2.3.10 Prolactinomas and hyperprolactinaemia (including macroprolactinaemia)

Prolactin promotes milk production in mammals. It was characterized as a hormone distinct from growth hormone, which also has lactogenic activity, as recently as 1971. In humans, the predominant prolactin species is a 23 kDa, 199 amino acid polypeptide synthesized and secreted by lactotroph cells in the anterior pituitary gland. Prolactin is produced also by other tissues including decidua, breast, T lymphocytes, and several regions of the brain, where its functions are largely unknown and its gene regulation different from that of the pituitary gene. Pituitary prolactin production is under tonic inhibitory control by hypothalamic dopamine, such that pituitary stalk interruption produces hyperprolactinaemia. The neuropeptides thyrotrophin-releasing hormone (TRH) and vasoactive intestinal peptide (VIP) exert less important stimulatory effects on pituitary prolactin release (1).

Following the discovery of prolactin as a separate hormone it became apparent that many apparently functionless ‘chromophobe’ pituitary adenomas were prolactinomas. Indeed, prolactinoma is the commonest type of functioning pituitary tumour diagnosed in humans. There is a marked female preponderance and prolactinoma is relatively rare in men. Several studies have revealed small prolactinomas in approximately 5% of autopsy pituitaries, most of which are undiagnosed during life. From a clinical standpoint, prolactinomas are divided arbitrarily into microprolactinomas (≤10 mm in diameter) and macroprolactinomas (>10 mm). This is a useful distinction which predicts tumour behaviour and indicates appropriate management strategies. Generally, microprolactinomas run a benign course. Some regress spontaneously, most stay unchanged over many years, and very few expand to cause local pressure effects. In contrast, macroprolactinomas may present with pressure symptoms, often increase in size if untreated and rarely disappear. Some clinicians find an intermediate category of meso-prolactinoma useful (10–20 mm in diameter), since this tumour group may have a more favourable treatment outcome than for larger macroprolactinomas.

Prolactinomas are usually sporadic tumours. Molecular genetics has shown nearly all to be monoclonal, suggesting that an intrinsic pituitary defect is likely to be responsible for pituitary tumorigenesis (see Chapter 2.3.2). Occasionally, prolactinoma may be part of a multiple endocrine neoplasia syndrome type I, but this occurs too infrequently to justify screening in every patient with a prolactinoma. Mixed growth hormone and prolactin-secreting tumours are well recognized and give rise to acromegaly in association with hyperprolactinaemia. Most contain separate growth hormone and prolactin-secreting cells whereas a minority secrete growth hormone and prolactin from a single population of cells, the mammosomatotroph adenomas. Prolactin-secreting adenomas may produce other hormones such as thyroid-stimulating hormone (TSH) or adrenocorticotropic hormone (ACTH), but such tumours are uncommon. Malignant prolactinomas are also very rare. A few cases have been described which have proved resistant to aggressive treatment with surgery, radiotherapy, and dopamine agonists. In a small proportion, extracranial metastases in liver, lungs, bone, and lymph nodes have been documented. The alkylating agent temozolomide is effective against some aggressive prolactinomas (2).

The clinical features of prolactinoma are attributable to three main factors: hyperprolactinaemia, space occupation by the tumour, and varying degrees of hypopituitarism (Box 2.3.10.1). The individual clinical picture will be determined by the sex and age of the patient, and the tumour size. In brief, hyperprolactinaemia stimulates milk production, particularly from the oestrogen-primed breast, and inhibits hypothalamic gonadotropin-releasing hormone release, which leads to hypogonadotropic hypogonadism.

Premenopausal women, most of whom will have microprolactinomas, usually present with oligomenorrhoea or amenorrhoea (90%) and/or galactorrhoea (up to 80%). Anovulatory infertility is common. Excluding pregnancy, hyperprolactinaemia accounts for 10–20% cases of secondary amenorrhoea. It should be noted that most women with galactorrhoea do not have menstrual disturbance, hyperprolactinaemia, or a pituitary tumour. Postmenopausal women are, by definition, already hypogonadal and markedly hypo-oestrogenaemic. Hyperprolactinaemia in this age group does not present with classic symptoms and a prolactinoma may be recognized only when it grows large enough to produce headache and/or visual disturbance.

