Abstract
Hyperprolactinaemia is a common cause of reproductive dysfunction, and is found in 14–39% of women with secondary amenorrhoea. Recent data from the Netherlands estimates that approximately 1 in 1000 women and 1 in 5000 men have been treated at some time with dopamine agonists for hyperprolactinaemia [1]. The prevalence of clinically significant prolactin secreting pituitary adenomas may be as high as 62 per 100 000 [2]. Drug-induced hyperprolactinaemia makes up a significant sub-set of total cases [3], and the most commonly encountered and clinically important drug class resulting in an elevated prolactin is the antipsychotics. In this review we summarize the regulation of prolactin and the clinical effects of hyperprolactinaemia, before looking specifically at the entity of antipsychotic-induced hyperprolactinaemia, its investigation and management.
Prolactin physiology and regulation
Human prolactin is a 199 amino acid polypeptide of molecular weight 23 kDa, similar in structure to growth hormone [4], secreted by the anterior pituitary gland. Prolactin-secreting cells are known as lactotrophs. During pregnancy there is considerable lactotroph hyperplasia which results in a doubling of pituitary size and physiological hyperprolactinaemia. The main target tissue of prolactin has traditionally been thought to be the breast, but prolactin receptors have been demonstrated in several other tissues including liver, ovary, testis and prostate [4]. The function of prolactin at these sites remains poorly understood.
The predominant hypothalamic factor regulating prolactin secretion is dopamine, which inhibits secretion from pituitary lactotrophs [5]. Thyrotropin releasing hormone (TRH) [6] and vasoactive intestinal peptide (VIP) [7] are stimulatory to prolactin secretion, and a further stimulatory factor known as prolactin releasing peptide has also been discovered [8]; however, the physiological role of these factors is still unclear.
Dopamine is delivered via the pituitary stalk to the lactotrophs. Acting via D2 receptors, dopamine inhibits prolactin gene expression and secretion. Any mechanism that might interfere with the delivery or action of dopamine will increase serum prolactin levels. In animal models where surgical disconnection of the hypothalamus and pituitary is undertaken, initial hyperprolactinaemia occurs [9]. Hypothalamic lesions or non-functioning pituitary adenomas which obstruct the delivery of dopamine down the pituitary stalk lead to hyperprolactinaemia: the so-called ‘stalk effect’. Usually serum prolactin does not exceed 2000 mIU/L in this setting [10], but exceptions do occur. It has been shown that successful surgical resection of non-functioning pituitary adenomas where dopamine delivery is restored, results in rapid normalization of prolactin levels [11]. The mechanism of antipsychotic-induced hyperprolactinaemia relates to the dopamine antagonist effect of these drugs on the D2 receptors of the tuberoinfundibular system and lactotrophs [3,12]. By blocking the interaction between dopamine and the D2 receptors, the tonic inhibition of prolactin secretion is negated and increased prolactin secretion results.
Elevated prolactin levels have a direct inhibitory effect on release of gonadotrophin releasing hormone (GnRH) from the hypothalamus. Lack of pulsatile GnRH release results in absence of luteinizing hormone (LH) secretion from the pituitary, and secondary gonadal failure. There is no specific level of prolactin which inevitably causes suppression of the gonadal axis, but levels exceeding twice the upper limit of the laboratory reference range generally do so. Prolactin levels less than two-fold elevated may not cause any adverse reproductive effects. In addition, the presence of hyperprolactinaemia may be due to ‘macroprolactin’, an immune complex between IgG and prolactin which is not biologically active, and generally not associated with any reproductive dysfunction. Macroprolactin may be responsible in 10% of cases of hyperprolactinaemia in unselected series [13]. When the macroprolactin is precipitated using polyethylene glycol (PEG), the residual prolactin level is normal in most subjects [14]. In patients with pathological hyperprolactinaemia, restoration of prolactin levels to normal rapidly allows GnRH pulsatile secretion to be re-established. This then leads to restoration of normal menstrual cyclicity in women and reversal of testosterone deficiency in men.
Clinical consequences of hyperprolactinaemia
Pathological hyperprolactinaemia, whether or not it arises from a structural pituitary or hypothalamic lesion, is an important cause of reproductive dysfunction in both genders [15]. Typically in women, menstrual disturbance occurs which may range in severity from luteal phase insufficiency with otherwise regular menses, to irregular or infrequent periods (oligomenorrhoea), to amenorrhoea. Galactorrhoea may also occur, alone or in combination with menstrual disturbance [4]. In men, hyperprolactinaemia usually results in loss of libido and erectile dysfunction as a result of testosterone deficiency [15,16]. The oligomenorrhoea in women and testosterone deficiency in men are associated with infertility which is generally reversible upon restoration of normal prolactin levels.
