4cola.oa

The new england journal of medicine Withdrawal of Long-Term Cabergoline Therapy Annamaria Colao, M.D., Ph.D., Antonella Di Sarno, M.D., Ph.D., Paolo Cappabianca, M.D., Carolina Di Somma, M.D., Ph.D., Rosario Pivonello, M.D., Ph.D., and Gaetano Lombardi, M.D., Ph.D.
b a c k g r o u n d
Whether the withdrawal of treatment in patients with nontumoral hyperprolactinemia, From the Departments of Molecular and Clinical Endocrinology and Oncology, Sec- microprolactinomas, or macroprolactinomas is safe and effective has been unclear. tion of Endocrinology (A.C., A.D.S., C.D.S., We performed an observational, prospective study of cabergoline (a dopamine-recep- R.P., G.L.), and Neurologic Sciences, Sec-tor agonist) withdrawal in such patients.
tion of Neurosurgery (P.C.) — all at Feder-ico II University of Naples, Naples, Italy. Ad-dress reprint requests to Dr. Colao at the The study population included 200 patients — 25 patients with nontumoral hyperpro- docrinology and Oncology, Federico II Uni-lactinemia, 105 with microprolactinomas, and 70 with macroprolactinomas. Withdraw- versity, via S. Pansini 5, 80131 Naples, Italy, al of cabergoline was considered if prolactin levels were normal, magnetic resonanceimaging (MRI) showed no tumor (or tumor reduction of 50 percent or more, with the tu- N Engl J Med 2003;349:2023-33.
mor at a distance of more than 5 mm from the optic chiasm, and no invasion of the Copyright 2003 Massachusetts Medical Society.
cavernous sinuses or other critical areas), and if follow-up after withdrawal could becontinued for at least 24 months.
Recurrence rates two to five years after the withdrawal of cabergoline were 24 percentin patients with nontumoral hyperprolactinemia, 31 percent in patients with micropro-lactinomas, and 36 percent in patients; with macroprolactinomas. Renewed tumorgrowth did not occur in any patient; in 10 female patients (22 percent) and 7 male pa-tients (39 percent) with recurrent hyperprolactinemia, gonadal dysfunction redeveloped.
In all diagnostic groups, prolactin levels at the time of recurrence were significantlylower than at diagnosis (P<0.001). The Kaplan–Meier estimated rate of recurrence atfive years was higher among patients with macroprolactinomas and those with micro-prolactinomas who had small remnant tumors visible on MRI at the time of treatmentwithdrawal than among patients whose MRI scans showed no evidence of tumor atthe time of withdrawal (patients with macroprolactinomas, 78 percent vs. 33 percent,P=0.001; patients with microprolactinomas, 42 percent vs. 26 percent, P=0.02).
c o n c l u s i o n s
Cabergoline can be safely withdrawn in patients with normalized prolactin levels andno evidence of tumor. However, because the length of follow-up in our study was insuf-ficient to rule out a delayed increase in the size of the tumor, we suggest that patients beclosely monitored, particularly those with macroprolactinomas, in whom renewedgrowth of the tumor may compromise vision.
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine tients, the levels of prolactin are in the normal range common type of pituitary tumor, has an es- one year after the withdrawal of cabergoline.22,23 ptimated prevalence of 100 per 1 million per- We report the results of a prospective study of ca- sons. In most cases, medical therapy with dopamine bergoline withdrawal in patients who were treatedagonists normalizes the level of prolactin, restores primarily with this compound.