The presentation of hyperprolactinaemia in men is with reduced libido, impotence (75%), and infertility associated with a reduced sperm count. Such symptoms are quite often concealed or ignored, particularly by older men, so men tend to present later with larger tumours causing pressure symptoms. Galactorrhoea is very uncommon in men but does occur occasionally. Weight gain is noted frequently by hyperprolactinaemic men.

Long-term hyperprolactinaemic hypogonadism may reduce bone mineral density (BMD) in either sex and is an important cause of secondary osteoporosis. Prolactinoma is an unusual cause of delayed puberty in both sexes and some advocate the routine measurement of serum prolactin in this situation.

The causes of hyperprolactinaemia can be divided simply into physiological, pharmacological, and pathological (Box 2.3.10.2). The normal prolactin range for nonpregnant women is below 500 mU/l (20 μg/l) and below 300 mU/l (12 μg/l) for men. The finding of mild hyperprolactinaemia should always be rechecked in a second blood sample to exclude possible venepuncture stress elevation, although the importance of this effect has probably been overemphasized. Pregnancy is the commonest cause of hyperprolactinaemic amenorrhoea and serum prolactin concentrations may rise as high as 8000 mU/l (320 μg/l) during the third trimester. Normal lactation is also associated with quite marked elevation of serum prolactin. As predicted from the physiological dopaminergic inhibition of prolactin secretion, treatment with dopamine receptor antagonist drugs commonly induces hyperprolactinaemia. Serum prolactin levels may rise as high as 5000 mU/l (200 μg/l). This is a particular problem with the major tranquillizers (for example, chlorpromazine) and antiemetics (such as metoclopramide) (3). It is a lesser problem with newer atypical antipsychotics such as quetiapine and olanzapine (4). There is potential for confusion if a patient does not reveal that he or she is taking an ‘over-the-counter’ preparation, such as a combined medication for the treatment of migraine which contains both an analgesic and an antiemetic. Similarly, some nonprescribed herbal or alternative remedies contain ingredients that cause prolactin elevation. Thus, a comprehensive drug history is essential. With regard to pathological causes of hyperprolactinaemia, it is important to exclude primary hypothyroidism. Modest hyperprolactinaemia is present in 40% of patients, although only 10% have levels above 600 mU/l (24 μg/l). Nevertheless, some hypothyroid young women may present with menstrual disturbance and galactorrhoea, together with very few ‘typical’ hypothyroid symptoms. Once venepuncture stress, pregnancy, interfering drugs, and primary hypothyroidism have been excluded, significant hyperprolactinaemia is usually associated with a pituitary adenoma (Box 2.3.10.2).

Prolactin in human serum exists in multiple molecular forms, with three dominant species identified by gel filtration chromatography: monomeric prolactin (23 kDa), big prolactin (50–60 kDa), and big-big prolactin (macro-prolactin, 150–170 kDa). Macroprolactin is a complex of prolactin with an IgG antibody which is detected to a greater or lesser extent by all prolactin immunoassays (5). This prolactin species is present in significant amounts in ∼25% of hyperprolactinaemic sera (Fig. 2.3.10.1) (5,6) and ∼1% of the normal population. In vivo, macroprolactin has little prolactin bioactivity and many patients with macroprolactinaemia do not have typical hyperprolactinaemic symptoms (7). Macroprolactinaemia is not associated with macroprolactinoma.

Failure to recognize that hyperprolactinaemia may be due to macroprolactinaemia may lead to unnecessary investigation, incorrect diagnosis, and inappropriate management in patients presenting with common symptoms suggesting possible prolactin excess, such as amenorrhoea or impotence. The presence of macroprolactin can be confirmed by polyethylene glycol (PEG) precipitation and most UK biochemistry laboratories screen hyperprolactinaemic sera using this simple method (8). PEG precipitates high-molecular-weight compounds, including immunoglobulins, and repeat assay of the treated serum gives the residual monomeric prolactin concentration. At the present state of knowledge, there is no justification for detailed pituitary investigation or long-term follow-up of an individual shown to have macroprolactinaemia.

