Introduction

Transsphenoidal surgery has long been considered the treatment of choice for acromegaly [1–3]. However, even in experienced hands, the surgical intervention is successful in only about 50% of cases as most patients have large or invasive tumors that cannot be removed completely [4]. Therefore, many patients require subsequent adjunct therapy, particularly medical treatment [5, 6]. Long-acting somatostatin analogs (SSAs) have been widely accepted as the best medical option to treat acromegaly [7]. Nevertheless, at least 35% of patients are considered to be resistant to the commercially available SSAs (Octreotide LAR and Lanreotide autogel) as they do not achieve insulin-like growth factor I (IGF-I) normalization while treated with these drugs [7, 8]. Pegvisomant, a new drug that acts as a growth hormone (GH) receptor antagonist, normalizes IGF-I levels in up to 97% of patients but it does not induce tumor shrinkage [9, 10]. A potent dopamine agonist, cabergoline, was shown to normalize plasma IGF-I levels in up to 39% of acromegalic patients treated with this drug as monotherapy [11]. Cabergoline efficacy is greater in subjects with concomitant hyperprolactinemia and when the elevation of IGF-I and GH levels is only mild or moderate [8, 11, 12].

There is still limited evidence towards the beneficial effects of adding cabergoline to the medical treatment regimen in acromegalic patients uncontrolled with long-acting SSAs. In four previous studies [13–16], totalizing 81 patients, the rate of IGF-I normalization ranged from 37% to 56% and was not dependent on prolactin (PRL) status (serum PRL levels and immunohistochemical staining for PRL in tumor cells).

The aim of this study was to evaluate the efficacy in normalizing IGF-I levels of the addition of cabergoline to the management of a large cohort of acromegalic patients partially responsive to longterm treatment with the long-acting SSA octreotide LAR (OCT-LAR).

Materials and methods

Patients

Fifty-two acromegalic patients (24 men and 28 women; mean age, 48.4 ± 14.5 years; age range, 23–71), routinely followed in 4 Brazilian neuroendocrine centers, were included in the current study. The main inclusion criteria for this prospective open trial was the persistence of increased IGF-I levels for age despite treatment with OCT-LAR (30 mg intramuscularly [IM]) for at least 12 months. The main exclusion criteria was the presence of optic chiasm compression by the tumor.

Study protocol and assays

The study was performed in accordance with the declaration of Helsinki and was approved by the local ethics committee. All study participants provided informed consent before enrollment had been reached.

The study protocol included the evaluation of the effectiveness of the addition of cabergoline to the medical treatment regimen in acromegalic patients partially responsive to OCT-LAR. The study was divided into three phases:

  • Phase I: evaluation of the response of IGF-I and GH levels to the short-term cabergoline treatment in addition to OCT-LAR

Cabergoline was given orally at night to the 52 patients for 6 months. The drug was started at the dose of 1.0 mg/week (0.5 mg twice weekly) which was progressively increased by 0.5–1.0 mg every six weeks until normalization of plasma IGF-I levels was obtained, unacceptable side-effects occurred or a maximum dose of 3 mg/week was reached.

Hormonal evaluations (IGF-I, GH, PRL) were performed before the first dose of cabergoline was administered and repeated at 6-week intervals after each cabergoline dose increment.

  • Phase II: evaluation of the long-term response of IGF-I levels to the addition of cabergoline to OCT-LAR treatment

This part of the study was designed to investigate whether the short-term responses observed in Phase I would be sustained during a long-term follow-up. Thus, all responsive patients were given cabergoline at the lowest doses that normalized their IGF-I levels in Phase I and were followed with clinical and hormonal evaluations at 6-month intervals for at least 12 months. Octreotide and cabergoline doses were kept unchanged during follow-up.

  • Phase III: evaluation of response of IGF-I levels to cabergoline as monotherapy after long-term combined treatment

This phase of the study was designed to assess whether patients who normalized IGF-I under treatment with cabergoline in addition to OCT-LAR would sustain normal IGF-I levels with cabergoline alone. Thus, after at least 12 months of continuous combined treatment and sustained normalization of IGF-I levels, whenever possible OCT LAR was withdrawn in responsive patients for at least 6 months and IGF-I levels were assessed at 3- to 6-month intervals.

