Prolactinoma and pregnancy
Hyperprolactinaemia, a common endocrine disorder, frequently leads to gonadal dysfunction which typically presents menstrual disturbance (amenorrhoea or oligomenorrhoea), galactorrhoea, dyspareunia, loss of libido and infertility. The oestrogen deficiency, which also predisposes women to premature osteoporosis1, is primarily a result of reduced gonadotrophin releasing hormone pulsatility2, and consequently suppressed luteinising hormone and follicle stimulating hormone release, but also in part from direct inhibition of gonadal steroid production by prolactin. Treatment of the elevated prolactin restores ovulation in 80%–90%3, and improves trabecular bone mineral density in just over 50% of women with hyperprolactinaemic amenorrhoea4. A number of factors can cause hyperprolactinaemia but the most frequent pathologic cause is a prolactin secreting tumour, prolactinoma.
Prolactinomas are the most common hormone-secreting (functioning) pituitary tumours in both autopsy and surgical series5. They are generally classified clinically by size: microadenomas < 10 mm in diameter and macroadenomas > 10 mm in diameter (with or without extrasellar extension). Prolactinomas occur more frequently in women than in men and differ not only in size, but also in clinical presentation, invasive growth and secretory activity. The vast majority (95%) of prolactinomas in women are microadenomas, which present with the clinical manifestations of hyperprolactinaemia described above, and rarely lead to hypopituitarism or neurologic dysfunction. In contrast, men with prolactinomas often present because of symptoms due to the size of the tumour rather than impotence, loss of libido, or infertility. Radiologic investigation demonstrates a macroadenoma in over 90% of cases in men6–8. Macroadenomas may cause—with local mass effects—hypopituitarism, headaches, visual field defects; or they may invade local structures to cause ophthalmoplegia and neurologic dysfunction (e.g. seizures or hemiparesis). Data on the natural course of prolactinomas comes from autopsy studies9 and studies where patients with microadenomas were followed up for long periods (2.5–8 years) without treatment10,11. From the post mortem studies, the risk of progression of microadenomas to macroadenomas is about 3% and that from the carefully followed patients is 6.9%. In other words, it appears that 93%–97% of microadenomas remain ‘small’.
The diagnosis of a prolactinoma is confirmed by sustained hyperprolactinaemia and cranial magnetic resonance imaging. There is no prolactin level that is diagnostic of a prolactinoma, although macroadenomas are usually associated with marked elevations of prolactin12. In their study, Bevan et al.12 correlated prolactin concentration with histological diagnosis, and found that a prolactin level of more than 8000 mu/L was always due to a prolactinoma. Furthermore, it was noted that a pituitary macroadenoma accompanied by a serum prolactin level of less than 3000 mu/L, was a result of a nonfunctional tumour, which interferes with dopamine synthesis, release or passage to the anterior pituitary, producing hyperprolactinaemia—a pituitary stalk effect, or ‘disconnection’ hyperprolactinaemia. The treatment options for prolactinomas include medical therapy with dopamine agonists, selective trans-sphenoidal adenectomy or radiotherapy.
Bromocriptine, which binds with high affinity to the dopamine receptors on both normal and adenomatous lactotroph cells is, like other dopamine agonists, the treatment of choice for prolactinomas of all sizes. It restores ovulatory menstrual cycles in 80%–90% of women of reproductive age, despite normal prolactin levels being achieved in only 75%–85% of patients3. In addition to reducing serum prolactin levels, it is well recognised that the size of even very large macroprolactinomas is markedly reduced by bromocriptine13,14. Tumour size reduction is often rapid, in some patients occurring in one to two weeks, but most shrinkage occurs during the first three months of treatment; however, in a minority (25%) tumour shrinkage is delayed for six months to one year. Although tumour shrinkage is usually accompanied by a fall in prolactin, there is no correlation between reduction in tumour size and prolactin levels measured either before or during therapy. In patients with macroprolactinomas and visual field deficits, improvement in vision is seen in about 90% of affected patients, often in the first 24–72 hours, and vision continues to improve over several months. For those intolerant of bromocriptine, the most promising alternative is cabergoline, a long-acting ergoline derivative. Because of its efficacy, tolerability and ease of administration, cabergoline may replace bromocriptine as the dopamine agonist of first choice15,16.