If serum prolactin concentrations are extremely high (as in some men with giant prolactinomas), the amount of prolactin antigen may cause antibody saturation in prolactin immunoradiometric assays (IRMAs), leading to artefactually low prolactin results. This is known as the high-dose ‘hook effect’ and has been recognized some time in other immunoassays (e.g. B-human chorionic gonadotrophin, hCG). This artefact may lead to misdiagnosis and inappropriate surgery for some patients with macroprolactinoma. Serum prolactin should always be assayed in dilution in any patient with a large pituitary lesion which might be a prolactinoma (9).

A number of dynamic tests have been proposed for the evaluation of hyperprolactinaemia but few UK clinical endocrinologists routinely use dynamic prolactin function tests. In my experience, the intravenous administration of a dopamine antagonist (such as 10 mg domperidone or metoclopramide) is a simple, well-tolerated procedure which provides clinically useful information, particularly for patients with modest elevations of serum prolactin. As illustrated in Fig. 2.3.10.2, dopamine antagonist administration to normal individuals results in a marked rise in serum prolactin concentration (to at least three times basal) together with little or no change in serum TSH (less than 2 mU/l rise). In contrast, patients with pituitary micro- and macro-lesions have blunted prolactin responses. Patients with microprolactinomas may, in addition, show exaggerated TSH responses due to enhanced dopaminergic tone on the anterior pituitary thyrotrophs (via short-loop hypothalamic feedback).

Sawers and co-workers reviewed 84 hyperprolactinaemic patients whose investigation had included a domperidone test and high-resolution MRI (10). Eighteen of 20 patients with normal prolactin responses to domperidone had normal MR scans and the other two had only microadenomas, possibly incidentalomas. In contrast, 18 of the remaining 64 patients with abnormal prolactin responses had lesions greater than 10 mm diameter. Of the remainder, 63% had microadenomas. Dopamine antagonist testing can therefore identify a subset of hyperprolactinaemic patients for whom detailed pituitary imaging is mandatory. Conversely, a normal prolactin response to domperidone identifies those who do not require pituitary imaging.

Dopamine antagonist testing can also be informative before and after surgery for microprolactinoma. Webster and colleagues described a series of 82 hyperprolactinaemic patients who underwent surgery for suspected prolactinoma (11). No tumour was found in three cases, including the only two patients with normal prolactin and TSH responses to domperidone. Overall, 79% of patients had early postoperative normalization of serum prolactin but there were three relapses during long-term follow-up. Two of these had persistently abnormal prolactin and TSH responses to domperidone, even when basal prolactin levels remained normal.

Thus, although few patients with microprolactinoma are now treated surgically, these data are important because they indicate that dopamine antagonist testing can confirm (or refute) the presence of a microprolactinoma with reasonable certainty. Clinicians may regard this confirmatory biochemical evidence to be helpful in the medical management of such patients when histological proof of the diagnosis is not forthcoming and MRI may be negative.TRH testing is less discriminatory and generally unhelpful in the investigation of hyperprolactinaemia. However, the test may have limited use in the evaluation of patients with growth hormone or gonadotropin-secreting tumours, a proportion of whom will show paradoxical stimulation of hormone release.

Pituitary imaging is best performed using MRI with gadolinium enhancement. Compared with high-resolution CT, this technique provides superior detail of the optic chiasm, suprasellar masses, and cavernous sinus invasion. It does not involve the use of ionizing radiation and has a limit of resolution of approximately 2 mm. With MRI, the majority of microadenomas appear as focal hypodense lesions within the pituitary on T₁-weighted images (Fig. 2.3.10.3a). It should be noted that microadenomas are present in a significant proportion of the normal population and small ‘incidentalomas’ may be revealed by high-resolution MRI in up to 20% of healthy subjects. Conversely, a normal MRI examination does not exclude a microadenoma. Macroadenomas have a variety of appearances but are usually obvious on MRI (Fig. 2.3.10.3b). Imaging provides no information on tumour function or pathology, and macroprolactinoma, nonfunctioning macroadenoma, and craniopharyngioma may have identical appearances on MRI.

Most patients with microprolactinomas have basal serum prolactin concentrations less than 5000 mU/l (200 μg/l). In patients with pituitary macrolesions, the basal serum prolactin is of considerable diagnostic value. A value greater than 5000 mU/l is virtually diagnostic of a macroprolactinoma and with a level greater than 10 000 mU/l there is no other possible diagnosis. A serum prolactin concentration lower than 2000 mU/l in a patient with a pituitary macrolesion usually indicates ‘disconnection’ hyperprolactinaemia rather than tumoural secretion of the hormone. This is due most commonly to a nonfunctioning pituitary macroadenoma, although intrasellar craniopharyngiomas and numerous other neoplastic and inflammatory pathologies may masquerade as ‘pseudo-pituitary adenomas’ (12).

