| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology (S.G.I.S., A.G., G.T., J.A.H.W.), Oxford Centre for Diabetes, Endocrinology, and Metabolism, Churchill Hospital, Oxford OX3 7LJ, United Kingdom; and Departments of Neuroradiology (J.V.B.), Neuropathology (A.G., N.S., O.A.), and Neurosurgery (S.C.), John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom
Address all correspondence and requests for reprints to: John A. H. Wass, Professor of Endocrinology, Department of Endocrinology, Oxford Centre for Diabetes, Endocrinology, and Metabolism, Churchill Hospital, Oxford OX3 7LJ, United Kingdom. E-mail: john.wass{at}noc.anglox.nhs.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of the study was to determine the incidence of nonsurgical rhinorrhea and identify biochemical, radiological, and histopathological factors associated with leakage.
Design, Setting, and Participants: A retrospective review of MPRL patients (n = 114) was compared with patients with nonfunctioning pituitary adenoma (NFA) (n = 181) seen over a 19-yr period (1985–2004).
Main Outcome Measures: Incidence of CSF rhinorrhea, factors predictive of leakage, and differential expression of candidate markers of invasiveness were measured.
Results: Nonsurgical CSF rhinorrhea occurred in 8.7% of MPRLs (10 of 114) [2.6% spontaneous (three of 114), 6.1% dopamine agonist induced (seven of 114)], whereas no NFAs developed nonsurgical rhinorrhea. There was a clear male preponderance in MPRLs with nonsurgical rhinorrhea (males to females, 9:1, P = 0.008). Dopamine agonist resistance was more frequent in MPRLs with rhinorrhea than with MPRLs without rhinorrhea [30% (n = 10) vs. 5% (n = 104) P = 0.003]. Baseline prolactin levels, rate of prolactin decline in response to dopamine agonists, and tumor volume at diagnosis did not predict CSF leakage. Candidate markers of invasiveness, specifically the protease-activated receptor 1 and e-cadherin expression scores and tumor macrophage density, were not significantly different between groups; MPRL+CSF rhinorrhea (n = 6), MPRL without CSF rhinorrhea (n = 9), and NFAs (n = 9).
Conclusions: The incidence of nonsurgical CSF rhinorrhea in MPRL patients (8.7%) is higher than expected. Dopamine agonist resistance is more common in MPRLs with CSF rhinorrhea; however, whether this is a mechanistic relationship requires further study. Protease-activated receptor 1 expression, e-cadherin expression, and macrophage infiltration rates do not distinguish tumors with from those without CSF rhinorrhea.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In the present study, a large series of patients with macroprolactinomas (MPRLs) and a large group of patients with nonfunctioning pituitary adenoma (NFA) were assessed for the incidence of nonsurgical CSF rhinorrhea and for biochemical and radiological features associated with spontaneous and dopamine agonist-induced CSF rhinorrhea.
This was complemented by the study of candidate markers of invasiveness, in a proportion of patients with suitable biopsy tissue, to test differential expression between tumors associated with CSF rhinorrhea, compared with tumors without rhinorrhea.
The protease-activated receptor 1 (PAR1) is emerging as a key protein in the control of cellular invasiveness and tumor progression (17). It may be activated by matrix metalloproteinase (MMP)-1, secreted by host fibroblasts (18) or act as a novel oncogene (19). Up-regulation of PAR1 expression in tumor cells is associated with tumor cell motility and an invasive phenotype in a range of human tumors (17, 18, 20). Whether PAR1 is associated with an invasive phenotype causing CSF leakage in pituitary adenomas is unknown.
Recent studies also revealed a direct correlation between the extent of tumor infiltration by macrophages, invasiveness, and metastatic potential (21, 22, 23). Tumor-associated macrophages secrete a number of factors (e.g. MMPs) affecting growth and survival of the tumor cells, matrix remodeling, and neoangiogenesis (24). A subgroup of macrophages may produce MMP-9 (22), and their abundance may be associated with invasiveness in prolactinomas (25). Whether total tumor macrophage content or MMP-9 expression levels play a role in the pathology of CSF rhinorrhea is unknown.
