2.3.6 Hypopituitarism: replacement of adrenal, thyroid, and gonadal axes

Hormone replacement of anterior pituitary hormone deficiency is one of the most frequent clinical interventions in pituitary disease, yet is an area which has rarely been the subject of rigorous scientific evaluation. Even in an era of ‘evidence-based’ medicine, recommendations for patient management are frequently based predominantly on clinical experience, consensus guidelines and occasional retrospective reviews rather than on controlled, prospective clinical trials. Within these limitations, this chapter will attempt to give a balanced view on current best management of adrenocorticotropic hormone (ACTH), thyroid-stimulating hormone (TSH) and gonadotropin deficiency.

Hydrocortisone, the generic pharmaceutical name for cortisol, is the standard form of glucocorticoid replacement for ACTH deficiency, and directly replaces the missing active hormone. Cortisone acetate was previously widely used, but is metabolized to cortisol to achieve its glucocorticoid activity, so that its onset of action is slower than hydrocortisone (a slight disadvantage) and biological half-life slightly longer (potentially a slight advantage).

The normal pattern of diurnal cortisol secretion is difficult to mimic precisely with oral therapy and there is no universal agreement regarding the appropriate dose, timing, and monitoring of hydrocortisone replacement, although the need for close attention has been highlighted (1). Normal individuals demonstrate undetectable cortisol and ACTH when asleep at midnight, with a sharp rise during the last hours of sleep to reach a peak at 08.00–09.00 h, followed by a steady decline throughout the rest of the waking day, with superimposition of variable peaks of cortisol secretion due other factors such as stress, meals, and exercise. Some approximation to this pattern can be achieved with thrice-daily regimens (hydrocortisone on rising, mid-day, and early evening) which appear to achieve more ‘physiological’ plasma cortisol levels (see below) compared with a traditional twice-daily regimen, which usually results in very low cortisol levels in late afternoon before the evening dose.

Two groups have reported progress in the development of a slow-release or tailored-release form of hydrocortisone, which might be capable of mimicking normal physiological cortisol profiles more accurately (2–4), but as yet neither form is available for routine clinical practice. Use of prednisolone or dexamethasone has been advocated on the basis of the longer half-life of these more potent synthetic glucocorticoids. However, these drugs cannot be monitored precisely and, in the case of dexamethasone, the potential for fine dose adjustment is limited by the pharmaceutical preparations available so that Cushingoid side effects are more frequent. Although these drugs may be preferred when suppression of an abnormal adrenal is required (e.g. in congenital adrenal hyperplasia), they have no advantage for routine replacement of hypopituitarism.

Criteria for deciding optimum hydrocortisone regimens are inevitably a compromise between theory, practicality, and patient convenience. Glucocorticoid replacement with hydrocortisone may be monitored using plasma cortisol measurements at multiple times throughout the day—a hydrocortisone day curve (HCDC). Studies using frequent sampling for plasma cortisol have identified wide interindividual variations in plasma cortisol levels obtained after the same dose of hydrocortisone and highlighted the need for individual adjustment of hydrocortisone dose, but such frequent sampling is rarely possible or necessary in routine practice. Simpler HCDC regimens are advocated by many centres to adjust and compare hydrocortisone replacement regimens (1, 5) although others have questioned their value and argued that clinical assessment may be equally effective (6).

My practice is to monitor hydrocortisone replacement with a simple HCDC involving collection of a 24-h urine for free cortisol (UFC) on the day prior to the test, and three plasma cortisols during a daycase attendance; patients take their morning hydrocortisone dose at the normal time, at home, on wakening and cortisol is measured at 09.00 h, 12.30 h (prior to any lunchtime dose),and 17.30 h (prior to the evening dose). For optimal replacement, I aim for a hydrocortisone dose which achieves a UFC and 09.00 h cortisol within the reference range for the normal population (28–220 nmol/24 h and 100–700 nmol/l, respectively, in our laboratory)—to avoid over-replacement—combined with 12.30 h and 17.30 h cortisol above 50 nmol/l, and ideally above 100 nmol/l—to avoid under-replacement. Using these criteria in a retrospective review of 130 patients (5) we demonstrated that thrice-daily regimens compared with twice-daily regimes are more likely to achieve 09.00 h plasma cortisol and UFC in the reference range for the normal population, more likely to avoid significant biochemical cortisol deficiency prior to the evening dose, and achieve the highest overall score for attainment of all four criteria. Overall, optimal replacement was most often achieved on hydrocortisone 10 mg on rising, 5 mg at lunchtime, and 5 mg in early evening, and this is therefore my usual starting dose for a new patient requiring replacement. This dose is lower than traditional recommendations but has also been derived by others using different empirical criteria for assessment and correlates with the oral hydrocortisone dosage equivalent of current estimates of the cortisol production rate in normal individuals (7).

