Diagnosis and Management of Hyperprolactinemia

1

Diagnosis and Management of Hyperprolactinemia: Results of a Brazilian Multicenter Study with 1234 Patients L. Vilar1, M.C. Freitas1, L.A. Naves2, L.A. Casulari2, M. Azevedo2, R. Montenegro3, A.I.Barros3, M. Faria4, G.C. Nascimento4, J. Lima5, L.H. Nóbrega5, T.P. Cruz6, A. Mota7, A. Ramos7, A. Violante8, A. Lamounier Filho8, M.R. Gadelha8, M.A. Czepielewski9, A.Glezer10 and M.D. Bronstein10 1

Hospital das Clínicas, Division of Endocrinology, Federal University of Pernambuco, Recife, Pernambuco, Brazil.

2

Brasilia University Hospital, Division of Endocrinology, Brasilia, Brazil.

3

Federal University of Ceará, Division of Endocrinology, Fortaleza, Ceara, Brazil.

4

Federal University of Maranhao, Division of Endocrinology, Sao Luiz, Maranhao, Brazil.

5

Natal Institute of Endocrinology, Division of Endocrinology, Natal, Rio Grande do Norte, Brazil.

6

Federal University of Bahia, Division of Endocrinology, Salvador, Bahia, Brazil.

7

Federal University of Paraiba, Division of Endocrinology, Campina Grande, Paraiba, Brazil.

8

9

Federal University of Rio de Janeiro, Division of Endocrinology, Rio de Janeiro, Brazil.

Federal University of Rio Grande do Sul, Division of Endocrinology, Porto Alegre, RioGrande do Sul, Brazil.

10

University of Sao Paulo Medical School, Neuroendocrine Unit, Division of Endocrinology and Metabolism, Sao Paulo, Sao Paulo, Brazil

Short title: Hyperprolactinemia: Brazilian Multicenter Study Key words: Prolactin; hyperprolactinemia; diagnosis; treatment; macroprolactin Address for correspondence: Lucio Vilar, M.D., Ph.D. Rua Clóvis Silveira Barros, 84/1202 – Boa Vista –Zip 50.050-270 - Recife –PE, Brazil. Telephone / Fax : (5581) 3222.42.59 E-mail:[email protected]

2

Abstract

Objective: To evaluate clinical and laboratorial features of 1234 patients with different etiologies of hyperprolactinemia, as well as the response of 388 patients with prolactinomas to dopamine agonists. Design/Setting/Patients: A total of 1234 hyperprolactinemic patients from 10 Brazilian endocrine centers were enrolled in this retrospective study. Main Outcome Measure: Prolactin (PRL) measurement, thyroid function tests and screening for macroprolactin were conducted. Results: Patients were subdivided as follows: 56.2% had prolactinomas, 14.5% drug-induced hyperprolactinemia, 9.3% macroprolactinemia, 6.6% non-functioning pituitary adenomas, 6.3% primary hypothyroidism, 3.6% idiopathic hyperprolactinemia and 3.2% acromegaly. Clinical manifestations were similar irrespective the etiology of the hyperprolactinemia. The highest PRL levels were observed in patients with prolactinomas but there was a great overlap in PRL values between all groups. However, PRL > 500 ng/mL allowed a clear distinction between prolactinomas and the other etiologies. Cabergoline was more effective than bromocriptine in normalizing PRL levels (81.9% vs. 67.1%, P < 0.0001) and in inducing significant tumor shrinkage and complete disappearance of tumor mass. Drug resistance was observed in 10% of patients treated with cabergoline and in 18.4% of those that used bromocriptine (P = 0.0006).Side-effects and intolerance were also more common in bromocriptine treated patients. Conclusion: Prolactinomas, drug-induced hyperprolactinemia and macroprolactinemia were the three most common causes of hyperprolactinemia. Although PRL levels could not reliably define the etiology of hyperprolactinemia, PRL values > 500 ng/mL were exclusively seen in patients with prolactinomas. Cabergoline was significantly more effective than bromocriptine in terms of prolactin normalization, tumor shrinkage and tolerability.