gonadal function and fertility, and substantiallyreduces the size of the tumor.1,2 Bromocriptine (at a dose of 2.5 to 15 mg daily) has been the tradition-
al drug used to manage prolactinoma3-5; it normal- inclusion criteria
izes prolactin levels in 80 to 90 percent of patients Patients were eligible for our study if after treatment
with microprolactinomas and in approximately 70 with cabergoline they had serum prolactin levels
percent of those with macroprolactinomas, decreas- that were in the normal range and the tumor had
es the size of the tumor, and improves visual-field disappeared or decreased in size by 50 percent or
defects. However, bromocriptine often has side ef- more from base line. Patients were considered for
fects that may prevent the administration of thera- withdrawal of cabergoline only if the outer border
peutic doses.6 Cabergoline, a selective dopamine of the tumor was 5 mm or more from the optic chi-
D2-receptor agonist with long-lasting action has asm, without magnetic resonance imaging (MRI)
been used as a highly effective treatment for mi- evidence of invasion of one or both cavernous sinus-
croprolactinoma and macroprolactinoma.7-10 The es or any other critical area. Patients were required
side effects of cabergoline appear to be less frequent to continue follow-up after withdrawal for at least
and less severe than those of bromocriptine.7,10 24 months. To minimize the risk of errors in read-
Primary cabergoline treatment has been associated ing MRI scans, all patients continued to receive ca-
with greater tumor shrinkage than has primary ther- bergoline therapy for 12 months after fulfilling the
apy with other dopamine agonists.10 In patients with withdrawal criteria and before withdrawal of the
microprolactinomas or macroprolactinomas who medication. The study was approved by the ethics
were treated primarily with either bromocriptine or committee of Federico II University of Naples. All
cabergoline, cabergoline appeared to be superior patients provided informed consent. In the period
in normalizing prolactin levels, restoring gonadal from 1994 to 1997, oral consent was obtained in the
function, and decreasing the size of the tumor.11
presence of a third party, and after 1997, written con- There is wide consensus that primary medical sent was obtained. All procedures were performed treatment of prolactinoma with dopamine agonists in accordance with the standard approach used to
is preferable to surgery, not only because of the ex- treat prolactinoma at Federico II University Hospital.
cellent clinical results of medical therapy but also
because of the risk of recurrent hyperprolactinemia patients
after unsuccessful surgery.12-14 A definitive cure of From January 1, 1994, through December 31, 1998,
these tumors is considered possible, however, only 354 patients (283 women and 71 men) in whom
with surgery or, in rare cases, with surgery plus ra- hyperprolactinemia was newly diagnosed received
diation therapy.2 The results of withdrawal of med- cabergoline as first-line therapy (194 patients with
ical therapy, based on scanty data obtained from microprolactinomas, 135 patients with macropro-
small patient cohorts that were followed for short lactinomas, and 25 patients with nontumoral hyper-
periods, have indicated that hyperprolactinemia re- prolactinemia) (Fig. 1). Subsequently, treatment
curs more often when treatment is withdrawn than was stopped in 57 patients (16 percent) because of
when surgery is performed — for a rate of approxi- pregnancy. Prolactin levels normalized in 273 of the
mately 90 percent in patients with macroprolactino- remaining 297 patients (92 percent); 200 of these
mas and 80 percent in patients with microprolacti- 273 patients (73 percent, 56 percent of the study co-
nomas.15-21 Renewed growth of the tumor appears hort) fulfilled the criteria for cabergoline withdraw-
to be uncommon but may occur after some delay al and were included in the study (Table 1). Diag-
and with a risk of compromised vision.15-21 To our nostic criteria for macroprolactinoma were serum
knowledge, no systematic studies have investigated prolactin levels of 200 µg per liter or more and ev-
the effect of cabergoline withdrawal on rates of re- idence on MRI of a pituitary tumor that was more
mission of prolactinoma, although preliminary re- than 10 mm in diameter. For microprolactinoma,
ports have indicated that in 17 to 30 percent of pa- the criteria were serum prolactin levels of 50 µg
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
c a b e r g o l i n e w i t h d r a w a l i n h y p e r p r o l a c t i n e m i a 354 Patients received cabergoline therapy 194 With microprolactinomas135 With macroprolactinomas 39 With microprolactinomas18 With macroprolactinomas 155 With microprolactinomas117 With macroprolactinomas 10 With microprolactinomas14 With macroprolactinomas 145 With microprolactinomas103 With macroprolactinomas shrinkage <50% or <5 mm fromthe optic chiasm or invasion of critical structures 40 With microprolactinomas33 With macroprolactinomas optic chiasm, no invasion of critical structures 200 Patients (67%) eligible for withdrawal Figure 1. Patient Population at Study Entry.
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine Table 1. Base-Line Characteristics of 200 Patients Eligible for Withdrawal of Cabergoline.*
Nontumoral
Hyperprolactinemia
Microprolactinomas
Macroprolactinomas
Characteristic
* Plus–minus values are means ±SD. P values are for comparison of the three groups by analysis of variance.
† P<0.001 for the comparison with the macroprolactinoma group.