Treatments prior to the addition of cabergoline included transsphenoidal surgery (TSS) and OCT-LAR in 25 patients (48.1%), primary therapy with OCT-LAR in 19 patients (36.5%), and combination of surgery, conventional radiotherapy and OCT-LAR in 8 subjects (15.4%). OCT-LAR was previously administered at a dose of 30 mg every 28 days IM, for a median time of 18 months (range 12–38). TSS and radiotherapy were performed 6–16 years (median, 8 years) and 9 to 14 years (median, 11.5 years), respectively, before the addition of cabergoline.

Patients were assessed clinically and by hormonal determinations every month. Patients who achieved normalization of IGF-I levels after the addition of cabergoline were labeled as responsive whereas those who persisted with increased IGF-I levels were considered non-responsive.

Serum GH levels were measured by a two-site chemiluminescent immunometric assay (Immulite 2000). The sensitivity of this assay is 0.01 ng/ml and the interassay coefficient of variation was <10%. Serum IGF-I was measured by a solid-phase enzyme labeled chemiluminescent immunometric assay with the sample pre-treatment in an on-bord dilution step (Immulite 2000). The manufacturer’s normal range for age was considered. The interassay coefficients of variation were <5%. IGF-I levels are expressed both as absolute values and as a percentage of the upper limit of normal age-matched range (%ULNR-IGF-I; normal: ≤100%). Serum PRL was measured using a two-site chemiluminescent immunometric assay (Immulite 2000). Normal values for PRL, as indicated by the assay method, ranged from 2.0 to 17.0 ng/ml in women and from 2.0 to 15.0 ng/ml in men. The interassay coefficient of variation was less than 5%. Hyperprolactinemia was defined as PRL levels above the normal range.

The radiological study included the evaluation of the sellar region by magnetic resonance imaging (MRI) [Sigma LX GE, Milwaukee, WI], 1.5T, and gradient of 23 mT/m. The slices were axial, coronal, and sagittal in T1, pre- and post-gadolinium, and in T2. The pituitary MRI was performed in all patients prior to the addition of cabergoline to the treatment regimen. Furthermore, in order to investigate the presence or to follow the development of a cardiac valve disease, an echocardiography was performed in 16 patients at study entry and after 6 months of combined treatment as well as in responsive patients every 6 months during the phase II of the study.

The hormonal assays were performed in the same centralized laboratory whereas MRI and ecocardiography were examined by highly skilled specialists in each center involved in the study.

Statistics

For comparison of categorical variables, the chi-squared test or the Fisher exact test were used where appropriate. The Student’s t test was performed for the comparative analysis of quantitative variables. Results are expressed as percentages and mean values ± SD or unless otherwise indicated. All results were considered significant if two-sided P value was less than 0.05.

Results

Characteristics of the patients

At enrollment in the current study, all patients had increased IGF-I levels and the %ULNR-IGF-I ranged from 107.14 to 428.57% (mean ± SD, 222.31 ± 85.20). Random serum GH levels ranged from 2.6 to 35.1 ng/ml (mean ± SD, 7.83 ± 7.11). Hyperprolactinemia with PRL values ranging from 27.9 to 252.8 ng/ml (mean ± SD, 117.80 ± 76.4) was also present in 17 patients (32.7%) (Tables 1, 2).

Table 1 Baseline characteristics of the 52 acromegalic patients under octreotide LAR treatment before the addition of cabergoline
Full size table
Table 2 Comparison of hormone levels during treatment with octreotide LAR (OCT-LAR) only and the combination OCT-LAR/cabergoline (CAB)
Full size table

The immunohistochemical (IHC) features of the tumors were available in only 15 patients previously submitted to TSS. All these adenomas stained positively for GH and 7 of them (46.7%) were also positive for PRL.

At the time of diagnosis, pituitary MRI depicted a macroadenoma (tumor diameter >1 cm) in all patients. Before the addition of cabergoline, pituitary MRI detected macroadenoma remnants in all patients who previously underwent transsphenoidal surgery and intrasellar macroadenomas in 19 patients previously submitted to primary therapy with OCT-LAR (4 of them with parasellar extension). None of the patients had visual field defects.