3

Introduction

Hyperprolactinemia is the most common abnormality of the hypothalamic-pituitary axis. Its prevalence varies from 0.4% in healthy adults to 70% in women with amenorrhea and galactorrhea (1). The usual clinical presentation of hyperprolactinemia includes menstrual disturbances, hypogonadism, galactorrhea, infertility, and decreased libido (2). Causes of hyperprolactinemia can be divided into physiological, drug-induced, pathological, and idiopathic ones. Drugs and prolactinomas are considered the main causes of non-physiological hyperprolactinemia (3).

Prolactin (PRL) exists in different forms in human serum. The predominant form in normal and hyperprolactinemic individuals is monomeric PRL (molecular mass 23 kDa), while big prolactin

(45-60 kDa) and big big PRL or macroprolactin (150–170 kDa) correspond to less than 20% of the total PRL (4-6). When the serum of a patient with hyperprolactinemia contains mostly big big PRL the condition is termed macroprolactinemia (3,5). As macroprolactin has a low biological activity, most patients with macroprolactinemia lack typical symptoms related to hyperprolactinemia (7,8). Medical therapy with the dopamine agonists cabergoline or bromocriptine is the initial treatment of choice in patients with micro- and macroprolactinoma as well as in those with the so-called "idiopathic" hyperprolactinemia (9). When infertility is the primary indication for treatment, bromocriptine use has an extensive safety experience and is preferred by some clinicians. However, for other indications, cabergoline appears to be more efficacious and better tolerated. Normalization of serum prolactin (PRL) levels and restoration of gonadal function is seen in up 90% of cases (10). Moreover, a clear-cut tumor mass shrinkage, even tumor disappearance at magnetic resonance imaging scan, is reported in 70–80% of patients (9, 11, 12, 13). However, a variable amount of patients ranging from 5–50% in different series does not achieve control of PRL hypersecretion and/or tumor shrinkage even after treatment with a variety of dopamine agonists at high doses (9, 10, 11, 14, 15, 16).

4 In the present study we retrospectively evaluate the clinical and biochemical features of 1234 patients with hyperprolactinemia investigated in 10 Brazilian endocrinology centers. The main objectives of this study are to determine the etiology of hyperprolactinemia in the studied patients and their clinical and demographic characteristics, to assess the usefulness of PRL levels in the definition of etiology of hyperprolactinemia and to evaluate the response of prolactinomas to dopamine agonists. To the best of our knowledge, such number of patients with different etiologies of hyperprolactinemia has never been evaluated in the same study. Materials and Methods

Patients and Study design A detailed questionnaire requesting demographic, clinical, biochemical, and radiological data, as well as the diagnosis and the chosen therapeutic option(s) of patients with hyperprolactinemia, was sent to the 10 participating centers. In particular, laboratory evaluation for the etiology of hyperprolactinemia included PRL levels, thyroid function and screening for macroprolactinemia through polyethylene glycol (PEG) precipitation. For baseline parameters, contributors had to state the normal range adopted by their laboratory. All patients with hyperprolactinemia routinely followed in the 10 participating centers from 2000 to 2006 were included in the study. Resistance of prolactinomas to dopamine agonists was defined as failure to normalize PRL levels despite the use for at least three months of the maximum tolerated dose of cabergoline or bromocriptine (10, 14). Patients with poor compliance with the treatment were not considered as drugresistant. All patients provided their informed consent to the inclusion of their clinical and laboratorial features in this study. The study was designed by the Department of Neuroendocrinology of Brazilian Society of Endocrinology and Metabolism. Assays

5 Hormones were measured in-house by each centre. PRL levels were measured by commercially available chemiluminescence immunoassays (normal value, men: 2.1–17.7 ng/mL; women: 2.8–29.5 ng/mL) in 8 centers or immunoradiometric assays (normal value, men: 2.3–11.5 µg/liter; women: 2.5– 14.5 µg/liter) in 2 centers. Macroprolactin was determined by measuring the serum PRL level before and after PEG precipitation. The latter was performed exactly the same way in each center. As suggested by Vieira et al. (17), PRL recoveries of 60% after PEG precipitation were used as the criteria diagnosis for macroprolactinemia and monomeric hyperprolactinemia, respectively. Patients with PRL recovery 50% reduction of pretreatment tumor volume. Statistical analysis In the analysis of qualitative variables, the

2

test or the Fisher’s exact test (when necessary) were

used. P values 0.999)

