low sodium levels can be found, but hyperkalaemia is an exclusive feature of primary hypoadrenalism due to mineralocorticoid deficiency (Arlt & Allolio 2003).
The diagnostic tools of routine use for ACTH deficiency include basal 0900 h cortisol, high-dose short Synacthen test (SST), low-dose SST, ITT and metyrapone stimulatory challenge; optimum use of each investigation requires careful assessment of their particular advantages and shortcomings (Table 4), together with periodic and detailed evaluation of clinical signs and symptoms suggesting adrenal failure.
TSH deficiency is the least commonly reported pituitary dysfunction in patients with RIH. When assessed by routine combination of TSH and free thyroxine (FT4) measurements, 5% of patients receiving radiation for childhood cancer and 6–50% treated with radiotherapy for pituitary adenomas report TSH deficiency (Littley et al. 1989, Darzy & Shalet 2005b). In the subgroup with FT4 levels in the lower quartile, 25% show abnormal circadian TSH variation and 32% have a blunted response to TRH stimulation (Rose et al. 1999b). This has been interpreted as the presence of hidden central hypothyroidism, an early stage in the development of thyrotroph dysfunction. However, it is unclear whether these findings illustrate a clinically significant degree of TSH secretory dysfunction, and whether T4 replacement improves any outcome measures in patients with these test results. Furthermore, measures of hidden central hypothyroidism have failed to correlate with FT4 levels and radiotherapy-dependent variables (Romijn & Wiersinga 1990, Adriaanse et al. 1992, Adriaanse et al. 1993, Darzy & Shalet 2005a,b). Globally, the standard of practice for the diagnosis and surveillance of suspected central hypothyroidism involves combined measurement of TSH and FT4 (Lania et al. 2008, Yamada & Mori 2008), while replacement therapy should be aimed at restoring FT4 levels into the upper quartile of the normal range. Although tri-iodothyronine (T3) represents the most important thyroid hormone in terms of biological activity, the diagnostic use of free T3 levels is hampered by substantial overlap between healthy and hypothyroid individuals (Ferretti et al. 1999, Alexopoulou et al. 2004). This overlap is attributed to preferential secretion of T3 when substantial impairment in TSH stimulatory capacity ensues (Alexopoulou et al. 2004).
Hyperprolactinaemia and diabetes insipidus
Radiation-induced hyperprolactinaemia is caused by a depletion of the hypothalamic inhibitory influence on PRL secretion. It is usually mild, more common in adult women receiving doses >40 Gy, with a 14–50% reported incidence of hyperprolactinaemia (Littley et al. 1989, Lam et al. 1991, Constine et al. 1993). In most cases, this is a subclinical finding that resolves spontaneously in the first 5 years after radiation exposure and rarely induces gonadotroph deficiency (Darzy & Shalet 2006). In acromegalic patients, radiation-related hyperprolactinaemia, which tends to resolve spontaneously, should be distinguished from PRL excess derived from somatotroph cell hypersecretion, usually accompanied by GH excess (Ciccarelli et al. 1989).
Despite the vicinity of vasopressin-secreting magnocellular neurons to the hypothalamic–adenohypophyseal functional unit, to our knowledge central diabetes insipidus has not been consistently reported as a direct consequence of RT for sellar or extrasellar brain neoplasms, which comes in sharp contrast with the common observation of this disorder after non-adenomatous tumoral or surgical insults to the hypothalamic–pituitary area (Sudhakar et al. 2004, Dumont et al. 2005). Interestingly, a series described the occurrence of diabetes insipidus in a significant minority of patients (8%) who received pituitary SFR using total doses as high as 140–180 Gy for the treatment of post-stroke intractable thalamic pain (Hayashi et al. 2007). Albeit impaired vasopressin secretion remained mostly a transient phenomenon in the series, these findings suggest a much higher dose threshold for radiation-induced diabetes insipidus. However, further research should be encouraged to confirm these findings and possibly shed more light into the pathophysiology of this phenomenon.
Associated disorders in patients with RIH
The follow-up of patients with RIH must also account for the presence of other associated morbidities that are highly dependent on the underlying neoplasm. In the case of tumours arising in the sellar region, these initially result from either surgical or tumour-induced compression of neural structures, most commonly the optic chiasm, cranial nerves and hypothalamus, leading to loss of visual function and symptoms of hypothalamic dysfunction (abnormal drinking and feeding behaviour, poikilothermia and memory and mood problems). Although hypopituitarism is the most common sequelae of radiotherapy for sellar and parasellar tumours, RT may also result in other relatively uncommon radiation-induced brain disorders listed in Table 6. These sequelae include radiation optic neuropathy, radiation brain necrosis, increased incidence of cerebrovascular accidents and second-brain tumours, and they tend to occur for months-to-years after RT. Although rare following conventional radiation doses used for sellar neoplasms, the clinician involved in the management of these patients needs to be aware of them.