A further mechanism involved in tumor cell invasion is complex alterations in adhesive cell-to-cell interactions. One of the key molecules implicated in this process is e-cadherin (26). It has been demonstrated that e-cadherin expression is significantly reduced in invasive prolactinomas, compared with noninvasive ones (27), whereas this phenomenon has not been observed in other functioning pituitary adenomas or in NFAs (28). We therefore chose e-cadherin as another potential marker that may be differentially expressed in tumors associated with and those without CSF rhinorrhea.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Clinical history, biochemical, and histological data were extracted from the patient’s hospital notes and laboratory records.
In this study we included both macroprolactinomas (>1 cm in diameter) and giant prolactinomas (>4 cm in diameter) in the MPRL group. The diagnosis of MPRL was based on tumor size greater than 1 cm in diameter with extrasellar extension and a raised prolactin (>3000 mU/liter) as well as histological verification when surgery was performed. The diagnosis of NFA was supported by the lack of clinical or biochemical evidence of pituitary hormone excess and with postoperative histology and immunostaining in all patients.
In this study we defined dopamine agonist resistance as failure to normalize prolactin levels within a year of dopamine agonist therapy, despite progressively increasing doses.
Biochemistry
The prolactin assay was performed in a single laboratory (Biochemistry Department, John Radcliffe Hospital, Oxford, UK) with a two-site sandwich immunoassay (Advia Centaur; Bayer Corp., Leverkusen, Germany), using direct chemiluminometric technology with constant amounts of two antibodies: first, a Lite reagent polyclonal goat antiprolactin antibody labeled with acridinium ester, and second, a solid-phase monoclonal mouse antiprolactin antibody covalently coupled to paramagnetic particles. The normal range in females is 60–620 mU/liter and in males 45–375 mU/liter.
CSF rhinorrhea was confirmed by the presence of CSF ß-transferrin.
Radiology
Pituitary magnetic resonance imaging (MRI) [or in a minority of patients computed tomography with multiplanar reformations] was performed in three orthogonal planes and read by a single observer (J.V.B.) using a standardized reporting form to collect data including pituitary and tumor dimensions and the direction and degree of extrasellar involvement (modified Hardy criteria). Volumes were calculated as the product of the maximum transverse, sagittal, and vertical tumor diameters in centimeters.
Immunohistochemistry (IHC)
Expression of potential invasiveness markers was tested in all MPRLs with CSF rhinorrhea for which suitable tumor tissue was available (six of 10). Tissue from MPRLs without CSF rhinorrhea and NFAs was selected from patients whose tumor volume on MRI most closely matched that of the index group. Because none of the six index patients had atypical adenomas as defined by the World Health Organization (numerous mitotic figures, MIB-1 index > 3% and widespread p53 oncoprotein overexpression), only typical adenomas were included in the control groups (n = 9 MPRL with no CSF rhinorrhea and n = 9 NFAs). Plurihormonal tumors were excluded.
The tissue was fixed in 4% buffered formalin, dehydrated, and embedded in paraffin wax. Four-micron sections were cut and mounted on SuperFrost Plus slides (VWR, Lutterworth, UK) and dried overnight at 37 C. Sections were dewaxed and rehydrated; the endogenous peroxidase activity was blocked using 3% hydrogen peroxide for 30 min. The primary mouse monoclonal antibodies used were as follows: PAR1 expression, WEDE15 Ab (Immunotech, High Wycombe, UK) (overnight at +4 C, 1:100); to identify tumor-associated macrophages by CD-68 marker, PG-M1 Ab (DakoCytomation, Ely, UK; 1 h at room temperature, 1:50); to identify e-cadherin, NCL-E-Cad Ab (Vision Biosystems Europe Ltd., Newcastle upon Tyne, UK; 1 h at room temperature, 1:50); and for MMP-9 expression, MMP-9 Ab (clone 4H3 and clone 36020; R&D Systems, Abingdon, UK; 1 h at room temperature, 1:30). PG-M1 antibody required antigen retrieval [microwaving with citrate buffer (pH 6.0), 2 x 5 min cycles] before primary antibody treatment. For e-cadherin, the sections underwent pressure cooking for 1 min in 0.01 mM citrate buffer (pH 6.0). To assess the proliferative