7 PRL levels

As shown in Table 2, the highest PRL levels were observed in patients with macroprolactinomas and the lowest in subjects with primary hypothyroidism. However, there was a great overlap in PRL values regardless the etiology of hyperprolactinemia. In fact, although most cases of hyperprolactinemia not related to prolactinoma presented with PRL levels < 100 ng/mL the same occurred in 25% of patients with microprolactinomas. Moreover, values as high as 490 ng/mL were found in a patient with a non-functioning pituitary adenoma. On the other hand, PRL levels > 500 ng/mL allowed a clear distinction between prolactinomas and the other etiologies of hyperprolactinemia. Table 3 details the distribution of patients according to the etiology of hyperprolactinemia and PRL levels.

Response to dopamine agonists Data on the treatment of 388 patients (320 women and 68 men) bearing prolactinomas (220 micro- and 168 macroadenomas) treated with cabergoline (CAB), bromocriptine (BCR) or both could be obtained. BCR was administered to patients with microprolactinomas for 12.3 ± 8.6 months (range, 2–34 months) and to patients with macroprolactinomas for 13.3 ± 9.6 months (range, 2–48 months). The mean BCR dose was 6.8 ± 3.0 mg/day (range, 2.5–15) for micro- and 8.8 ± 3.3 mg/day (range, 5–15) for macroprolactinomas. CAB treatment was given for 11.9 ± 8.7 months (range, 2–36 months) in patients with microprolactinomas and for 12.5 ± 8.9 months (range, 2–34 months) in patients with macroprolactinomas. The mean dose of CAB was 1.2 ± 0.7 mg/week (range, 0.2–3.5) for micro- and 1.7 ± 0.7 mg/week (range, 1.0–3.5) for macroprolactinomas. The therapy with BCR in 304 patients with prolactinomas resulted in normalization of PRL levels in 67.1% of cases (72.4% of 178 microprolactinomas and 59.4% of 128 macroprolactinomas, P = 0.011).

8 Cabergoline was administered to 121 patients with microprolactinomas and 117 with macroprolactinomas, of whom 154 (64.7%) had been previously treated with bromocriptine. Normoprolactinemia was achieved in 83–91% of naive patients, in 81–88% of BCR intolerant patients, in 50–61% of BCR resistant patients and in 100% of BCR responsive patients (Fig.2). Overall, normalization of PRL levels was found in 85.9% and 77.8% cases of micro-and macroprolactinomas respectively (P = 0.101). CAB was more effective than BCR in inducing normoprolactinemia (81.9% vs. 67.1%, P < 0.0001), without any difference according to the gender. Overall, the normalization of PRL levels with CAB and BCR was similar in men and women (70.5% vs. 75.9%, P = 0.329). The therapy with dopamine agonists in 320 women resulted in disappearance of galactorrhea in 100% and resumption of normal menses in 79%%. Among 68 men, improvement of hypogonadism symptoms occurred in 60.3%. As shown in Fig. 3, among the patients treated with CAB, the prevalence of significant macroprolactinoma shrinkage was higher in naive patients (90%) than in BCR intolerant (40.4%), BCR resistant (30%), and responsive patients (40%). Moreover, CAB was more effective than BCR in inducing both significant tumor shrinkage (80% vs. 58.7%, P = 0.048) and complete disappearance of the tumor mass (57.5% vs. 34.7%, P = 0.034) in naive patients with macroprolactinomas. Drug resistance was observed in 24 patients (10%) treated with CAB (14 with a macro- and 9 with a microadenoma) and in 56 patients (18.5%) that used BCR (36 with a macro- and 20 with a microadenoma) (P = 0.006). Side-effects did not differ with BCR and CAB. The most common were headaches, dizziness, postural hypotension and nausea. They were more often observed in BCR treated patients (42.7% vs. 21.8%, P < 0.0001). Moreover, the rate of intolerant patients was also higher in the BCR group (14.5% vs. 8.0%, P = 0.019). Of the 98 macroprolactinemic women, 25 (25.5%) were treated with bromocriptine (n = 15) or cabergoline (n = 10) for 9.7 ± 7.1 months (2–36 months), before the realization that they had

9 macroprolactinemia. Normalization of PRL levels occurred in 32% of cases, galactorrhea disappeared in all patients but in none of them reversion of menstrual irregularities was observed.

