Prolactinomas are the most common pituitary tumors and account for about half of cases (1, 2). Their prevalence in the general population ranges from 6 to 50 per 100,000 (3, 4). Without appropriate treatment, prolactinomas may cause hypogonadism, infertility, bone loss, headaches, and visual fields defects from mass effect (5), as well as metabolic disorders (6). Based on their size, prolactinomas are classified into macro adenoma (>1 cm) or micro adenoma (<1 cm) (5). The mainstay of therapy involves the use of dopamine agonists (DA), even for patients with optic chiasm compression by large tumors (7, 8). Cabergoline (CAB) has been largely recommended as the first line agent, due to its better tolerability and higher efficacy in normalizing prolactin (PRL) levels and inducing tumor shrinkage, compared to bromocriptine (2, 5, 7, 8).

Studies of patients treated with CAB for microprolactinomas and macroprolactinomas have reported normalization of PRL levels in 75–90%, associated with an average decrease in tumor volume of 72–92% (9–11). In the Brazilian multicenter study on hyperprolactinemia, CAB therapy was able to induce significant tumor shrinkage and complete disappearance of tumor in 80 and 57.5% of patients, respectively (7).

A major drawback of DA therapy is the potential need to keep the medication indefinitely in many patients (2, 5). Indeed, despite the widespread use of DA for patients with prolactinomas and symptomatic idiopathic hyperprolactinemia for many decades, the optimal treatment strategy and duration of treatment is still not evident (12).

The 2011 Endocrine Society Clinical Practice Guideline recommends that, with careful clinical and biochemical follow-up, therapy may be tapered and perhaps discontinued in patients who have been treated with DA for at least 2 years, provided they no longer have elevated PRL levels, nor detectable tumor remnant on magnetic resonance imaging (MRI) (13). Accordingly, four recent studies (n = 292) have demonstrated that CAB therapy can be successfully discontinued, although hyperprolactinemia recurrence rates ranged from 25 to 69% (mean, 49%) (14–17).

On the other hand, it is not well established whether CAB therapy can be successfully withdrawn after a failed first attempt (18). This prospective study was conducted to address this topic and to try to identify possible predictor factors.

The optimal duration of therapy with DA for patients with prolactinomas or non-tumoral hyperprolactinemia remains controversial (12). There is however strong and growing evidence that discontinuation of dopamine agonist treatment may be successfully achieved in a selected group of patients treated for at least 2 years (14–17, 19, 20). A recent systematic review and meta-analysis has shown that the pooled proportion of patients with persisting normoprolactinemia after CAB withdrawal was 35% in a random effects model (21). In the series by Karlip et al. (15), which involved 46 patients, the overall recurrence was 54% and the median time to recurrence was 3 months (range, 1(18 months) (15).

In contrast, it is currently unknown how valid and how useful could be a second try to discontinue CAB administration in patients whose first withdrawal did not succeed. Presently, there are only data from a recent pilot prospective two-center study that evaluated 17 patients who had undertaken a second course of CAB treatment for at least 24 additional months (18). During a median follow-up of 6.1 months (ranging from 1 to 60 months) after CAB withdrawal, 11 patients (64.7%) recurred. The estimated overall recurrence rate was 44 events per 100 person-years. Moreover, the estimated cumulative hazard of recurrence was 40% at 6 months and 82% at 12 months (18).

We conducted a prospective trial to evaluate the outcome of a second attempt of CAB withdrawal in 34 patients who met strict selection criteria and have been treated with CAB for 24(30 months (median, 27 months). All patients presented with normal PRL levels, were given CAB doses (1 mg weekly, and had either no visible tumors or small tumor remnants (<1 cm) on MRI. We found that 10 patients (29.4%) were able to maintain normal PRL levels for up to 26 months following CAB withdrawal.

Concerning the 24 recurring subjects, all recurred within 15 months after CAB withdrawal, whereas 18 (75%) did so within 12 months. The median time to recurrence was 10.5 months. Similarly, in the study by Kwancharoen et al. (18), most of the recurrence (59%) also occurred within 1 year, whereas the median time to recurrence was 6 months. Likewise, Karlip et al. (15) had previously reported that 91% of the recurrences after the first withdrawal were observed within the first year of discontinuation. Thus, close clinical monitoring, especially within the first year after withdrawal, should be carried out in all patients in whom CAB therapy is discontinued.