After commencing hydrocortisone, individual dose adjustment is essential and a HCDC should be performed, doses adjusted, and HCDC repeated until the optimal hydrocortisone dose for an individual patient is identified. Thereafter I do not perform repeated measurement of cortisol unless clinical conditions change or symptoms indicate the need; others have advocated repeating measurements on a regular basis but there is no objective evidence comparing these two approaches.

Even thrice-daily regimens fail to accurately mimic the normal physiological circadian rhythm of cortisol and attention has been focused on the differences in circulating cortisol levels overnight in patients on hydrocortisone replacement (where levels are low or undetectable throughout the night) compared with normal individuals (where levels rise substantially during the last hours of sleep). These differences would be hard to avoid using standard preparations of hydrocortisone, since few patients would be prepared to wake to take a tablet during the night, but certainly represent an unphysiological feature of current replacement regimens, may contribute to an overnight deficiency of a variety of metabolic fuels (8), and are potentially an area where slow-release preparations could have advantages in future (although clinical trial evidence is still awaited). In practical terms, patients should certainly be advised to take hydrocortisone as soon as possible after wakening to avoid prolonged activity with low circulating cortisol levels, and those who habitually wake in the early hours some time before rising might benefit from taking their hydrocortisone dose at that time.

Patients with ACTH deficiency should not require mineralocorticoid replacement, since the renin–angiotensin–aldosterone axis is not disrupted by pituitary disease. Studies have confirmed normal aldosterone levels in hypopituitary patients on replacement—although there may be subtle differences in dynamic responses to physiological stimuli.

Glucocorticoid replacement is essential during intercurrent illness, and doses need to be increased for all but the most minor illness in order to mimic the normal increase in ACTH and cortisol secretion that occurs during stress and illness. Appropriate patient education on this aspect of replacement is a vital part of management of hypoadrenalism. Patients must be advised to double their normal oral dose of hydrocortisone during common pyrexial illnesses, and understand the need for parenteral glucocorticoid replacement if illness, operation, vomiting, or diarrhoea prevents the effective administration or absorption of oral glucocorticoid. Patients should seek medical advice if symptoms worsen in spite of increased oral hydrocortisone, should keep an ‘emergency’ ampoule of hydrocortisone at home, and if possible they or their family should be taught to give the injection if medical help is unavailable. Some patients also find a symptomatic need for increased glucocorticoid replacement during psychological stress, but this is much more difficult to define or regulate. During severe intercurrent illness or major surgery, hydrocortisone 100 mg, intramuscularly every 6 h will provide consistent, high levels of circulating cortisol comparable with those found in normal individuals during such stress; intravenous boluses of hydrocortisone produce wide swings in cortisol levels and are therefore less desirable, but stable, high plasma cortisol levels can be achieved by an IV infusion of hydrocortisone 5 mg/h (preceded by a 25 mg IV bolus) (9), although this is only appropriate in circumstances where an IV infusion can be reliably maintained and monitored.

Patient support groups have developed clearly formatted guidance on replacement for intercurrent illness and during surgery and other procedures which are readily available on the internet and invaluable when providing advice to patients and surgical colleagues (10).

Adverse effects of supraphysiological doses of glucocorticoid are serious and well known, as are those of severe deficiency. In theory perfect physiological replacement should lead to no adverse effects, since circulating cortisol levels would be no different from normal subjects, but close attention to replacement doses is essential in order to achieve this.