Efficacy of the short-term combination therapy (octreotide LAR + cabergoline)—phase I of the study

As shown in Table 2, the combination therapy resulted in a significant reduction in mean GH, IGF-I and PRL levels. Mean GH levels decreased from 7.83 ± 7.11 to 4.74 ± 5.51 ng/ml (P < 0.001) whereas IGF-I decreased from 584.66 ± 169.58 to 349.73 ± 161.74 ng/ml (P < 0.001). After the addition of cabergoline at doses ranging from 1.0 to 3.0 mg/week (mean, 2.19 ± 0.64) normalization of IGF-I levels was achieved in 21 patients (40.4%) whereas in the remaining patients the mean reduction of IGF-I levels was 24.8 ± 19.9% (range, 3.9–56.9). Moreover, safe serum GH levels (GH < 2.5 ng/ml) were observed in 24 patients (46.1%). Finally, both findings (normal IGF-I levels and safe GH levels) were depicted in 15 patients (28.8%). Among responsive patients the final mean %ULNR-IGF-I levels were 57.7 ± 16.7% (range, 24.0–88.44). The corresponding values for non-responsive patients were 175.76 ± 54.51% (range, 111.74–279.54). It should also be noted that no escape phenomenon was observed after cabergoline addition.

PRL levels also changed significantly (P = 0.001) during cabergoline treatment, decreasing from 49.51 ± 65.92 to 10.35 ± 12.65 ng/ml. However, serum PRL remained above the normal range (between 25.8 and 71.58 ng/ml) in 3 patients (5.7%) despite the use of cabergoline in doses up to 3 mg/week.

Among responsive patients, the final dose of cabergoline was 1.0 mg/week in 1 (4.7%), 1.5 mg/week in 4 (19.1%), 2.0 mg/week in 9 (42.8%) and 3.0 mg/week in 7 (33.4%). As mentioned, the mean final dose of cabergoline was 2.19 ± 0.64 mg/week.

Efficacy of the long-term combination therapy (octreotide LAR + cabergoline)—phase II of the study

As shown in Fig. 1, all 21 short-term responsive patients sustained normal IGF-I levels during long-term follow-ups ranging from 12 to 24 months (median: 15). Indeed, the final mean %ULNR-IGF-I levels were 57.9 ± 21.2% (range, 23.5–94.37). No escape of treatment was observed during long-term follow-ups.

Fig. 1
figure 1

Behaviour of IGF-I levels in responsive patients during short-term and long-term follow-up under combination treatment (OCT octreotide LAR)

Full size image

IGF-I levels after octreotide LAR withdrawal in long-term responders to combined treatment—phase III of the study

The efficacy of cabergoline as monotherapy, after OCT-LAR withdrawal, was assessed in 11 patients who reached normalization of IGF-I levels while on combined therapy. The other patients were kept on OCT-LAR and the reason was heart failure in 3 cases, poorly controlled diabetes mellitus in 4 and patients’ refusal in 3 cases. Six months after OCT-LAR withdrawal, IGF-I levels were increased in all patients (mean %ULNR-IGF-I of 148.27 ± 23.54%).

Predictors of response to cabergoline addition to octreotide LAR

Compared to non-responsive subjects, responsive patients (the ones who had IGF-I levels normalized) had lower mean %ULNR-IGF-I baseline values (167.39 ± 55.86 vs. 256.39 ± 83.44%; P < 0.0001) and lower mean baseline GH levels (5.99 ± 4.94 vs. 10.37 ± 6.16 ng/ml; P < 0.0001). Although PRL levels were higher in responsive patients (80.21 ± 90.96 vs. 27.44 ± 28.44 ng/ml; P < 0.0001), the rate of hyperprolactinemia did not significantly differ (38.1% vs. 29.0%; P = 0.4361) when both groups were compared. Furthermore, the rate of positivity for PRL in immunohistochemical analysis was similar in responsive and non-responsive patients (50.0% vs. 42.8%; P = 0.5952) (Table 3).