An intermediate serum prolactin level (2000–5000 mU/l) in a patient with a large pituitary lesion produces an area of some diagnostic uncertainty which dynamic prolactin function tests cannot resolve. Most of these patients will have true prolactinomas and the remainder ‘disconnection’ hyperprolactinaemia (13). This has implications for management strategies as discussed below in the section on macroprolactinoma. Figure 2.3.10.4 shows the range of serum prolactin levels in a group of patients with pituitary macroadenomas who underwent surgery and thus provided corroborative immunohistochemistry (12).

There is usually about 10 mm between the top of the normal pituitary and the optic chiasm. All patients with pituitary macrolesions and suprasellar extension should therefore undergo specialist ophthalmological assessment, including Goldman perimetry. Since there is great variation in the pattern of suprasellar tumour growth (and the position of the chiasm may also vary) the pattern of visual impairment may range from the classic bitemporal hemianopia to partial quadrantic defects or scotomas. No pattern of visual loss is specific to prolactinoma, compared with other tumour types.

Larger pituitary masses may cause hypopituitarism either by direct pituitary compression or by disruption of hypothalamic control mechanisms. Patients with microprolactinomas usually have normal growth hormone, ACTH, and TSH function. However, with macroprolactinomas the degree of hypopituitarism is likely to be proportional to the size of the tumour. With the largest tumours, ACTH and TSH deficits may be present at diagnosis in approximately 20% of patients and growth hormone deficiency is almost invariable. All patients with macroprolactinomas should have full pituitary function testing.

Most patients with prolactinoma require treatment. Infertility, menstrual disturbance with longstanding hypogonadism (risk of secondary osteoporosis), troublesome galactorrhoea, an enlarging pituitary tumour, and tumour pressure effects (particularly visual failure) are all indications for treatment. Dopamine agonist drugs are now indicated as primary medical therapy for patients with prolactinomas of all sizes. However, an important exception is the patient with a pituitary macrolesion and minor elevation of prolactin, who is most likely to have a nonfunctioning pituitary adenoma requiring surgery for decompression and histological diagnosis. It may be reasonable to simply observe some patients with microprolactinomas, particularly if circulating sex steroid concentrations are judged to be adequate.

The introduction of medical therapy with dopamine agonists revolutionized the treatment of patients with prolactinoma. The first dopamine agonist was bromocriptine, a semisynthetic ergopeptine derivative, introduced in 1971. On a global basis, this probably remains the most commonly used dopamine agonist, but other longer-acting and better tolerated drugs, such as cabergoline and quinagolide, are now widely available. Most endocrinologists use cabergoline as first-choice dopamine agonist; a large comparative study with bromocriptine convincingly demonstrated its superiority in terms of tolerability, patient convenience and efficacy (14). Similar data favour the use of quinagolide over bromocriptine (15). Pregnancy is a special situation, since there are many more safety data for bromocriptine than for the newer agents. Many endocrinologists select bromocriptine as first-line treatment for hyperprolactinaemic infertility.

Bromocriptine is used in a dose of 2.5 mg twice or thrice daily. The doses of 20–40 mg per day used in early studies are no more efficacious and produce more side-effects. Cabergoline is usually effective in a dose of 0.5–1.0 mg once or twice weekly and quinagolide in a once-daily dose of 75–150 μg. In order to minimize side-effects, patients can be advised to take these two drugs, together with a supper snack, just before retiring to bed.

All dopamine agonists may produce unwanted side effects including, in decreasing order of importance, upper gastrointestinal disturbance (especially nausea), postural hypotension, constipation, nasal stuffiness and Raynaud’s phenomenon (Box 2.3.10.3). These can be minimized by using an incremental dosage schedule and taking tablets with food. In a double-blind comparison of bromocriptine and cabergoline, 12% of patients stopped bromocriptine because of intolerance whereas only 3% stopped cabergoline (14).