In both genders, hyperprolactinaemia and the resulting deficiency of sex steroid may lead to bone loss and osteoporosis [17,18]. While correction of the elevated prolactin improves bone density [19], it has been demonstrated that the bone loss is not directly mediated by the elevated prolactin. Restoration of sex steroid concentrations is required for subsequent improvement in bone density, even in the presence of ongoing hyperprolactinaemia [20].
Uncommonly, prolactin secreting pituitary tumours may present in adolescents with delayed puberty, and in girls with primary amenorrhoea [21]. Other pituitary and hypothalamic lesions in individuals under 18 which may cause mild hyperprolactinaemia (such as craniopharyngioma, germinoma) usually present with features of mass effect and hypopituitarism with or without diabetes insipidus. In such cases, hypogonadism likely results from the direct tumour effect rather than the modestly elevated prolactin.
There is controversy regarding the effect of hyperprolactinaemia on breast cancer. In vitro studies have shown that prolactin is mitogenic [22], but a recent Dutch study showed no increase in breast cancer rate in women with idiopathic hyperprolactinaemia or prolactinomas [23]. Furthermore, surgical series of breast cancer patients noted hyperprolactinaemia in a similar number of women with benign breast conditions as breast cancer [24]. While some studies have shown an increased rate of breast cancer in women with schizophrenia, others have shown lower rates [25]. Factors other than hyperprolactinaemia may be involved, and more data is required to determine whether antipsychotic-induced hyperprolactinaemia is an independent risk factor for breast cancer.
Antipsychotic-induced hyperprolactinaemia
Hyperprolactinaemia is a common finding among patients with psychiatric disease. Overall in patients with schizophrenia, more than 65% of women of reproductive age and 40–70% of men have an elevated prolactin [26,27]. Age does not affect prolactin levels in men, but older age is associated with lower prolactin levels in women [27], with a lower prevalence of hyperprolactinaemia of approximately 45% in post-menopausal women [26]. Other studies in mixed populations of people with schizophrenia and bipolar disorder treated with antipsychotics show a prevalence of hyperprolactinaemia of approximately 50% in women and 25% in men [28].
The strongest predictor of hyperprolactinaemia in patients with disorders such as schizophrenia is the type of antipsychotic they are prescribed, and the dose (higher levels with higher doses) [27]. The older ‘typical’ antipsychotic drugs include three broad classes: the phenothiazines (chlorpromazine, fluphenazine, thioridazine, trifluoperazine and perphenazine), the thioxanthenes (thiothixene) and butyrophenones (haloperidol). The extent of the antipsychotic effect of these drugs parallels the rise in prolactin they cause, with haloperidol being the most potent in both regards. After oral administration, prolactin levels peak after approximately 1 week and remain elevated [29]. Normalization of the prolactin occurs within 4 days following cessation [30]. The newer atypical antipsychotic agents vary in their propensity to cause hyperprolactinaemia due to differences in D2 receptor binding activity and duration, as well as some having partial agonist activity. Of the atypicals, risperidone and amisulpride are associated with the highest rates of hyperprolactinaemia (70–100%), exceeding that observed with many of the typical agents. Other atypical agents such as olanzapine and quetiapine are less commonly associated with hyperprolactinaemia (10–40%), while clozapine has very weak affinity for the D2 receptor and hardly ever elevates prolactin (<5%) [3,31–33]. Of particular interest is aripiprazole which, whilst having high affinity for the D2 receptor, also has low rates of hyperprolactinaemia (<5%), presumably because it is a partial agonist at the D2 receptor [33–35].
Reproductive and sexual dysfunction
In pre-menopausal women, menstrual disturbance is a common consequence of drug-induced hyperprolactinaemia. Most studies show that 40–50% of women treated with antipsychotics have menstrual disturbance, which may take the form of amenorrhoea, oligomenorrhoea or polymenorrhoea (dysfunctional uterine bleeding (DUB)) [26]. A menstrual history is therefore a crucial component of the medical evaluation in such patients. Over 30% of women on antipsychotics with hyperprolactinaemia have serum oestradiol levels in the post-menopausal range, and there is an association between the prolactin elevation and the subsequent sex hormone deficiency [26]. Symptomatic galactorrhoea may occur in up to 20% of patients [36]. Among patients treated with different agents, the prevalence of symptomatic reproductive dysfunction parallels the hyperprolactinaemia, with 44% of risperidone-treated patients compared to just 3% of olanzapine-treated patients reporting such symptoms [31].