‡ P=0.03 for the comparison with the microprolactinoma group. § P<0.001 for the comparison with the microprolactinoma and macroprolactinoma groups.
per liter or more and evidence on MRI of a pituitary treatment, the dose was adjusted every two months
tumor that was 10 mm or less in diameter. For non- on the basis of suppression of serum prolactin. The
tumoral hyperprolactinemia, the criteria were serum dose of cabergoline was increased to 5 to 7 mg per
prolactin levels above the normal range and evi- week in patients in whom prolactin levels did not
dence on MRI of a normal pituitary, without anoth- normalize11 and was reduced in patients in whom
er explanation for an increased prolactin level, such the prolactin levels declined to less than 5 µg per liter.
as primary hypothyroidism or drug-induced hyper- Before the withdrawal of cabergoline, the dose was
prolactinemia.24 No patient with nontumoral hyper- reduced to 0.5 mg per week in all patients; caber-
prolactinemia or a microprolactinoma and only 31 goline was withdrawn only in patients whose prolac-
of 103 (30 percent) of those with macroprolacti- tin levels remained normal after dose reduction.
nomas had panhypopituitarism. All male patients
had a history of decreased libido and impaired sex- study protocol
ual potency, and all female patients had a history The few patients who had hypopituitarism received
of menstrual disturbances. Of 208 female patients, standard replacement therapy with recombinant
111 (53 percent) had spontaneous or expressible human growth hormone (5 to 8 µg per kilogram of
galactorrhea. Of 103 patients with macroprolac- body weight per day subcutaneously), levothyroxine
tinomas, 39 (38 percent) had visual-field defects, (50 to 100 µg orally per day), cortisone acetate (25
and visual loss occurred in 4 of these patients (10 to 37.5 mg orally per day), and either estrogen–
percent).
progestin (orally each day) or testosterone (250 mgby intramuscular injection monthly), as necessary.11 t r e a t m e n t p r o t o c o l
Serum insulin-like growth factor I, free thyroid Cabergoline was administered orally at a single hormone, and testosterone and serum and urinarystarting dose of 0.5 mg in the first week and then at a sodium and potassium were measured periodicallydose of 0.5 mg twice per week. After two months of to assess the adequacy of the hormone-replacement Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
c a b e r g o l i n e w i t h d r a w a l i n h y p e r p r o l a c t i n e m i a therapy. At the time of enrollment, the serum pro- peated every six months the first year and then year-
lactin level was calculated as the average of a 6-hour ly. If recurrent hyperprolactinemia was diagnosed,
profile during which blood was sampled every 30 the patient underwent MRI at the time of diagnosis.
minutes from 8 a.m. to 2 p.m.; prolactin levels were
measured at 8, 8:15, and 8:30 in the morning dur- ophthalmologic examination
ing treatment, and the average value was recorded. All patients with macroprolactinomas underwent
Prolactin levels were measured by radioimmuno- testing for visual-field defects with the use of Gold-
assay (intraassay and interassay coefficients of vari- mann–Friedmann perimetry, as well as visual acu-
ation, 5 percent and 7 percent, respectively; normal ity, at base line. The ophthalmologic examination
range, 5 to 25 µg per liter in women and 5 to 15 µg was repeated every six months during cabergoline
per liter in men). General clinical examinations were treatment in patients with visual disturbances. After
performed throughout follow-up. After the with- the withdrawal of cabergoline, two patients were re-
drawal of cabergoline, prolactin levels were meas- examined because of visual abnormalities due to at-
ured every 15 days during the first month, then once rophy of the optic nerve, despite the disappearance
a month for 5 months, quarterly during the second of tumor during treatment with cabergoline.
half of the first year, and every 6 months thereafter.
This scheme was used for all the study subjects dur- statistical analysis
ing follow-up, unless there was a recurrence. For pa- Data are reported as means ±SD. Data analysis
tients in whom a recurrence developed, treatment was performed with the use of SPSS software (SAS
was provided according to our current clinical pro- Institute). Mean values in the three study groups
tocol on the basis of prolactin levels.
were compared with the use of paired and unpairedt-tests and analyses of variance. Categorical varia- d i a g n o s i s o f r e c u r r e n c e
bles were compared with the use of Fisher’s exact Recurrence was considered to have occurred if pro- test. A Cox proportional-hazards regression analy-lactin levels were above the normal range. To min- sis was used to determine which variables indepen-imize the risk of renewed symptoms, cabergoline dently predicted recurrence of hyperprolactinemia,treatment was immediately restarted if clinical evaluated as the average prolactin value at the lastsymptoms related to hyperprolactinemia reappeared follow-up visit after the withdrawal of cabergoline.
or if repeated measurements of prolactin after 7 to We entered into the model only variables that had10 days confirmed hyperprolactinemia. Otherwise, a P value of less than 0.01 in the univariate analysis.
follow-up was continued according to the study The Kaplan–Meier method was used to analyze theprotocol.
primary end point of recurrent hyperprolactinemiaduring long-term follow-up. Recurrence-free sur- m r i s t u d i e s
vival was measured from the date of cabergoline Tumor mass was evaluated with the use of MRI as withdrawal to the date of relapse, and the data werepreviously reported.10 The MRI studies were per- censored at the date of the last follow-up visit. Theformed with clinical 0.5-Tesla scanners from 1994 log-rank test was used to compare recurrence-freeto 1999, 1-Tesla scanners from 2000 to 2001, and survival curves. P values of less than 0.05 were con-1.5-Tesla scanners from 2002 on, with the use of sidered to indicate statistical significance.