Acute psychotic reactions have been described with quinagolide, albeit rarely. It is unclear whether this important side effect is drug-specific since acute psychosis was encountered occasionally in earlier patients treated with large bromocriptine doses. Sleep and mild mood disturbances can occur with all the dopamine agonists (Box 2.3.10.3).

Recent studies of parkinsonian patients treated with dopamine agonists revealed restrictive valvular heart disease in about one third of patients taking pergolide. The valvulopathy was mostly mild, but correlated with the cumulative dose, and a similar effect was demonstrated with cabergoline. However, much higher doses of dopamine agonist are used in Parkinson’s disease (typically 20–30 times the dose used to treat prolactinoma), patients tend to be older (perhaps with altered cardiac susceptibility) and a large cumulative dose is attained more quickly than in prolactinoma patients. Nevertheless, European medicines regulatory agencies issued a drug alert in late 2008 related to potential valvulopathy in endocrine patients on ergot-derived dopamine agonists (16–18). Of seven published studies of endocrine patients treated with cabergoline, with appropriate age and sex matched controls, only one showed a significantly increased risk of valve disease. This study also showed rates of moderate tricuspid regurgitation in controls that were sixfold greater those in the other studies, suggesting the investigators may have used more stringent echocardiographic criteria. The results from the controlled studies show valvular abnormalities in 50 of 450 (11%) of endocrine patients taking cabergoline compared with 33 of 416 (8%) for controls (p = 0.13). If uncontrolled data are added for the patients on cabergoline, the percentage with valvular abnormalities falls to 8% (61/645). Overall, the cardiac risks associated with low-dose cabergoline seem to be low but further studies are required for reassurance. The need for echocardiographic surveillance in endocrine patients remains unproven. At the present state of knowledge, it would seem reasonable to focus on patients taking more than 2 mg cabergoline per week but there are no data to inform the best screening protocol.

Medical therapy is remarkably effective in the treatment of microprolactinoma (19–21). In early studies of patients treated with bromocriptine, normoprolactinaemia or ovulatory cycles were restored in 80–90% of patients. Fertility returned within two months in 70% of women. Galactorrhoea disappeared or was greatly reduced in the majority, usually within a few days or weeks. In comparative studies of cabergoline and bromocriptine, resumption of ovulatory cycles or occurrence of pregnancy was documented in 72% of cabergoline patients (up to 1.0 mg twice weekly) compared with 52% in the bromocriptine group (up to 5.0 mg twice daily) (14). The number of women with stable normoprolactinaemia was also higher in the cabergoline group (83% vs 58%). Bone mineral density (BMD) has been shown to increase during long-term dopamine agonist therapy, presumably in response to restoration of normal ovarian oestrogen secretion, although there are no prospective data on fracture reduction (19).

Tumour shrinkage occurs during long-term treatment, although this is less critical than for patients with macroprolactinomas. Importantly, a minority of patients may be ‘cured’ after a period of dopamine agonist treatment. The mechanism is unknown. The probability of ‘cure’ remains unclear but at least one-third of microprolactinomas seem to remit with time (22–23). A dopamine-agonist induced pregnancy may increase the chances of remission (24). For these reasons, most endocrinologists interrupt dopamine agonist treatment every 3 years, for further clinical assessment and prolactin testing. In doing so, one should remember that women may continue to have ovulatory cycles for 3–6 months after withdrawal of the long-acting drug cabergoline.

In some centres, transsphenoidal surgery may be offered as an alternative to medical therapy. Indeed, surgery may be essential if the patient is intolerant of or resistant to dopamine receptor agonists. Surgical success is critically dependent on surgical experience and the size of the tumour. In most large centres, normoprolactinaemia is achieved post operatively in 60–90% of patients, with results for larger microprolactinomas (4–9 mm diameter) being significantly better than for smaller ones. Previous dopamine agonist therapy may hamper surgery but this is less troublesome for micro- than it is for macroprolactinomas. Recurrence of hyperprolactinaemia, usually without radiologically evident tumour, is well recognized. Using normoprolactinaemia as the main criterion of cure, it is probably reasonable to speak of a long-term surgical cure rate of between 50% and 70% when counselling patients with respect to choice of therapy. It is important to mention the small but measurable morbidity of transsphenoidal surgery (discussed elsewhere in this volume), together with the small risk of loss of normal pituitary function. The latter would be particularly important if the patient wished fertility.