Table 3 IGF-I, GH and prolactin (PRL) levels, as well as the rate of positivity for PRL in immunohistochemical analysis, in responsive and non-responsive atients to combined therapy
Full size table

As shown in Fig. 2, there was a great overlap in IGF-I levels between responsive and non-responsive patients before the addition of cabergoline. However, no patients with %ULNR-IGF-I greater than 250% benefited from combined therapy. Moreover, better results concerning normalization of IGF-I levels were found in patients with %ULNR-IGF-I up to 220% (63.6%), in comparison with those with %ULNR-IGF-I ranging from 221 and 250% (38.9%) and the whole group of patients (40.4%). We also found that 12 of the 21 responsive patients (57.1%) had GH levels up to 5 ng/ml. By contrast, 74.2% of non-responsive subjects presented with GH levels >5 ng/ml. Furthermore, among the 20 patients with GH levels up to 5.0 ng/ml, 12 (60%) normalized IGF-I during cabergoline. Finally, the highest rate of IGF-I normalization (73.3%) was found in the subgroup of patients with %ULNR-IGF-I ≤220% and random GH levels ≤5 ng/ml (Fig. 3).

Fig. 2
figure 2

Comparison of IGF-I levels in responsive and non-responsive patients to combination treatment

Full size image
Fig. 3
figure 3

Rates of IGF-I normalization after the addition of cabergoline according to different criteria

Full size image

As shown in Table 4, the effectiveness of the addition of cabergoline was apparently not influenced by the type of previous therapy. Indeed, the rate of normalization of IGF-I was 40.0%, 42.1% and 33.1% in patients previously submitted to TSS and OCT-LAR, OCT-LAR primary therapy or the combination of TSS, radiotherapy and OCT-LAR and, respectively.

Table 4 The number of responsive patients (with normalization of IGF-I) and non-responsive patients (without normalization of IGF-I) after the addition of cabergoline according to the previous type of therapy
Full size table

Tolerability

The combination of cabergoline and OCT-LAR was well tolerated and no patients had to interrupt the treatment. After the addition of cabergoline, transient moderate dizziness and/or nausea were reported in 10 patients (19.2%), whereas 5 patients (9.6%) presented with transient headache. Moreover, none of the patients who underwent echocardiograms developed cardiac valve regurgitation.

Discussion

The dopamine agonist bromocriptine was the first drug used in the management of acromegaly but the results were disappointing and normalization of IGF-I levels was only observed in about 10% of patients, despite the use of very high doses [17]. More recently, some studies, involving about 130 patients, have examined cabergoline therapy for acromegaly and its efficacy varied considerably in these reports [11, 18–21]. In the largest study published to date, 64 unselected patients were treated for up 40 months and the use of cabergoline (at doses ranging 1.0–3.5 mg weekly) resulted in IGF-I normalization in 39% of cases [11]. Better results were found in cases of GH-PRL co-secreting adenomas [11].

The rational for using combination drug therapy in acromegaly is the fact that at least one third of patients are resistant to the treatment with somatostatin analogs (SSAs) which are considered the best medical therapy for the disease [5–7]. In the current study, we observed that the addition of relatively low doses of cabergoline (1.0–3.0 mg/week; mean, 2.19 ± 0.64) to the treatment of acromegalic patients partially responsive to longterm therapy with OCT-LAR was able to induce normalization of IGF-I levels in 40.4% of cases. Moreover, that response was not dependent on the presence of hyperprolactinemia or PRL staining in tumors cells by immunohistochemistry. The best predictors of benefit from combination therapy were IGF-I levels that were significantly lower among responsive patients. Furthermore, no patients with %ULNR-IGF-I >250% normalized IGF-I levels. Normalization of IGF-I levels after the addition of cabergoline was strongly associated with %ULNR-IGF-I ≤ 220% and/or serum random GH levels ≤5.0 ng/ml under octreotide treatment. The efficacy of cabergoline in normalizing IGF-I levels in our study increased progressively according to the criteria adopted to select and analyze patients: 40.4% (overall), 60% (only GH ≤5 ng/ml), 63.6% (only %ULNR-IGF-I ≤220%) and 73.3% (GH ≤5 ng/ml and %ULNR-IGF-I ≤220%). Similar results were reported by Mattar et al. [15] On the other hand, the effectiveness of the addition of cabergoline was not influenced by the type of previous therapy.