Gross Cushingoid side effects and symptoms of severe hypoadrenalism are certainly clinically obvious and any competent clinical endocrinologist can avoid such extremes of inappropriate replacement, but minor degrees of over- or under-replacement could easily be clinically undetectable, yet give rise to important morbidity or even mortality. Several studies support this view: glucose tolerance and insulin secretion alter with hydrocortisone replacement (11), and blood pressure rises with replacement therapy (12) and may show qualitative differences between regimens although some workers have found no obvious changes related to hydrocortisone dose (13). Therefore, if minor over-replacement caused a slight worsening cardiovascular risk factors such as glucose intolerance, central obesity or blood pressure, then this might be undetectable in an individual yet have a significant influence on overall cardiovascular morbidity.

Bone mineral density may be subnormal in some patients on glucocorticoid replacement (14) although this is not confirmed by other more recent studies (6); excessive steroid replacement is a possible aetiological factor, and markers of bone metabolism normalize after reduction of hydrocortisone replacement dose to more ‘appropriate’ levels, although subsequent changes in bone mineral density are variable (7, 15). These factors indicate the need to avoid even subclinical glucocorticoid over-replacement and aim for the lowest total dose of hydrocortisone replacement compatible with good health. Conversely, avoidance of very low cortisol levels before the next dose seems advisable to minimize the risk of hypoadrenalism if intercurrent illness or stress occurs at that time.

Drugs that induce CYP3A4 liver enzyme drug metabolism (e.g. many anticonvulsants, rifampicin) increase hydrocortisone metabolism and patients taking these drugs may require higher doses of hydrocortisone to achieve adequate circulating levels. Concomitant growth hormone replacement therapy has been shown to reduce cortisol levels on a constant dose of hydrocortisone. This effect appears to be mediated by a reduction in levels of cortisol-binding globulin and so may not be of clinical importance, but may indicate the need for revised criteria for assessment in these patients.

Dehydroepiandrosterone (DHEA) is normally the most abundant circulating adrenal steroid and levels are low in ACTH deficiency. Several clinical trials (16–18) have suggested improvements in general wellbeing, mood, subjective health status, and bone density with DHEA replacement (25–50 mg daily), particularly in women. However, some other studies have shown no benefit and the detail of effects on mood and wellbeing appears to vary between studies. Although DHEA is now regularly advocated as an important part of full adrenal replacement (19), the area remains controversial. Furthermore DHEA is not available as a licensed pharmaceutical preparation but only as a ‘nutritional supplement’ (which can therefore be readily purchased directly by the general public). No long-term follow-up data are available and routine replacement cannot currently be advocated—but the possibility of an individual therapeutic trial of replacement may be reasonably discussed with patients who have unresolved symptoms despite full and satisfactory conventional pituitary replacement therapy.

Levothyroxine is the routine replacement used for treatment of TSH deficiency. Its long plasma half-life ensures stable levels of thyroid hormones on once-daily administration, and conversion to T₃in vivo results in appropriate blood levels of both T₄ and T₃. Liothyronine (T₃; triiodothyronine) can be used, but has no advantage in most circumstances. Combined T₄ and T₃ replacement and use of ‘natural’ thyroid extracts has been advocated in print and on the internet by a variety of ‘alternative’ clinicians and groups, but clinical trial evidence is limited and where available suggests no benefit (20).

Starting dose and regimen of levothyroxine replacement depends on the clinical circumstances. Prior to commencing levothyroxine it is essential to know the status of the ACTH–adrenal axis, because starting levothyroxine without glucocorticoid replacement in a patient with severe ACTH deficiency may precipitate a hypoadrenal crisis. If ACTH deficiency is present, hydrocortisone must be started before levothyroxine. Thereafter, many patients with TSH deficiency have serum free T₄ levels only slightly below the reference range, and in patients with such mild deficiency and no evidence of cardiovascular disease, replacement can be simply commenced with a near-full replacement dose of levothyroxine 100 μg once daily. In patients with more profound reductions of serum free T₄, a lower starting dose of 50 μg daily, increased after a few weeks to 100 μg may be better tolerated. In elderly patients, or any patient with known cardiovascular disease—particularly ischaemic heart disease, greater care is required: in most cases a starting dose of 25 μg will be well tolerated, increased slowly in 25 μg increments over several weeks until the target dose is achieved.

