**2. Radiation-induced hypothalamic-pituitary dysfunction**

#### **2.1 Pathophysiology**

Radiation damage is a potent cause of hypothalamic-pituitary (h-p) axis dysfunction. Deficiency of one or more of the anterior pituitary hormones may occur following radiotherapy for tumours of the head and neck when the h-p axis falls within the field of radiation. The pathophysiology of radiation-induced damage remains poorly understood. Neuronal cell death and degeneration due to the direct effects of radiation appear to play a major role (Hochberg et al., 1983); however, vascular damage has also been proposed (Chieng et al., 1991).

The onset and severity of radiation-induced hypopituitarism is primarily determined by the total radiation dose, the fraction size and the time allowed between fractions for tissue

Endocrine Complications Following Radiotherapy

follow up post radiotherapy (Littley et al., 1989a).

in the dysfunction of the irradiated h-p axis.

hormone deficiency.

and Chemotherapy for Nasopharyngeal Carcinoma 135

leukaemia. For example, isolated GH deficiency is frequently seen in children who received radiation doses of less than 24 Gy (Ogilvy-Stuart et al., 1992) but none in the adults (Littley et al., 1991). In a study of 56 patients irradiated for non-pituitary brain tumours in adulthood, Agha et al (Agha et al., 2005) reported variable degrees of hypopituitarism in 41% of patients. In this study (Agha et al., 2005), GH deficiency (32%) was less frequent that that reported in irradiated children (Clayton & Shalet, 1991; Livesey et al., 1990; Samaan et al., 1987), but ACTH (21%), TSH (9%) and gonadotropin (27%) deficiencies were relatively more common than or similar to that reported in cancer survivors irradiated during childhood (Constine et al., 1993; Livesey et al., 1990; Samaan et al., 1987). The differential influence of age is less clearly defined with intensive irradiation, but it appears to follow the same pattern. Samaan et al (Samaan et al., 1987) in their study of 166 patients aged 6-80 years, who had received high dose irradiation for NPC, showed that children younger than 15 years of age had a higher incidence of GH deficiency soon after radiotherapy than older patients; however, the older age group showed more adrenocortical and luteinizing

Irrespective of the intensity of radiation schedule, radiation-induced h-p dysfunction is also time dependent. Both increased incidence and severity of hormonal deficits are seen with longer post-irradiation follow-up intervals (Achermann et al., 2000; Clayton & Shalet, 1991; Lam et al., 1991; Littley et al., 1989b; Samaan et al., 1987; Schmiegelow et al., 2000) (Fig 1 &2). Secondary pituitary atrophy consequent upon lack of hypothalamic releasing/trophic factors accounts for the progressive nature of the hormonal deficits, in addition, to the delayed direct effects of radiotherapy on the axis. There is a belief that radiation may cause delayed brain tissue damage and dysfunction through chronic inflammation and/or enhanced release of proinflammatory cytokines (Chiang et al., 1997; Kyrkanides et al., 1999). The delayed direct radiation damage to the pituitary gland is supported by the gradual decline in the elevated prolactin levels seen in some patients after prolonged periods of

The predominant site of radiation damage, pituitary vs. hypothalamic, has attracted some controversy. Contrary to what had been believed that the hypothalamus is more radiosensitive than the pituitary and that hypothalamic damage predominates following less intensive radiation schedules (<50 Gy), recent studies by the author et al (Darzy et al., 2005, 2006, 2007, 2009) have strongly suggested the opposite with robust evidence that direct radiation-induced damage to the pituitary still occurs even with low radiation doses and that the pituitary may be the predominant site of radiation damage. However, with higher range of conventional irradiation, i.e. doses in excess of 60 Gy, there is robust clinical evidence to suggest that intensive radiotherapy inflicts dual damage to both the pituitary as well as the hypothalamus resulting in early multiple anterior pituitary hormone deficiencies (Chen et al., 1989; Lam et al., 1991; Pai et al., 2001; Samaan et al., 1982). In addition to direct hypothalamic damage, neuropharmacological studies have suggested that radiationinduced hypothalamic dysfunction may be secondary to radiation damage of the suprahypothalamic neurotransmitter pathways (Jorgensen et al., 1993; Ogilvy-Stuart et al., 1994). Radiation-induced changes at cellular and molecular levels most certainly play a role