Pregnancy in women with prolactinomas
During pregnancy, the normal pituitary gland increases in weight and size by 35% and 50%–70%, respectively, predominantly due to an increase in the number and size of lactotrophs. The percentage of lactotrophs in the pituitary gland rises from 15%–20% of total pituitary cells in nulliparous women (and in men) to approximately 50% at the end of normal gestation17. When visualised radiologically18, the pituitary develops an upward convexity of its superior surface mainly as a consequence of lactotroph hyperplasia, with a resultant increase in prolactin synthesis and secretion19. To prepare the breast for lactation, prolactin levels rise and may be six to ten times higher at term than before conception. In those women who do not breastfeed, serum prolactin falls to preconception levels by three to six weeks postpartum20. The massive hyperoestrogenaemia of pregnancy is considered responsible for this increase in prolactin secretion, since oestrogen is a known stimulus to both prolactin synthesis and secretion via prolactin mRNA, as well as lactotroph proliferation21.
Therefore, once ovulation and fertility are restored, usually with dopamine agonist therapy, there are two major issues that arise in a women who harbours a prolactinoma: firstly, the ‘stimulatory’ effect of pregnancy on tumour size, and secondly, the effects of medical therapy, dopamine agonists, on the fetus.
Effect of pregnancy on prolactinomas
The hormonal milieu of pregnancy (hyperoestrogenaemia), not only has stimulatory effects on lactotrophs, but also oestrogen receptors in prolactinomas, leading to symptomatic tumour enlargement in some patients. In 1979, Gemzell and Wang22 summarised 217 pregnancies in 187 women with prolactinomas, who were categorised according to tumour size and previous treatment. There were 91 pregnancies in 85 women with microadenomas that had not received prior surgery or irradiation. Only five of these pregnancies (5.5%) developed systematic tumour enlargement, defined as headaches or visual disturbance. There were 56 pregnancies in 46 women with previously untreated macroadenomas; severe headaches or visual impairment were reported in 20 pregnancies (35.7%). In comparison, there were 70 pregnancies in 67 women with macroadenomas who had received irradiation, surgery or combined therapy prior to pregnancy; two women developed headaches with visual field defects, and three reported headache alone, a complication rate of 7%.
The review by Molitch23 summarised 16 series between 1979 and 1985, with 246 women with microadenomas and 91 women with macroadenomas (45 had no prior treatment; 46 had prior surgery or radiation therapy) who became pregnant. In women with microadenomas, only four (1.6%) had symptoms of tumour enlargement, and eleven (4.5%) had asymptomatic tumour enlargement (radiological evidence). All women were managed conservatively. In those with macroadenomas with no prior surgical or radiation therapy, seven (15.5%) had symptomatic tumour enlargement and four (8.5%) asymptomatic enlargement, necessitating surgery in four women and bromocriptine therapy in two. In contrast, of 46 women with macroadenomas treated with prior surgery or radiation therapy, only two (4.3%) had symptomatic tumour enlargement, and none had evidence of asymptomatic tumour growth.
Effects of dopamine agonist therapy on the fetus
Although the current evidence suggests that bromocriptine is probably safe in pregnancy, it is advised that fetal exposure to bromocriptine be limited to as short a period as possible. In women who had conceived while taking bromocriptine, but in whom treatment was discontinued three to four weeks post conception, there was no increase in the incidence of spontaneous abortions, ectopic or multiple pregnancies, or congenital malformations24. In addition, a long term follow up study by Raymond et al.25 of 64 infants and children (six months to nine years) whose mothers took bromocriptine for the first four weeks of gestation have shown no increase in abnormalities. Furthermore, in the event of tumour expansion in pregnancy, bromocriptine has been used successfully, with no ill effects on the infant23,26. The information on safety in pregnancy for other dopamine agonists, such as cabergoline and quinagolide, is limited at present.