Defining the optimal replacement dose of levothyroxine in TSH deficiency is problematical, and little scientific evidence is available to guide the clinician. Unlike primary hypothyroidism, where serum TSH is a sensitive marker of under- or over-replacement, there is no biochemical marker to indicate precise physiological levels of replacement for an individual patient—indeed serum TSH may be low, normal, or even slightly elevated in untreated TSH deficiency. Therefore, adjustment is based on the clinical response and on measurement of circulating thyroid hormone levels, which are limited by the very wide reference ranges in the normal population. Serum free T₄ appears the most appropriate marker with which to adjust the levothyroxine dose and a conventional recommendation is to maintain free T₄ in the middle or upper part of the reference range for normal individuals, although this is certainly not ‘evidence based’ and begs the question of the criteria used to define TSH deficiency since it implies that patients with pituitary disease (and indeed 50% of the normal population!) with a free T₄ in the lower half of the normal range might benefit from levothyroxine replacement therapy. Some workers have advocated using a levothyroxine dose based on body weight (1.6 µg/kg) (21) but in doing so increased free T₄ levels close to the upper limit of the reference range.

We have recently audited free T₄ levels in over 340 patients with pituitary disease at risk of TSH deficiency in our clinic (defined as evidence of macroadenoma and/or pituitary surgery, and/or pituitary radiotherapy) and compared them with those in 1800 patients with primary thyroid disease being monitored via our ‘thyroid shared-care’ register. Over 95% of pituitary patients had a free T₄ within the reference range at latest follow-up and 38% of patients were taking levothyroxine treatment. In contrast, using samples in patients with primary thyroid disease with a serum TSH in the laboratory normal range (and therefore assumed euthyroid) as controls, serum free T₄ was below the 10th centile of controls on no treatment in 17% of pituitary patients who were not on levothyroxine and below the 10th centile of controls on levothyroxine in 39% of pituitary patients on thyroid replacement (note that controls on levothyroxine treatment with a normal TSH had considerably higher free T₄ than controls who were not) (22) (Fig. 2.3.6.1). This audit suggests that TSH deficiency may be substantially underdiagnosed and undertreated in routine clinical practice. The 20–80th centile range for controls on levothyroxine was a free T₄ level of 14–19 pmol/l in our laboratory and we propose to use this as target range for replacement levels in TSH deficiency in the future.

Ultimately, such controversies regarding appropriate levothyroxine replacement dosage and monitoring can only be answered by a controlled trial. In the meantime, I accept a serum free T₄ anywhere in the middle centiles of the normal range when the patient is asymptomatic, but I will push the free T₄ into the upper part of the reference range if the patient continues to have symptoms suggestive of hypothyroidism.

There are no specific data on adverse effects of excessive thyroid replacement in hypopituitarism, but data in patients with primary hypothyroidism are reassuring. Although thyrotoxicosis is a well-documented risk factor for osteoporosis, bone density remains normal even in patients on deliberate supraphysiological replacement with levothyroxine (23). However, thyrotoxicosis is also a risk factor for cardiovascular disease and, although there is no direct evidence of such adverse effects of levothyroxine replacement, the association of suppressed TSH levels and risk of atrial fibrillation in older patients in population-based studies (24) indicates the need for caution to avoid unnecessary over-replacement.

Other than changes induced by the restoration of the euthyroid state and normal metabolic rate, levothyroxine has few interactions with other therapy. A variety of other drugs raise (e.g. oral oestrogen therapy) or lower (e.g. anticonvulsant therapy) the levels of total T₄ in serum by altering binding to Thyroid-Binding Globulin (TBG) which could lead to inappropriate dose adjustment, but free T₄ is rarely affected in reliable assays. Concomitant growth hormone replacement causes a fall in total and free T₄ (usually with a rise in T₃ levels due to increased extrathyroidal conversion); levels usually remain within the normal range but dose adjustment may be required.