Pituitary damage is demonstrated by impaired GH, LH/FSH, and TSH responses to direct stimulation with exogenous GHRH, LHRH or TRH, respectively. Hypothalamic damage, on

repair (duration of the radiation schedule) (Thames & Hendry, 1987; Littley et al., 1989a). Radiation schedules utilising the same total dose administered over a shorter duration (larger fraction size) inflict more damage to the h-p axis. To minimise the damage to healthy neuronal tissues (including h-p axis), most radiation schedules have not used more than 2 Gy per fraction and no more than 5 fractions per week. Increasing the fraction size above 2 Gy per fraction (for the same total dose) can induce relatively more injury to the late responding (neuronal) than the early responding (tumour) tissues (Withers, 1994).

Intensive external fractionated radiotherapy in doses exceeding 60 Gy remains the primary treatment for NPC. The radiotherapy field normally covers the nasopharynx and both sides of the neck. The h-p axis is routinely included in the irradiated volume. Consequently, the rate and intensity of neuro-endocrine disturbances complicating treatment of NPC far exceed that seen following less intense therapeutic radiation schedules (18-45 Gy) used for the treatment of brain tumours or haematological malignancies. With modern technological advances in computed tomography and magnetic resonance imaging, it has become possible to use conformal radiotherapy to deliver a higher radiation dose to the main bulk of the tumour while sparing the important nearby structure to reduce long-term complications (Wei, 2001). In addition, shielding the pituitary gland during radiotherapy has been shown to reduce disturbances in pituitary function without compromising tumour control (Sham et al., 1994).

The nature of the neuro-endocrine disturbance following h-p axis irradiation is also determined by the differential radiosensitivity of hypothalamic-pituitary function. This has been shown in animal models (Hochberg et al., 1983; Robinson et al., 2001) and reflected in clinical observations in irradiated patients. Epidemiological studies reveal that the growth hormone (GH) axis is the most radiosensitive followed by the gonadotrophin (FSH & LH), adrenocorticotrophic hormone (ACTH) and thyroid stimulating hormone (TSH) axes (Clayton & Shalet, 1991; Constine et al., 1993; Duffner et al., 1985; Lam et al., 1991; Littley et al., 1989a) (Figures 1& 2).

Low radiation doses of less than 40 Gy mostly affects the most vulnerable GH axis in isolation resulting in variable degrees of GH deficiency (Clayton & Shalet, 1991; Constine et al., 1993; Duffner et al., 1985; Littley et al., 1989a). Deficiencies of other anterior pituitary hormones start to occur when the total radiation dose delivered to the h-p axis exceeds 40 Gy, but much less frequently than GH deficiency. Panhypopituitarism is mostly seen following intensive irradiation with doses exceeding 60 Gy, typically used for the treatment of nasopharyngeal carcinoma and skull base tumours (Chen et al., 1989; Lam et al., 1991; Pai et al., 2001; Samaan et al., 1987, 1982). In contrast, posterior pituitary dysfunction with diabetes insipidus has not been reported even after the most intensive irradiation schedules (Pai et al., 2001).

With less intensive radiation schedules utilising doses of less than 40 Gy, it would appear that age at irradiation influences differential impact on various h-p axes susceptibility to radiation damage. The somatotrophic (GH) axis is more vulnerable to radiation damage in children than adults, while the ACTH axis seems to be more vulnerable to damage in adults than children. These conclusions are based on the relative frequencies of various anterior pituitary hormones deficiencies reported with various radiation schedules, with a dose range of 18-50 Gy, administered to children and adults for non-pituitary brain tumours and

repair (duration of the radiation schedule) (Thames & Hendry, 1987; Littley et al., 1989a). Radiation schedules utilising the same total dose administered over a shorter duration (larger fraction size) inflict more damage to the h-p axis. To minimise the damage to healthy neuronal tissues (including h-p axis), most radiation schedules have not used more than 2 Gy per fraction and no more than 5 fractions per week. Increasing the fraction size above 2 Gy per fraction (for the same total dose) can induce relatively more injury to the late