T -weighted gradient-recalled echo in the sagittal and coronal planes. For each measurement, 7 to 11 slices were obtained, at a thickness of 2 to 3 mm and
an in-plane spatial resolution of 0.70 to 0.97 mm. nontumoral hyperprolactinemia
Images were obtained before and after the admin- In the group of patients with nontumoral hyper-
istration of 0.1 mmol of gadolinium chelate (dieth- prolactinemia, cabergoline was withdrawn after
ylene-triamine pentacetate). The maximal tumor a median treatment period of 36 months at a me-
diameter was calculated in millimeters. The MRI dian dose of 0.5 mg per week (Table 1). Of 25 fe-
studies were performed before treatment with ca- male patients with nontumoral hyperprolactinemia,
bergoline; at 3, 6, and 12 months during the first 6 (24 percent) had a recurrence after a median of
year of treatment; and then every 6 to 12 months on 18 months (Table 2), but in none of them did symp-
the basis of the reduction in the size of the tumor. toms reappear. In the remaining 19 of these pa-
After the withdrawal of cabergoline, MRI was re- tients (76 percent), hyperprolactinemia remained
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine controlled at a median of 48 months after the with- mained controlled in the remaining 73 patients (70drawal of cabergoline.
percent) at a median of 36 months after the with-drawal of cabergoline.
m i c r o p r o l a c t i n o m a
In the group with microprolactinomas, cabergoline macroprolactinoma
was withdrawn after a median treatment period of In the macroprolactinoma group, cabergoline was
48 months at a median dose of 1 mg per week (Ta- withdrawn after a median treatment period of 42
ble 1). In 63 of 105 patients, MRI studies of the pitu- months at a median dose of 1 mg per week (Table 1).
itary gland showed no evidence of renewed tumor Of 70 patients with macroprolactinomas, MRI stud-
growth. In the remaining 42 patients, the tumor ies of the pituitary gland in 46 showed no evidence
decreased by 55.5±4.8 percent (from 7.0±1.4 to of renewed tumor growth. In the remaining 24 pa-
3.1±0.6 mm). In 32 patients (30 percent), hyper- tients, tumor size decreased by 60.7±6.2 percent
prolactinemia recurred after a median of 12 months; (from 16.9±4.1 to 6.6±1.6 mm). Twenty-five of the
none of these 32 patients had MRI evidence of re- 70 patients (36 percent) had a recurrence of hyper-
current microprolactinomas. In 5 of 25 female pa- prolactinemia after a median of 18 months; none
tients (20 percent) oligomenorrhea developed; none of these 25 had MRI evidence of recurrent growth
of the 25 female patients had galactorrhea. Three of of the macroprolactinoma. Of 14 female patients,
seven male patients (43 percent) noted decreases in 5 noted renewed oligomenorrhea (36 percent), and
sexual potency and libido, although their testoster- 1 had recurrent galactorrhea (7 percent). Of 11 male
one levels did not change. Hyperprolactinemia re- patients, 4 (36 percent) reported decreased sexual
Table 2. Characteristics of the Patients According to Whether Hyperprolactinemia Recurred after the Withdrawal of Cabergoline.*
Nontumoral
Characteristic
Hyperprolactinemia
Microprolactinomas
Macroprolactinomas
Recurrence Recurrence Value Recurrence Recurrence Value Recurrence Recurrence Value Average prolactin level at last follow-up Range of follow-up after withdrawal — mo * Plus–minus values are means ±SD. P values were calculated with the use of Student’s t-test for unpaired (individual) data and the chi-square test or Fisher’s exact test for proportions.
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
c a b e r g o l i n e w i t h d r a w a l i n h y p e r p r o l a c t i n e m i a potency and decreased libido, which were accom- panied by a decrease in testosterone levels (from 4.2±0.5 to 3.2±0.4 µg per liter). In 45 patients (64percent), hyperprolactinemia remained controlledat a median of 48 months after the withdrawal ofcabergoline.