activity of the tumor cells, we also evaluated the Ki-67 labeling index and p53 expression in each specimen studied. For this purpose, we used MIB-1 (DakoCytomation) and p53 (DakoCytomation) primary antibodies. The primary antibodies were incubated for 1 h at 25 C (MIB-1, 1:40 and p53, 1:100 concentrations, respectively) after antigen retrieval [microwaving with citrate buffer (pH 6.0) for 2x 5 min cycles]. Positive controls for the antibodies studied were as follows: breast carcinoma cells (WEDE15), squamous epithelial cells (e-cadherin), and tonsil germinal center macrophages (PGM-1). Negative controls were obtained in cases in which primary antibodies were not applied. After the application of primary antibodies WEDE15, PGM1, MIB-1, and p53, sections were washed twice with Tris-buffered saline (pH 7.6) (containing 0.005% Tween 20). DakoCytomation REAL EnVision peroxidase polymer was applied as a secondary antibody at 25 C for 35 min. The sections were buffer washed and then treated with DakoCytomation REAL EnVision DAB as per the manufacturer’s guidelines for color development. The slides were lightly counterstained with hematoxylin, and the procedure was completed by dehydration, clearance, and mounting the slides in DPX for examination. The detection of e-cadherin was performed with VECTASTAIN Elite ABC kit (Vision Biosystems).
Assessment of IHC results
All the slides were blinded by labeling (N.S.) with random case numbers to avoid bias. All assessments were done by two independent observers (A.G. and O.A.) who had no knowledge of the tumor type or the group to which the specimen belonged. Complete concordance existed between the observations made. PAR1 expression pattern was cytoplasmic in all samples (Fig. 1A
). The intensity of PAR1 expression was semiquantitatively graded as nil (0), weak (1+), moderate (2+), or intense (3+). Positive staining for PAR1 was considered as either patchy (presence of islands of tissue with and without staining) or diffuse (all islands showing positive staining). E-cadherin expression pattern was primarily membranous and to a lesser extent, cytoplasmic. The membranous pattern was considered in the evaluation (Fig. 1B). The intensity of staining was semiquantitatively evaluated as described above. The extent of tumor macrophage infiltration (shown in Fig. 1C), was assessed using a quantitative method. The surface area (square millimeters) of a representative part of the tumor was calculated using digital image software (Axiovision, release 4.4, Carl Zeiss, Jena, Germany). After that, all PGM-1 positive cells in that particular area (in high power fields) were counted. Tumor macrophage density (cell number per square millimeter) was obtained by dividing the number of positive cells by the surface area. MMP-9-positive macrophages were also assessed in high-power fields. The Ki-67/MIB-1 labeling index and p53 expression were calculated as the percentage of positive nuclei observed in several representative high-power fields.
|
The students’ t test (to compare means), 
2 test (to compare proportions), and Kruskal Wallis test (to compare the medians among three groups who underwent IHC studies) was used to compare values. The level of significance was set at P 
0.05.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
One hundred fourteen patients with MPRL (57 males, 57 females) and 181 patients with NFA (110 males, 71 females) were identified. Ten patients with MPRL (8.7%) developed nonsurgical CSF rhinorrhea, whereas no patient with a NFA developed nonsurgical rhinorrhea (Table 1).
|
Despite no male preponderance in our series of 114 patients with MPRL (male/female, 1:1), patients with nonsurgical CSF rhinorrhea had a clear male preponderance (nine of 10, P = 0.008).
Tumor volume at diagnosis was not significantly different between MPRL with rhinorrhea and MPRL without rhinorrhea [3.18 cm3 (SD 2.7) vs. 4.16 cm3 (SD 3.7) P = 0.17]. Sphenoid sinus invasion is significantly higher in MPRL with no CSF rhinorrhea when compared with those with CSF rhinorrhea (100 vs. 60%, P = 0.02). Skull base invasion, as assessed by MRI evidence of infiltration of the sphenoid, ethmoid, clivus, or infrasellar extension (without sphenoid infiltration), was not significantly different between MPRL with and without CSF rhinorrhea because all tumors matched for volume had skull base infiltration (100 vs. 100%, P = 1.00).