Intensive external fractionated radiotherapy in doses exceeding 60 Gy remains the primary treatment for NPC. The radiotherapy field normally covers the nasopharynx and both sides of the neck. The h-p axis is routinely included in the irradiated volume. Consequently, the rate and intensity of neuro-endocrine disturbances complicating treatment of NPC far exceed that seen following less intense therapeutic radiation schedules (18-45 Gy) used for the treatment of brain tumours or haematological malignancies. With modern technological advances in computed tomography and magnetic resonance imaging, it has become possible to use conformal radiotherapy to deliver a higher radiation dose to the main bulk of the tumour while sparing the important nearby structure to reduce long-term complications (Wei, 2001). In addition, shielding the pituitary gland during radiotherapy has been shown to reduce disturbances in pituitary function without compromising tumour control (Sham et

The nature of the neuro-endocrine disturbance following h-p axis irradiation is also determined by the differential radiosensitivity of hypothalamic-pituitary function. This has been shown in animal models (Hochberg et al., 1983; Robinson et al., 2001) and reflected in clinical observations in irradiated patients. Epidemiological studies reveal that the growth hormone (GH) axis is the most radiosensitive followed by the gonadotrophin (FSH & LH), adrenocorticotrophic hormone (ACTH) and thyroid stimulating hormone (TSH) axes (Clayton & Shalet, 1991; Constine et al., 1993; Duffner et al., 1985; Lam et al., 1991; Littley et

Low radiation doses of less than 40 Gy mostly affects the most vulnerable GH axis in isolation resulting in variable degrees of GH deficiency (Clayton & Shalet, 1991; Constine et al., 1993; Duffner et al., 1985; Littley et al., 1989a). Deficiencies of other anterior pituitary hormones start to occur when the total radiation dose delivered to the h-p axis exceeds 40 Gy, but much less frequently than GH deficiency. Panhypopituitarism is mostly seen following intensive irradiation with doses exceeding 60 Gy, typically used for the treatment of nasopharyngeal carcinoma and skull base tumours (Chen et al., 1989; Lam et al., 1991; Pai et al., 2001; Samaan et al., 1987, 1982). In contrast, posterior pituitary dysfunction with diabetes insipidus has not been reported even after the most intensive irradiation schedules

With less intensive radiation schedules utilising doses of less than 40 Gy, it would appear that age at irradiation influences differential impact on various h-p axes susceptibility to radiation damage. The somatotrophic (GH) axis is more vulnerable to radiation damage in children than adults, while the ACTH axis seems to be more vulnerable to damage in adults than children. These conclusions are based on the relative frequencies of various anterior pituitary hormones deficiencies reported with various radiation schedules, with a dose range of 18-50 Gy, administered to children and adults for non-pituitary brain tumours and

responding (neuronal) than the early responding (tumour) tissues (Withers, 1994).

al., 1994).

al., 1989a) (Figures 1& 2).

(Pai et al., 2001).

leukaemia. For example, isolated GH deficiency is frequently seen in children who received radiation doses of less than 24 Gy (Ogilvy-Stuart et al., 1992) but none in the adults (Littley et al., 1991). In a study of 56 patients irradiated for non-pituitary brain tumours in adulthood, Agha et al (Agha et al., 2005) reported variable degrees of hypopituitarism in 41% of patients. In this study (Agha et al., 2005), GH deficiency (32%) was less frequent that that reported in irradiated children (Clayton & Shalet, 1991; Livesey et al., 1990; Samaan et al., 1987), but ACTH (21%), TSH (9%) and gonadotropin (27%) deficiencies were relatively more common than or similar to that reported in cancer survivors irradiated during childhood (Constine et al., 1993; Livesey et al., 1990; Samaan et al., 1987). The differential influence of age is less clearly defined with intensive irradiation, but it appears to follow the same pattern. Samaan et al (Samaan et al., 1987) in their study of 166 patients aged 6-80 years, who had received high dose irradiation for NPC, showed that children younger than 15 years of age had a higher incidence of GH deficiency soon after radiotherapy than older patients; however, the older age group showed more adrenocortical and luteinizing hormone deficiency.