Sexual dysfunction in patients with psychotic illnesses has historically been difficult to study and its prevalence under-reported [37]. Studies relying on spontaneous reporting provide much lower estimates of sexual dysfunction than those making specific enquiry regarding such symptoms [37]. Both the psychotic illness itself and its treatment may contribute to sexual dysfunction. It has been shown that medication-free men with schizophrenia have lower libido compared to both normal controls and also to antipsychotic-treated men with schizophrenia [38]. However, the medicated men with schizophrenia had more erectile and orgasmic difficulties and a lower level of sexual satisfaction than those medication-free [38]. A number of studies have revealed significant sexual dysfunction in patients treated with antipsychotic drugs, with rates of 50–68% in men and 30–52% in women [39,40].
As in studies of menstrual disturbance, it has been noted in an open label study comparing risperidone to olanzapine that there is a higher rate of sexual dysfunction in those treated with the prolactin-elevating drug (52.4% in the risperidone group versus 12% in the olanzapine group) [41]. Furthermore, in a study where patients were switched from a prolactin-elevating agent to olanzapine, both male and female subjects noted improved sexual function [42]. According to Smith, hyperprolactinaemia, whilst not the sole cause of sexual dysfunction in these patients, does appear to be the predominant factor in both genders [37]. In men, other issues such as age, depression, medication dose and anticholinergic or antiadrenergic side effects contribute to sexual dysfunction in the absence of hyperprolactinaemia, while in women, the only significant factor apart from hyperprolactinaemia is depression [37]. In contrast to these findings, however, Howes and colleagues found no relationship between prolactin levels, hypogonadism and sexual dysfunction [40]. Similarly, a lack of correlation between risperidone-induced hyperprolactinaemia and erectile dysfunction has been reported [43], and while males treated with quetiapine versus risperidone had lower prolactin levels and less sexual impairment, the prolactin level was not correlated with the sexual dysfunction [44]. Other factors must contribute to the sexual dysfunction attributed to antipsychotic medication, such as direct and indirect CNS effects of antipsychotic medications, including interactions with other receptors (histaminic, muscarinic, serotonergic and alpha-1-adrenergic) [37].
Osteoporosis/osteopaenia
Osteoporosis is defined either by a bone mineral density (BMD) less than 2.5 standard deviations below the mean for a young person (T-score <2.5) or the presence of minimal trauma fracture, while osteopaenia is defined by a T-score of between −1.0 and −2.5. As previously outlined, bone loss is a consequence of sex steroid deficiency rather than elevated prolactin per se. It should also be noted that other factors may contribute to bone loss in patients with psychiatric illness, including diet, smoking, lack of exercise, hypercortisolism, polydipsia, and thyroid disorders [45,46]. As in other patients with bone loss, calcium supplementation is indicated where there is evidence of inadequate dietary intake, while vitamin D should be given in standard replacement doses to maintain levels > 50 nmol/L [47].
The prevalence of bone loss seems to vary with ethnicity and gender. In one Korean study of patients with schizophrenia, only females demonstrated lower BMD compared to controls [48]. Seventeen out of 18 women with bone loss in that study had hyperprolactinaemia, and overall prolactin levels were significantly higher in the group with bone loss than those with normal BMD [48]. In a predominantly Caucasian group from Ireland, 57% of men and 32% of women on long-term prolactin elevating antipsychotics had osteopaenia or osteoporosis [49]. The degree of bone loss correlated with medication dose, and in the men it related inversely to plasma testosterone levels [49]. In a UK study of mixed ethnicity, spine BMD was reduced in male patients of African or African-Caribbean origin, while there was no such finding among female patients [50]. A further small study which compared bone density in women treated with olanzapine versus prolactin-elevating agents showed the lowest rates of hypogonadism and bone pathology in those on olanzapine, intermediate on typical antipsychotics and highest on atypical prolactin-elevating agents (risperidone and amisulpride) [51]. A follow-up of these subjects longitudinally showed that those on olanzapine increased lumbar spine BMD over a 1-year period, while those on prolactin-elevating agents lost lumbar spine BMD [52]. Where a bone-sparing intervention had been initiated in patients on prolactin-elevating agents, lumbar spine BMD improved, compared to those left untreated [52].