According to previous studies, predictor factors for higher chance of successful CAB withdrawn include lower PRL levels, longer duration of treatment, tumor size (micro- > macroadenomas), previous pituitary radiotherapy or surgery, and pregnancy (5, 12, 13, 15, 16, 22). It has been shown that women with prolactinomas who became pregnant have a higher rate of remission than women without previous pregnancy (22, 23). Moreover, PRL levels are lower after delivery as compared to levels before conception and complete remission of hyperprolactinemia has been reported in 17–37% of women after pregnancy (23). For that reason, we did not include in our study women with history of pregnancy over the last 3 years.

Among our patients, a stratified analysis on clinical, pituitary imaging, and biochemical characteristics, at the beginning of the study and at CAB withdrawal, was unable to detect parameters that could quite accurately identify individuals most likely to respond to treatment discontinuation. However, non-recurring patients had significantly lower mean PRL levels at withdrawal. Furthermore, the recurrence rate was lower in patients without visible tumor remnant on pituitary MRI than in subjects with small tumor remnants (60 vs. 79%), though the difference was not statistically significant (P = 0.20). Accordingly, Hu et al. (21) reported that patients who received the lowest CAB dose and presented a significant reduction in tumor size before withdrawal were more likely to achieve the best success (P < 0.001). Likewise, in the series by Karlip et al. (15), size of tumor remnant prior to withdrawal predicted recurrence (18% increase in risk for each millimeter). By contrast, Kwancharoen et al. (18) could not depict any statistically significant clinical predictors of recurrence, but this may have been influenced by the small number of subjects enrolled in their trial (n = 17).

It was previously reported that a nadir PRL at the time of first withdrawal below 5.4 ng/ml was associated with a lower risk of recurrence (16). In our series, although mean PRL levels before withdrawal were lower in non-recurring subjects, individual PRL values in both groups greatly overlapped (8.2–22.5 vs. 9.6–25.2), as shown in Table 3.

Our study has some limitations. Notably, because of the small number of patients, some or most clinical predictors may not have reached statistically significance. It is also possible that some of the patients presently on remission may recur later on. However, recurrences beyond 2 years after CAB withdrawal have only seldom been reported (15–17, 24). In the largest series to date (n = 200), 63 patients (31.5%) had recurrent hyperprolactinemia, 56% during the first year, 33% during the second year, 11% during the third year, and none thereafter (P < 0.001) (16, 24).

As shown in Table 5, a combined analysis of our findings and those from the study by Kwancharoen et al. (18) would indicate that about 31% of a selected group of patients with prolactinomas treated with CAB for at least additional 24 months could benefit from a second drug discontinuation. In four recent previous studies, which involved 292 patients with micro- or macroprolactinomas, the mean recurrence rate after first CAB withdrawal was 49% [range, 25–69% (14–16)]. Therefore, the chance of achieving a successful second withdrawal seems to be similar to the first one. However, further studies with a greater number of patients and longer-term follow-up are clearly warranted.

A likely explanation for the successful second withdrawal could be a longer exposure to CAB therapy and consequently to its antiproliferative and pro-apoptotic effects on pituitary tumor cells, described in several studies (25).

An additional advantage of CAB withdrawal would be to reduce the risk of fibrotic cardiac valvulopathy, which has been associated with the use of DA therapy (26–28), particularly with the higher dosage given to Parkinson’s disease patients (29, 30). However, most studies have not shown an increased prevalence of significant cardiac valve regurgitation in patients with pituitary diseases receiving CAB (13, 31–33).

In conclusion, the findings of the current study yield increasing awareness that a second trial of CAB discontinuation should be attempted in well-selected patients with prolactinomas who display normal PRL levels and tumor remnant <1 cm at low doses (≤1 mg/week). Although our patients have been treated for two or more years, the optimal length of therapy before attempting a second CAB withdrawal still needs to be established. Patients with lower PRL levels and no visible tumor on MRI seemed to be the most likely to benefit from this approach but it was also effective in subjects with small tumor remnants. It is essential however to closely follow-up these patients and reinstitute therapy promptly in symptomatic recurrent ones.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.