Discussion

In this retrospective multicenter study the clinical and biochemical features of 1234 patients with hyperprolactinemia were evaluated. The three more frequent causes were prolactinomas (56.2%), drug-induced hyperprolactinemia (14.6%) and macroprolactinemia (9.3%). Regardless its etiology, hyperprolactinemia predominated in females (78.4% of cases). According to other studies, microprolactinomas are found, almost invariably, more frequently in women, while there are discordant data whether macroprolactinomas are more frequent in men (18), in women (13) or have an equal distribution between genders (15, 19). Clinical manifestations were similar irrespective the etiology of the hyperprolactinemia. Moreover, the frequency of menstrual disturbances and galactorrhea did not differ when women with macroprolactinemia and monomeric hyperprolactinemia were compared. In contrast, the combination of galactorrhea and amenorrhea was more often found in the latter. The prevalence of hypogonadism symptoms in men (particularly erectile dysfunction) was also comparable in both groups. Macroprolactinemia has been recognized for many years in asymptomatic patients or research volunteers (2, 5, 20). In more recent series, however, a significant proportion of patients with macroprolactinemia appeared to suffer from symptoms commonly associated with hyperprolactinemia. Among our macroprolactinemic patients, 49.5% were asymptomatic while hypogonadism symptoms and/or galactorrhea were found in 51.5%. In contrast, only 9.5% of subjects with monomeric hyperprolactinemia were asymptomatic (P < 0.0001). Our findings are in agreement with those reported by other groups. In the series of Hauache et al. (21), 90% (36/40) of patients with predominance of monomeric prolactin presented with symptoms related do PRL excess. Conversely, 46.4% (13/28) of subjects with macroprolactinemia had no symptoms. Similarly, the prevalence of

10 irregular menses in patients evaluated by Alfonso et al. (22) was similar in cases of macroprolactinemia (57%) and monomeric hyperprolactinemia (46%). The real prevalence of macroprolactinemia in patients with hyperprolactinemia is not fully established yet. In the present study, 115 patients (9.3%) had macroprolactinemia, a percentage comparable to that reported in five recent studies where that prevalence ranged from 10% to 23% (12, 22-26). However, it should be noted that the routine screening for macroprolactin has not been carried out in all centers involved in the present study. This could have contributed to underestimate the prevalence of macroprolatinemia in our study population. In fact, it was recently reported by a Brazilian group that, among 115 consecutive patients referred for the investigation of hyperprolactinemia, 19 (16.5%) had macroprolactinemia (25). The highest PRL levels were found in patients with macroprolactinomas and the lowest in subjects with primary hypothyroidism. Classically, PRL levels > 250 ng/mL have been considered to be highly suggestive of the presence of a prolactinoma (2, 3, 11). Among our 540 patients without a prolactinoma, 23 (4.2%) had PRL levels > 250 ng/mL. Although most cases of hyperprolactinemia not related to prolactinoma presented with PRL levels < 100 ng/mL the same occurred in 25% of patients with microprolactinomas. There was, therefore, a great overlap in PRL values regardless the etiology of hyperprolactinemia. Nevertheless, PRL levels

500 ng/mL were exclusively seen in patients with

prolactinomas. In the management of patients with a pituitary macroadenoma and hyperprolactinemia, it is absolutely crutial to distinguish prolactinomas and pseudoprolactinomas (mainly represented by non-functioning pituitary adenomas [NFPA]), once they require different treatments. Pseudoprolactinomas increase PRL levels by disrupting the normal delivery of dopamine to the adenohypophysis and usually cause mild to moderate PRL elevation, levels almost always being 150 ng/mL, and three patients (7%) had a PRL level of 200 ng/mL or more. In one patient with a plasmacytoma PRL values reached 504 ng/mL. In most acromegalic patients, the PRL elevation usually results of stalk compression and is mild to moderate (2, 3). However, in about one third of cases there is PRL co-secretion by the tumor. In that situation, the hyperprolactinemia rarely may be severe and PRL values as high as 5,250 ng/mL (30) or 6,000 µg/L (31) have already been reported. In the present study, PRL levels ranged from 28-275 ng/mL in cases of acromegaly (< 100 ng/mL in 69% of patients). In our study, PRL values in macroprolactinemic patients ranged from 30-404 ng/mL (mean, 114.5 ± 92.9) and, although they were