The overall message from these studies is that where antipsychotic-induced hyperprolactinaemia results in hypogonadism, significant bone loss may occur in both genders. Restoration of normal sex hormone levels, either by a change in medication which leads to normoprolactinaemia or gonadal steroid replacement has the potential to reverse this process. Therefore the treatment strategy with respect to bone loss should involve attempts at lowering the prolactin to normal, thereby normalizing endogenous sex steroids or treatment with exogenous oestrogen or testosterone respectively. Monitoring of bone mineral density after 2 years of treatment will provide an objective measure of improvement. Specific anti-resorptive therapies such as the bisphosphonates should be reserved for cases of established osteoporosis poorly responsive to the above measures.
Investigation
In the face of finding a raised serum prolactin in a patient with a psychiatric disorder, the key issue is to differentiate drug-induced hyperprolactinaemia from that caused by a structural pituitary lesion. It should not be assumed that an elevated prolactin necessarily results from medication, especially in the setting of those atypical antipsychotic drugs where hyperprolactinaemia is less prevalent. As previously outlined, hyperprolactinaemia usually normalizes after 3–4 days off the offending medication [30], but in a patient with an unstable psychiatric state, cessation of an effective agent is not advisable. Where risperidone, amisulpride, or a typical antipsychotic agent is being prescribed with a view to long term therapy, measurement of prolactin prior to commencement creates a baseline for later comparison. If a previously normal prolactin level has not been noted and the potentially offending medication must be continued, the patient should undergo pituitary imaging, preferably with MRI scanning, to exclude a structural pituitary lesion. There have been a number of case reports of prolactinomas in patients with schizophrenia [53–56], and mild to moderate hyperprolactinaemia may be seen in non-functioning pituitary macroadenomas at similar levels to those observed in drug-induced cases [10]. A review of the US FDA Adverse Event Reporting System (AERS) suggested a higher prevalence of pituitary adenomas associated with the use of some atypical antipsychotic agents, particularly risperidone [57], a finding confirmed in the WHO database [58]. Community-based MRI studies show that approximately 10% of normal individuals may harbour an asymptomatic pituitary microadenoma (usually non-functioning) [59], and this must be taken into account when interpreting the imaging findings.
All patients with hyperprolactinaemia should also have a check of their reproductive hormones, notably follicle stimulating hormone (FSH), LH, oestradiol (women) and testosterone (men). Where a structural pituitary abnormality is found on imaging, full anterior pituitary function testing should be carried out, encompassing, as a minimum, morning cortisol (before 0900 h), T4, thyroid stimulating hormone (TSH) and insulin-like growth factor-1 (IGF-1) in addition to the reproductive hormones. A finding of a borderline morning cortisol (<350 nmol/L) requires further investigation with a suitable stimulation test of the entire hypothalamic–pituitary–adrenal axis. Plasma IGF-1 screens for both growth hormone (GH) excess and deficiency; mixed GH/prolactin secreting pituitary tumours do occur, while a large pituitary tumour may be associated with GH deficiency. The discovery of a structural pituitary lesion should prompt referral to an endocrinologist experienced in the management of pituitary disease, while the finding of normal pituitary anatomy is reassuring.
Management
Broadly there are three management strategies for antipsychotic-induced hyperprolactinaemia [36,60]:
Switch to an alternative antipsychotic agent
This should be considered the initial management of choice for patients with antipsychotic-induced hyperprolactinaemia. We have reviewed the literature concerning the lower rates of hyperprolactinaemia, hypogonadism and sexual dysfunction in cross-sectional studies of patients treated with prolactin-sparing compared to prolactin elevating agents. A randomized double blind study has also shown that the rate of hyperprolactinaemia was lower in quetiapine versus risperidone, with similar clinical efficacy [61]. In a further study of patients with schizophrenia randomized to olanzapine versus typical antipsychotics, those treated with olanzapine had lower prolactin levels at 3 months [62]. Thus, it would seem prudent to factor into clinical decision-making regarding initial choice of antipsychotic, the propensity or otherwise to raise prolactin. In those patients who have a raised prolactin, expressly if there are sexual or other side effects, switching to a prolactin-sparing agent is warranted unless other clinical parameters pertain such that the prolactin-elevating agent is preferred (e.g. differential clinical response or risk of side effects such as metabolic syndrome with other agents). As outlined above, several studies have demonstrated that switching between an agent such as risperidone to a prolactin-sparing drug lowers serum prolactin [42,63,64]. There has also been recent data regarding the addition of the partial agonist aripiprazole as adjunctive treatment in patients treated with prolactin-raising antipsychotics [65]. In a randomized double blind study, the addition of aripiprazole to either risperidone or quetiapine was safe and well tolerated [66]. This strategy appears to be effective in patients on risperidone [66–68] but less so for those taking amisulpride [68]. In risperidone treated patients, aripiprazole in doses greater than 6 mg per day reduced prolactin by >50% but fewer than 50% of women regained normal menses [67].