Simultaneous use of proton pump inhibitors or antacids, calcium carbonate, and ferrous sulfate may reduce absorption of levothyroxine and result in altered dose requirements.

Unlike adrenal and thyroid replacement, replacement of gonadotropin deficiency offers an extensive choice of gonadal steroid replacement therapy for both sexes, and the choice of gonadotropin or gonadotropin-releasing hormone (GnRH) therapy if and when fertility is desired. Direct, oral replacement of the missing gonadal steroids is not possible due to rapid first-pass liver metabolism and short half-life. Again, in most cases the choice between different treatment modalities is largely a matter of patient and physician preference rather than ‘evidence based’.

All female patients of premenopausal age with gonadotropin deficiency require oestrogen replacement to avoid the long-term consequences of oestrogen deficiency; cyclical progestagen is also essential in patients with an intact uterus to avoid endometrial hyperplasia and neoplasia.

Choice of oral replacement is extensive, including all forms of oral oestrogen marketed for postmenopausal oestrogen deficiency and all combined oestrogen–progesterone contraceptive pills. Almost any such preparation represents acceptable replacement, and there is little or no objective evidence to choose between regimens. In younger women the low dose (20–35 μg ethinylestradiol) oestrogen-progesterone pill is often preferred since there are extensive data on its safety in long-term use in women of this age, due to its lower cost, and since in some cases taking the ‘pill’ may feel more ‘normal’ psychologically than taking ‘HRT’. Monitoring of blood levels is impossible or unreliable with most preparations, but all give adequate levels of oestrogen to avoid effects of deficiency, and in an individual patient a regular menstrual withdrawal bleed is considered an adequate bioassay of oestrogen effect. When future pregnancy is planned it may be appropriate to monitor uterine size and endometrial thickness by ultrasonography to ensure that the uterus is indeed adequately oestrogenized.

A variety of transdermal oestradiol delivery systems are available—some of which also deliver transdermal progestagen. Gel preparations are available but have not been widely used. Such regimens have the advantage of achieving lower physiological levels of natural oestradiol, but the disadvantage of skin reactions and unsatisfactory skin adherence in some patients, and of increased cost in all. This form of oestrogen replacement has advantages as first line in patients with complex pituitary disease since it avoids the effects of oral oestrogen on other hormone-binding proteins and shows less interaction with growth hormone replacement levels (see below). Transdermal oestrogen may also be particularly appropriate in patients where direct hepatic effects need to be avoided (e.g. in rare patients who develop hepatic adenomas on oral oestrogen, or in young patients with known thrombophilia or previous thromboembolic disease).

Oestradiol implants achieve high circulating levels of oestrogen over long periods, but the problems with tachyphylaxis, which are well established in routine postmenopausal oestrogen replacement, make this form of replacement suboptimal.

In recent years large studies of hormone replacement in postmenopausal women have suggested that, although HRT certainly reduces the risk of osteoporosis and probably bowel cancer, the increased risk of thromboembolism and cardiovascular disease at postmenopausal age means that the overall effect on health may be negative. These conclusions do not, however apply to younger women with oestrogen deficiency, including those with gonadotropin deficiency, since the risk of osteoporosis is greater and the risk of cardiovascular events lower due to younger age; indeed it is still hoped that oestrogen replacement should reduce the risk of premature vascular disease. Resolution of local and systemic symptoms of oestrogen deficiency provides additional drivers for routine oestrogen replacement in all gonadotropin-deficient women of premenopausal age. Balanced against these positive effects is a slight increase in risk of venous thromboembolism compared with the deficient state (which nevertheless remains a very low absolute risk), and mixed evidence on a possible slight increase in risk of breast cancer. In addition, some women will suffer from unwelcome cyclical changes similar to those which may be experienced the normal cycle. In the younger women the balance of risks seems overwhelmingly in favour of routine oestrogen replacement. In women of postmenopausal age the choice of whether or not to take oestrogen replacement is ultimately no different from that in women without pituitary disease.