Irrespective of the intensity of radiation schedule, radiation-induced h-p dysfunction is also time dependent. Both increased incidence and severity of hormonal deficits are seen with longer post-irradiation follow-up intervals (Achermann et al., 2000; Clayton & Shalet, 1991; Lam et al., 1991; Littley et al., 1989b; Samaan et al., 1987; Schmiegelow et al., 2000) (Fig 1 &2). Secondary pituitary atrophy consequent upon lack of hypothalamic releasing/trophic factors accounts for the progressive nature of the hormonal deficits, in addition, to the delayed direct effects of radiotherapy on the axis. There is a belief that radiation may cause delayed brain tissue damage and dysfunction through chronic inflammation and/or enhanced release of proinflammatory cytokines (Chiang et al., 1997; Kyrkanides et al., 1999). The delayed direct radiation damage to the pituitary gland is supported by the gradual decline in the elevated prolactin levels seen in some patients after prolonged periods of follow up post radiotherapy (Littley et al., 1989a).

The predominant site of radiation damage, pituitary vs. hypothalamic, has attracted some controversy. Contrary to what had been believed that the hypothalamus is more radiosensitive than the pituitary and that hypothalamic damage predominates following less intensive radiation schedules (<50 Gy), recent studies by the author et al (Darzy et al., 2005, 2006, 2007, 2009) have strongly suggested the opposite with robust evidence that direct radiation-induced damage to the pituitary still occurs even with low radiation doses and that the pituitary may be the predominant site of radiation damage. However, with higher range of conventional irradiation, i.e. doses in excess of 60 Gy, there is robust clinical evidence to suggest that intensive radiotherapy inflicts dual damage to both the pituitary as well as the hypothalamus resulting in early multiple anterior pituitary hormone deficiencies (Chen et al., 1989; Lam et al., 1991; Pai et al., 2001; Samaan et al., 1982). In addition to direct hypothalamic damage, neuropharmacological studies have suggested that radiationinduced hypothalamic dysfunction may be secondary to radiation damage of the suprahypothalamic neurotransmitter pathways (Jorgensen et al., 1993; Ogilvy-Stuart et al., 1994). Radiation-induced changes at cellular and molecular levels most certainly play a role in the dysfunction of the irradiated h-p axis.

Pituitary damage is demonstrated by impaired GH, LH/FSH, and TSH responses to direct stimulation with exogenous GHRH, LHRH or TRH, respectively. Hypothalamic damage, on

Fig. 1. Cumulative probability of normal endocrine function following radiotherapy for nosopharyngeal carcinoma. Adapted from Lam et al 1991, with permission.

Fig. 2. Percentages of 166 patients with abnormal hormonal levels according to years after radiotherapy for NPC. Adapted from Samaan et al 1987 – Table II, with permission.

Endocrine Complications Following Radiotherapy

**2.2 Growth Hormone (GH) deficiency** 

deficiency soon after irradiation.