Due to the excellent therapeutic responses to either dopamine agonists or transsphenoidal surgery, radiotherapy is no longer considered acceptable primary therapy for microprolactinoma.

Longitudinal studies suggest only 5% of microprolactinomas progress to larger lesions. Hence, in a woman with a microprolactinoma who has normal menses and libido, non-troublesome galactorrhoea, and who does not wish to become pregnant, there may be no clear indication for antiprolactinoma therapy. Before recommending simple observation of a microprolactinoma, most endocrinologists would wish to confirm ‘adequate’ circulating sex steroid concentrations (mean oestradiol above 200 pmol/l in a woman and testosterone above 7 nmol/l in a man), together with BMD within 1 SD of age-related mean values. In this situation it would be reasonable to monitor the patient with 6–12 monthly serum prolactin and oestradiol/testosterone estimations, supplemented with bone densitometry every 3–5 years, thus enabling individualized timing of any intervention. The question of oral contraceptive safety often arises. There are good data confirming safety of the oral contraceptive in combination with a dopamine agonist in women with microprolactinomas but no satisfactory prospective studies of treatment with an oral contraceptive alone. If the latter course of action is taken, serum prolactin should be checked every 3–6 months, with the addition of dopamine agonist therapy should the serum prolactin level rise above an arbitrary target level (e.g. twice the basal level).

These drugs directly activate pituitary D2 dopamine receptors, mimicking the action of endogenous hypothalamic dopamine. In addition to reducing prolactin secretion, D2 receptor stimulation results in rapid involution of the cellular protein synthetic machinery and thus marked reduction in lactotroph cell size. This effect, together with an antimitotic action, accounts for the rapid and sustained tumour shrinkage which enables these drugs to be used as primary therapy for patients with larger prolactinomas, even those with pressure effects. Dopamine agonist treatment is followed typically by a rapid fall in serum prolactin (within hours) and tumour shrinkage (within days or weeks). Tumour regression is often followed by an improvement in visual function over a (short) time-course which rivals that seen after surgical decompression of the chiasm. Thus, patients with macroprolactinomas presenting with visual failure are no longer the neurosurgical emergencies they were previously regarded to be. Nevertheless, it is vitally important that all patients with a pituitary macrolesion producing chiasmal compression should have serum prolactin measured urgently (and checked in dilution–see section on prolactin immunoassay). Four illustrative patients are shown in Fig. 2.3.10.5, Fig. 2.3.10.6, Fig. 2.3.10.7, and Fig. 2.3.10.8.

A meta-analysis of 271 well-characterized macroprolactinomas treated with dopamine agonists showed that 79% of tumours shrank by more than a quarter, and 89% shrank to some degree. The pretreatment prolactin level is not a reliable predictor of tumour shrinkage, since 83% of tumours showed significant tumour shrinkage in both the ‘above 100 000 mU/l’ and ‘5000–10 000 mU/l’ groups. Of the macroprolactinomas large enough to produce chiasmal compression, 85% showed significant tumour shrinkage (25).

Tumour shrinkage can be demonstrated within a week or two of starting dopamine agonist therapy and most shrinkage takes place during the first three months of treatment (Fig. 2.3.10.9) (25–26). However, in many patients, shrinkage continues at a slower rate over many months (see tumours in Fig. 2.3.10.6 and Fig. 2.3.10.7). It is recommended to repeat MRI 3 months after commencing dopamine agonist therapy and, if there has been an acceptable response, again at 1 and 2 years.

Approximately 40% of macroprolactinomas treated with dopamine agonists for 1–3 months show tumour size reduction by at least one half. Of those treated for 1 year or longer, almost 90% show such shrinkage (25). Visual field defects improve in approximately 90% of patients in whom they were abnormal before treatment. Although early visual improvement occurs frequently, it may be several months before maximum benefit accrues. Thus, persistence of a visual field defect is not an absolute indication to proceed to surgery.

Suppression of serum prolactin usually accompanies successful tumour shrinkage. Indeed, all of the responsive patients in the meta-analysis showed a fall in serum prolactin of at least 50%, and in 58% of patients serum prolactin became entirely normal (25).