In agreement with our results, 4 previous studies [13–16], involving 81 patients, have shown that the rate of IGF-I normalization with combined therapy with OCT-LAR and cabergoline ranged from 37 to 56% and this response was also not influenced by PRL status. The largest of these series evaluated 34 patients with a mean baseline %ULNR-IGF-I of 152 ± 0.62% and found that, using cabergoline doses ranging from 1.5 to 3.5 mg/week (mean, 2.38 ± 1.00), 56% of patients reached normal IGF-I levels [16]. The lowest response rate observed in our study would probably be related to the highest IGF-I levels depicted in our patients (mean %ULNR-IGF-I of 222.31 ± 85.20%). Moreover, we evaluated a larger cohort of patients and our maximal dose of cabergoline was 3.0 mg/week. In these series, as shown in Table 5, patients with baseline %ULNR-IGF-I up to 220% were those who benefited the most from combination treatment. In a large series of acromegalic patients treated with cabergoline alone, pretreatment IGF-I levels below 750 ng/ml, which roughly corresponds to our cutoff of 220% for the %ULNR IGF-I, were also associated with a better response to cabergoline.

Table 5 Effectiveness of adding cabergoline (CAB) to medical treatment of acromegalic patients partially responsive to octreotide LAR (data from 5 studies)
Full size table

In our protocol, we limited the maximal dose of cabergoline to 3 mg/week as there is evidence that high doses of this drug may be associated with increased risk of cardiac valvular dysfunction, as shown in patients with Parkinson’s disease [22]. Moreover, in the series by Colao et al., [23] moderate tricuspide regurgitation without clinical significance was found in patients with prolactinomas chronically treated with cabergoline, particularly those who were given higher doses of cabergoline. However, four other studies showed no differences in valvular abnormalities in patients with hyperprolactinaemia treated with cabergoline compared with controls [24–27]. On the other hand, it has been shown that patients with uncontrolled acromegaly may be at high risk of developing progressive mitral valvular regurgitation [28]. In our study, no valvular cardiac disease was observed in patients treated with cabergoline.

The mechanism underlying the improvement of hormonal response to SSA/cabergoline combined treatment has not been defined yet [16]. Some explanations have been proposed such as a ligand-induced heterodimerization between somatostatin receptors (SSTRs) and the dopamine receptor subtype 2 (D2R) [16]. Indeed, hetero-oligomerizations between SSTRs or between SSTRs and other G protein-coupled receptors, such as D2R, have already been reported [29]. Furthermore, some adenomas expressing D2R and the subtype 2 of SSTRs (SSTR2) show a better inhibition of GH secretion after the addition of cabergoline to SSAs treatment [30].Moreover, recently developed chimeric compounds with high affinity for both dopamine and somatostatin receptor subtypes have shown increased GH-suppressing potency as compared to the combination of SSTR and DR subtype-preferential compounds [31–33].

In a previous study, 35% of patients reached normalization of IGF-I levels after increasing the dose of octreotide LAR from 30 to 40 mg monthly [34]. However, in the great majority of studies the maximum dose of octreotide LAR was 30 mg/month [5–7, 13–16, 35–38]. An alternative to OCT-LAR/cabergoline combined therapy would be the combination of OCT-LAR with pegvisomant [39, 40] that is the most effective drug in normalizing plasma IGF-I levels [8–10]. However, pegvisomant is a very expensive drug. Moreover, transient liver enzyme elevations occur in up to one third of patients who receive that treatment regimen [39].

In conclusion, our findings demonstrated that the addition of cabergoline, even at relatively low doses (1.0–3.0 mg/week; mean, 2.19 ± 0.64), is effective in both short- and long-term control of IGF-I levels in acromegalic patients partially responsive to octreotide LAR, particularly in those with mild/moderately elevated GH/IGF-levels, irrespective of prolactin status.