Treatment of prolactinomas during pregnancy
The treatment of prolactinomas in women desiring pregnancy is contentious, both prior to conception and during pregnancy itself. The treatment options for anovulatory women who wish to conceive are: dopamine agonist therapy, trans-sphenoidal surgery or radiotherapy, or a combination of these.
Given the efficacy of ovulation induction with bromocriptine (80%–90%) and the low risk of tumour enlargement during pregnancy (1.6%–5.5%) in a woman with a microprolactinoma22,23, there is general agreement that bromocriptine should be used to restore ovulation and allow conception. Following the diagnosis of pregnancy the drug is stopped. With this in mind and the observation that women with microprolactinomas who have had at least one pregnancy are more likely to demonstrate normoprolactinaemia and tumour regression compared with those who have not had any pregnancy27,28, it seems illogical to advise surgery or radiotherapy as treatment. Furthermore, trans-sphenoidal surgery, even in experienced hands, has initial cure rates between 65%–80% of women, with postoperative recurrence of hyperprolactinaemia in 20% of women, usually within one to two years of surgery. The morbidity and mortality rates of this operation are 0.4% and 0.27%, respectively29,30.
It is in women with macroprolactinomas where there is major controversy over treatment. The risk of clinically important enlargement of the macroprolactinomas during pregnancy is between 15.5% and 35.7%23,24. Given this fact, and the observation that women with macroprolactinomas treated surgically or with radiotherapy prior to pregnancy had a lower risk (4.3%) of tumour enlargement23, many clinicians advocate prophylactic debulking surgery or radiotherapy for macroprolactinomas in women of reproductive age who desire pregnancy. However, it is important to note that the above observations of tumour expansion during pregnancy22,23 relate to an era when bromocriptine was not widely available. In the review by Molitch23 fewer than half of the women received bromocriptine prior to pregnancy; the number of women who required surgical intervention for symptomatic tumour enlargement (just under 50%) would have consequently been considerably less had it been appreciated that bromocriptine could successfully reduce tumour size in pregnant. Furthermore, Molitch23 was able to demonstrate a difference in tumour growth in patients with macroadenomas who underwent surgery or radiotherapy prior to pregnancy compared with those who did not. The difference was not statistically significant if only women with symptoms of tumour enlargement were compared. In other words, surgical resection of a macroadenoma prior to conception may not prevent symptomatic tumour enlargement during pregnancy.
The surgical cure rates for macroadenomas are 10%–40%, with recurrence rates of 15%–80%, and morbidity and mortality rates of 6.5% and 0.9%, respectively29,30. Apart from the morbidity and mortality associated with surgery, and its poor cure rates, women who have had surgery, for both microadenomas and macroadenomas, have lower prolactin levels and consequently decreased milk production and poorer breastfeeding, compared with those who are treated with bromocriptine alone31. However, selective transsphenoidal adenomectomy may be indicated in certain circumstances: macroprolactinomas with visual field defects unresponsive to dopamine agonist therapy, rapid re-expansion of prolactinomas on cessation of dopamine agonist therapy, intolerance of medical therapy, presence of pituitary apoplexy, and the preference of the woman.
Radiotherapy is rarely curative and has a very slow and inadequate response32. More importantly, radiotherapy comes at some ‘expense’. Tsagarakis et al.33 reported that 50% of women were normoprolactinaemic after a mean of 8.5 years, with inevitable pituitary dysfunction (94% growth hormone deficiency at two to three years and 15%–23% thyrotrophin, gonadotrophin and corticotrophin deficiency occurring three to eleven years later). With this in mind and the implications of pituitary hormone replacement—in particular growth hormone—radiotherapy appears to be an ill-advised therapeutic option for women wishing to conceive. However, radiotherapy should be considered in women with macroprolactinomas, which continue to grow despite dopamine agonist therapy or after incomplete surgical removal (invasive prolactinoma).