Oral oestrogen raises plasma total T₄ and cortisol levels by inducing increases in thyroid- and cortisol-binding globulin, respectively. This does not influence active levels of these hormones or change the necessary dose of replacement, but does make biochemical monitoring of thyroid and adrenal reserve and/or replacement doses of these hormones more difficult. It is therefore usually best to fully assess thyroid and adrenal axes and/or optimize replacement levels before oral oestrogen is started (or during a 2–3-month break in therapy), or to use transdermal oestradiol as an alternative. Oral oestrogen also causes a slight reduction in insulin-like growth factor 1 (IGF-1) and rise in growth hormone levels, which may require dose adjustment to concomitant growth hormone replacement therapy to maintain IGF-1 in the target range.

Men with gonadotropin deficiency usually require androgen replacement not only for relief of the symptoms of hypogonadism but also to prevent the long-term consequences of gonadotropin deficiency—particularly osteoporosis. More recent evidence that low testosterone levels may be associated with increased risk of cardiovascular disease and which may be reduced by replacement (25), provides an additional reason to advocate replacement even if the patient is apparently asymptomatic.

In recent years a wide range of forms of testosterone replacement have become available (26).

Depot testosterone ester preparations were the traditional form of androgen replacement, typically given as testosterone enantate, 250 mg IM every 3 weeks (range 2–4 weeks) or mixed testosterone esters (Sustanon) IM every 3 weeks. These regimens are certainly the cheapest preparations but do not mimic normal physiology, resulting in high or high normal levels of plasma testosterone in the days immediately following injection falling to low normal or subnormal levels before the next dose. Most patients will notice some changes in mood, libido, muscle strength, or general ‘drive’ consistent with these hormonal fluctuations, and some find this troublesome. The need for IM injections also usually necessitates regular visits to the surgery which may be inconvenient, although occasionally patients may self-administer.

These disadvantages mean that newer forms of testosterone replacement, which have become available in recent years have become the preferred methods of replacement in many countries. A depot preparation of testosterone undecanoate 1000 mg administered every 3 months (Nebido) is now a popular alternative that gives much more stable and physiological testosterone levels, achieved more rapidly if the second injection is given after 6 weeks (27). Timing of doses can be adjusted by measuring trough levels of testosterone prior to injection. Long-term experience of this formulation is still awaited.

Testosterone undecanoate is a 17α-hydroxyl ester which is active orally since its highly polar side chain and oily vehicle allows direct absorption into the lymphatic system bypassing hepatic metabolism. The half-life remains relatively short, and multiple daily doses are necessary—typically 40 mg, 2–3 times daily—but oral therapy is preferred by some patients. This preparation is not currently approved for use in the USA but is available widely elsewhere. Other oral formulations of testosterone are either ineffective or have an unacceptable incidence of hepatic side effects and are no longer used.

Testosterone pellets can be implanted subcutaneously by trocar and provide stable testosterone levels, typically using testosterone 600 mg every 6 months. The need for repeated surgical procedures means that this form of replacement is now largely superseded by long-acting depots.

Transdermal gel preparations (typically 50 mg in 5 ml of 1% gel) have become increasingly popular in recent years. They are available in a variety of formulations including sachet, tube, and pump dispenser—all of which have their advantages and disadvantages in terms of convenience and ability to adjust doses. In each case the gel is applied by the patient to the shoulders, arms, or abdomen after bathing or showering, typically first thing in the morning. Contact of gel, or gel-treated areas, with females or children needs to be avoided. This form of treatment is very convenient for many patients but can be problematic for individuals who need to wash or shower frequently. Transdermal patch preparations are still available, but their niche has been largely overtaken by gel preparations. The patch may cause troublesome skin reactions in 5–10% of patients. All forms of transdermal therapy achieve physiological peak levels of testosterone and can even mimic the normal testosterone circadian rhythm.