As shown in Table 2, in all three groups, the over- all recurrence rate, the median time to recurrence,and the prolactin level at recurrence were similar.
The average prolactin level was significantly lowerat the last follow-up visit than at diagnosis in allgroups (Table 2). The Kaplan–Meier estimate of the five-year recurrence rate was higher among patients Microprolactinomas, negative MRIMicroprolactinomas, positive MRI with either macroprolactinomas or microprolacti- Lack of Recurrence after Withdrawal (%)
nomas who had remnant tumors on MRI before the withdrawal of cabergoline than among those who had no evidence of tumor on MRI (patients with Follow-up after Withdrawal (mo)
macroprolactinomas, 78 percent vs. 33 percent, P=0.001; patients with microprolactinomas, 42 per- cent vs. 26 percent, P=0.02) (Fig. 2). Sixty-three pa- tients had recurrent hyperprolactinemia, more thanhalf of them (56 percent) during the first year afterthe withdrawal of cabergoline, 33 percent during the second year, 11 percent during the third year,and none thereafter (P<0.001).
Recurrences of hyperprolactinemia were not as- sociated with male or female sex (P=0.93) or with the presence of menopause (28 percent of studypatients were in premenopause and 36 percent inpostmenopause, P=0.50). Age, prolactin levels atbase line, nadir prolactin levels, percentage of pro- lactin suppression, smallest tumor diameter after Lack of Recurrence after Withdrawal (%)
cabergoline therapy, duration of treatment, and dose of cabergoline were higher in patients with a recurrence than in those in whom persistent con- trol was achieved, though there were some differ- Follow-up after Withdrawal (mo)
ences among the three groups (Table 2). Cox regres- Figure 2. Kaplan–Meier Estimates of Recurrence Rates after 60 Months in the
sion analysis indicated that the maximal diameter Three Groups of Patients (Panel A) and in the Overall Study Population (Panel B)
of a tumor during cabergoline treatment was the According to Whether MRI Studies during Treatment Were Positive or Negative.
best predictor of the prolactin level at the last fol- Panel A shows recurrence rates in 25 patients with nontumoral hyperpro- low-up visit after the withdrawal of cabergoline (chi- lactinemia, 63 patients with microprolactinomas and negative MRI studies, square=12, P<0.001). The hazard rate for the re- 42 patients with microprolactinomas and persistent abnormalities on MRI, 46 patients with macroprolactinomas and negative MRI studies, and 24 pa- currence of hyperprolactinemia was 19 percent for tients with macroprolactinomas and persistent abnormalities on MRI. Panel each millimeter increment in the maximal tumor B shows rates of recurrence in two groups of patients divided on the basis of diameter. Figure 3 summarizes the prevalence of MRI findings. The first group included the 25 patients with nontumoral hyper- normoprolactinemia after cabergoline withdrawal prolactinemia and the 63 patients with microprolactinomas and 46 patients in different categories of patients.
with macroprolactinomas who had negative MRI studies. The second group included the 42 patients with microprolactinomas and the 24 patients with macroprolactinomas who had persistent abnormalities on MRI. The P values were calculated with the log-rank test. In none of the patients with a recur-rence of hyperprolactinemia after cabergoline withdrawal did the tumor reap- Our study showed that, in general, remission of hy- perprolactinemia persisted after the withdrawal Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine uctures. Kaplan–Meier estimates of recurrence at five- abergoline Therapy
Patient Population at the Time of Withdrawal of C
Figure 3.
MRI positive denotes small tumor remnants (shrinkage ≥50 percent), ≥5 mm from the optic chiasm, and no invasion of critical str year follow-up for each group are given in parentheses. Percentages shown have been rounded.
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
c a b e r g o l i n e w i t h d r a w a l i n h y p e r p r o l a c t i n e m i a of cabergoline, without any evidence of renewed Because the possibility of inducing long-last- tumor growth. Neither sex nor age was associated ing control of hyperprolactinemia without continu-with the recurrence of hyperprolactinemia, and the ing pharmacologic treatment has profound conse-maximal tumor diameter during treatment with cab- quences not only for patients’ compliance but alsoergoline was the best predictor of the prolactin lev- for the costs of treatment, we designed our study toel at the last follow-up visit after withdrawal, with determine the rate of success of cabergoline with-a hazard rate to predict recurrence of 19 percent.