Paired MRI scans (baseline and after leakage but before surgery) were available in four of seven patients with dopamine agonist-induced CSF rhinorrhea. Mean tumor volumes at baseline were 50 cm3, whereas after leakage the mean volume was 39.1 cm3 (P = 0.54).
Prolactin at diagnosis and the rate of drop of prolactin after dopamine agonist therapy were not significantly different between MPRL patients with and those without nonsurgical rhinorrhea (Table 2). Rate of drop of prolactin was not significantly different between subjects who developed dopamine agonist-induced CSF rhinorrhea and patients with no CSF leak (52,178 mU/liter/month vs. 16,572 mU/liter/month, P = 0.099) (Table 2).
|
The demographic characteristics of patients with MPRL and CSF rhinorrhea are displayed in Table 3. In three MPRL patients (2.6%), CSF rhinorrhea occurred spontaneously and was the presenting clinical feature; two of these patients initially presented with meningitis requiring high dose iv antibiotics. In two patients (67%), the leakage stopped within 6 wk of starting dopamine agonist therapy. The third patient required a transsphenoidal adenomectomy and sphenoid packing to resolve the CSF rhinorrhea.
|
Macrophage density and PAR1, e-cadherin, and MMP-9 expression
PAR1 and e-cadherin scoring and macrophage densities in all three groups (MPRL with CSF rhinorrhea, MPRL without CSF rhinorrhea, and NFAs) are given in Table 4. PAR1 was expressed in all samples studied, scores ranging between 1 and 3. Median PAR1 and e-cadherin scores were not significantly different among all three groups (P = 0.36 and 0.20, respectively) (Table 4). The frequency of patchy distribution of PAR1 was also similar among all groups (MPRL+CSF rhinorrhea, 50%; MPRL without CSF rhinorrhea, 33%; NFA, 55.5%, P = 0.68). MPRLs (both with and without CSF rhinorrhea) tended to have higher macrophage densities than the NFA group, but the difference was not statistically significant (P = 0.58). Too few macrophages expressed MMP-9 to perform any meaningful quantification and data analysis. Tumor cells were consistently MMP-9 negative.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This is the first study to assess systematically the incidence of nonsurgical CSF rhinorrhea (8.7%) in a large series of subjects with MPRL. Most cases were secondary to dopamine agonist therapy (6.1%), but some occurred spontaneously (2.6%). Of the latter group, two patients presented with meningitis and rhinorrhea as the initial presentation.
In our series of 114 MPRL patients, the incidence of dopamine agonist-induced CSF rhinorrhea was 6.1%; in agreement with a reported incidence of 7% in a smaller series of 42 patients (6). Dopamine agonist resistance was significantly higher in MPRL patients with CSF rhinorrhea (30 vs. 5%, P = 0.003). These factors have not been previously reported. Paired MRI scans at baseline vs. at dopamine agonist-induced leakage (50 vs. 39.1 cm3, P = 0.54) were not significantly different; however, only four paired scans were available for review.
An unexpected finding is that tumors with CSF leakage have less sphenoid sinus invasion on MRI than tumors with no CSF leakage matched for size. This statistical difference is almost certainly due to the small sample size especially in the group of tumors with CSF leakage (n = 10).
Macroprolactinomas are reported to occur more frequently in men, and giant prolactinomas occur almost exclusively in males. Our series of MPRL patients shows an equal sex distribution at presentation (male to female, 1:1). This is in agreement with a large series of prolactinomas reported by Colao et al. (30). Despite the identical sex distribution within the whole group of MPRLs, there was a clear male preponderance within the group with CSF rhinorrhea (nine of 10: P = 0.008).
The reported male preponderance of MPRL has been put down to a combination of delay in diagnosis because symptoms are often more subtle and to histologically more aggressive tumors in men with higher indices of cell proliferation as measured by Ki-67 immunoreactivity using MIB-1 antibody staining (31) and lower cell adhesion molecules such as e-cadherin and 
-catenin (27). E-cadherin and ß-catenin expression is reduced in invasive MPRLs, and this correlates with invasiveness and more aggressive behavior, especially in men, compared with women (27). MPRLs reportedly grow more rapidly in men and are often invasive with a higher incidence of bromocriptine resistance (31).