Testosterone levels can be measured in blood on all forms of replacement. With conventional intramuscular depot, I advocate measuring levels 1 week after injection (which should be in the upper part of the reference range for normal individuals) and just prior to injection (which will usually be low normal); frankly subnormal levels prior to injection may indicate the need for more frequent injection, while very high peak levels may necessitate a reduction in dose. On the 3-monthly depot a preinjection level is most informative. Testosterone levels in the morning on transdermal preparations should be in the mid-reference range. Random testosterone measurements are often low or low normal on oral testosterone undecanoate, but the short half-life makes these measurements less reliable, and dihydrotestosterone levels may be preferentially raised.

Severe androgen deficiency causes a reduction in libido and potency, which are usually restored by appropriate replacement therapy. In patients who have been hypogonadal for many years, counselling of the patient and their partner will be necessary before starting therapy. However, a proportion of men presenting with sexual dysfunction will also be found to have testosterone levels in the borderline low range, without elevation of gonadotropins, but with no other evidence of pituitary disease; improvement in libido and particularly potency is far less certain with replacement in these patients.

Excess androgen replacement will cause polycythaemia, and this can also occur in elderly men, in chronic obstructive airways disease and in sleep apnoea with high-normal replacement doses, so that haemoglobin levels need to be monitored in such patients. Sleep apnoea may also be worsened by testosterone replacement. The possibility of long-term adverse effects on the prostate remains controversial. Hypogonadal men have lower prostate volume and prostate-specific antigen levels than the normal population, and induction of hypogonadism in patients with established prostate cancer induces temporary disease remission and reduction in prostate volume—thus raising the possibility that testosterone replacement might increase the incidence of prostate cancer and prostatic hypertrophy. While it seems likely that replacement will raise the incidence of these conditions to those of the normal population there is currently no evidence to suggest an absolute increased risk of either condition with appropriate replacement. In spite of this, and the current lack of evidence that screening for prostate cancer is beneficial in the normal population, some form of prostate monitoring is usually recommended in older men on testosterone replacement; I favour monitoring serum prostate-specific antigen (PSA), with further investigation only if PSA levels rise above normal, but others recommend repeated rectal examination and/or prostate imaging.

Previous concerns about the possibility of increased cardiovascular risk arose from the increased risk of cardiovascular disease in men compared with women—but evidence now suggests that low testosterone itself may be associated with the metabolic syndrome and increased cardiovascular risk and that testosterone replacement may therefore be beneficial to cardiovascular health (25). Long-term clinical trial evidence is still awaited.

Where gonadotropin deficiency develops before puberty, special care is required to induce pubertal developmental at an appropriate speed in both sexes. Simply commencing full adult replacement doses will result in inappropriately rapid pubertal development with insufficient time for usual psychological adaptation, in less satisfactory secondary sexual development and in possible attenuation of the pubertal growth spurt and final height.

Gonadal steroid replacement is commenced in low doses in both sexes. In females, low doses of oral oestrogen (e.g. ethinyloestradiol 10 μg daily, or even lower doses where available) or transdermal oestradiol (e.g. 25 μg or less, twice weekly) is commenced, and usually continued at very low dose for 6–12 months before increasing steadily to full replacement dosage with eventual addition of cyclical progestagen. Males may commence with testosterone enantate 100 mg IM every 3–4 weeks, again continued at this low dose for many months before increasing towards adult replacement dosage; similar low-dose regimens can be devised using transdermal testosterone or oral testosterone undecanoate. Other factors sometimes require attention: girls with prepubertal panhypopituitarism may fail to develop pubic hair, and use of topical testosterone creams has been described if normal hair growth is desired; boys may need reassurance that pubertal gynaecomastia is common, and that normal adult facial and body hair develop slowly over many years once testosterone levels have reached the normal adult range.

Normal individuals show a pubertal growth spurt associated with gonadal steroids, particularly in males, yet gonadal steroids are also responsible for epiphyseal fusion and the cessation of linear growth. Appropriate adjustment of replacement doses of gonadal steroid and growth hormone are therefore essential in patients with hypopituitarism during induction of puberty. Overall, studies indicate that although early puberty (spontaneous or induced) reduces final height, deliberately delaying puberty induction to allow increased time for growth probably does not increase final height significantly, and may be associated with obvious adverse psychological consequences.