drawal and potentially useful criteria for identify- The efficacy of the primary treatment of both mi- ing patients with the highest likelihood of prolac- croprolactinoma and macroprolactinoma with do- tin control after the discontinuation of cabergolinepamine-agonist compounds and, more specifically, therapy. According to the literature on bromocrip-with cabergoline, is widely documented.7-11,22-30 tine withdrawal,15-21,31 achieving normoprolacti-Cabergoline therapy has also been successful in nemia is the first criterion and a mandatory one.
patients whose prolactinomas were resistant to We chose tumor shrinkage as an additional criterionbromocriptine or who could not tolerate bromo- and divided our patients on the basis of the extentcriptine,26,27 with a success rate of over 90 per- of tumor shrinkage into two groups, those whosecent in patients with newly diagnosed prolactino- tumor disappeared and those in whom a 50 per-mas.10,23,29 The most important shortcoming of cent or greater reduction occurred from base linemedical treatment has been considered to be its in a tumor mass that was not near the optic chiasmtheoretically lifelong requirement.
or invading the cavernous sinuses or other critical The results of bromocriptine withdrawal have cerebral areas. Previous morphologic studies of tu- been reported after a wide range of follow-up peri- mor specimens obtained after long-term treatmentods (8 to 240 months; median, 24). The results have with bromocriptine have shown atrophic tumor-cellbeen variable, with normoprolactinemia sustained nests, pyknosis, and cytolysis, as well as karyorrhe-in 7 to 38 percent of patients.15-21,31 Notably, an in- xis, necrosis, fibrosis, hyalinosis, and inflammato-crease in tumor volume with clear reexpansion has ry-cell infiltration, suggesting a cytocidal effect ofbeen found in approximately 10 percent of cas- the drug.37 We attribute our high rate of stable nor-es.15,17 Renewed growth of the tumor after dis- moprolactinemia without tumor recurrence, evencontinuation of bromocriptine therapy, however, among patients with macroprolactinomas, to anseems to depend on the duration of treatment be- effective antitumoral effect of cabergoline.
fore drug withdrawal: the tumors appear to be most It should also be noted that even though patients susceptible to renewed growth after a course of whose MRI studies showed no evidence of tumortreatment of less than 12 months.32-35 had a significantly lower rate of recurrence than Withdrawal of cabergoline has been reported in those with visible tumors on MRI, 59 percent of pa- only a few studies. One study reported persistent tients with small, remnant microprolactinomas andnormoprolactinemia in 1 patient of 9 with a macro- 23 percent of those with small, remnant macropro-prolactinoma (11 percent) and 4 patients of 18 with lactinomas had persistent normoprolactinemia af-microprolactinomas (22 percent) 12 months after ter cabergoline withdrawal. This finding suggestswithdrawal23; another reported that there was no that in some of these patients the abnormalities de-change in prolactin levels in 24 percent of 25 pa- tected on MRI may have been small, nonfunction-tients after 3 to 60 months36 and in 31 percent of ing lesions, fibrotic scars, or other nontumoral ab-32 patients who were treated for 3 to 24 months normalities (incidentalomas).
12 months after withdrawal.22 In an earlier study, We excluded patients with certain conditions, we found that normoprolactinemia persisted 12 such as pregnancy,38,39 previous surgery, and ra-months after withdrawal in 17 percent of 23 pa- diotherapy, that are known to facilitate the occur-tients with microprolactinomas who were treated rence of normoprolactinemia after dopamine-ago-with cabergoline for 12 months.25 An important nist withdrawal. Menopause might be consideredshortcoming of previous studies was the lack of cri- a factor that influences the reduction of hyperpro-teria for timing withdrawal: in those studies caber- lactinemia,40 but the rate of recurrence of hyper-goline was apparently withdrawn after varying peri- prolactinemia was similar among patients who wereods (12 months, on average), without systematic premenopausal and those who were postmeno-evaluation of the outcome of withdrawal with re- pausal. Remission of prolactinomas has also beenspect to prolactin suppression and tumor shrinkage. described as part of the natural history of untreated Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine tumors.41-43 However, even allowing for the pos- was 19 percent. However, until data from a studysible role of this natural history in the outcome of with a longer follow-up period are available, closemicroprolactinomas, it seems unlikely that macro- monitoring for recurrent hyperprolactinemia andprolactinomas would spontaneously regress. In our renewed tumor growth is important, particularly instudy, the overall rates of remission at five years patients with macroprolactinomas, in whom re-as estimated by the Kaplan–Meier method were newed growth may compromise vision.