Sex steroid replacement
With the exception of galactorrhoea, the adverse consequences of hyperprolactinaemia relate to the deficiency of oestrogen in women and testosterone in men. Where either no structural pituitary abnormality or a small microadenoma (probable ‘incidentaloma’) is found, and the prolactin-elevating agent is continued or prolactin remains elevated despite a switch to a prolactin-neutral antipsychotic, a valid therapeutic option is to commence the patient on sex hormone replacement [69]. While there has been controversy surrounding the use of post-menopausal hormone therapy (HT) in the form of oestrogen/progestin in women, it is generally accepted that such treatment is routinely indicated in women of pre-menopausal age with hypogonadism [70]. Furthermore, there has been some evidence of a positive benefit of transdermal oestrogen therapy on psychotic symptoms [71,72]. The use of oral oestrogen does result in a small but significant increased risk of venous thromboembolism, although this risk is diminished by the use of topical oestrogen preparations [73]. Women with an intact uterus require the use of a cyclical progestin to induce a withdrawal bleed or continuous combined oestrogen/progestin to provide endometrial protection. Several suitable proprietary agents are available including low dose combined oral contraceptives and HT preparations.
In men, testosterone replacement is generally given either via a trans-dermal preparation or intra-muscular injection. Oral testosterone has poor bioavailability and is not recommended. The availability of a long-acting testosterone ester – testosterone undecanoate has enabled a 3 monthly injection for most men with hypogonadism, eliminating the need for frequent injections and the ensuing peaks and troughs of the shorter acting esters [74]. Of the transdermal preparations, a gel applied daily is generally better tolerated than the patch which has a high rate of skin irritation.
Dopamine agonists
Early reports using bromocriptine in people with schizophrenia demonstrated an exacerbation of psychosis in some of them [75]. However, a number of case reports and small case series show that bromocriptine [76–79] and the more potent and D2 receptor specific agent cabergoline [80,81] have been successfully and safely used in the management of antipsychotic-induced hyperprolactinaemia [82]. Dopamine agonists have also been used in cases where psychotic illness co-exists with a prolactinoma [53,55]. Dopamine agonists may exacerbate psychosis [83,84], and there are no published randomized controlled trials of dopamine agonist therapy in the setting of antipsychotic-induced hyperprolactinaemia, which means any use of a dopamine agonist in this setting must be approached with caution and be considered a third line strategy [60]. However, where fertility is desired or galactorrhoea is a troublesome symptom, a dopamine agonist is essentially the only effective treatment if the patient cannot be switched to a prolactin-sparing agent. If a dopamine agonist is used, the patient's mental state should be monitored closely, and there should be close communication between psychiatrist, endocrinologist and primary care physician/general practitioner. The lowest dose which results in normalization of the patient's prolactin and restoration of a functional hypothalamic–pituitary–gonadal axis should be used. A suggested starting dose of cabergoline is 0.25 mg twice weekly, with titration depending on response. The duration of treatment should be to the normal age of menopause for women and probably lifelong for men if the antipsychotic agent use is ongoing.
Summary of management strategies
Where possible, a patient with antipsychotic-induced hyperprolactinaemia should be switched to an alternative atypical antipsychotic agent. Otherwise, sex steroid replacement may be used as treatment of secondary hypogonadism and the resultant bone loss. A randomized controlled trial comparing sex steroid replacement with low dose dopamine agonist therapy is required to formally determine the safety and efficacy of each approach.
Acknowledgement
The authors would like to thank Helen Schultz for her assistance with the literature review.