**2.2.1 Epidemiology and pathophysiology** 

radiation schedules (Constine et al., 1993; Rose et al., 1999).

diagnosis of GH deficiency in adults (Shalet et al., 1998).

and Chemotherapy for Nasopharyngeal Carcinoma 137

the other hand, is characterized by hypothalamic pattern of responses (delayed responses) to LHRH and TRH tests. A robust sign for hypothalamic damage is the occurrence of hyperprolactinaemia due to a reduction in hypothalamic release of the inhibitory neurotransmitter, dopamine. These abnormalities in hypothalamic functions have been clearly described in those intensively irradiated for nasopharyngeal carcinoma (Chen et al., 1989; Lam et al., 1991; Samaan et al., 1987) and skull base tumours (Pai et al., 2001) but much less frequently in those treated for other brain tumours or leukaemia with less intensive

GH deficiency is the earliest manifestation of neuro-endocrine injury following cranial irradiation. With intensive irradiation used for NPC, the cumulative frequency of GH deficiency is well above 60% after 5 years (Lam et al., 1991). Higher incidence of severe GH deficiency is seen with longer follow up periods reaching well above 80% (Samaan et al., 1987). Given the higher radiosensitivity of the GH axis, GH deficiency is almost always present if deficiencies of one or more of the other anterior pituitary hormones are confirmed. Studies of stimulated GH secretion in children treated for brain tumours indicate that almost all children treated with doses in excess of 35 Gy will have blunted GH secretion within 2-5 years of treatment (Clayton & Shalet, 1991). With the more intensive radiation used for NPC, all children treated for this condition will undoubtedly manifest features of severe GH

Apart from the higher radiosensitivity of the GH axis in children, the higher frequency of severe GH deficiency in children may be explained by the much higher threshold of peak GH response to stimulation used to diagnose GH deficiency in this age group. Children who have been categorised as having severe GH deficiency may in fact be categorised as having normal GH status when retested in adult life. This apparent discrepancy is not related to recovery of the GH axis, but can be attributed to the use of more strict thresholds for the

GH is secreted in a pulsatile manner with a diurnal variation. The latter is characterised by nocturnal increase in GH secretion. This complex pattern of secretion is under hypothalamic control. Recent pathophysiological studies by the author et al of stimulated and spontaneous GH secretion in a cohort of adult cancer survivors irradiated for brain tumours with doses of less than 50 Gy, suggested that hypothalamic regulation of GH secretion in patients with severe GH deficiency is maintained with preserved pulsatility and diurnal variation (Darzy et al., 2005, 2006). The reduction in GH levels appears to be related to a predominant quantitative damage to the pituitary somatotrophs leading to reduced GH pulse amplitude but not frequency. Another study by the author et al (Darzy et al., 2007) has suggested the presence of a compensatory increase in hypothalamic GHRH release to maintain a normal spontaneous GH secretion in patients with reduced pituitary somatotrophs reserve indicated by reduced peak GH responses to direct stimulation with the most potent GHRH and Arginine stimulation test. There has also been a suggestion for the presence of 'compensated GH deficiency' in some patients who would otherwise have been diagnosed with GH deficiency due to impaired peak GH responses to insulin-induced

12345

Fig. 1. Cumulative probability of normal endocrine function following radiotherapy for

nosopharyngeal carcinoma. Adapted from Lam et al 1991, with permission.

LH deficiency

GH; Gonadotrophins; ACTH; TSH; Prolactin

FSH deficiency

1-4 yrs; 5-9 yrs; 10-14 yrs; ≥15 yrs

Fig. 2. Percentages of 166 patients with abnormal hormonal levels according to years after radiotherapy for NPC. Adapted from Samaan et al 1987 – Table II, with permission.

Increased TSH

Hyperprolactinaemia

0

Percentage

GH deficiency

ACTH

deficiency

10

20

30

40 50

Probability of normal endocrine function

60

70

80

90

100

the other hand, is characterized by hypothalamic pattern of responses (delayed responses) to LHRH and TRH tests. A robust sign for hypothalamic damage is the occurrence of hyperprolactinaemia due to a reduction in hypothalamic release of the inhibitory neurotransmitter, dopamine. These abnormalities in hypothalamic functions have been clearly described in those intensively irradiated for nasopharyngeal carcinoma (Chen et al., 1989; Lam et al., 1991; Samaan et al., 1987) and skull base tumours (Pai et al., 2001) but much less frequently in those treated for other brain tumours or leukaemia with less intensive radiation schedules (Constine et al., 1993; Rose et al., 1999).