76 percent among patients with nontumoral hyper- Supported in part by a grant ( 2003068735) from the Italian Minis- prolactinemia, 67 percent among those with micro- ter of University and Research, Rome. We are indebted to Giovanni Vitale, Maria Luisa Landi, and Nicola prolactinomas, and 57 percent among those with Milano (Department of Molecular and Clinical Endocrinology andmacroprolactinomas, rates that are higher than Oncology, Federico II University of Naples) for contributing to pa-those generally reported in the literature as sponta- tients’ care; to Francesco Briganti, Sossio Cirillo, and Francesco Di Salle (Department of Biomorphologic and Functional Sciences, Fe- derico II University of Naples) for reading MRI studies of the sella; to Our data support the concept of periodic with- Mario Petretta (Department of Internal Medicine, Federico II Uni- drawal of cabergoline therapy, especially in patients versity of Naples) for his help in performing an accurate statistical analysis of the data; and to Edward Laws, Jr. (Department of Neuro- with negative MRI studies during treatment. The logical Surgery, University of Virginia, Charlottesville), for his criti-risk of recurrent hyperprolactinemia with each mil- cal evaluation and linguistic revision of the manuscript.
limeter increment in the size of the tumor remnant r e f e r e n c e s
12. Schlechte JA, Sherman BM, Chapler FK,
20. Zarate A, Canales ES, Cano C, Pilonieta
and prolactin excess. Lancet 1998;352:1455- VanGilder J. Long term follow-up of women CJ. Follow-up of patients with prolactino- with surgically treated prolactin-secreting mas after discontinuation of long-term ther- pituitary tumors. J Clin Endocrinol Metab Management of prolactinomas. J Clin Endo- 13. Bevan JS, Adams CBT, Burke CW, et al.
21. Passos VQ, Souza JJS, Musolino NRC,
Factors in the outcome of transsphenoidal Bronstein MD. Long-term follow-up of pro- Drugs five years later: bromocriptine. Ann surgery for prolactinoma and non-function- lactinomas: normoprolactinemia after bro- ing pituitary tumour, including pre-opera- tive bromocriptine therapy. Clin Endocrinol 22. Ferrari C, Paracchi A, Mattei A, de Vin-
prolactin-secreting macroadenomas: results 14. Losa M, Mortini P, Barzaghi R, Gioia L,
centiis S, D’Alberton A, Crosignani P. Caber- of a prospective multicenter study. J Clin En- Giovanelli M. Surgical treatment of prolac- goline in the long-term therapy of hyper- tin-secreting pituitary adenomas: early re- sults and long-term outcome. J Clin Endo- MF. Dopamine agonists and pituitary tumor 23. Cannavo S, Curto L, Squadrito S, Almoto
shrinkage. Endocr Rev 1992;13:220-40.
15. Johnston DG, Hall K, Kendall-Taylor P,
B, Vieni A, Trimarchi F. Cabergoline: a first- Parkes D. Side effects of bromocriptine.
Patrick D, Watson M, Cook DB. Effect of do- choice treatment in patients with previously pamine agonist withdrawal after long-term untreated prolactin-secreting pituitary ade- Webster J, Piscitelli G, Polli A, et al.
therapy in prolactinomas: studies with high- nomas. J Endocrinol Invest 1999;22:354-9.
definition computerised tomography. Lancet 24. Colao A, Di Sarno A, Cappabianca P, et
criptine in the treatment of hyperprolactine- al. Gender differences in the prevalence, 16. Wang C, Lam KSL, Ma JT, Chan T, Liu
clinical features and response to cabergo- line in hyperprolactinemia. Eur J Endocrinol Ferrari CI, Abs R, Bevan JS, et al. Treat- effect of drug withdrawal. Clin Endocrinol 25. Di Sarno A, Landi ML, Marzullo P, et al.
line: a study of 85 patients. Clin Endocrinol The effect of quinagolide and cabergoline, 17. van ’t Verlaat JW, Croughs RJ. With-
two selective dopamine receptor type 2 ago- Verhelst J, Abs R, Maiter D, et al. Caber- nists, in the treatment of prolactinomas.
goline in the treatment of hyperprolactine- therapy for macroprolactinomas: effect on Clin Endocrinol (Oxf ) 2000;53:53-60.
mia: a study in 455 patients. J Clin Endo- plasma prolactin and tumour size. Clin En- 26. Colao A, Di Sarno A, Sarnacchiaro F, et
al. Prolactinomas resistant to standard do- 10. Colao A, Di Sarno A, Landi ML, et al.