We had expected that increased macrophage density, increased PAR1 expression, and reduced staining for cell adhesion molecule e-cadherin might be predictive of tissue invasiveness and therefore CSF leakage. However, we could not demonstrate significant differences in these markers among the three groups. Although macrophage density tended to be higher in prolactinomas, compared with NFAs, the difference did not reach statistical significance. It should be noted, however, that the low number of cases with CSF rhinorrhea for which tissue was available might have precluded us from finding a significant difference among groups; this warrants further investigations in a larger series.
Macrophages expressing MMP-9 were rarely found; we were therefore unable to assign invasiveness of these tumors to enhanced MMP-9 expression. However, it is known from studies in other tumors that molecules mediating tumor invasion may be expressed preferentially at the invasive front (periphery) rather than the tumor core (32). The nature of selective adenomectomy or incomplete resection due to surgical inaccessibility of the invasive component introduces assessment bias toward the tumor core. It is also noteworthy that the majority of our CSF rhinorrhea cases with available tissue were treatment-induced; only two samples from patients with spontaneous CSF rhinorrhea could be studied. It is likely that the mechanisms of treatment-induced leakage (see below) differ from those associated with spontaneous CSF leakage, with only the latter potentially being solely associated with differential degrees (molecular or microanatomical) of invasiveness. Ideally, a more comprehensive, gene chip-based molecular study should be conducted once sufficient fresh tissue of prolactinomas with and without CSF leakage becomes available. Such an approach may identify new candidate markers able to differentiate these groups.
Pretreatment prolactin levels have been reported to be positively correlated with microvessel density, suggesting increased angiogenesis in MPRLs, which often have very high prolactin levels (33). We hypothesized that pretreatment prolactin may predict the risk of CSF rhinorrhea; however, in our series there was no significant difference in baseline prolactin levels between groups of MPRL patients with or without rhinorrhea.
CSF rhinorrhea is a recognized side effect of dopamine-agonist treatment of prolactinomas. This is assumed to be due to tumor shrinkage unplugging defects in the dura and bone, resulting in fistulae and CSF leakage. Discontinuation of dopamine agonist therapy may stop the leak presumably by tumor reexpansion plugging the defects in the dura and bone (6, 9). As dopamine agonist-induced rhinorrhea is an accepted phenomenon, presumed to be due to rapid tumor shrinkage, we hypothesized that a rapid reduction in serum prolactin may predict CSF rhinorrhea. However, the rate of fall of prolactin was not significantly different among the groups of MPRL patients with and without CSF rhinorrhea (P = 0.07).
Dopamine agonists have now been in use for more than three decades and are the mainstay of treatment for prolactinomas (34, 35, 36). They are successful in reducing prolactin levels and inducing tumor shrinkage in most MPRL patients. Dopamine agonist resistance is defined as failure to normalize prolactin levels or failure to decrease tumor size by 50% or more and occurs in about 10% of prolactinomas (37). Dopamine agonist resistance is attributed to lower dopamine (type 2 iodothyronine deiodinase) receptor expression on the tumor cell surface (38). In our series, MPRLs resulting in CSF rhinorrhea had a significantly larger proportion that was resistant to dopamine agonists (P < 0.003). The mechanism underlying dopamine agonist resistance and whether it relates etiologically to CSF rhinorrhea is unclear and requires further investigation.
Zornitzki et al. (39) reported a case of MPRL with secondary dopamine agonist resistance associated with aggressive tumor behavior and increasing markers of cell division, suggesting that dopamine agonist resistance may signify tumor dedifferentiation. Increased tumor invasiveness in such patients may therefore result in an increased risk of CSF rhinorrhea. This is further supported by a study by Delgrange et al. (40) reporting that invasive MPRL as assessed by intracavernous internal carotid artery encasement may predict a negative response to dopamine agonists. It is therefore important to assess dopamine agonist resistance in MPRL subjects because this appears to not only increase the risk of CSF rhinorrhea but also may signify more aggressive tumors. We should note, however, that tumor volume was not different between MPRL patients with and without rhinorrhea, and none of the patients with dopamine agonist resistance in our cohort has had an aggressive course with multiple recurrences.