Gonadal steroid replacement does not induce fertility, but ovulation and spermatogenesis can be stimulated by therapy with gonadotropin injections (initially human menopausal, now increasingly recombinant) or by pulsatile GnRH administered subcutaneously by infusion pump.

Use of gonadotropin therapy and pulsatile GnRH are both highly successful, resulting in fertility rates approximating normal levels with repeated cycles of treatment in expert hands (28). GnRH therapy is mostly used where the gonadotropin deficiency is considered primarily ‘hypothalamic’, but may also be successful in patients considered to have primary pituitary disease. As always, such therapy should only be undertaken with close biochemical and ultrasound monitoring of the ovarian response, and in centres with extensive experience of ovarian stimulation techniques; precise description of treatment regimens is beyond the scope of this book.

Gonadotropin and GnRH therapy can both induce spermatogenesis, but induction of adequate spermatogenesis takes a minimum of 3 months and may require 1–2 years. Luteinizing hormone therapy is usually used first in the form of human chorionic gonadotropin (hCG) (typically 1000–2000 IU, 2–3 times/week); this should result in adequate testosterone levels and may sometimes be sufficient to allow spermatogenesis, but follicle-stimulating hormone activity is usually required for adequate fertility. A wide variety of regimens have been recommended with successful fertility in a majority of patients which may occur at surprisingly low total sperm counts. Coexistent primary testicular defects (e.g. related to cryptorchidism) may cause failure. The need to wear an infusion pump or attend for regular intramuscular injections over many months is clearly a disadvantage, but self-administration of low doses of gonadotropins subcutaneously may also be successful in both induction of fertility and increase in testicular size (29). When pregnancy is achieved, spermatogenesis may occasionally be maintained by testosterone replacement alone although usually continued or repeated gonadotropin therapy is required. Sperm may also be frozen after successful treatment for use in future attempts at fertility.

Pituitary coma is a rare, but life-threatening, presentation of severe, longstanding, untreated hypopituitarism, usually precipitated by stress including infection, trauma, surgery, or infarction, or by an acute pituitary insult such as apoplexy. Treatment is firstly full replacement with parenteral hydrocortisone (as above), followed by correction of other factors which may precipitate or worsen coma, including hypothermia, salt and water depletion due to hypoadrenalism, and/or diabetes insipidus, hyponatraemia due to excessive desmopressin or hypothalamic dysfunction, and (slow) replacement of hypothyroidism.

Pituitary disease and the investigation and replacement of pituitary deficiencies are complex and optimal management is only possible when the patient is fully aware of the nature and consequences of the disorder and can participate actively in the adjustment of replacement therapy. Patient education is therefore an essential part of management, and patients should also be encouraged to obtain further information and support from groups such as the Pituitary Foundation (UK) (www.pituitary.org.uk) or Pituitary Network Association (USA) (www.pituitary.org).

Replacement therapy may be very complex in patients with panhypopituitarism and patients must be aware of the nature, indications, and administration schedule for all replacement therapies that they require. For those on glucocorticoid replacement it is essential that they understand the need to increase doses during intercurrent illness (see above), and carry a steroid card and MedicAlert/SOS bracelet or pendant. When injection or infusion pump therapy is indicated patients must be fully conversant with the techniques. Gonadal steroid replacement is often neglected by both physicians and patients, and patients should be fully aware of the benefits and risks of replacement. Patients in the U.K should also be made aware than hypopituitarism is currently an indication for exemption from National Health Service (NHS) prescription charges.

The biochemical assessment of pituitary deficiency remains regrettably imprecise and many patients will have evidence of equivocal or borderline deficiency. Other patients will have normal tests but have a pituitary disorder in which progressive hypopituitarism is possible or likely (e.g. following pituitary radiotherapy). All such patients must be fully aware of the symptoms and consequences of hypopituitarism (particularly hypoadrenalism which might be life-threatening), should participate actively in the decision to commence or withhold replacement therapy, and report promptly the development of new symptoms that might indicate worsening deficiency. Asymptomatic patients with borderline or equivocal ACTH deficiency who are not on regular glucocorticoid replacement should have a supply of hydrocortisone for intercurrent illness with clear information about when and how it should be used.