Several investigators have demonstrated recovery of impaired anterior pituitary function in association with tumour shrinkage. Importantly, these data have been extended to include recovery of growth hormone reserve, which may obviate the need for expensive growth hormone replacement in a proportion of these patients (27). In contrast, it is worth noting that at least two-thirds of men with successfully treated prolactinomas have persistently subnormal testosterone levels and require androgen supplementation (25). In premenopausal women with medically treated macroprolactinomas cyclical menses return in over 90%.

Overall, the acquisition of dopamine agonist resistance during therapy appears to be very rare, even with treatment periods of 10 or more years (28). A handful of cases have, however, been described (25). Primary resistance to cabergoline occurs in fewer than 10% of patients (19) but most patients will normalize prolactin if the drug is tolerated and the dose can be increased (29–31).

Although prolactinomas usually remain sensitive to dopamine agonists, the drugs do not provide a definitive cure for most patients with macroprolactinoma and many have to remain on long-term therapy. Immediate tumour re-expansion may occur after drug withdrawal following medium-term therapy (up to 1 year) but re-expansion is less common after long-term treatment (several years) (23, 32). Recent withdrawal studies have suggested that up to 40% of macroprolactinomas may remain in remission after withdrawal of long-term cabergoline therapy, particularly in those patients who achieved prolactin normalization and near tumour disappearance during treatment (Fig. 2.3.10.10) (23). In patients who need to remain on treatment, the dose of dopamine agonist can often be reduced considerably once initial tumour regression has been achieved, with ongoing satisfactory control of tumour size.

Approximately 10% of genuine macroprolactinomas fail to regress during dopamine agonist therapy. The mechanism is obscure since most patients with nonshrinking tumours have marked suppression of serum prolactin levels. However, patients with little or no fall in serum prolactin often show minimal reductions in tumour size and a few continue to grow. Some nonshrinking tumours have large cystic components, some have atypical histology and some appear to have a deficiency of membrane-bound D2 dopamine receptors (25).

Macroprolactinoma shrinkage has been demonstrated with all of the clinically available dopamine agonists including bromocriptine, quinagolide, and cabergoline. Studies of cabergoline show that over 80% of previously untreated macroprolactinomas undergo significant tumour regression. Significant success rates were recorded also in patients previously resistant to or intolerant of other dopamine agonists, including bromocriptine (19). Some examples of cabergoline-induced shrinkage are shown in Fig. 2.3.10.6, Fig. 2.3.10.7, and Fig. 2.3.10.8. It is worth trying an alternative dopamine agonist in the event of drug resistance or intolerance (15).

Macroprolactinoma is virtually certain if serum prolactin is greater than 5000 mU/l in a patient with a pituitary macrolesion and primary treatment with a dopamine agonist has an excellent chance of tumour volume reduction. As noted earlier, a serum prolactin level between 2000 and 5000 mU/l presents some diagnostic uncertainty. Closely supervised dopamine agonist therapy is appropriate, provided surgery is performed in the event of any visual deterioration. Dopamine agonists reduce prolactin secretion from both normal and tumorous lactotrophs; therefore, serum prolactin is likely to fall irrespective of the cause of the hyperprolactinaemia. Pituitary macrolesions associated with prolactin levels less than 2000 mU/l are rarely prolactinomas and surgery should be undertaken to decompress the lesion and provide a histological diagnosis. Some of these important practice points are illustrated by the case shown in Fig. 2.3.10.11.

Medical treatment alone is an acceptable option for most patients with macroprolactinoma, particularly those with fertility needs in whom adjunctive therapy might compromise gonadotropin function. Physicians should be aware of the infrequent complication of cerebrospinal fluid (CSF) rhinorrhoea, which may occur after shrinkage of inferiorly invasive tumours and may be very difficult to correct surgically (33). Traction ophthalmopathy may occur rarely if the optic chiasm is adherent to the upper part of a shrinking tumour (34). Pituitary apoplexy may occur in a cystic tumour as it shrinks (35) (Box 2.3.10.3).

A minority of endocrinologists consider that dopamine agonist therapy alone is unsuitable for long-term management of macroprolactinoma and recommend external beam radiotherapy. Although prolactin levels fall over a period of several years after radiotherapy, enabling dopamine agonist withdrawal in a proportion of patients, this treatment is likely to be followed by varying degrees of hypopituitarism (see Chapter 2.3.6).