Management of subjects with macroprolactinoma and CSF rhinorrhea
The majority of patients with dopamine agonist-induced rhinorrhea (71%) required one or more surgical procedures to stop the leakage; however, two patients had resolution of CSF rhinorrhea on continuing dopamine agonist therapy while awaiting surgical treatment (within 2 wk of the diagnosis of rhinorrhea). Two patients with spontaneous CSF rhinorrhea also had resolution of rhinorrhea on dopamine agonist therapy, although the mechanism of resolution of CSF rhinorrhea is unclear. Therefore, contrary to previous reports (41), surgical therapy may not always be necessary.
Although counterintuitive, the resolution of CSF rhinorrhea on continued dopamine agonist therapy has been reported in isolated reports (9, 42, 43). A plausible hypothesis would be alteration in CSF pressure dynamics reducing the raised intracranial pressure resulting from a large MPRL, thereby allowing spontaneous healing of the CSF fistula (5). Our data importantly suggest therefore that it is not necessary to stop dopamine agonist therapy in subjects with MPRL who develop CSF rhinorrhea.
Conclusion
This is the first large series to ascertain the incidence of spontaneous CSF rhinorrhea in 114 patients with macroprolactinoma (8.7%). This is also the largest series to report the 6.1% incidence of dopamine agonist-induced CSF rhinorrhea to date. This suggests that dopamine agonist-induced CSF rhinorrhea is either underrecognized or underreported.
Males are significantly more likely to develop spontaneous and dopamine agonist-induced CSF rhinorrhea. Dopamine agonist resistance was higher in MPRL patients with CSF rhinorrhea. Whether the difference in dopamine agonist resistance is related to the mechanism of prolactinomas inducing CSF rhinorrhea requires further study.
Candidate molecules associated with tumor invasiveness PAR1, e-cadherin, and MMP-9, as well as total tumorassociated macrophage number, failed to distinguish, in our cohort, between adenomas with CSF rhinorrhea from those without. Classical markers of tumor aggressiveness such as proliferative activity (Ki-67) and p53 oncoprotein expression also did not differ between groups. CSF rhinorrhea is a potentially serious complication of MPRL and dopamine agonist therapy; therefore, all patients diagnosed with a MPRL should be forewarned of this risk and monitored closely. It is not necessary to stop dopamine agonist therapy in subjects with MPRLs who develop CSF rhinorrhea.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Summary: S.G.I.S., A.G., N.S., G.T., S.C., O.A., and J.A.H.W. have nothing to declare. J.V.B. consults for Cordis Neurovascular and Boston Scientific Corp.
First Published Online July 10, 2007
1 S.G.I.S. and A.G. contributed equally to this work. 
Abbreviations: CSF, Cerebrospinal fluid; IHC, immunohistochemistry; MMP, matrix metalloproteinase; MPRL, macroprolactinoma; MRI, magnetic resonance imaging; NFA, nonfunctioning pituitary adenoma; PAR1, protease-activated receptor 1.
Received February 20, 2007.
Accepted July 2, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
(HIF-2) expression in tumor-infiltrative macrophages to tumor angiogenesis and the oxidative thymidine phosphorylase pathway in human breast cancer. Cancer Res 62:1326–1329-, ß-, and 
-catenins, and p120 (cell adhesion molecules) in prolactinoma behavior. Mod Pathol 15:1357–1365[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  S. K. Mankia, R. A. Weerakkody, S. Wijesuriya, N. Kandasamy, F. Finucane, M. Guilfoyle, N. Antoun, J. Pickard, and M. Gurnell Spontaneous cerebrospinal fluid rhinorrhoea as the presenting feature of an invasive macroprolactinoma BMJ Case Reports, May 21, 2009; 2009(may21_1): bcr1220081383 - bcr1220081383. [Abstract] [Full Text]
|
|
|
|
 |
|
 E. Delgrange, T. Daems, J. Verhelst, R. Abs, and D. Maiter Characterization of resistance to the prolactin-lowering effects of cabergoline in macroprolactinomas: a study in 122 patients Eur. J. Endocrinol., May 1, 2009; 160(5): 747 - 752. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