Bromocriptine should therefore be recommended as primary therapy in women with macroprolactinomas34,35. However, women with macroprolactinomas should be advised against conception until normal visual fields are documented and a repeat pituitary magnetic resonance scan has shown tumour shrinkage to within the pituitary fossa, which may take three to six months. As with microprolactinomas, bromocriptine therapy should be withdrawn as soon as pregnancy is confirmed and recommenced if significant tumour expansion is evident (i.e. visual field defect or headache). In women where suprasellar extension persists after six months of bromocriptine or cabergoline treatment, selective trans-sphenoidal adenomectomy may be considered. Breastfeeding with its suckling stimulus does not appear to elicit a significant increase in prolactin concentrations or tumour growth23; however, women with macroadenomas should be monitored closely and bromocriptine started if there is symptomatic tumour growth.
Monitoring during pregnancy and possible complications
Following conception women with prolactinomas should be monitored closely for symptoms and signs suggestive of tumour enlargement. Although some physicians recommend that microadenomas do not require close follow up during pregnancy, most endocrinologists recommend visual field assessment in each trimester of pregnancy by direct confrontation with a red pin, and by formal Goldmann perimetric visual field testing in women who became symptomatic. In women with macroadenomas six-weekly tests of visual acuity and Goldmann perimetry is advised. Since tumour expansion can occur rapidly it is important to advise women with prolactinomas to report any visual symptoms or headache and the pituitary magnetic resonance imaging should then be performed. Because of the physiological increase in prolactin during pregnancy, periodic checking of prolactin levels is of no benefit. In addition, prolactin levels do not always rise during pregnancy in women with prolactinomas, as they do in normal pregnancy, and may also not rise with tumour enlargement.
If there is evidence of symptomatic tumour enlargement during pregnancy, it is best treated with bromocriptine rather than surgery, which has a 1.5-fold increased risk of fetal loss in the first trimester and a fivefold risk in the second trimester36. Emergency debulking of the tumour may be indicated if there is progressive visual field loss despite medical therapy. These women need daily visual acuity and field testing on a magnetic resonance scan on admission. The possible consequences of rapid tumour enlargement are hypopituitarism and pituitary apoplexy. Compression of the normal pituitary or distortion of the hypothalamo-hypophyseal portal system by the tumour may result in hypopituitarism. If this is suspected clinically a pituitary hormone profile should be performed and hormone replacement therapy started immediately.
Pituitary apoplexy, which may be subclinical, usually presents with acute onset of headache, meningism, visual disturbance and hypopituitarism. It is caused by haemorrhagic infarction into the tumour, although non-adenomatous tissue may also be involved, and may present during pregnancy, or after childbirth, the classical Sheehan's syndrome. In order to prevent the possibility of pituitary apoplexy during labour elective caesarean section should be considered in women with symptomatic tumour enlargement.
One should consider the possibility of lymphocytic hypophysitis, an important differential diagnosis of pituitary mass during pregnancy and which may also lead to headache, visual disturbance and hypopituitarism37. Most cases of this rare autoimmune disorder present late in pregnancy or up to one year after childbirth. The diagnosis often requires a biopsy of the lesion to distinguish it from a tumour. About 40% of cases have hyperprolactinaemia. Partial pituitary adrenocorticotrophin hormone insufficiency may occur, most commonly deficient of adrenocorticotrophin.
Dopamine agonist therapy is the primary treatment for prolactinomas of all sizes in women of reproductive age. Unlike women with microprolactinomas, conception should be delayed in those with macroprolactinomas until magnetic resonance imaging has demonstrated tumour shrinkage to within the sella. Once pregnancy is confirmed dopamine agonist therapy can be stopped. The risk of tumour enlargement during pregnancy is low in women with microadenomas but is greater than those with macroadenomas. In the event of symptomatic tumour enlargement bromocriptine can be used safely, and close monitoring of visual acuity and usual fields is required. In addition, the complications of acute tumour enlargement, hypopituitarism and pituitary apoplexy, must be carefully looked for.