Radiation-induced brain disorders
The effects of cancer treatments involve systemic as well as local effects. As a result of the direct effect of the tumour, surgical and radiation treatments and chemotherapy regimens, as many as 41% of childhood cancer survivors suffer endocrine sequelae, together with non-endocrine side effects, such as congestive heart failure, second malignancies, cognitive dysfunction, renal failure, coronary artery disease or cerebrovascular accidents (Stevens et al. 1998, Oeffinger et al. 2006, Skinner et al. 2006). The most common endocrine side effects of cancer treatments are shown in Table 7. Awareness of these risks among physicians involved in the long-term care of childhood cancer survivors is an important tool for optimising outcomes.
Endocrine late effects of cancer treatments
In most endocrine disorders presenting in cancer survivors, clear correlations have been found between treatment-related variables (i.e. radiation dose, chemotherapy scheme) and their development. Diabetes mellitus represents an exception to this rule. Cancer survivors seem to be at risk of being diagnosed with this disorder (Neville et al. 2006, Stava et al. 2007), a finding that has been particularly well documented following haematopoietic cell transplant (Taskinen et al. 2000, Hoffmeister et al. 2004). Increased rates of obesity, sedentary lifestyle, effect of some chemotherapy regimens, glucocorticoid treatments and hormonal deficiencies have been advocated as risk factors for diabetes mellitus in cancer survivors (Neville et al. 2006). Radiation-induced β-cell dysfunction was postulated as a causative mechanism in a series of nephroblastoma survivors (Teinturier et al. 1996), but to our knowledge these findings have not been reproduced (Hawkins et al. 1996). Total body irradiation has been inconsistently associated with increased risk of diabetes mellitus (Lorini et al. 1995, Taskinen et al. 2000, Hoffmeister et al. 2004, Neville et al. 2006), but the available evidence does not firmly link part of the diabetes risk to a direct deleterious effect of RT on β-cells. Notwithstanding the aetiology, the burden of diabetes-associated morbidities strongly advocates for routine screening strategies in cancer survivors.
Hypopituitarism and mortality
The prognosis of patients with RIH is in great measure influenced by the underlying disease and other effects of cancer treatment. However, hypopituitarism may be an independent risk factor for mortality, particularly due to cerebrovascular causes (Rosen & Bengtsson 1990, Bates et al. 1996, 1999, Bulow et al. 1997, Erfurth et al. 2000, 2001, Tomlinson et al. 2001, Brada et al. 2002). Cerebrovascular accidents in cohorts with hypopituitarism are more common in women and in untreated hypogonadal patients of both sexes (Tomlinson et al. 2001, Lindholm et al. 2006, Nielsen et al. 2007). Higher mortality rates in women with hypopituitarism have been attributed to suboptimal sex-hormone replacement strategies and possibly an overlooked decrease in glucocorticoid requirements as compared with their male counterparts. The use of radiotherapy has also been associated with increased cerebrovascular mortality in most (Bates et al. 1996, Tomlinson et al. 2001, Brada et al. 2002) but not all studies (Erfurth et al. 2002). Radiation-induced damage to the vascular wall may be the cause of the enhanced cerebrovascular atherogenicity in irradiated patients (Gittoes 2005). These epidemiological findings represent a note of caution in the use of radiotherapy for pituitary tumours until further evidence defines the long-term risk profile of more modern radiation techniques and hormonal replacement practice.
Conclusions
The advent of modern radiotherapy has contributed to improved survival and QoL in patients with a variety of malignancies and benign tumours, albeit with an increased risk of treatment related side effects. RIH is to be one of the most common sequelae of RT. While hypothalamic–pituitary sparing techniques are being developed, strategies for risk-assessment and early recognition of RIH and other adverse effects should be initiated at completion of radiation treatments. As regards RIH, pathophysiological tailoring of biochemical assessment and current availability of replacement therapies closer to the physiology of the normal pituitary gland is expected to further improve outcomes in this population. However, other possible side effects of the primary neoplasm and its treatments on the pituitary function, its target organs and the general health outcome should also be considered. The recognition and management of radiation-induced brain disorders, endocrine and systemic effects of radiation and chemotherapy and the sequelae of the primary tumour are important tools for optimising outcomes. As the population of cancer survivors continues to increase and cancer treatments evolve in a constant endeavour to improve survival rates, renewed efforts for defining and treating the sequelae of future cancer survivors should be undertaken.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
- © 2009 Society for Endocrinology
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