18. Moriondo P, Travaglini P, Nissim M,
pamine agonists respond to chronic caber- Macroprolactinoma shrinkage during caber- Conti A, Faglia G. Bromocriptine treatment goline treatment. J Clin Endocrinol Metab goline treatment is greater in naive patients of microprolactinomas: evidence of stable than in patients pretreated with other dopa- prolactin decrease after drug withdrawal.
27. Delgrange E, Maiter D, Donckier J. Ef-
mine agonists: a prospective study in 110 J Clin Endocrinol Metab 1985;60:764-72.
fects of the dopamine agonist cabergoline in patients. J Clin Endocrinol Metab 2000;85: 19. Winkelmann W, Allolio B, Deuss U,
patients with prolactinoma intolerant or re- sistant to bromocriptine. Eur J Endocrinol 11. Di Sarno A, Landi ML, Cappabianca P, et
lactinemia after withdrawal of bromocrip- al. Resistance to cabergoline as compared tine long-term therapy in patients with pro- 28. Biller BM, Molitch ME, Vance ML, et al.
with bromocriptine in hyperprolactinemia: Treatment of prolactin-secreting macroade- prevalence, clinical definition, and therapeu- Scapagnini U, eds. Prolactin: basic and clin- tic strategy. J Clin Endocrinol Metab 2001; ical correlates. Padova, Italy: Liviana Press, nist cabergoline. J Clin Endocrinol Metab Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.
c a b e r g o l i n e w i t h d r a w a l i n h y p e r p r o l a c t i n e m i a 29. Colao A, Di Sarno A, Landi ML, et al.
34. Orrego JJ, Chandler WF, Barkan AL.
effects of time, medical treatment and preg- Long-term and low-dose treatment with cab- Rapid re-expansion of a macroprolactinoma nancy in 176 hyperprolactinemic women.
ergoline induces macroprolactinoma shrink- after early discontinuation of bromocriptine.
Eur J Obstet Gynecol Reprod Biol 1992;44: age. J Clin Endocrinol Metab 1997;82:3574- 35. Cunnah D, Besser M. Management of
40. Karunakaran S, Page RC, Wass JA. The
30. Pinzone JJ, Katznelson L, Danila DC,
prolactinomas. Clin Endocrinol (Oxf ) 1991; effect of the menopause on prolactin levels Pauler DK, Miller CS, Klibanski A. Primary in patients with hyperprolactinaemia. Clin medical therapy of micro- and macroprolac- 36. Muratori M, Arosio M, Gambino G,
41. Koppelman MC, Jaffe MJ, Rieth KG,
bergoline in the long-term treatment of hy- 31. Bergh T, Nillius SJ, Wide L. Menstrual
perprolactinemic and acromegalic patients.
mia, amenorrhea, and galactorrhea: a retro- function and prolactin levels after long-term spective assessment of twenty-five cases.
bromocriptine treatment of hyperprolacti- 37. Gen M, Uozumi T, Ohta M, Ito A, Kaji-
naemic amenorrhoea. Clin Endocrinol (Oxf ) wara H, Mori S. Necrotic changes in prolac- 42. Schlechte J, Dolan K, Sherman B, Chap-
tinomas after long term administration of ler F, Luciano A. The natural history of un- 32. Molitch ME. Medical treatment of pro-
treated hyperprolactinemia: a prospective analysis. J Clin Endocrinol Metab 1989;68: 38. Jeffcoate WJ, Pound N, Sturrock ND,
33. Thorner MO, Perryman RL, Rogol AD,
43. Sisam DA, Sheehan JP, Sheeler LR. The
et al. Rapid changes of prolactinoma vol- tients with hyperprolactinaemia. Clin Endo- natural history of untreated microprolacti- ume after withdrawal and reinstitution of 39. Crosignani PG, Mattei AM, Severini V,
Copyright 2003 Massachusetts Medical Society. Cavioni V, Maggioni P, Testa G. Long-term images in clinical medicine
The Journal welcomes consideration of new submissions for Images in Clinical
Medicine. Instructions for authors and procedures for submissions can be found on
the Journal’s Web site at www.nejm.org. At the discretion of the editor, images that
are accepted for publication may appear in the print version of the Journal, the elec-
tronic version, or both.
Downloaded from www.nejm.org at IMPERIAL COLL SCH MED on January 16, 2004.
Copyright 2003 Massachusetts Medical Society. All rights reserved.

Source: http://ds.dial.pipex.com/town/estate/aquc35/book/colao.pdf