A meta-analysis of 2226 macroprolactinomas treated with primary surgery showed prolactin normalization in only 34% of patients (19). Certainly, one would not anticipate a curative surgical procedure in patients with giant, invasive macroprolactinomas, such as that illustrated in Fig. 2.3.10.3b. Consequently, in view of the effectiveness of medical treatment, only a minority of patients with large tumours should now require surgical intervention. There are a few selected situations in which some clinicians might consider surgery and a cautionary note on the effect of dopamine agonists on macroprolactinoma fibrosis is necessary. There is a direct relationship between tumour fibrosis and duration of medical treatment such that surgery is made much more difficult—and may even be hazardous—if dopamine agonists have been given for longer than 3 months (26, 36). Overall, it is prudent to limit preoperative dopamine agonist therapy to a maximum of 3 months if surgery is to be undertaken.

Oestrogens have a stimulatory effect on prolactin synthesis and secretion, and the hormonal changes of normal pregnancy cause marked lactotroph hyperplasia. MRI studies have confirmed a gradual doubling in pituitary volume during the course of gestation. In view of these effects of pregnancy on normal lactotrophs it is not surprising that prolactinomas may also increase in size.

The potential risk to the patient depends on the prepregnancy size of the prolactinoma. For women with microprolactinomas the risk of clinically relevant tumour expansion is very small indeed—less than 2%. Dopamine agonists can be safely stopped in such patients as soon as pregnancy has been confirmed. Nevertheless, patients should be advised to report for urgent assessment in the event of severe headache or any visual disturbance. Routine endocrine review may be arranged on two or three occasions during the pregnancy, but formal charting of visual fields is unnecessary and measurement of serum prolactin provides no useful information, given the considerable prolactin rise during normal gestation. Women can safely breastfeed their infants.

There has been some controversy concerning the risk of pregnancy for women with larger prolactinomas. In early reviews, macroprolactinoma expansion was reported to occur in nearly 40%, but many of these women received ovulation induction with gonadotropins and not dopamine agonists. More recent reviews suggest that symptomatic macroprolactinoma expansion occurs in fewer than 20% of women. The figure is probably 10% or lower in women given a several-month course of dopamine agonist prior to conception

Some clinicians continue to recommend conservative debulking surgery or even radiotherapy before pregnancy in women with macroprolactinomas to reduce the likelihood of major tumour expansion. However, dopamine agonists may be safely employed as sole therapy, using the following strategy. Medical treatment should be used for a minimum of 6 months, and preferably 12 months, together with follow-up MRI to assess residual suprasellar extension, before conception is attempted. If the tumour has shrunk to within the fossa, the dopamine agonist can be withdrawn once pregnancy is confirmed, with a less than 10% chance of re-expansion problems. If neurological problems do occur, bromocriptine should be started during the pregnancy and this will restore tumour control in nearly all cases. If there is significant suprasellar tumour before conception, the choice is between debulking surgery or continuing bromocriptine throughout the pregnancy. The latter seems to be effective but present experience is still relatively limited.

There is no evidence of teratogenicity in the offspring of women treated with simple bromocriptine-induced ovulation or those treated throughout pregnancy with the drug. Nevertheless, it is prudent not to use the drug during pregnancy unless absolutely necessary. Safety data for cabergoline and quinagolide are limited to a few hundred pregnancies, compared with several thousand for bromocriptine. Outcomes of 380 pregnancies following cabergoline treatment during a 12-year observational study have been reported recently (37). The spontaneous abortion rate in 329 pregnancies with known outcome was 9.1%, well within the expected range. The fetal malformation rate also fell within reported ranges for the general population with no pattern of type or severity. Since clinical experience is limited in relation to pregnancy and since the drug has a long half-life, the manufacturer still recommends that cabergoline be stopped 1 month prior to intended conception. However, this is clinically inconvenient and requires repeated monitoring of prolactin and ovarian status. There seems to be little risk in women who become pregnant while taking cabergoline. Pregnancy safety data on quinagolide are limited and perhaps less reassuring than those for cabergoline. In a recent review of 176 pregnancies in women treated with the drug, 14% ended in spontaneous abortion. Nine fetal malformations were diagnosed, including two infants with Down’s syndrome (38). Quinagolide has an intermediate duration of action and, in acknowledgement of the limited pregnancy experience, the manufacturer recommends that the drug be withdrawn as soon as pregnancy is confirmed.